Flow Cytometry in Hematopathology (2nd Edition)

Flow Cytometry in Hematopathology: A Visual Approach to Data Analysis and Interpretation

D.N. and L.W.D.: With all of our love to Vladimir, Petrushka, Fidelio, Feivel, Amadeus, Cherubino, Tamino, Wolfgang (“Bum Jr.”), Belmonte, Apollo, Sergei, “Post Auto,” Paddington, “Cookie Jar,” Sasha, Misha, Clipper, Beaker, Lanipo and Hiapo.

Flow Cytometry in Hematopathology: A Visual Approach to Data Analysis and Interpretation Second Edition

Doyen Nguyen, MD Lawrence W. Diamond, MD Raul C. Braylan, MD University of Florida College of Medicine Gainesville, FL

© 2007 Humana Press Inc., a part of Springer Science+Business Media, LLC 999 Riverview Drive, Suite 208 Totowa, New Jersey, 07512 www.humanapress.com All rights reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording, or otherwise without written permission from the Publisher. The content and opinions expressed in this book are the sole work of the authors and editors, who have warranted due diligence in the creation and issuance of their work. The publisher, editors, and authors are not responsible for errors or omissions or for any consequences arising from the information or opinions presented in this book and make no warranty, express or implied, with respect to its content. Due diligence has been taken by the publishers, editors, and authors of this book to assure the accuracy of the information published and to describe generally accepted practices. The contributors herein have carefully checked to ensure that the drug selections and dosages set forth in this text are accurate and in accord with the standards accepted at the time of publication. Notwithstanding, as new research, changes in government regulations, and knowledge from clinical experience relating to drug therapy and drug reactions constantly occur, the reader is advised to check the product information provided by the manufacturer of each drug for any change in dosages or for additional warnings and contraindications. This is of utmost importance when the recommended drug herein is a new or infrequently used drug. It is the responsibility of the treating physician to determine dosages and treatment strategies for individual patients. Further it is the responsibility of the health care provider to ascertain the Food and Drug Administration status of each drug or device used in their clinical practice. The publisher, editors, and authors are not responsible for errors or omissions or for any consequences from the application of the information presented in this book and make no warranty, express or implied, with respect to the contents in this publication. For additional copies, pricing for bulk purchases, and/or information about other Humana titles, contact Humana at the above address or at any of the following numbers: Tel.: 973-256-1699; Fax: 973-256-8341; E-mail: [email protected]; or visit our Website: www.humanapress.com This publication is printed on acid-free paper. ANSI Z39.48-1984 (American National Standards Institute) Permanence of Paper for Printed Library Materials. Photocopy Authorization Policy: Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted by Humana Press Inc., provided that the base fee of US $30.00 per copy is paid directly to the Copyright Clearance Center at 222 Rosewood Drive, Danvers, MA 01923. For those organizations that have been granted a photocopy license from the CCC, a separate system of payment has been arranged and is acceptable to Humana Press Inc. The fee code for users of the Transactional Reporting Service is: [978-1-58829-855-3/07 $30.00]. 10 9 8 7 6 5 4 3 2 1 eISBN: 978-1-59745-162-8 Library of Congress Control Number: 2007922048

Contents

Preface to the Second Edition Preface to the First Edition Acknowledgments List of Abbreviations List of Case Studies Color Plates

ix xi xiii xv xvii follow p. 174

Chapter 1 Approach to flow cytometry: General considerations 1.1 1.1.1 1.2 1.3 1.4

Reasons for the necessity of proper data analysis The pitfalls of the FCM data format of “percent positive” per antibody tested General aspects of FCM data analysis and interpretation Other applications of FCM in hematopathology Maturation and differentiation of hematopoietic elements: An overview based on the immunologic markers currently in use in the FCM laboratory

2 2 4 7

9

Chapter 2 FCM immunophenotyping and DNA analysis: Practical aspects that can affect data analysis and interpretation 2.1 2.1.1 2.1.2 2.2 2.3 2.4 2.4.1 2.4.2 2.4.3 2.5 2.5.1 2.5.2 2.5.3 2.5.4 2.6 2.6.1 2.6.1.1 2.6.2

Sample selection Liquid specimens Solid tissue specimens Preparing nucleated cell suspensions Cell yield and viability Sample staining Surface antigens Intracellular antigens DNA content Data acquisition Calibration Color compensation List mode data collection Exclusion of nonviable cells Antibody panel design Antibody selection Anti-light chain antibodies Fluorochrome conjugation

13 14 15 16 16 16 17 17 18 19 19 19 20 21 22 23 23 26

vi

CONTENTS

2.7 2.7.1 2.7.2 2.8 2.8.1 2.9 2.9.1 2.9.2 2.10 2.10.1 2.10.2

Comprehensive antibody panels Disease-oriented antibody panels Antibody panels oriented by specimen type Tailored panels and add-on testing Minimal residual disease FCM immunophenotyping data representation Analysis panels Color display Approach to DNA data analysis DNA ploidy S-phase

31 31 32 34 36 37 37 38 41 42 42

Chapter 3 FCM data analysis on nearly homogeneous samples 3.1 3.1.1 3.1.2 3.1.3 3.1.3.1 3.2 3.3 3.4 3.4.1 3.4.2 3.4.3 3.5 3.5.1 3.5.1.1 3.5.2 3.5.3 3.6 3.6.1 3.6.2 3.6.3 3.6.3.1 3.6.3.2 3.6.3.3 3.6.3.4 3.6.4 3.6.5 3.7 3.8

FCM parameters Forward scatter Side scatter Fluorescence Heterogeneous fluorescence intensity (bimodal, variable) Fluorescence dynamic range Strategy to the visual review of FCM immunophenotyping data Common SSC/CD45 patterns Assessment of the blast population Immature neoplastic cells with downregulated CD45 SSC/CD45 in mature lymphoid disorders Other dot plot patterns useful in acute leukemia diagnosis Useful antigenic features in AML Myeloid phenotypic abnormalities and MRD detection Precursor B-ALL versus bone marrow B-cell progenitors Useful antigenic features in precursor T-lymphoma/leukemia Evaluation of mature lymphoid malignancies Assessment of surface light chain expression Assessment of pan B-cell antigens Useful antigenic features in mature B-cell malignancies CD10 expression: Follicular center cell lymphomas Pattern of CD20 and CD11c coexpression CD5 expression Aberrant B-cell profile Identification of abnormal mature T-cells Useful antigenic features in mature T-cell malignancies Assessing the biological behavior of mature lymphoid neoplasms Dot plot patterns in histiocytic proliferations and nonhematopoietic malignancies

50 50 53 54 64 71 71 73 73 81 83 86 86 95 98 99 106 106 108 112 112 117 123 126 131 150 166 171

Chapter 4 FCM data analysis on heterogeneous specimens 4.1 4.1.1

Identifying normal FCM samples Benign/reactive solid lymphoid tissue (e.g., lymph nodes, tonsils)

175 175

CONTENTS

4.1.1.1 4.1.2 4.1.2.1 4.1.2.2 4.1.2.3 4.1.2.4 4.1.2.5 4.1.2.6 4.1.2.7 4.2 4.2.1 4.2.2 4.2.2.1 4.2.2.2 4.3 4.4 4.4.1 4.4.1.1 4.4.1.2 4.4.2 4.4.3 4.5 4.5.1 4.5.1.1 4.5.1.2 4.5.1.3 4.5.1.4 4.5.2 4.5.2.1 4.5.2.2 4.5.2.3 4.6

Pattern of CD10/CD20 coexpression: Distinction between FRFH and FCC lymphoma Normal peripheral blood and normal bone marrow Blast region Bone marrow B-cell precursors Lymphocytes Monocytes Plasma cells Erythroid precursors Maturing myeloid cells Abnormal heterogeneous samples with a detectable immature neoplastic population Blasts of lymphoid lineage Blasts of myeloid lineage AML High-grade MDS and MPD with increased blasts Minimal residual disease Abnormal heterogeneous samples with detectable mature neoplastic populations Abnormal mature B-cells Evaluation of CD5 and CD23 FCM features suggestive of anti-CD20 therapy Abnormal mature T-cells and NK cells Abnormal plasma cells present Abnormal blood or bone marrow samples with no detectable neoplastic cells Altered cellular composition and abnormal SSC Increased monocytic elements Increased eosinophils Conspicuous basophils or mast cells Abnormal SSC in granulocytes Abnormal antigenic maturation in myeloid or erythroid precursors Antigenic abnormalities in myeloid precursors Antigenic abnormalities in erythroid precursors Identifying low-grade MDS Coexisting malignancies

vii

179 187 190 192 192 208 209 213 216 220 221 222 224 226 233 233 238 238 242 245 255 260 262 262 262 265 265 273 273 280 280 285

Chapter 5 FCM interpretation and reporting 5.1 5.1.1 5.1.2 5.1.2.1 5.1.2.2 5.1.2.3 5.1.2.4 5.1.2.5 5.1.2.6 5.1.2.7

Immature hematopoietic malignancies ALL/lymphoblastic lymphoma Myeloid malignancies AML-M3 AML with minimal maturation AML with maturation AML with monocytic differentiation AML with erythroid hyperplasia AML with megakaryocytic differentiation MPD and MDS

289 290 292 292 292 292 294 295 295 295

viii

CONTENTS

5.1.3 5.2 5.2.1 5.2.1.1 5.2.1.2 5.2.1.3 5.2.1.4 5.2.1.5 5.2.1.6 5.2.1.7 5.2.1.8 5.2.2 5.2.3 5.2.3.1 5.2.3.2 5.2.3.3 5.3

Acute leukemias with a multilineage antigenic profile Mature lymphoid malignancies B-cell LPD/NHL CD10 expression Coexpression of CD11c, CD25 and CD103 Coexpression of CD5 and CD23 CD5+ CD23¯ B-cell neoplasms CD45 and/or pan B-cell antigens markedly downregulated Nondescript B-cell phenotype and high FSC Nondescript B-cell phenotype and low FSC Monoclonal B-cells of undetermined significance Plasma cell dyscrasias T-cell LPD/NHL CD4+ T-cell LPD/NHL CD8+ disorders CD30+ lymphoma FCM reporting

Suggested reading Appendix: Using the case studies CD-ROM Index

300 301 302 302 306 306 307 310 311 314 318 318 320 321 323 329 333 335 341 343

Preface to the Second Edition

During the time that has elapsed between the first edition and the second edition of this book, there has been considerable improvement in the incorporation of flow cytometry immunophenotyping into hematology and pathology laboratories, including institutions where previous practice had relied heavily on traditional morphology and paraffin-based immunostaining. In addition, flow cytometry (FCM) immunophenotyping has also gained acceptance as a useful diagnostic tool for the identification of not only acute myeloid leukemias but also other myeloid disorders, both malignant and premalignant. During the same time, advances have been made in terms of instrumentation and commercially available reagents. For instance, the introduction of the T-cell receptor (TCR)-Vβ eight-tube kit has greatly facilitated the evaluation of some difficult to diagnose mature T-cell disorders, and the implementation of the DNA dye DRAQ5 has improved the grading of lymphoma subpopulations present in heterogeneous samples. The role of FCM analysis has also progressed beyond that of establishing a diagnosis to that of monitoring disease and providing prognostic information. The second edition of this book reflects the recent advances in the FCM analysis of hematopoietic disorders. To this end, the chapters have been revised to incorporate additional text and figures. The focus of the book and its organization remain unchanged, however. The availability of new software tools has made it possible to add more case studies to the new companion CD-ROM, as well as to render the disk easier to use without the need to install a database engine. The listing of the case studies (and their diagnoses) is provided at the beginning of the book. The reader will do well not to omit the case studies from consideration as they supplement the information provided in the book. It is hoped that this book will bring a better appreciation of the important role of FCM analysis in the diagnosis and management of hematopoietic disorders. When FCM is applied systematically, the potential exists to reduce the confusion that still exists in the current classification of certain malignant lymphomas and lymphoproliferative disorders.

Preface to the First Edition

Flow cytometry immunophenotyping of hematopoietic disorders is a complex and demanding exercise that requires a good understanding of cell lineages, developmental pathways, and physiological changes, as well as broad experience in hematopathology. The process includes several interrelated stages, from the initial medical decision regarding which hematologic condition is appropriate for FCM assay, to the final step of diagnosis whereby the FCM data is correlated with other relevant clinical and laboratory information. The actual FCM testing involves three major steps: preanalytical (specimen processing, antibody staining), analytical (acquiring data on the flow cytometer) and postanalytical (data analysis and interpretation). The literature, including the latest FCM textbooks, provides ample information on the technical principles of FCM such as instrumentation, reagents and laboratory methods, as well as quality control and quality assurance. Similarly, correlations of morphologic findings and phenotypic profiles have been well covered in many publications. In contrast, much less attention has been given to the other equally important aspects of FCM immunophenotyping, especially data analysis. The latter is a crucial step by which a phenotypic profile is established. To bridge this gap in the literature, the focus of this book is more on FCM data analysis than laboratory methods and technical details. For the reader to become familiar with our data analysis strategy, an overview of our approach to the preanalytical and analytical steps is also presented, with an emphasis on the preanalytical aspects, which have been rarely touched upon in the literature. The process of data analysis follows a practical and systematic approach, utilizing the visual patterns of the dual parameter displays rather than calculating a “percent positive” for each individual antibody. The FCM graphic displays presented throughout the book, together with the clinical case studies contained in the companion CD-ROM should facilitate the readers to gain an in-depth appreciation of this visual approach to data analysis. Via the case studies, the topics discussed in the textbook can be illustrated in greater detail, and the FCM diagnostic subtleties will become more apparent. The book is designed for all laboratory professionals involved in the immunophenotyping of hematologic disorders, including pathologists, PhDs, and technologists working in FCM laboratories, residents and fellows in pathology and hematopathology training programs, as well as clinical hematologists with a special interest in this subspecialty. In terms of organization, this book breaks away from the traditional mold used in other textbooks. The chapters are not arranged by specific diagnosis (i.e., the end point of a diagnostic workup) but by how the data presents at the time of the diagnostic consultation. This organization reflects the reallife problem-solving methods applied daily in the laboratory, whereby the strategies employed differ depending on whether the cell population in the sample analyzed is heterogeneous or nearly homogeneous. The few available books covering FCM phenotypes in hematologic malignancies have tended to focus more on leukemias than lymphomas. In this book, equal emphasis is given to both categories of disease, thereby providing considerably more information on lymphomas and chronic lymphoproliferative disorders. Furthermore, DNA cell cycle analysis is also

xii

PREFACE TO THE FIRST EDITION

included in the FCM study of mature lymphoid malignancies, in which the DNA data have been proven to be of prognostic significance, thus permitting a more objective and reproducible grading of these tumors. The approach to the classification of hematologic neoplasms employed in this book also departs from that used in the various existing classifications. The antigenic profiles of leukemias and lymphomas have been incorporated into the more recent classification schemes. However, the phenotypes of many disorders, in particular malignant lymphomas, have been derived from paraffin-based immunostaining instead of FCM studies, thus not taking into consideration the large amounts of valuable information provided by FCM immunophenotyping (e.g., a better appreciation of the pattern of antigenic density distribution and coexpression). The approach taken in this book is to simplify the classification (which should facilitate the comparison of results between different institutions) by utilizing the graphical patterns of phenotypic expression and the results of DNA cell cycle analysis where appropriate, together with other relevant clinical/laboratory data including the morphology of the submitted specimen. A more detailed discussion on the morphology of the bone marrow and peripheral blood manifestations of hematologic disorders can be found in our previous textbook (and its companion CD-ROM) entitled Diagnostic Hematology: A Pattern Approach (Arnold Publishers; distributed in the United States by Oxford University Press; ISBN 0-7506-4247-5). For practical reasons, most of the FCM graphics in the book are presented as black-andwhite illustrations. The dot plots in many of the case studies contained in the CD-ROM are, on the other hand, presented in color to facilitate the viewing of the cell cluster(s) of interest, especially for educational purposes. The use of color dot plots is popular in some laboratories. In our opinion, laboratory staff involved in FCM data analysis should be familiar with both black-and-white and color FCM displays however, rather than relying solely on the color format. The list of suggested readings is not meant to be exhaustive. Many of the references were chosen mainly for the readers to obtain more depth on certain topics, for example, the maturation and differentiation of hematopoietic cells.

Acknowledgments

The authors would like to acknowledge David Novo, president of De Novo Software (www. denovosoftware.com), for providing several copies of his FCS Express software to create the FCM illustrations in the book and the case studies contained in the CD-ROM. We would also like to thank our medical colleagues, in particular Wolfgang Huebl and Samer Al Quran, for the contribution of hematologic samples and/or unusual cases. The authors also thank the many FCM technologists we have been associated with for their excellent work, in particular Mikhail Mazharov and the crew at the University of Florida Shand’s Hospital hematopathology laboratory, especially John B. Anderson, Heath I. Bailey, Sandra M. Campbell, Catherine R. Charleston, William D. Dixon, Gabriel R. Luchetta, and Darin J. Ryder.

List of Abbreviations

7-AAD AIDS ALCL ALL(s) AML(s) APC ATLL BC CD cKappa cLambda CLL CLL/PL CML CMMoL CMV CSF CV DI DLCL DNA EBV EDTA ET FCC(s) FCM FITC FNA FRFH FSC GCC(s) G-CSF HCL H&E HIV HTLV IM LCL(s) LGL(s) LL

7-Amino-actinomycin D Acquired immunodeficiency syndrome Anaplastic large cell lymphoma Acute lymphoblastic leukemia(s) Acute myeloid leukemia(s) Allophycocyanin Adult T-cell leukemia–lymphoma Blast crisis Cluster designation Cytoplasmic kappa Cytoplasmic lambda Chronic lymphocytic leukemia Chronic lymphocytic leukemia with increased prolymphocytes Chronic myeloid leukemia Chronic myelomonocytic leukemia Cytomegalovirus Cerebrospinal fluid Coefficient of variation DNA index Diffuse large-cell lymphoma Deoxyribonucleic acid Epstein-Barr virus Ethylenediamine tetraacetic acid Essential thrombocythemia Follicular center cell(s) Flow cytometry Fluorescein isothiocyanate Fine-needle aspiration Florid reactive follicular hyperplasia Forward scatter Germinal center cell(s) Granulocyte colony-stimulating factor Hairy cell leukemia Hematoxylin and eosin Human immunodeficiency virus Human T-cell lymphotropic virus Infectious mononucleosis Large cell lymphoma(s) Large granular lymphocyte(s) Lymphoblastic lymphoma

xvi

LIST OF ABBREVIATIONS

LPC LPD(s) MALT MBCL MCL MDS M:E ratio MF MGUS MM MNC(s) MoAb(s) MPD(s) MPO MRD MUD-BMT MZBL N/C NHL(s) NK NOS NSE PBSCT PCR PE PerCP PLL PMT(s) PRV PTCL(s) RBC(s) SLVL

Lymphoplasmacytoid Lymphoproliferative disorder(s) Mucosa-associated lymphoid tissue Monocytoid B-cell lymphoma Mantle cell lymphoma Myelodysplastic syndrome(s) Myeloid:erythroid ratio Mycosis fungoides Monoclonal gammopathy of undetermined significance Multiple myeloma Mononuclear cell(s) Monoclonal antibody (antibodies) Myeloproliferative disorder(s) Myeloperoxidase Minimal residual disease Matched unrelated donor bone marrow transplant Marginal zone B-cell lymphoma Nuclear/cytoplasmic Non–Hodgkin lymphoma(s) Natural killer Not otherwise specified Nonspecific esterase Peripheral blood stem cell transplant Polymerase chain reaction Phycoerythrin Peridinium chlorophyll protein complex Prolymphocytic leukemia Photomultiplier tube(s) Polycythemia rubra vera Peripheral T-cell lymphoma(s) Red blood cell(s) Splenic lymphoma with villous lymphocytes

List of Case Studies

Case 1 Case 2 Case 3 Case 4 Case 5 Case 6 Case 7 Case 8 Case 9 Case 10 Case 11 Case 12 Case 13 Case 14 Case 15 Case 16 Case 17 Case 18 Case 19 Case 20 Case 21 Case 22 Case 23 Case 24 Case 25 Case 26 Case 27 Case 28 Case 29 Case 30 Case 31 Case 32 Case 33 Case 34 Case 35 Case 36 Case 37 Case 38 Case 39 Case 40 Case 41

Precursor B-ALL, CD15+ AML-M3 AML-M3v AML-M3v, agranular Mature CD4+ T-cell LPD/NHL, CD45– AML, non-M3 with bimodal CD45 AML with minimal maturation, negative for both CD34 and HLA-DR AML with monocytic differentiation AML with monocytic differentiation, maturing monocytic elements present Precursor B-ALL, CD56+ AML with monocytic differentiation, unusual phenotype Precursor B-ALL Precursor B-ALL with phenotypic variation during follow-up Precursor B-ALL, CD10–, and MRD detection Precursor T-ALL, CD3+ and CD8+ Activated CLL/SLL, CD2+ Lymphoplasmacytoid leukemia–lymphoma, CD5+, CD23– Mantle cell lymphoma with bimodal CD5 FCC lymphoma, low-grade, bimodal CD20 FCC lymphoma, bimodal FSC, increased S-phase fraction FCC lymphoma (morphologically FCC III) Hairy cell leukemia, CD10+ High-grade B-cell NHL with plasmablastic differentiation High-grade B-cell lymphoma, CD2+ Large B-cell lymphoma, CD10+ and CD56+ Low-grade T-cell LPD, CD4+ and CD8+ Peripheral T-cell lymphoma Thymoma Adult T-cell leukemia–lymphoma Adult T-cell leukemia–lymphoma (abnormal TCR-Vβ profile) T-prolymphocytic leukemia T-prolymphocytic leukemia (abnormal TCR-Vβ profile) True NK lymphoma–leukemia, aggressive, CD56+ and CD103+ True NK proliferation, indolent, CD16+ and CD56+ T-NK leukemia, CD57+ T-gamma-delta lymphoma–leukemia T-NK leukemia–lymphoma, CD56+ Langerhans cell histiocytosis Florid reactive follicular hyperplasia (FRFH) Florid reactive follicular hyperplasia, high CD4:CD8 ratio FCC lymphoma, low grade, partial involvement

xviii

LIST OF CASE STUDIES

Case 42 Case 43 Case 44 Case 45 Case 46 Case 47 Case 48 Case 49 Case 50 Case 51 Case 52 Case 53 Case 54 Case 55 Case 56 Case 57 Case 58 Case 59 Case 60 Case 61 Case 62 Case 63 Case 64 Case 65 Case 66 Case 67 Case 68 Case 69 Case 70 Case 71 Case 72 Case 73 Case 74 Case 75 Case 76 Case 77 Case 78 Case 79 Case 80 Case 81 Case 82 Case 83 Case 84 Case 85 Case 86 Case 87 Case 88 Case 89 Case 90 Case 91

Normal pediatric bone marrow; and FRFH (tonsil) Precursor B-ALL (relapse) with coexisting B-cell precursors Congenital neutropenia Viral lymphadenitis EBV infection (infectious mononucleosis) HIV infection FRFH and reactive plasmacytosis in HIV infection HIV infection and G-CSF effect T-lymphoblastic leukemia–lymphoma, bimodal CD3 and CD8 Precursor B-ALL AML with erythroid hyperplasia AML-M4 AML-M4E AML-M4E AML with maturation, CD19+ High-grade MDS in a child Hemodilute bone marrow with increased blasts, eosinophils and basophils CML (peripheral basophilia and thrombocytosis) CLL, minimal residual disease, negative morphology Sézary syndrome with follow-up Sézary syndrome (peripheral blood and lymph node) Sézary syndrome T-NK leukemia, indolent, low involvement T-gamma-delta proliferation, indolent Plasma cell neoplasm, CD13+ and CD117+ Plasma cell dyscrasias (two patients) Plasma cell neoplasm, CD20+ (bright) Chronic myelomonocytic leukemia (CMMoL) Hypereosinophilic syndrome Mast cell disease AML and G-CSF effect CML with increased blasts Low-grade MDS, CD56+ granulocytes (two patients) G-CSF effect, severe Low-grade MDS with erythroid hyperplasia Low-grade MDS secondary to lymphoma treatment Coexisting CLL and AML, non-M3 Coexisting CML and CLL Coexisting CLL and large cell lymphoma, consistent with Richter syndrome Biclonal CLL (two patients) Biclonal FCC lymphoma Precursor B-lymphoblastic lymphoma, complex phenotype AML with minimal maturation, coexpressing CD5, CD7, and CD56 AML with megakaryocytic differentiation (AML-M7) AML with megakaryocytic differentiation (AML-M7), CD56+ Multiclonal acute leukemia Acute leukemia with a complex phenotype AML with maturation, CD19+ and CD10+ in a subset of blasts Paraimmunoblastic lymphoma (PLL equivalent) Mantle cell lymphoma, bimodal surface light chain, low S-phase fraction

LIST OF CASE STUDIES

Case 92 Case 93 Case 94 Case 95 Case 96 Case 97 Case 98 Case 99 Case 100

Mantle cell lymphoma with an increased S-phase fraction Large B-cell lymphoma, CD5+ and CD8+ Human herpes virus-8 body cavity-based lymphoma Waldenstrom macroglobulinemia Low-grade B-cell lymphoproliferative disorder, CD5+, CD23– Plasmacytoma Peripheral T-cell lymphoma, CD4+, CD103+ Peripheral T-cell lymphoma, CD30+ (ALK-1+ by immunohistochemistry) Peripheral T-cell lymphoma CD30+, ALK-1+

xix

CHAPTER

1

Approach to flow cytometry: General considerations

The application of flow cytometry (FCM) in diagnostic hematopathology has gained much momentum since its introduction in the late 1970s. FCM immunophenotyping is now an established and necessary laboratory test in the clinical evaluation of any suspected hematologic malignancy. Several factors have contributed to the progress in FCM analysis, including the expanding repertoire and commercial availability of monoclonal antibodies and fluorochromes, as well as technological advances in both the hardware and software of FCM instruments. As a result, three- and four-color FCM studies have become the norm in hematopathology laboratories during this past decade. Because it is a multiparameter analysis, FCM immunophenotyping offers the advantage of efficiency coupled with a high degree of sensitivity. Not only can an extensive array of markers be evaluated by FCM, but the expression of several antigens can also be assessed simultaneously on any given cell population. In contrast, the number of markers that can be used in immunostaining on tissue sections or smears is limited, and fewer cells can be evaluated using routine morphologic criteria. Although dual staining may be achieved by immunohistochemistry in selected situations, the technique is essentially limited to a single antibody per slide. Using FCM, the relative proportion of the population of interest within a sample can be quantitated, and its DNA content can be measured, when appropriate. The high level of sensitivity of FCM also allows for the detection of rare neoplastic cells (based on their specific characteristics) coexisting with other benign subpopulations. The usefulness of FCM immunophenotyping is multifold, as it facilitates (1) the distinction between neoplastic and benign conditions, (2) the diagnosis and characterization of lymphomas and leukemias, (3) the assessment of other neoplastic and preneoplastic disorders such as plasma cell dyscrasias and myelodysplastic syndromes, and (4) the detection of minimal residual disease in patients with acute leukemia or chronic lymphoid malignancies. In some groups of lymphoid neoplasms, FCM study also provides prognostic information. In many institutions, there is a tendency to perform immunological studies only when the lesion is considered difficult to diagnose by conventional morphology. It is preferable, however, not to delay FCM testing but to perform it routinely, even when the morphology is apparently typical, because the findings help to confirm the diagnosis and may provide prognostic or other useful biological information. In addition, the data are valuable for follow-up purposes, especially when samples of tumor recurrences are very small (e.g., cerebrospinal fluid [CSF], needle aspirates) where morphologic examination may fail to detect neoplastic cells. Proper data analysis is a critical step in FCM immunophenotyping. In this process, the phenotypic profile of the cells of interest is derived from the light scatter and fluorescence intensity signals recorded from each individual cell on a cell-by-cell basis (the data thus collected is referred to as list mode data). Although the literature contains numerous publications on the characteristic immunophenotypes associated with different hematologic malignancies, few publications describe how the data analysis was performed.

2

FLOW CYTOMETRY IN HEMATOPATHOLOGY

1.1 Reasons for the necessity of proper data analysis Even today, many laboratories still continue the less than desirable practice of reporting antigenic expression as the percentage of positive cells. In this approach to FCM analysis, (1) the cell population is gated first by light scatter, then the antibody fluorescence analyzed on single parameter histograms or dual parameter plots; (2) for each marker, a cursor is moved and set to measure the fraction of cells with fluorescence greater than that of the control sample (in which cells are exposed to an irrelevant immunoglobulin); (3) the results are then reported as percent positive per antibody tested. The origins of this approach to data reporting can be traced back to the microscopic evaluation of immunostaining performed on glass slides (smears, cytospin preparations) and the reporting techniques used for lymphocyte subset analysis (e.g., CD4 counts in human immunodeficiency virus [HIV]-infected patients). In many laboratories, the larger share of the FCM workload is composed of T- (or other) cell subset determinations on peripheral blood samples that do not harbor malignant cells. In this setting, a change in the number of cells in each subset is clinically important. Furthermore, the cells analyzed are discrete subpopulations of normal lymphoid cells with relatively bright fluorescence. Therefore, it is appropriate to report each subset as a percent positive for each antibody and, where applicable, to calculate the CD4 : CD8 ratio. The numerical values thus generated are reminiscent of those obtained for chemistry tests, in which the abnormalities consist of altered levels of the normal components in the blood. In samples suspected of harboring a hematopoietic malignancy, however, determining the exact number of neoplastic cells is less important than determining whether or not neoplastic cells are present and, if present, the type of hematopoietic neoplasm they represent. Unfortunately, this information is not always apparent from the “percent-positive” data format. The percent-positive format assumes, incorrectly, that within a leukemia or lymphoma, all of the tumor cells uniformly either lack or exhibit the same degree of clear-cut expression for a given antigen. However, in contrast with benign lymphocytes, neoplastic hematopoietic cells of the same clone often do not express the same amount of a given antigen on their cell surface and, therefore, display variability in the fluorescence intensity for that marker. The degree of variability depends on the particular surface antigen. For instance, reporting a case of leukemia as being 40% CD20 positive is ambiguous. This number could represent either (1) a case in which 40% of the cells formed a distinct population with a fluorescence intensity well above the negative control or (2) a single population in which 100% of the cells displayed a shifted fluorescence intensity, but only 40% of the cells were brighter than the background. The latter occurrence is frequently observed when the tumor cell expression for a surface antigen is weak.

1.1.1 The pitfalls of the FCM data format of “percent positive” per antibody tested In the context of a leukemia–lymphoma workup, it is important to express the immunophenotyping data in ways that avoid ambiguity and offer the optimal information for correlation with other clinical and laboratory data. Expressing the FCM data as “percent positive” per antibody tested is rarely relevant and may even be misleading. Many institutions have used the 20% level as an arbitrary cutoff value for a marker to be considered positive. None of the publications has described how this number became established, however. The following reallife examples (obtained in 1997 in London, UK) illustrate why the approach of reporting FCM results as percent positive and omitting the fluorescence data is inappropriate in leukemialymphoma immunophenotyping, and can lead to erroneous interpretations.

APPROACH TO FLOW CYTOMETRY

3

Flow cytometry results on a bone marrow from a patient with suspected chronic myeloid leukemia in blast crisis (CML-BC), from an institution where the FCM laboratory is not part of the hematopathology laboratory, are shown below. The specimen was processed by FicollHypaque. Other procedure-related information was not made available. CD19 CD20 CD10 CD13 CD33

31% 20% 29% 32% 26%

CD34 30% CD14 7% Kappa 8% Lambda 10% HLA-DR 29%

CD2 CD3 CD7 CD5

11% 13% 16% 19%

Based on this format of data reporting and the 20% threshold, the case was interpreted as a biphenotypic blast crisis of CML (positive for CD19, CD10, CD13, CD33). However, when the list mode data was visualized on dual parameter dot plots, correlating the forward scatter (FSC) and antibody fluorescence, it became clear that (1) the neoplastic cells constituted 30% of the cell population in the FCM sample and (2) they were of medium cell size and had the following phenotype: CD19 moderate, CD20 dim, CD10 moderate, CD34 weak, HLA-DR moderate. Other antigens, (i.e., CD13, CD33, CD14, CD2, CD3, CD5, CD7, kappa, and lambda) were not expressed by the tumor cells. The CD13 (32%) and CD33 (26%) were present on mononuclear myeloid precursors (promyelocytes, myelocytes, and metamyelocytes) and not on the neoplastic population. The correct phenotype is that of a precursor B-cell acute lymphoblastic leukemia (ALL) and not biphenotypic leukemia. Correlation with the bone marrow aspirate morphology further confirmed a lymphoid blast crisis of CML. A limited (follow-up) panel was performed on the peripheral blood of a patient with known chronic lymphocytic leukemia (CLL), to assess the efficacy of anti-CD20 therapy as part of a clinical trial. The lymphocyte count was 3.1 × 107/L. The blood film was unremarkable except for a mild increase in large granular lymphocytes. The FCM data were reported as follows: CD2 75% CD3 62%

CD19 23% CD20 18%

Kappa 17% Lambda 8%

Based on these results, it was concluded that there was no residual CLL in the patient’s peripheral blood, especially as the kappa:lambda ratio was within the normal range. However, subsequent reevaluation of the list mode data, using simple correlated displays of FSC and antibody fluorescence, was sufficient to demonstrate the presence of a small population of monoclonal B-cells (CD19 moderate, CD20 weak) with weak kappa expression, in a background of benign T-cells and polyclonal B-cells. Contrary to the initial interpretation, residual CLL was present in the patient’s peripheral blood. It is apparent from the above examples that reporting FCM data as percent positive per antibody tested can negate the usefulness of FCM and easily lead to confusing or erroneous interpretations, which may impact therapeutic decisions. Some laboratories do include fluorescence data in the FCM reports. However, the data may still be expressed in a suboptimal (and, therefore, inappropriate) manner, as shown in the following case example. Below are the FCM results on a peripheral blood specimen studied (in the mid-1990s) at a teaching hospital:

4

FLOW CYTOMETRY IN HEMATOPATHOLOGY CD2 CD3 CD4 CD7 CD8 CD13 CD33 CD34 TdT

48% moderate 45% moderate 21% moderate 47% moderate 20% moderate 3% moderate 1% moderate 1% weak 55% moderate

CD19 CD20 CD22 sIgM Kappa Lambda CD10 CD45 HLA-DR

47% moderate 26% moderate 47% moderate 48% moderate 3% moderate 2% moderate 36% moderate 100% strong 55% moderate

The results indicated a proliferation of immature cells (TdT+). The case was interpreted as ALL with a mixed (B-cell and T-cell) lineage. Because of the data-reporting format, it is unclear whether the immature cells are of B- or T-cell lineage, however. Although fluorescence intensities were mentioned, data interpretation in this particular laboratory was actually based on percent positive with an arbitrary 20% cutoff. When proper visual data analysis was subsequently applied to the raw data, it became apparent that the blood sample contained a clearly identifiable neoplastic population of precursor B-ALL, admixed with a high number of normal T-cells.

1.2 General aspects of FCM data analysis and interpretation The above-described situations indicate the necessity of a comprehensive approach to FCM data analysis and interpretation. In the authors’ experience, the optimal method is for the laboratory medical staff to apply a visual approach to FCM data analysis rather than relying on percentages. In other words, data interpretation is based on a visual appraisal of the FCM graphics, assessing the complex patterns formed by the shape and relative position of the cell clusters observed on various dot plots such as FSC versus fluorescence, side scatter (SSC) versus CD45, and correlated fluorescence dot plots. Any other approach to FCM data interpretation, using a scoring system or percent positive per antibody, underutilizes the full potential of FCM. Laboratory professionals, as well as clinicians, should realize that visual FCM data analysis is a process reminiscent of the microscopic examination of morphologic material (e.g., bone marrow aspirate smears, lymph node sections) in which the data form a pattern and are reported in a qualitative and quantitative (where appropriate) format. Although microscopic examination encompasses all elements in the sample, reporting the data focuses only on the abnormal component. Similarly, the FCM interpretative report should be based on the cells of interest, even though the list mode data should be collected unselected (i.e., it includes all cells in the sample). Collecting list mode data ungated ensures that no abnormal cells are lost, because in many instances, the nature of the abnormal population is not yet known at the time the specimen is run. Restricting the initial data collection to certain preset criteria (i.e., a “live gating” approach such as the use of a live light scatter gate) may easily result in “throwing the critical cells away.” A specific example is missing a small number of circulating hairy cells when the analysis is live-gated on cells with the light scatter characteristics of normal lymphocytes. An additional advantage of the ungated approach is that the presence of other cells serves as internal positive and negative controls. After the data have been acquired ungated, certain gating procedures can be applied during the analysis step. Some of the most useful gating strategies include (1) gating on B-cells, to determine clonality (Figure 1.1) and the coexpression of other critical antigens, and (2) gating

APPROACH TO FLOW CYTOMETRY

a

b

c

d

5

Figure 1.1 (a) Lymph node sample with two B-cell populations differing in FSC signals and CD20 intensities. (b) Overlay kappa/lambda histograms gated on R2: The B-cells with dimmer CD20 and lower FSC are polyclonal. (c, d) Gated on R3: The B-cells with brighter CD20 and higher FSC are monoclonal for lambda.

on CD45 to characterize leukemic blasts (Figure 1.2). These strategies require the use of multicolor (two-color, at the very least) antibody combinations. Following the recommendations by the U.S.–Canadian Consensus on the Use of Flow Cytometry Immunophenotyping in Leukemia and Lymphoma in 1997, the multiparameter approach to FCM testing has become a standard routine in clinical laboratories, taking advantage of more sophisticated instrumentation and a larger repertoire of fluorochromes. Since the publication of the consensus recommendations, there has been an increase in the awareness of the visual approach to FCM data analysis. The literature contains very little information on this approach, however. The purpose of the first edition of this book was to fill this void. More recent developments in the field are added in this second edition. The FCM dot plots and histograms displayed in this book, using FCS Express software, are derived from clinical samples analyzed primarily on Becton-Dickinson instruments, using commercially available antibody reagents (see Chapter 2). Other current state-of-the-art instruments are equipped with a similar capability for multicolor FCM testing and mechanisms for

6

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 1.2 (a) Peripheral blood sample with a distinct cluster (R2) in the blast region. (b–d) Gated on R2: Blasts are positive for CD13 and CD33. CD19 is negative. CD34 is expressed with a bimodal distribution.

color compensation. The principles of FCM data analysis presented in this book are applicable to all brands of flow cytometers. Interpretation of the FCM immunophenotyping results is one step in the diagnosis of malignant lymphoma and leukemia. Although, in many cases the diagnosis is apparent after a visual inspection of the FCM immunophenotyping data together with the DNA cell cycle histogram, in other instances the antigenic profile and the pattern of the cell clusters suggest only a differential diagnosis instead of a specific disorder. In such cases, it is critical that the diagnostic interpretation takes into account the other clinical and laboratory data, such as the hemogram findings and the cytologic/morphologic features. The synthesis of the pertinent results requires the responsible medical staff in the laboratory to be well versed in the different subdisciplines of hematopathology. Irrespective of whether a case is straightforward or complex, the authors advocate a routine systematic approach to FCM diagnostic interpretation. This will ensure that no relevant information is omitted. A correlation between the FCM findings and the available morphologic data should be performed in all cases. Wright-Giemsa-stained cytospins made from the cell suspension of the

APPROACH TO FLOW CYTOMETRY

7

tissue or fluid submitted for FCM study must be reviewed, to correlate the findings with those derived from the FCM plots. This is especially helpful when abnormal (neoplastic) cells are few or the FCM data cannot be clearly interpreted. For peripheral blood specimens, the FCM data are correlated with the hemogram and cytologic features from a fresh blood film. Similarly, FCM interpretation on bone marrow specimens should include a review of the hemogram, peripheral blood film, bone marrow aspirate smear or imprint, and cytochemistries, where appropriate. It cannot be emphasized enough that hemogram findings, along with fresh peripheral blood and bone marrow smears, must accompany the specimen when bone marrow is sent to a referral laboratory for immunophenotyping, so that a proper, thorough diagnostic evaluation of the case can be conducted. For interpretation on solid tissue (e.g., lymph nodes), the FCM data are correlated with the morphologic features on the imprints and hematoxylin and eosin (H&E) sections (where available). In addition to the above-mentioned minimum correlation with the morphologic findings, it is also important to review immunoelectrophoresis results in suspected lymphoplasmacytoid neoplasms or plasma cell dyscrasias. Knowledge of the pertinent clinical history, especially the type of therapy (e.g., immunotherapy, growth factors), is also useful to further refine FCM diagnostic interpretation. This necessitates a dialogue between the medical staff caring for the patient and the FCM laboratory. Because of the time delay associated with molecular genotyping and cytogenetics, these techniques play a minimal role at the time of rendering the diagnosis. Correlation of those results with the FCM data is useful, however, for confirming the diagnosis or providing additional information.

1.3 Other applications of FCM in hematopathology In addition to being a diagnostic tool, FCM analysis has also been used for prognostic purposes. The main caveat, when determining the prognostic significance of biological parameters of neoplastic cells, is that the validity of the results is influenced by various factors such as laboratory methodologies, clinical staging procedures, and therapeutic protocols. Despite such drawbacks, studies have shown that the DNA index may be of prognostic significance in childhood ALL and the S-phase fraction is useful in grading a lymphoproliferative disorder/nonHodgkin lymphoma (LPD/NHL). The current classification of LPD/NHL according to the WHO scheme is partly based on cellular ontogeny and differentiation rather than on the biological behavior of the tumor. The choice of therapeutic regimens in lymphomas is still based on the grade of the tumor, however, because cell cycle-dependent drugs continue to feature prominently in the arsenal of chemotherapy for LPD/NHL. Therefore, it would be helpful that FCM testing on lymphomas includes DNA and cell cycle analysis, as the S-fraction gives an indication as to whether the tumor has a high growth-fraction (i.e., aggressive, high grade) or a low growth-fraction (i.e., indolent, low grade), which in turn influences the patient’s response to therapy and survival. Determination of the S-fraction by multiparameter DNA analysis is preferable to paraffin-based immunostaining for Ki67 or PCNA (proliferating cell nuclear antigen), where it is not always possible to distinguish proliferating lymphoma cells from the intimately admixed proliferating benign T-cells. The value of DNA ploidy in LPD/NHL as a prognostic factor still remains controversial. The presence of DNA aneuploidy is helpful, however, to identify those cases of suspected peripheral T-cell malignancies in which the hematologic and immunophenotyping data reveal no abnormalities. In Sézary syndrome, DNA ploidy by FCM has been shown recently to be useful for diagnosis and minimal residual disease (MRD) monitoring. Furthermore, the presence of aneuploidy is associated with increased numbers of large cells in the involved tissues.

8

FLOW CYTOMETRY IN HEMATOPATHOLOGY

In addition to cell cycle analysis, there have also been attempts to correlate certain antigenic features with the patient’s response to therapy or survival. While there is only limited evidence that the expression of any particular antigen could serve as a reliable predictor of prognosis, it appears that the intensity of CD45 expression affects the outcome in pediatric ALL, and CD38 positivity in CLL is associated with an unfavorable clinical course. More recently, the advent of gene expression profiling by DNA microarray analysis has led to the identification of genes that may discriminate the subtypes of CLL. Most notably, the differential expression of the Zap-70 gene has been shown to correlate with the mutational status of immunoglobulin heavy-chain variable-region (IgVH) genes, the latter being an important prognostic factor in CLL. Because Zap-70 protein can be readily determined by FCM, in contrast with the more labor intensive and costly DNA sequencing analysis for IgVH mutations, Zap-70 is currently viewed as the best surrogate marker of IgVH mutational status in CLL. According to recent studies, the expression of Zap-70 protein identifies a subgroup of CLL with a more rapidly progressive clinical course and poorer outcome. Note that testing for Zap-70 is limited to peripheral blood samples only, however. An important application of FCM analysis is MRD monitoring. With the emergence of new treatment protocols combining high-dose chemotherapy, autologous stem cell transplants and immunotherapy, MRD detection is becoming necessary in clinical laboratories. The applications of MRD monitoring are multifold, depending on the particular hematologic malignancy. For instance, in childhood ALL, the patient’s MRD status during the initial phase of therapy is a powerful prognostic indicator based on which patients can be stratified into different risk groups. Routine monitoring of MRD during clinical remission in acute leukemias also facilitates early therapeutic intervention, so as to reduce the morbidity/mortality associated with overt clinical relapse. Furthermore, MRD assay is a helpful tool for assessing tumor response to new treatment modalities, such as humanized anti-CD33 conjugated to calicheamicin (gemtuzumab ozogamicin) for AML, and the tyrosine kinase inhibitor, imatinib, in Ph1-CML. The use of MRD detection in mature lymphoid malignancies differs from that in acute leukemia, in that it is usually not applicable at the time of front-line treatment because most of the patients, especially those with low-grade disease, still harbor MRD while in “complete” clinical remission after conventional therapy. MRD detection is therefore applied mainly to patients who, because of relapse, receive high-dose chemotherapy with stem cell transplant, and/or immunotherapy such as anti-CD20 (rituximab) or, in the case of CLL, anti-CD52 (alemtuzumab). FCM was thought not to be as sensitive a technique for MRD detection as polymerase chain reaction (PCR)-based methodologies. However, this apparent lack of sensitivity is most likely due to the fact that the number of cells acquired in a standard FCM clinical assay is far less than that used in PCR analysis. Studies have shown that one leukemic cell in 104 to 105 bone marrow mononuclear cells can be detected by FCM when a large number of cells are analyzed, thus achieving a sensitivity level comparable with that of molecular analysis. These two techniques complement each other and are best applied in tandem to reduce any potential falsenegative results. The FCM approach has the advantage of being less labor intensive and achieving a faster turnaround time. Furthermore, the ability of FCM to separate viable from dying cells permits a more accurate quantitation of MRD levels. Irrespective of the methodology, it appears that the clinically significant MRD level is 0.01% (i.e., 10–4). The presence of residual leukemic cells above this level at the end of therapy or an increase in MRD levels in consecutive bone marrow samples during clinical remission has been shown to be associated with a higher risk of relapse and a poorer overall survival, and it tends to correlate with adverse cytogenetic abnormalities. MRD studies have not been feasible in all patients with acute leukemia, however. Oligoclonality, clonal evolution, lack of specific leukemia sequences or absence of nonrandom

APPROACH TO FLOW CYTOMETRY

9

genetic abnormalities are some of the limiting factors to PCR-based MRD detection. Similarly, monitoring MRD status by FCM analysis can only be achieved if the leukemic blasts exhibit specific antigenic features differentiating them from their normal counterparts. In precursor B-ALL, it has been shown recently by DNA microarray analysis that a significant number of genes are overexpressed in leukemic cells in comparison with normal B-cell progenitors in the bone marrow. Of the several proteins encoded by the overexpressed genes, CD58 has become the marker of choice for MRD monitoring by FCM analysis because the protein is consistently overexpressed in a large number of patients with precursor B-ALL, and fluorochrome-conjugated anti-CD58 antibodies are commercially readily available.

1.4 Maturation and differentiation of hematopoietic elements: An overview based on the immunologic markers currently in use in the FCM laboratory Adequate immunophenotyping of hematologic malignancies requires a large battery of cellular markers. These markers can be broadly categorized into the following groups: B-cell, T-cell, natural killer (NK) cell, myeloid/monocytic, erythroid, megakaryocytic, and non-lineage-associated markers (including activation markers such as CD38 and HLA-DR). The establishment of these markers was derived from studies on the differentiation and maturation of hematopoietic cells. Noncommitted hematopoietic stem cells express CD34. There is considerable heterogeneity within the population of CD34+ progenitors, however, and those in the later stages also express HLA-DR and CD38. Another marker of immature cells is TdT, present predominantly although not exclusively, among lymphoid precursors. The earliest B-cell precursors can be identified by cCD22, the hallmark of the B-cell lineage. Cytoplasmic CD22 is present even before any detectable rearrangement of the immunoglobulin (Ig) genes. Also found on early B-cell precursors are CD19 and CD10. The subsequent maturation and differentiation of the B-cell precursors in the bone marrow is characterized by a gradual decrease in CD10 together with a gradual gain of CD20. Cytoplasmic µ chain can be detected in the late stages of B-cell precursors. The appearance of surface Ig expression, along with the disappearance of immature markers (TdT, CD34), defines a mature B-cell. Mature naïve (resting) B-cells leave the bone marrow to enter the circulation and are characterized by the presence of well-expressed surface IgM, IgD and either kappa or lambda, along with CD20 and CD22. The further differentiation into different subtypes of mature B-cells (expressing IgA or IgG) and plasma cells occurs mainly in peripheral lymphoid tissues, where the cells home into different microenvironments depending on their stage of maturation. The identifier of a T-cell lineage is cCD3, which appears in the earliest T-cell precursors in the thymus prior to rearrangements of the T-cell receptor (TCR) genes. Pre-T-cells are derived from a common T/NK progenitor coexpressing CD34, CD2 and CD5. Thymic maturation is characterized by acquisition of CD1a, and somatic recombination of the TCR genes to produce a diversity of TCR required for the recognition of a large variety of peptide antigens presented on major histocompatibility complex (MHC) molecules. The TCR molecule is part of a larger signaling complex that also includes the costimulatory molecule CD4 or CD8, and the signal transduction module formed by the various subunits of CD3. Until recently, studies of the TCR have been based on molecular techniques, either by Southern blot analysis of the β chain genes or PCR assay of the γ chain genes. The recent availability of a large panel of antibodies identifying 70% of the different variable (V) regions of the β chain has made it possible to study the TCR-Vβ repertoire by FCM, and thereby detect clonal expansions of abnormal T-cells. The TCR molecule is a heterodimer composed of either α and β chains, or γ and δ chains. The structure of each chain is similar to that of Ig, consisting of constant and variable regions.

10

FLOW CYTOMETRY IN HEMATOPATHOLOGY

There are at least 65 Vβ segments on the β-chain gene. The TCR genes, located on chromosome 7 (β- and γ-chain genes) and chromosome 14 (α-chain gene, and δ-chain gene within the α-chain cluster) are rearranged in an orderly fashion. TCR-δ rearrangement occurs first as thymocytes progress from the CD34+ CD1a– to the CD34+ CD1a+ stage. As these cells rearrange their γ genes, CD4 is upregulated producing CD4 (dim) single positive thymocytes. TCR-β gene rearrangements become detectable as the pre-T-cells become double positive for CD4 and CD8. The TCR-α gene is rearranged at the very late stage of pre-T-cell development. The prevailing thinking is that if γ and δ are productively rearranged first, then the T-cell precursor will become a γδ T-cell. Conversely, successful rearrangement of the β-gene will most likely be followed by recombination of the α-gene to produce an αβ T-cell. The subsequent maturation involves three processes: (1) precursor T-cells are exposed to thymic epithelial cells expressing either MHC class I or MHC class II molecules; (2) they undergo the process of negative and positive selection; and, (3) differentiate into either cytotoxic or helper T-cells, respectively, prior to entering the circulation. The selection process helps to prevent autoimmunity and to ensure that the mature T-cells are capable of recognizing foreign peptides. Only a very small number of mature T-cells remain double positive for CD4 and CD8. Immunological recognition of a mature T-cell is based on the absence of TdT expression or other early thymic markers (e.g., CD1) and the overt expression of surface CD3. The great majority of mature T-cells are TCR-αβ positive, coexpressing CD2, CD5, CD7 and either CD4 or CD8. A minor population of T-cells express γδ TCR instead, accounting for about 3% to 5% of CD3+ T-cells in the blood. The γδ T-cells emigrate from the fetal thymus to reside in cutaneous and mucosal (especially gastrointestinal) sites. Current knowledge about the function of γδ T-cells still remains limited. The majority of γδ T-cells are CD4- and CD8-negative; approximately one-third are CD8-positive. Cells of the myeloid lineage can be identified by the expression of CD13, CD33, and CD117. The maturation process of myeloblasts into promyelocytes is accompanied by a loss of CD34 and HLA-DR, followed by the expression of CD15, CD11b, and CD16 at the myelocyte and metamyelocyte stages. Although CD13 and CD33 are expressed at all stages of granulocytic differentiation, the level of CD33 gradually decreases as the cells mature, whereas CD13 exhibits a bimodal distribution, being more brightly expressed on blasts and neutrophils, but less intensely on intermediate myeloid precursors. Intense CD14 is a characteristic of mature monocytes. An additional identifying feature of monocytic lineage is the expression of other myeloid markers (CD33, CD64), which differ in intensity and distribution from that on myeloid cells. Erythroid precursors are characterized by downregulated CD45 and high expression of CD71 (transferrin receptor). CD71 is also present on cells of other lineages, however its levels are highest in erythroid cells, presumably because of the large amount of iron required for hemoglobin synthesis. According to studies performed on erythroid colonies, the levels of CD71 peak in the earliest erythroid precursors (prior to the proerythroblast stage), and decrease as the nucleated erythroid cells mature into reticulocytes. Erythroid maturation is accompanied by further loss of CD45 antigen and acquisition of glycophorin A (Gly-A), that is, CD235a. The levels of Gly-A become highest at the basophilic erythroblast stage, and remain unchanged from there onward. The maturation of reticulocytes into red blood cells (RBCs) is marked by a concomitant loss of reticulum and CD71 antigen. The evaluation of CD55 and CD59 is performed infrequently in the hematopathology laboratory, only when paroxysmal nocturnal hemoglobinuria (PNH) is suspected. These antigens are involved in the protection of cells by inhibiting the formation of membrane-attack complex. Decrease or loss of CD55 and CD59 in RBCs is characteristic of PNH. Platelets and megakaryocytes can be identified by the presence of glycoprotein (Gp) IIb/IIIa, a heterodimer formed by integrin α2β (CD41) and integrin β3 (CD61). In addition, CD36,

APPROACH TO FLOW CYTOMETRY

11

known as GpIV and GpIIIb, is expressed. This antigen is also found on erythrocytes and monocytes, however. It is important to be aware that some markers, although considered to be associated with a specific lineage, may be expressed by neoplastic cells of a different lineage. For example, CD15 is typically associated with myeloid processes, but it can also be present in a substantial number of T-ALL and precursor B-ALL, as well as monocytic leukemias. Similarly CD7, a T-cell-associated antigen, is often present in acute myeloid leukemia (AML). Such instances have given rise to the concept of aberrant immunophenotypes. The use of the label “lineage infidelity” warrants some caution, however, because not all lineage-associated antigens share the same degree of specificity. For example, antigens such as CD11b, CD16, and CD15 do not constitute sufficient evidence of myeloid differentiation. Furthermore, the concept of aberrancy is based on our current understanding of the normal sequence of phenotypic development, which may still be incomplete and evolving. The great majority of currently known hematopoietic antigens are not lineage associated. For instance, CD10 is not only found on precursor B-cells and germinal center cells, but also on neutrophils, and may be expressed in some cases of precursor T-cell lymphoma/leukemia. Some of the non-lineage-associated antigens are activation markers that may be newly expressed on B- or T-cells after antigen stimulation and cellular activation. Common activation markers include HLA-DR, CD25, CD30, and CD38. Some non-lineage-associated antigens are adhesion molecules, which serve to facilitate cell-to-cell interaction. For instance, CD58, also known as lymphocyte function-associated antigen-3 (LFA-3) is an adhesion molecule widely expressed on hematopoietic and non-hematopoietic cells. The interaction of CD58 with CD2 (its only known ligand) mediates T-cell activation and cytokine production, as well as cytolytic activity in T-cells and NK cells. The value of CD58 in diagnostic FCM is, however, in the distinction between precursor B-ALL and benign B-cell progenitors in the bone marrow. Another example of a non-lineage-associated marker routinely used in FCM immunophenotyping is CD56 (a neural cell adhesion molecule). Although CD56 is considered to be associated with the NK lineage, it is present in several other types of disorders including plasma cell tumors, small-cell neuroendocrine tumors, and a significant number of AMLs. Similarly, CD57, though expressed by most NK cells and a small subset of T-cells (also referred to as memory cytotoxic T-cells), is also present in many non-hematopoietic tissues or neoplasms. True NK cells and NK-like T-cells can be putatively identified based on the combined pattern of expression of one or more “NK markers” (CD16, CD56, or CD57) together with some of the T-cell-associated antigens. More recently, the study of NK cells and NK-like T-cells has been further facilitated by the commercial availability of antibodies against NK receptors. NK receptors are membrane proteins involved in the recognition of MHC class I molecules (HLA-A, -B, -C, and -G) or their homologues. Functionally, NK receptors are either activating or inhibitory. The latter is involved in the recognition of class I MHC molecules with self-antigens, thus protecting healthy cells from NK cytotoxicity. Conversely, no inhibitory signal is generated in NK cells when the expression of MHC class I molecules is altered, which then leads to NK lysis of the target cells. Lysis is also mediated by the binding of activating NK receptors with specific class I ligands on the viral-infected cells or tumor cells. Structurally, NK receptors are subdivided into two broad categories: the killer cell Ig-like receptors (KIR) encoded by a family of 14 polymorphic genes on chromosome 19, and the C-lectin-like receptors encoded by a family of six conserved genes on chromosome 12. The lectin-like NK receptors are composed of either heterodimers such as the CD94:NKG2 family of receptors specific for HLA-E, or homodimers such as CD161, which recognize MHC class I-like products. In contrast with the lectin-like receptors, KIR receptors are much more

12

FLOW CYTOMETRY IN HEMATOPATHOLOGY

complex structurally because they are involved in the recognition of the polymorphic determinants on MHC class I molecules. The KIR receptors are further subdivided into subfamilies according to the number of their Ig-like domains (two vs. three) and the length of their cytoplasmic tails (long vs. short). Each subfamily consists of one or more receptors. Receptors with long cytoplasmic tails have inhibitory function. KIR antigens are highly polymorphic. Their expression in NK cells is complex in that individual cell clones may express several members of the KIR family, and a given KIR receptor (for instance, CD158e) may be present only on a particular subset of NK cells. In a normal individual, the pattern of the usage of the KIR repertoire by NK cells remains stable over time. Currently, only a very small number of KIR antibodies (CD158a, CD158b, and CD158e) are commercially available, which does not yet permit adequate evaluation of the large KIR repertoire.

CHAPTER

2

FCM immunophenotyping and DNA analysis: Practical aspects that can affect data analysis and interpretation

In the optimal setting, the FCM lymphoma–leukemia immunophenotyping laboratory is an integral component of the diagnostic hematopathology service. Flow cytometric analysis involves three stages: preanalytical (specimen handling and processing, including antibody staining), analytical (running the sample through the flow cytometer and acquiring data), and postanalytical (data analysis and interpretation). The quality and performance of the preanalytical and analytical steps impact on the resulting fluorescence data and thereby the interpretation. Deficiencies such as suboptimal instrument performance, poor reagent quality (antibodies and/or fluorochromes), or poor specimen quality can all result in inadequate resolution of positive and negative immunofluorescence. Integrating the FCM and hematopathology laboratories facilitates both the preanalytical steps (because the specimen can be processed simultaneously for other related technologies) and the postanalytical steps (during which the FCM results are correlated with other data prior to establishing a diagnosis). The methodologies used in the authors’ laboratories follow closely the recommendations of the 1997 U.S.–Canadian Consensus on the Use of Flow Cytometry Immunophenotyping in Leukemia and Lymphoma. The general aspects of the methodologies for the preanalytical and analytical steps are presented in this chapter. Discussions on quality control are not included, however, as they have been presented at length in previous textbooks and manuals.

2.1 Sample selection The laboratory has little control over certain factors, such as specimen collection and transportation, which can adversely affect the sample prior to its arrival. Although rigorous quality control applied to the various intralaboratory procedures can ensure the accuracy and reproducibility of the FCM results, poor specimen collection remains a major source of potential unsatisfactory FCM analysis. The time elapsed between specimen acquisition and delivery to the laboratory, and the environmental conditions during transport are critical factors affecting the viability of the cells in the sample. As a rule, specimens cannot be held for more than 48 hours in the fresh state after collection. This time window is much narrower for samples harboring a tumor with a high turnover rate (e.g., Burkitt lymphoma). For these reasons, specimen requirements and acceptance guidelines should be thoroughly communicated to the clinical services as well as to referring institutions. Exposure to extreme temperatures and the presence of blood clots (or gross hemolysis) are conditions that can render a blood or bone marrow specimen unacceptable for analysis. Fresh specimens for FCM processing and analysis fall into two broad categories: liquid samples (peripheral blood, bone marrow, body fluids) and solid tissue (lymph nodes, tonsils/ adenoids, spleen, bone marrow biopsies, and extranodal infiltrates). The size and cellularity of the sample, as well as the viability of the cells therein, are the main factors determining the final cell yield available for FCM study. Lower cell yield limits the number of markers that

14

FLOW CYTOMETRY IN HEMATOPATHOLOGY

can be analyzed. In such instances, knowledge of the clinical setting and prior FCM results are critical in choosing the appropriate markers to be tested.

2.1.1 Liquid specimens Peripheral blood can be collected in either ethylenediamine tetraacetic acid (EDTA) or heparin. Collection in EDTA is preferred, however, because a hemogram and a blood smear can be obtained from the same sample. The volume of blood required depends on the white blood cell (WBC) count; 10 mL of blood is adequate in most instances. The blood specimen is preferably maintained at room temperature. Referred blood specimens from outside institutions should be accompanied by a hemogram and a fresh blood smear (i.e., free of storage artifacts), either unstained or stained with Wright-Giemsa, for cytologic evaluation. For quality control, however, the FCM laboratory should also make a smear from the FCM blood sample. Similarly, EDTA is the anticoagulant of choice for bone marrow specimens sent to the FCM–hematopathology laboratory. Bone marrow smears, cytochemistries where appropriate, and FCM studies can all be performed from the same tube. Ideally, the hematopathology staff should be personally involved with the collection of the bone marrow specimen to ensure that the sample obtained is adequate enough to cover all the desired tests (e.g., FCM, cytogenetics, morphology). Two of the authors routinely performed the bone marrow procedure themselves. Using the previously described recommended approach (Nguyen and Diamond, Diagnostic Hematology: A Pattern Approach), an ample quantity of marrow aspirate and admixed blood is collected in EDTA tubes. The marrow spicules are allowed to rise to the top, from where they can easily be harvested virtually free of blood contamination and allocated in appropriate amounts for FCM, morphologic smears, and any other necessary studies. This optimal approach eliminates one of the most frequent problems encountered in the laboratory, the marked discrepancy between the cellular bone marrow smears made at the bedside and those made from the aspicular, severely hemodilute marrow sample submitted for FCM analysis. Referred bone marrow aspirate samples received from outside institutions should be accompanied by a fresh bone marrow smear containing an adequate number of spicules. Preferably, a fresh blood smear and hemogram should also be included, so that a complete diagnostic evaluation of the bone marrow can be carried out properly. Approximately 3 to 5 mL of representative marrow aspirate is usually sufficient for a comprehensive FCM analysis, unless the marrow is severely hypocellular. Because bone marrow aspirates have a much higher cell density than peripheral blood specimens, degenerative changes tend to occur more quickly. Refrigeration and the addition of nutrient media containing serum proteins to the aspirate will help to maintain cell viability in bone marrow samples that cannot be processed soon after collection. Once received in the laboratory, the bone marrow sample is poured out onto a Petri dish to check if spicules are present. A small portion of the spicules is taken to prepare an “in-house” marrow smear for quality control. Two scalpels are applied to mince the remaining spicules to release cellular elements (especially neoplastic lymphoid cells and plasma cells) that tend to adhere to the spicules. In the case of an aspicular aspirate (“dry tap”), the FCM marrow sample should consist of at least two 2-cm-long core biopsies submitted in sterile tissue culture media, preferably RPMI supplemented with fetal calf serum and a mixture of antibiotics. Cell suspensions can be obtained from the biopsies by the same mechanical dissociation procedure applied to solid tissue. In body cavity effusions, it is important to collect samples with good cell viability. This can be achieved by draining off the existing effusion, and later obtaining the reaccumulated “fresh” effusion.

FCM IMMUNOPHENOTYPING AND DNA ANALYSIS

15

For deep-seated lesions (e.g., mediastinal or retroperitoneal), the preferred specimen for FCM analysis is a fine-needle aspiration rather than a core biopsy. Multiple passes (at least 3 to 4), performed by an experienced person, usually provide a higher cell yield than a tissue core biopsy, and minimize sampling error. Body fluids such as CSF, vitreous humor or pericardial fluid carry a scant number of cells and do not require much initial processing. A cell count can be obtained and cytospins prepared directly from the submitted sample.

2.1.2 Solid tissue specimens Solid tissue specimens (e.g., lymph nodes, spleen, tonsil/adenoids, or extranodal sites) should be submitted as thin slices, less than 2 mm thick, in a generous amount of sterile tissue culture media at 4ºC (i.e., on ice). This helps to reduce the rate of autolysis and degradation of cellular proteins and DNA. For in-house cases, where specimens are delivered immediately to the laboratory, it is only required to keep the tissue moistened with a culture media- (or saline-) soaked gauze. There are no set rules for the required amount of sample because this is dependent on several factors, including the cellularity in the sample, the fragility of the cells, and the susceptibility to apoptosis. More is always better, especially in the case of extranodal specimens where there may be a significant proportion of nonlymphoid tissue or fibrocollagenous stroma. It is important, however, to submit a generous amount of fresh solid tissue to the FCM–hematopathology laboratory, so that an adequate amount of sample may be allocated to various other procedures in addition to the preparation of cell suspensions for FCM analysis. These procedures, potentially necessary for the complete characterization of a particular lymphoid tumor, include the following: • Snap-freezing for immunohistochemistry or molecular genetics. • Fixation in a 1 : 1 mixture of RPMI and ethanol for molecular genetics (optional). • Wright-Giemsa-stained air-dried touch imprints for cytologic evaluation. The cut surface of the tissue slice should be blotted to remove excess fluid prior to making imprints. The imprints would be otherwise unreadable as slow drying produces severe shrinkage artifacts on the cellular elements. • Fixation for histology, preferably with B-5 or a fixative with a heavy metal component (e.g., barium chloride) for morphologic correlation. For referral cases from an outside pathology laboratory, a small fraction of the sample submitted to the FCM laboratory can be used for histology (if possible). Efforts should be made to obtain H&E sections from the referring institutions, so as to achieve immunophenotypic-morphologic correlation in the FCM hematopathology reports.

To ensure that the sample is representative, the slices sent to the FCM laboratory should be adjacent to those submitted for routine histology (Figure 2.1). A practice to be avoided is that of submitting the tip of the lymph node for FCM while allocating the central portion for routine histology. It is also not advisable to hold the fresh tissue for FCM analysis until after the histologic sections are ready, because the time delay often adversely affects the viability of the sample.

Figure 2.1 Diagram of lymph node slicing and the allocation of the slices to different studies. F, FCM analysis; H, histology; I, immunohistochemistry; M, molecular studies. Each slice is less than 2 mm thick.

16

FLOW CYTOMETRY IN HEMATOPATHOLOGY

Cell suspensions are obtained from the solid tissue by mechanical dissociation whereby the tissue is minced with two scalpels in a Petri dish containing a small volume of culture media, and then passed through a fine-wire-mesh screen. Alternative techniques, such as repeated aspiration of the tissue using an 18-gauge needle or scraping the cut surface of the tissue section with a scalpel blade or glass slide at a 45-degree angle, can also be employed.

2.2 Preparing nucleated cell suspensions Separating nucleated cells from red blood cells in liquid specimens is achieved by red cell lysis (using ammonium chloride or other solutions). The ghost red cells are removed in the subsequent washing steps. White cells can be stained with antibodies prior to or after red cell lysis. In rare instances, red cells fail to lyse. This may be the result of increased numbers of reticulocytes (e.g., specimens harboring a red cell disorder such as a hemoglobinopathy or thalassemia) or increased lipids in the serum. The unlysed red cells can be electronically removed from analysis, however, by using antiglycophorin antibody and a gating procedure. The red cell lysis procedure is preferred to density gradient methods (e.g., Ficoll-Hypaque) because it allows cells to be maintained close to their native state. Density gradient methods are based primarily on the buoyant density of normal lymphocytes. Because neoplastic cells do not necessarily share the same density as normal lymphocytes, the density gradient methods, despite their ability to remove erythrocytes, mature granulocytes, and dead cells, can result in excessive loss of critical cells. This effect is especially undesirable for samples with a low content of neoplastic cells. In addition, selective population losses in the CD8 subsets can also occur with density gradient techniques.

2.3 Cell yield and viability Following the red cell lysis step, the cell yield is determined with an automated cell counter. Cytospins are made from the cell suspension to serve as morphologic controls, to ensure that the critical cells have been retained for analysis. When necessary, it is helpful to prepare additional cytospins for cytochemical stains (e.g., myeloperoxidase, nonspecific esterase). The viability of the cell suspension can be assessed by FCM, based on the uptake of DNA-binding dyes such as propidium iodide, 7-amino-actinomycin D (7-AAD) or TOPRO-3 by dead cells. A manual alternative technique using hematocytometer counting is trypan blue exclusion. As a rule, the cell yield and viability tend to be lower in aggressive tumors composed of large cells than in low-grade tumors composed of small cells. The larger neoplastic cells are more fragile and therefore more susceptible to damage and cell loss during the washing and centrifugation steps of specimen processing. Solid tissue specimens often have lower viability than liquid specimens in which most of the cells are already in a disassociated state and exposed equally to the surrounding nutrients. There is no set rule concerning the viability level below which a specimen yields uninterpretable data and therefore becomes unacceptable for FCM analysis. A lower viability can be better tolerated in specimens composed of nearly all neoplastic cells than in samples with a scanty proportion of tumor cells.

2.4 Sample staining Sample staining should be carried out as soon as possible after the nucleated cell suspension has been prepared. Delaying this step will only reduce viability and induce cell clumping,

FCM IMMUNOPHENOTYPING AND DNA ANALYSIS

17

especially if the tubes holding the cell suspensions are stored in an upright position. With the exception of cases with low cell yield, a portion of the cell suspension should be kept aside for potential repeats or add-on testing.

2.4.1 Surface antigens The multicolor direct immunofluorescence-staining technique using commercially available antibodies is employed for the simultaneous detection of multiple cell surface markers. Cell surface antigen staining is performed on viable unfixed cells. All staining is performed at 4ºC to minimize capping and antigen shedding. Appropriate isotype controls are included. The usual number of cells recommended for immunostaining is 106 cells for each test (i.e., each tube of antibody reagent cocktail). In situations with low cell yield, it is possible to perform the staining with as few as 1 × 105 to 2 × 105 cells/tube, however. The procedure should be carried out gently so as to minimize any further cell loss. An efficient and cost-saving strategy is the use of a microtiter plate-based method, which reduces the number of cells required for each test along with the volume of reagents, while maintaining the same proportions as in the conventional tube method. Each well of the microtiter plate requires only one-fifth of the reagents and cells. Batches of antibody reagent panels in microtiter plates can be prepared in advance, stored frozen, and thawed for use as needed. The microtiter plate method is further enhanced by the use of automated cell handlers. Computer-controlled devices to resuspend and introduce cells into the fluidic system of the flow cytometer are currently available, thus achieving a highly efficient and virtually “hands-free” operation.

2.4.2 Intracellular antigens Although the testing for intracellular antigens is performed up front in some laboratories, the authors’ preference is to have these done as “add-on” tests when it is medically necessary (see Section 2.8). The staining procedure is more laborious than cell surface antigen staining and calls for cell fixation and permeabilization. Cells are fixed to maintain structural integrity and are permeabilized to allow antibodies to reach the appropriate intracellular targets. The fixative should preserve the epitope of the intracellular antigen in question without causing aggregation of the cell suspension. Intracellular targets include TdT, cytoplasmic light chains, cCD3, cCD22, myeloperoxidase, bcl-2, cyclin-D1, and Zap-70. For increased sensitivity, the detection of intracellular antigens is done in conjunction with cell surface antigens (e.g., CD38 [for cytoplasmic immunoglobulins], CD19, or CD10 [for TdT]). To evaluate surface and intracellular antigens simultaneously, the cell surface antigens are stained first, followed by the fixation and permeabilization step, then staining of the intracellular marker. As with all staining procedures, appropriate background controls are included. In addition, a cell line expressing the targeted intracellular antigen is run in parallel with the patient’s sample. It may be necessary to empirically determine the optimal staining conditions for the various surface antigen(s)–intracellular antigen combinations to ensure the stability of the antigenantibody-fluorochrome complexes on the cell surface and the preservation of the targeted intracellular antigens. Because the success of intracellular antigen staining depends on the use of small-sized fluorochromes, antibodies targeted against intracellular antigens are most often conjugated to fluorescein isothiocyanate (FITC). Other fluorochromes, such as phycoerythrin (PE), may be used for simultaneous surface antigen labeling. In the authors’ laboratory, this staining technique was also applied to the simultaneous analysis of a surface antigen and DNA content using FITC and propidium iodide (PI) labeling, as there is a good separation between the emission signals of these two fluorochromes. The recent advent of DRAQ5 (deep-red

18

FLOW CYTOMETRY IN HEMATOPATHOLOGY

fluorescence bisalkylaminoanthraquinone no. 5) has replaced PI and alleviated the fixation and permeabilization steps for this procedure, however.

2.4.3 DNA content In the authors’ laboratory, DNA ploidy and cell cycle analysis are routinely assessed in acute lymphoblastic leukemia–lymphoma and in tissues involved by LPD/NHL. Prior to the millennium, the analysis was carried out with PI as the DNA dye. In selective instances where the proportion of ALL cells was low (e.g., partial involvement of the peripheral blood or bone marrow), DNA analysis was performed in tandem with TdT or cell surface antigen (e.g., CD19) staining. A similar approach was applied to solid tissues with partial involvement by LPD/NHL where DNA analysis was gated on the critical cells by utilizing the cell size (FSC) parameter or, when necessary, the concomitant staining for a surface label. In recent years, however, DRAQ5 has replaced PI as the DNA dye of choice. As a result, tandem staining with two surface antigens has become the norm (i.e., irrespective of the proportion of neoplastic cells in the FCM sample) in the author’s laboratory. DNA staining is performed on fresh cell suspensions. A large number of cells need to be analyzed to obtain robust measurements of the S-phase, an area that contains relatively fewer cells than the other phases of the cell cycle. The staining procedure is started promptly after the nucleated cell suspension is ready to minimize degradation of DNA. DNA degradation can alter the stoichiometry of dye binding to DNA, thereby affecting interpretation. The authors do not recommend DNA analysis on formalin-fixed, paraffin-embedded tissues. Although this approach makes retrospective studies possible, it often results in poorer quality analysis and an apparently higher S-phase as an artifact. Separation of normal lymphocytes closely admixed with tumor cells cannot be performed reliably on fixed tissue. Furthermore, the archival material precludes simultaneous DNA and cell surface antigen staining. There exist a variety of DNA-binding fluorochromes with excitation/emission wavelengths ranging from the ultraviolet to the visible regions of the spectrum. Before the availability of DRAQ5, the most common DNA dyes for clinical applications were PI and 7-AAD. The former binds to double-stranded nucleic acids by intercalating between the base pairs. Combining PI staining with RNAse treatment (to eliminate nonspecific staining of folded back single-strand RNA) had produced consistent high-quality results in the authors’ laboratory. The permeabilization step utilizes a fixative, either ethanol or paraformaldehyde. Ethanol fixation provides excellent preservation of DNA for long periods of time, but may cause cell shrinkage and loss of cell surface staining. It is best suited for stand-alone DNA staining. Paraformaldehyde fixation is more appropriate for simultaneous antigen detection (using FITC–conjugated antibodies) and DNA analysis, because it can preserve cell light scatter properties and antigen staining. The concentration of paraformaldehyde and the duration of fixation are critical factors to be considered, because a slight excess of this fixative may induce DNA cross-linking, thereby increasing the coefficient of variation (CV) of the DNA histograms. Either PI or 7-AAD can be employed in the simultaneous evaluation of DNA content and surface antigen(s). A limitation of PI is that only one surface antigen can be tested using an FITC–conjugated antibody because the spectral emission of PI overlaps with that of PE, whereas that of 7-AAD does not. A more costly reagent than PI, 7-AAD is also a bulky molecule, leading to the problem of nonstoichiometric DNA binding, which in turn may result in false aneuploidy. The commercial availability of DRAQ5 in recent years has revolutionized the simultaneous study of DNA with multiple surface antigens. The optimal excitation wavelength for this synthetic fluorochrome is at 647 nm, the emission line of a krypton laser. DRAQ5 is also excitable at the wavelengths emitted by the argon laser (488 nm) and helium-neon laser (633 nm) present

FCM IMMUNOPHENOTYPING AND DNA ANALYSIS

19

in standard benchtop FCM instruments, however. The emission spectrum of DRAQ5 extends from 670 nm into the far-red visible light, and is therefore distinct from that of FITC, PE and Texas red. DRAQ5 staining can thus be combined with FITC– and PE–conjugated antibodies, and the stained sample can be analyzed without much need of color compensation. DRAQ5 shares some of the desirable features found in PI, namely stoichiometric DNA binding and a low CV, to permit the detection of near-diploid aneuploidy. The most useful feature of DRAQ5 is its rapid penetration into live cells to bind to DNA without the need of fixation or permeabilization, thus preserving the light scatter and antigenic properties of the cells in the sample. This, in turn, facilitates the measurements of cell cycle phases on different cell subpopulations, especially the tumor population, within a heterogeneous sample.

2.5 Data acquisition The data acquired from the flow cytometer consist of light scatter and fluorescence measurements on single cells suspended in a liquid stream passing single file through a monochromatic light beam produced by a laser. Each cell is an event of a particular light intensity, recorded in an appropriate channel proportional to that intensity. The light scatter measurements reflect the physical properties of the cells (cell size or internal complexities), whereas the fluorescence data give information on the membrane or intracellular molecules (proteins, DNA) depending on the antibodies and dyes used for labeling the cells. As each dye-labeled cell passes through the light beam, the fluorochrome bound to the cell absorbs light, is excited to a higher energy state, and quickly returns to its relaxed state by emitting a fluorescence signal of a longer wavelength. The fluorescence signals are collected and amplified by photodetectors.

2.5.1 Calibration The various elements of the flow cytometer, namely light scatter detectors and fluorescence detectors, which include the voltage settings and spectral compensations of the photomultiplier tubes, must be monitored daily to ensure proper data acquisition. Standardized fluorescent beads are used for calibrating the instrument to determine that there have not been significant variations in the instrument settings from day to day. In this procedure, the electronics (i.e., the high-voltage settings of the photomultiplier tubes) are adjusted to place the fluorescence peaks of the beads in the same channels every time, thereby documenting any minor change in the settings. Another set of beads is used to monitor the mean channel and the CV for the FSC, SSC, and each of the fluorescence signals. These parameters are considered to be operating within acceptable limits if the CVs are <3.0 for FSC and SSC and <2.0 for each of the fluorescence signals.

2.5.2 Color compensation The initial setup of compensation settings is an empirical process, which depends on the fluorochrome combination and the particular instrument (i.e., the optical filters and photomultiplier tubes [PMTs]). Because of the broad emission spectra of fluorochromes, the light collected by an optical filter of a specific wavelength range reaching a PMT detector, consists of not only the signals from the intended fluorochrome but also signals from the other fluorescent dye(s). For example, in the case of an FITC and PE combination, the overlap in their emission spectra is such that a higher percentage of FITC signals are detected by the PE detector than the other way around. To correct this “spillover,” a fraction of the FITC fluorescence gets subtracted from the total fluorescence measured by the PE detector. The most appropriate

20

FLOW CYTOMETRY IN HEMATOPATHOLOGY

material for the initial setup and subsequent monitoring of color compensation is a normal control cell preparation, which consists of mononuclear cells from the buffy coat of a unit of donated blood. Each buffy coat can yield approximately 50 to 100 vials of 2 × 107 mononuclear cells that can be stored frozen, then thawed for daily use. The control cells are stained with mutually exclusive markers bearing the fluorochromes of interest, such as FITC–CD4, PE– CD8, and peridinium chlorophyll protein complex (PerCP)–CD20. Compensation adjustments are made while examining the data from each tube on dual fluorescence dot plots. Once compensation has been adjusted for each of the PMTs, the three aliquots are mixed into a single tube and analyzed. Any further modifications to the PMT high voltage or gain setting will require recompensation.

2.5.3 List mode data collection The goal of data acquisition is to collect measurements (light scatter, fluorescence) from each cell individually. The presence of doublets or aggregates can corrupt the acquired data. The adverse effect of doublet contamination is most serious when the cells of interest are few and the likelihood for confusion with a doublet event is high. The doublets may either originate from the cell suspension or result from an inappropriately high rate of sample throughput. In the latter instance, a doublet is the result of two single particles being so close to each other in the sample core stream that the flow cytometer sees them as a single event. The diameter of the core stream and thereby the alignment of the cells is affected by the sample delivery rate into the flow cell; the higher the sample delivery rate, the wider the core stream and the lower the precision and accuracy of the data collected. Doublets and cell aggregates in the cell suspension can be the result of either delay in analysis or suboptimal tissue dissociation, in which case they are also present on the corresponding cytospins. When running the sample for DNA analysis, the concentration of cells should be kept high (between 5 × 105 and 2 × 106/mL) and the flow rate low to achieve a low CV. The lower the CV is, the better the measurement of both ploidy and the calculations of the different phases of the cell cycle. In the authors’ laboratory, the maximum threshold for the flow rate is 200 cells/second. If the sample contains two different populations with very close DNA contents, the resolution can be improved by lowering the flow rate further. The optimal number of events to be acquired is between 30,000 and 70,000 cells, to achieve adequate statistics for cell cycle determinations. More cells should be acquired when the fraction of neoplastic cells is low. The data are collected ungated (i.e., no data, including debris, is discarded). The extent of debris signals present in the channels below the G0 peak is a clue to the magnitude of the debris in the channels underlying the S-phase and G2 peak. A slightly higher flow rate, 400 to 500 cells/second, is acceptable for immunophenotyping. In the authors’ laboratory, 20,000 to 30,000 events are routinely collected ungated from each sample tube after exclusion of the nonviable cells. Low levels (below 1%) of abnormal cells can be detected by an experienced observer when the phenotype is quite specific, such as that displayed by hairy cells. In other instances, where the phenotype is not so pathognomonic, acquiring a higher number of cells will improve the sensitivity of detection. The ungated data collection ensures that no critical cells are missed. After the initial acquisition of ungated data, in selected cases additional data can be collected, gating on the particular minute population of interest (e.g., live-gating to enrich a small population of potentially monoclonal B-cells). All other necessary gating procedures, such as gating on CD45 to identify leukemic blasts, are performed during the subsequent analysis of the originally ungated list mode data. For cases with poor red cell lysis, as evidenced by the presence of abundant RBCs on the cytospin and a sizable cluster of CD45-negative cells with very low FSC, the data collection should be on 20,000 “non-RBC” events.

FCM IMMUNOPHENOTYPING AND DNA ANALYSIS

21

A different approach to data collection is applied for the assessment of minimal residual disease because the sensitivity of the detection depends on the number of cells analyzed. According to previous studies on MRD (many using PCR-based methods), it appears that the clinically significant level of MRD is at 10−4 and above. For assessing MRD by FCM, detailed information on the phenotypic characteristics of the patient’s leukemic cells should be available from a previous FCM analysis at the time of diagnosis. Using a set of appropriate markers, the data are acquired with a “live gate.” Greater than 105 cells need to be collected in order to achieve a sensitivity of detecting one leukemic cell per 104 normal cells.

2.5.4 Exclusion of nonviable cells Flow cytometric identification of nonviable cells is preferably done with PI incorporation, especially for solid tissue specimens, as the viability is often lower than that in liquid specimens (Figure 2.2). Other possible reagents include TOPRO-3 or 7-AAD. An alternative

a

b

c

d

Figure 2.2 Liver with involvement by peripheral T-cell lymphoma. (a) The high uptake of PI by dead and dying cells facilitates the separation of viable (R1 gate) from nonviable cells. The viability in this sample is 15%. (b) The high content of nonviable cells imparts a disarrayed appearance to the ungated FSC/SSC dot plot. (c, d) Gated on R1: The tumor (arrow) demonstrates high FSC, downregulated CD3 and a loss of CD7 expression. Residual T-cells (low FSC, CD3+, CD7+) are present.

22

FLOW CYTOMETRY IN HEMATOPATHOLOGY

method is to gate dead cells out by the forward scatter parameter. Although dead and dying cells appear larger under the microscope, they have a lower refractive index than viable cells and therefore scatter less light in the forward direction. As a result, they can be removed electronically by excluding the lower FSC channels, where dead cells and debris accumulate. This FSC gating approach is less accurate than PI exclusion, however, especially on samples with cell populations of heterogeneous cell size, because the larger dead cells will fall in the same region as the smaller viable cells.

2.6 Antibody panel design For the analysis of leukemias and lymphomas, the antibodies are assembled into panels. There has been no agreement between laboratories on a uniform and standardized panel. It is not unusual for the panels in use in a particular laboratory to be based on practical (cost concerns and number of samples) rather than medical considerations. The design of antibody panels reflects one of two opposite approaches to the FCM workup of leukemias and lymphomas: comprehensive versus stepwise. In the first approach, the sample is analyzed with a large panel from the start. The number of antibodies in the panel is sufficiently extensive to permit a full characterization of the neoplasm, including any aberrant expression of the cell surface antigens. In the second approach, the FCM study starts with a limited panel. Then, based on the initial results, further analysis with appropriate antibodies is performed on the remaining sample, if necessary. Not infrequently, more than one round of additional testing may be required before reaching a final diagnosis. The stepwise strategy is more economical in terms of reagent cost, but the turnaround time is slower. The high frequency of additional testing can be disruptive to the laboratory workflow, especially in a laboratory with a high volume of FCM specimens. This can end up being expensive in terms of the technologists’ efficiency. A more important factor to consider is the decreased cell viability in the sample between the initial and subsequent additional steps, which, in turn, can affect the results adversely. The comprehensive approach, because of its large battery of reagents, is more costly. Using the microtiter plate technique and/or processing a large number of specimens can offset the cost, however. The availability of additional fluorochromes and newer generations of flow cytometers, being more sophisticated and equipped with three (or more) detectors should also help, in the long run, in decreasing the cost by reducing the number of redundant antibodies and shortening the analysis time. The clinical impression or the morphologic features of the specimens should not dictate the design and selection of an antibody panel. Despite efforts to improve the communication between the clinical and laboratory services, the clinical information on the FCM request forms is often scanty, vague, and potentially misleading. Furthermore, to have an antibody panel for each specific group of hematologic neoplasms (e.g., an ALL panel, AML panel, B-cell LPD panel, and T-cell LPD panel) would be inappropriate and defeat the purpose of FCM immunophenotyping. For instance, the presence of many large mononuclear cells in the blood or bone marrow does not necessarily indicate acute leukemia. Involvement by a large cell lymphoma can give a similar morphologic picture and, consequently, lead to the selection of the incorrect antibody panel (i.e., an acute leukemia panel instead of a lymphoma panel). Conversely, small ALL cells in sheets in the bone marrow can easily be mistaken as involvement by a mature B-cell LPD. Errors in panel selection can be circumvented by the comprehensive approach to FCM analysis, and in the absence of economic constraints, it is preferable to apply such an approach. A certain degree of redundancy of some critical antibodies (e.g., CD34 and CD20) between different tubes is necessary to optimize the sensitivity of FCM analysis and thereby permit the detection of aberrant antigenic expression on neoplastic cells. In the context of a

FCM IMMUNOPHENOTYPING AND DNA ANALYSIS

23

scanty specimen with a low cell yield, however, the morphologic findings or clinical information can be useful for guiding the selection of a limited battery of reagents.

2.6.1 Antibody selection Given the current large repertoire of reagents available from various manufacturers, careful antibody selection and panel design is important to achieve the best possible diagnostic information. Several technical factors need to be considered, including the type of antibody (monoclonal vs. polyclonal), antibody isotype (IgG1, IgG2, IgM), antibody clone, dye conjugation, and preparation by the manufacturer (i.e., the stoichiometry between antibody and fluorochrome). Except for the analysis of immunoglobulins (Igs), FCM immunophenotyping relies on monoclonal antibodies. Compared with the polyclonal antibodies used in the early days of immunophenotyping, monoclonal reagents are cleaner, with less background and crossreactivity. Monoclonal antibodies targeted against the same antigen structure but produced by different manufacturers do not necessarily have similar antibody reactivity, however. A specific example is the difference in reactivity between CD14–Leu M3 (Becton Dickinson), CD14– Mo2 (Beckman-Coulter), and CD14–My4 (Beckman-Coulter). Another factor to consider when constructing a panel is that antibodies (and fluorochrome conjugation) are more likely to be optimized (and thus best suited for a particular brand of flow cytometer) if the reagents and instruments are from the same manufacturer. One advantage of monoclonal antibodies is consistency in titer and affinity from one lot of antibody of the same clone to the next. Subtle changes in the preparation may alter the reactivity however, which, in turn, may affect the usefulness of the antibody in characterizing a particular disorder. A specific example is anti-CD20. This reagent has a wide dynamic range in fluorescence intensity, a useful feature to discriminate CLL/SLL (chronic lymphocytic leukemia/small lymphocytic lymphoma) from other B-cell neoplasms, as well as distinguishing neoplastic B-cells from a background of benign B-cells. A seemingly minor change in the manufacturer’s preparation of the same clone of CD20 antibody may produce a batch of CD20 with decreased brightness and thereby an altered fluorescence dynamic range, with the resulting loss of this important discriminating function. 2.6.1.1 Anti-light chain antibodies The high specificity of monoclonal antibodies, each recognizing distinctly defined epitopes, can be a disadvantage in the detection of surface Igs. Immunoglobulins have many epitopes and are, therefore, more easily detectable by polyclonal Fab′2 fragments than by monoclonal reagents. The surface immunoglobulins in some mature B-cell malignancies may not be produced correctly; one or more epitopes may be deleted or altered. Staining with a monoclonal antibody may yield a false-negative result if the reagent happens to be specific for the missing or modified epitope, whereas polyclonal antibodies, especially those with broader specificities, will give a positive result by reacting with the other epitopes of the immunoglobulin. The reactivities of anti-light chain (kappa, lambda) antibodies can vary widely between manufacturers. To select the optimal brand of kappa and lambda reagents, several candidate pairs are tested at various dilutions with samples from a variety of known disorders, including reactive lymphoid hyperplasia in lymph nodes or tonsils, CLL/SLL, mantle cell lymphoma (MCL), follicular center cell (FCC) lymphoma, and a neoplasm not producing immunoglobulin (e.g., AML). The samples with reactive hyperplasia, in which both light chains are expressed, are used for titration, to determine the optimal dilution for each reagent. For any given brand of kappa and lambda, the optimal dilution is usually the same for both antibodies. In some instances, however, because of preparation and lot-to-lot variation, one of the two light chain

24

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

Figure 2.3 Titration of anti-kappa and anti-lambda antibodies using a reactive lymph node. (a, b) CD20positive B-cells are polyclonal for kappa and lambda. For this particular pair of reagents, kappa had to be one step more dilute than lambda to achieve the desired results.

antibodies may need to be diluted more than the other so as to give identical fluorescence signals (Figure 2.3). Based on this comparative evaluation of different brands of kappa and lambda, a given brand is considered optimal if the fluorescence signals fall in the expected range for any given disorder, namely weak intensity in CLL/SLL and moderate to strong in MCL, along with the least background staining. With some brands of kappa/lambda reagents, it may not be possible to achieve the appropriate fluorescence intensity. The signals, irrespective of the dilutions, are inappropriately bright for samples with CLL/SLL (Figure 2.4) and may fall in the same fluorescence range observed in FCC and MCL. The intensity of surface light chain expression is one of several criteria critical in the diagnosis and subclassification of mature B-cell malignancies. Therefore, in selecting the optimal polyclonal kappa and lambda antibodies, the appropriate fluorescence signal is a more important consideration than the ability to detect extremely dim light chain expression. Lack of detectable light chains in a mature B-cell population invariably signifies an abnormal or neoplastic B-cell population, with the exception of benign plasma cells and reactive large germinal center cells. For this reason, it is quite unnecessary to conjugate the light chain antibodies to a fluorochrome with a high quantum yield such as PE. Furthermore, because the kappa antibody serves as a control for lambda and vice versa, it is advisable to conjugate these antibodies to the same type of fluorochrome (i.e., FITC), rather then having one light chain conjugated to FITC and the other to PE. In the authors’ experience, this strategy is much more sensitive than the reagent kit “kappa–FITC/lambda–PE” provided by the manufacturers. The combination kappa–FITC/lambda–PE may yield ambiguous kappa-lambda results, especially in the context of extreme follicular hyperplasia, or when monoclonal B-cells coexist with a larger population of benign B-cells (see Sections 4.1.1 and 4.4.1) (Figures 2.5 and 2.6). Preferably, the antibody panel should contain two sets of kappa and lambda reagents from different manufacturers, each pair combined with a different B-cell marker in the following configuration: kappa (1)–FITC/CD20–PE and lambda (1)–FITC/CD20–PE; kappa (2)–FITC/CD19–PE and lambda (2)–FITC/CD19–PE. This configuration is most helpful in the evaluation of two or more coexisting B-cell subpopulations, benign and neoplastic. Another advantage of this approach is the possibility of assessing the relationship between CD19 and

FCM IMMUNOPHENOTYPING AND DNA ANALYSIS

a

b

c

d

e

f

25

Figure 2.4 Selection of kappa and lambda reagents tested on the same specimen. (a, b) First set of antibodies: The intensity of the positive kappa light chain on CLL cells is inappropriately bright. (c) The peak fluorescence of kappa is 1 decalog brighter than that of lambda. (e, f) The second set of reagents from a different manufacturer yields appropriate results (i.e., dim positive kappa on CLL cells). (d) The same results are seen on the corresponding kappa/lambda overlay histograms.

26

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 2.5 Lymph node with FRFH. Benign germinal center cells (R2) display the brightest CD20 intensity (a), but typically lack surface light chains (c, d). The results are clear-cut, using kappa (2) and lambda (2), both labeled with FITC. Staining with the reagent combination kappa (1)–FITC and lambda (1)–PE produces ambiguous results, with an apparent excess of kappa (b).

CD20, a valuable feature in the characterization of low-grade B-cell neoplasms (see Section 3.6.2).

2.6.2 Fluorochrome conjugation For certain antibodies, conjugation to a particular fluorochrome can affect how much information can be derived from the test results. A specific example is the conjugation of FITC to CD10, used in combination with CD20–PE. The pattern of cell clusters on the CD10–FITC/CD20–PE dot plot is useful for distinguishing the following: (1) normal precursor B-cells (hematogones) from precursor-B ALL cells and (2) follicular lymphoma from florid reactive follicular hyperplasia (FRFH), either of which can present with no detectable light chain. In FRFH, the different cell clusters are seen in close continuity with each other (see Section 4.1.1.1). In contrast, when CD10 is conjugated to PE, the dot plot pattern often does not yield the same useful information because the apparent

FCM IMMUNOPHENOTYPING AND DNA ANALYSIS

a

b

c

d

27

Figure 2.6 Lymph node with FRFH. Benign germinal center cells are CD10 positive (a). bcl-2 is not overexpressed (c). Surface light chains are absent (d). The results are not clear-cut with the reagent combination kappa (1)–FITC and lambda (1)–PE, however (b).

CD10 expression on reactive germinal center cells is much more intense. The resulting doublepositive CD10/CD20 cell cluster becomes distinctly separated from the other cell clusters (Figures 2.7, 2.8 and 2.9), an appearance simulating that of FCC lymphoma, which then necessitates additional testing for bcl-2. There exist instances where the conjugation of CD10 to PE is diagnostically useful, however. A specific example is the use of CD10–PE, in combination with kappa–FITC and lambda–FITC for detecting residual/relapsed follicular lymphoma in patients receiving anti-CD20 therapy (rituximab). The selection of fluorochromes to be conjugated to the monoclonal antibodies (MoAbs) is based on the principle that a fluorochrome with a high quantum yield (e.g., PE, allophycocyanin [APC]) should be used if the antigen sought after is expressed at a low level. Otherwise, the MoAb can be conjugated to a fluorochrome with lower quantum yield, such as FITC. For antigens known to be present at high density (e.g., CD45), the corresponding MoAb can be conjugated to PerCP. For example, CD13 and CD33 should be conjugated to PE so as to maximize the separation of cells expressing these antigens. This principle does not apply to the surface light chains on B-cells, however, because weak expression of monoclonal light chains is a useful diagnostic criterion for the subclassification of B-cell LPD/NHL.

28

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 2.7 Conjugation of CD10 to PE produces an apparent increase in CD10 brightness on the benign germinal center cells (thin arrow) in FRFH (a). The resulting CD10/CD20 staining pattern mimics that on tumor cells (arrow) in an FNA of an FCC lymphoma (b). The latter is monoclonal for lambda (e, f). B-cells in FRFH are polyclonal for kappa and lambda (c, d). The CD10+ CD20− cluster in the FNA consists of granulocytes (b).

FCM IMMUNOPHENOTYPING AND DNA ANALYSIS

a

b

c

d

e

f

29

Figure 2.8 Conjugation of CD10 to PE results in a similar CD10/CD20 staining pattern between an FCC lymphoma (arrow) with partial lymph node involvement (a), and benign germinal center cells (thin arrow) in FRFH (b). Residual B-cells are polyclonal (open arrow) and the lymphoma cells are monoclonal for lambda (c, d). Benign germinal center cells show no light chain expression (e, f).

30

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 2.9 FCC lymphoma versus FRFH (continuation of Figure 2.8). Overexpression of bcl-2 in lymphoma cells (arrow) in comparison with residual B- and T-cells (a, c). Germinal center cells (thin arrow) in FRFH are bcl-2 negative (b, d).

The higher the number of fluorochromes for simultaneous analysis, the lower the number of tubes in the antibody panel and the lower the total number of cells required for the entire FCM study; reagents and cells are thus utilized more efficiently. Multiparameter analysis with three or four fluorochromes is currently the norm. Four-color labeling is more desirable if there are no associated technical difficulties. The four-color assay is most suitable for evaluating MRD, where the sample may be scanty and the number of critical cells rare. When testing solid tissue samples that often contain a significant number of dead cells, it is prudent that one of the dyes be used for excluding nonviable cells. In designing reagent cocktails for multicolor staining, it is important to be aware that one antibody can interfere with the reactivity of another; this can be due to either the isotype of the antibody and/or the type of dye, to which the antibody is conjugated. The electronic process of color compensation is based on the assumption that there is no steric hindrance, enhancement of binding, or dyeto-dye interaction (such as energy transfer or quenching) between the antibodies present in the

FCM IMMUNOPHENOTYPING AND DNA ANALYSIS

31

cocktail. In general, IgG antibodies are preferable to IgM because there is less likelihood for nonspecific binding and steric hindrance. Similarly, a very large size fluorochrome can cause steric interference. The authors prefer the single molecule type of fluorochrome, namely FITC, PE, APC, and PerCP to the large tandem conjugates (energy transfer dyes) such as PerCP–Cy5.5, PE–Texas red, or PE–Cy5. Proper energy transfer (i.e., the enhancement of the fluorescence of the acceptor molecule) simultaneous to the quenching of fluorescence of the donor is critical to the function of a tandem conjugate. When there is inefficient energy transfer between the donor and the acceptor molecules, or nonspecific binding to the Fc receptor of monocytes (e.g., PE– Cy5), the use of tandem conjugates may lead to misleading results. The argument for using tandem conjugates is the possibility to perform four-color immunophenotyping on instruments equipped with a single laser (488 nm). Potential problems, namely steric interference and the difficulties in achieving optimal color compensation among the four fluorochromes, especially when two of the four are energy transfer dyes (e.g., PE–Cy5 and PE–Texas red in the cocktail), must be carefully considered in this approach, however. Furthermore, some of the tandem dyes, namely Cy7-containing reagents, require additional precautions during sample preparation and analysis because of their susceptibility to light-induced degradation.

2.7 Comprehensive antibody panels In the authors’ laboratory, the antibody panels have been designed based on a comprehensive approach to FCM analysis. The rationale is to optimize the detection and characterization of the critical cells for determining (1) the lineage of the cells of interest (e.g., myeloid, B-cell, T-cell), (2) their maturity status, (3) the clonality, where appropriate, (4) the specific subtype of hematopoietic malignancy and (5) the status of the normal elements present. The use of large comprehensive panels also facilitates the detection of two or more unrelated neoplastic processes present in the same specimen. Appropriate isotype controls are included in the panels. The evaluation of the FCM data also relies on internal controls, however (e.g., T-cells serve as internal control for B-cells and vice versa) (Figure 2.10).

2.7.1 Disease-oriented antibody panels Over the years, the panels have evolved to incorporate new monoclonal antibodies of diagnostic significance. In addition, the authors modified the design of the panels from a diseaseoriented approach to one based on specimen type. The former strategy includes panels directed toward acute leukemia, lymphoproliferative disorders, or both types of disease. The differences between the acute leukemia and lymphoproliferative panels are that the former includes myeloid markers (e.g., CD13, CD64), whereas the latter contains additional antibodies (e.g., CD25, CD103) necessary for subclassifying lymphoproliferative disorders (LPDs) and lymphomas. Antibodies needed to identify the lymphoid lineages as well as maturity status are present in both panels. The more comprehensive panel combines all of the antibodies of the acute leukemia and lymphoproliferative panels and is most useful when no relevant clinical information is available, and either the number of critical cells is so low as to escape routine microscopic screening, or the nature of the abnormal cells is undetermined by morphologic criteria. Furthermore, the maturity status of the neoplastic cells (e.g., blasts vs. large lymphoma cells in the blood or bone marrow) may not be apparent on morphologic examination, because critical cytologic features such as nuclear chromatin can be easily altered by a slight degree of suboptimal staining and processing.

32

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 2.10 Isotype-matched negative controls (a) and internal controls (b–d). The normal residual T-cells (arrow) serve as a negative internal control for CD20, and a positive control for CD2 and CD3. The tumor cells, which comprise the larger cell cluster, are positive for CD20, and negative for CD2 and CD3.

2.7.2 Antibody panels oriented by specimen type The main drawback to the disease-oriented strategy is that in order to select the proper panel (acute leukemia vs. lymphoproliferative), morphologic screening of the specimen by an adequately experienced professional is required. In the authors’ experience from several laboratories (United States and Europe), incorrect panel selection because of morphologic misinterpretation is not an infrequent occurrence. To circumvent these issues, the authors have replaced the above antibody panels with panels oriented by specimen type. Thus, the two new panels, applied since the late 1990s, are a blood/bone marrow/spleen (BBS) panel and a tissue/ fluid (TF) panel. This strategy is based on the following rationale: (1) There is an equal predilection for immature hematopoietic malignancies (namely AML and ALL) as well as mature lymphoid neoplasms (B-cell or T-cell LPD/NHL) to involve the blood, bone marrow, and/or spleen, whereas (2) solid tissue and body fluids are much more often involved by lymphoid malignancies (irrespective of the maturity status of the neoplastic cells) than by myeloid neoplasms. In addition, disorders such as plasma cell tumors and hairy cell leukemia (HCL) occur

FCM IMMUNOPHENOTYPING AND DNA ANALYSIS

33

infrequently in the tissue and fluid compartments. The BBS panel, similar to the combined panel under the disease-oriented strategy, contains a large battery of surface antibodies up front, to permit a full characterization of the neoplasm. The TF panel is smaller, because several of the myeloid-associated markers, as well as CD103 and CD138, are not included. In the unusual event where the initial data reveal involvement of the tissue or fluid by AML, HCL, or plasma cell tumors, then the necessary antibodies are run as add-on tests using portions of the cell suspensions that are routinely set aside for such eventualities. The design of antibody cocktails for the different “tubes” in the BBS and TF panels, whether in three- or four-color versions (and soon, a six-color version), is aimed at obtaining the maximum diagnostic information from the antibodies. The composition of the cocktails corresponds broadly with the types of cells, normal and abnormal, which may be encountered in any BBS or TF specimen. For instance, the combination CD10 (FITC), CD20 (PE), CD19 (PerCP–Cy5.5), and CD5 (APC) permits the identification of CD5+ B-cells (benign or neoplastic), germinal center cells, or follicular lymphoma cells in lymphoid tissues. In view of the higher frequency of B-cell malignancies (in the Western world) than T-cell disorders, the TF panel is geared more toward B-cells and includes the use of heavy chain antibodies, which gives information on the differentiation stage of any given neoplastic B-cell population. The TF panel also contains one single “myeloid tube” (CD34, CD13/33, CD45, CD14) to cover the eventuality of a myeloid disorder involving the TF compartment (e.g., extramedullary hematopoiesis). In contrast, several myeloid tubes are required in the BBS panel, so as to achieve the full characterization of AMLs and any other abnormalities in granulocytic maturation. Redundancy in antibodies between “tubes” is necessary for several reasons, including the fact that at least one of the markers serves as an anchor-marker across several tubes. For instance, CD45 is the main anchor marker for most, if not all tubes in the BBS panel. CD20 and CD3 are the secondary anchor markers for the evaluation of mature B- and T-cells in the BBS compartment, respectively. Furthermore, some of the antigens are expressed on more than one cell lineage. For example, CD56 not only helps to identify NK cells and NK-like Tcells, it is also an important marker for the detection of abnormal plasma cells. CD56 is therefore included in one of the T-cell “tubes” (CD7, CD56, CD45, CD3), and in the plasma cell “tube” (CD38, CD56, CD45, CD19) of the BBS panel. In contrast with the BBS panel, the TF panel is not built around a single anchor-marker. Rather the B-cell “tubes” are anchored using CD20 and/or CD19, and the T-cell “tubes” anchored with CD3. The design of some of the cocktails also takes into account the fact that certain antibody– fluorochrome combinations, namely those with a wide dynamic range, are more informative than others. A specific example is the previously mentioned CD10–FITC/CD20–PE (or –APC) configuration, which has proved to be more useful than its counterpart CD10–PE/CD20–FITC (or –APC) in discriminating germinal center cells from follicular lymphoma cells in the TF compartment. Another important combination is CD11c–FITC/CD20–PE. Because of the wide dynamic range of CD11c, the pattern and the position of the critical cells on the dot plot permits one to distinguish HCL from other CD11c+ B-cell LPDs (see Section 3.6.3.2). This combination is especially helpful because a small number of B-cell LPDs may exhibit CD103 reactivity identical to HCL. In general, markers with a wide dynamic range are more useful, as cell populations positive for the same marker can be easily distinguished by their different fluorescence reactivities with that marker. Compared with other B-cell antibodies (e.g., CD19, CD22, FMC-7), CD20 has an optimal wide dynamic range. Therefore, most of the antibodies needed for the characterization of B-cell LPD/NHL are anchored to CD20 instead of CD19. The panels do not include FMC-7 or CD22 because, in the authors’ experience, these provide no additional diagnostic information to CD19 and CD20. These antibodies may be helpful for the evaluation of B-cells, neoplastic and benign, after anti-CD20 therapy, however.

34

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

Figure 2.11 Relationship of FMC-7 to CD20. (a) Downregulated CD20 and absent FMC-7 in CLL/SLL; (b) FMC-7 and CD20 coexpression in Burkitt lymphoma.

The main utility of FMC-7 is its absence; lack of FMC-7 expression is a typical finding in CLL. However, this feature is rather redundant for the diagnosis of CLL in light of the characteristic CD20 fluorescence pattern and the relationship of CD20 to CD19 (see Section 3.6.2) in this disorder. On a dual fluorescence display of FMC-7 and CD20, a clear linear relationship can be demonstrated (Figure 2.11).

2.8 Tailored panels and add-on testing In addition to these large routine panels, smaller panels can be tailored to analyze follow-up specimens in patients with a recent diagnosis of hematopoietic malignancy if the original graphical FCM data from the diagnostic sample is available for review. The smaller panel is especially applicable if the follow-up specimen (e.g., CSF, fine-needle aspiration [FNA]) has a low cell yield. For instance, a so-called clonal excess detection panel, which includes CD19, CD20, kappa, and lambda, and a few other markers (CD10, CD5, CD23, CD103), can be applied to follow patients with B-cell LPD/NHL using two or three B-cell tubes. The antibody composition of these tubes can be either identical to their counterparts in the TF panel, or tailored according to the known phenotype of the tumor. Because of its scantiness, CSF is handled differently from other specimens. Most CSF samples are submitted as follow-up specimens, to rule out involvement by acute leukemia (primarily ALL) or, less commonly, high-grade LPD/NHL. The cytospins may be reviewed first to determine if FCM is applicable. In general, when suspicious cells are present, the selection of the key antibodies to analyze the CSF is based on the FCM results from an earlier diagnostic specimen (e.g., lymph node, bone marrow). Often, the neoplastic cells can be detected by a four-antibody cocktail in a single tube. Antibodies to detect intracellular antigens (TdT, myeloperoxidase [MPO], cytoplasmic light chains, cCD3, cCD22, bcl-2) are not included in the authors’ standard panels. The analysis is performed as an add-on test because antibody staining for intracellular antigens is more timeconsuming than that for surface antigens. Because T-cell LPD/NHLs occur infrequently in the

FCM IMMUNOPHENOTYPING AND DNA ANALYSIS

35

Western world, testing for the surface antigens TCR α/β and TCR γ/δ, NK receptors (CD94, CD161, and KIR antigens) and the T-cell Vβ repertoire are also performed only on a case-bycase basis. Analysis of the Vβ repertoire employs a commercial eight-tube kit, identifying 24 different TCR-Vβ specificities (i.e., about 70% of the normal human Vβ repertoire). The mixture of Vβ antibodies in each tube is designed in such a way that each would react with 10% to 15% of T-cells. Each tube of the kit contains three different Vβ antibodies conjugated to FITC and/or PE, which then allows for two other T-cell markers to be added to the same tube. Note that TCR-αβ cannot be added into any of the tubes of the kit, however. The combined staining for TCR-Vβ/TCR-αβ is not technically feasible, possibly because of steric hindrance. Therefore, to determine the usage of the Vβ repertoire within the TCR-αβ+ T-cell population, the analysis should include testing for TCR-αβ and TCR-γδ (e.g., using the combinations, TCR-αβ/CD8/CD3/CD4 and TCR-γδ/CD8/CD3/CD4). The selection of which T-cell markers to use depends on the abnormalities detected on the initial FCM run. For example, CD3 and CD4 are employed if the initial data demonstrate a CD4+ T-cell population with either downregulated or upregulated CD3 expression. To maintain the optimal efficiency in a busy FCM–hematopathology laboratory, the following guidelines can be applied to “automate” the decision of when and which intracellular antigen staining to perform: TdT: Reactivity with either TdT or CD34 indicates that the neoplasm is composed of immature cells. Therefore, testing for TdT may be omitted for maturity assessment if the leukemia is already CD34+. TdT testing is most appropriate when the results from the standard panels indicate a lymphoid neoplasm with no CD34, no surface light chain expression, and no evidence of plasma cell differentiation. In that case, TdT is necessary to establish the maturity status of the tumor cells, which affects the diagnosis and therapy. A useful approach to assess TdT in ALL is to combine the TdT assay with DNA analysis. Aneuploidy is not only helpful as a prognostic marker but the TdT/ DNA combination will also serve as a useful fingerprint for the detection of residual/relapsed disease in the patient’s follow-up specimens. Another approach to monitoring MRD of T-ALL or precursor B-ALL is the combination of TdT with T-cell (e.g., CD7, CD3) or B-cell (e.g., CD19, CD10) markers, respectively. The combination TdT/CD19 offers good discrimination between benign B-cell progenitors and residual/relapsed precursor B-ALL (see Section 3.5.2). Whereas the immature neoplastic cells in ALL can be confused morphologically with mature neoplastic lymphoid cells in LPD/NHL, blasts in AML are morphologically distinctive from the maturing myeloid precursors. Therefore, testing for TdT in AML is not necessary irrespective of whether CD34 is expressed or not. The expression of TdT in AML is noncontributory for diagnostic and prognostic purposes. Furthermore, because of the high frequency of antigenic shift in AML, it is unlikely that TdT can be useful as a fingerprint at relapse. MPO antibody: The demonstration of MPO activity or CD13 and CD33 expression constitutes firm evidence of myeloid differentiation. MPO activity can be detected cytochemically (MPO cyto) or immunologically (MPO Ab). Therefore, MPO Ab testing may not always be necessary if the leukemia is either positive for MPO cyto or expresses both CD13 and CD33. Acute leukemias expressing only one of these two antigens (either CD13 or CD33) but with no lymphoid markers are most likely AML, in which case MPO Ab testing may help to confirm the diagnosis. MPO Ab testing is most informative when the blast population does not demonstrate a clear lineage (e.g., only one lymphoid and one myeloid marker are expressed). At the same time, staining for the appropriate cytoplasmic lymphoid marker, either cCD22 or cCD3, should also be performed. The frequency of cases where additional testing for MPO Ab and cCD3 or cCD22 is needed is relatively low. Cytoplasmic CD3, CD22, or mu chain: With the use of multiple antibodies in the panel and the multicolor approach, the need to stain for cCD3 and cCD22 rarely arises in our laboratory. In most cases of T-ALL, the presence of CD2, CD5, and CD7 is sufficient to infer the T-cell lineage.

36

FLOW CYTOMETRY IN HEMATOPATHOLOGY Previously, cytoplasmic IgM (cmu) was used for subclassifying precursor B-ALL. It is no longer necessary to perform this staining because the presence or absence of cmu has been shown to be of no relevance to prognosis and therapy. Testing for cCD3 may be helpful when the neoplastic cells express fewer than three pan T-cell-associated markers and lack surface CD3 and other lineage markers. Occasional high-grade T-cell lymphomas, in which only CD2 and CD5 or CD2 and CD7 are present, fall into this category. bcl-2: Testing for bcl-2 in combination with CD20 is appropriate when the results from the TF panel indicate a population of slightly larger cells, CD10+ and CD20+ (intense), and with poor or no surface light chain expression. In these cases, the differential diagnosis is FCC lymphoma versus FRFH. In addition to the pattern of CD10–FITC/CD20–PE coexpression (see Section 4.1.1.1), bcl-2 testing helps to resolve the differential diagnosis. This testing is particularly useful in cases of fine needle aspirates or when tissue samples are small and bcl-2 staining by immunohistochemistry may not be easily interpretable. Cytoplasmic light chains: This assay is performed to assess plasma cell clonality and is most often done in combination with surface CD38 or CD138 staining. Therefore, testing for cytoplasmic kappa (cKappa) and cytoplasmic lambda (cLambda) are performed when the results from the standard panel reveal a relative increase in the number of plasma cells (a distinct population with bright CD38 and negative CD45) but without phenotypic aberrancies. The assay can be omitted if the plasma cells are overtly abnormal, expressing CD56 or CD117. Staining for cytoplasmic light chains can also be paired with CD20 or CD19 to detect lymphoid malignancies with plasmacytic differentiation.

2.8.1 Minimal residual disease Detection of MRD has been performed primarily in acute leukemias for guiding therapy as well as for prognostic purposes. More recently, it has also been applied to patients with low-grade B-cell malignancies undergoing high-dose chemotherapy, stem cell transplant and immunotherapy, in which case the “clonal excess” tubes are employed for MRD detection. In acute leukemia, MRD assay by FCM analysis takes advantage of the immunophenotypic abnormalities frequently exhibited by leukemic blasts. The abnormalities may be overt, such as the expression of a marker from a different cell lineage, or subtle, in the form of downregulated or upregulated expression of a number of antigens when compared with normal counterparts. For instance, leukemic cells of most precursor-B ALL can be distinguished from bone marrow B-cell progenitors based on the differences in the expression of several antigens, including TdT, CD45, CD19, CD20, CD10, CD38, CD34 and CD58 (see Section 3.5.2). Based on the brighter expression of CD10 and CD58 in the leukemic blasts and using a four-antibody cocktail (e.g., CD10, CD58, CD45, CD19), it is possible to achieve a high resolution of detecting residual/relapsed disease to the level of one blast in 10,000 cells (10−4). In patients with T-ALL, knowledge of the antigenic profile of the leukemic cells would help to construct a tailored cocktail (e.g., TdT, CD8, CD45, CD4) for MRD detection purposes. The combination of a pan-T-cell antigen with either TdT or CD34 (e.g., TdT/CD3) has also proved to be useful, as combined expression of TdT/CD3 or CD34/CD3 is virtually never encountered in normal bone marrow cells. The presence of other abnormalities such as aneuploidy, the expression of CD56, or an aberrant myeloid antigen, further facilitates the detection of MRD in ALL of either lineage. The evaluation of MRD in AML by FCM relies on the phenotypic aberrancies (see Section 3.5.1.1) present on the patient’s leukemic blasts at the time of diagnosis. Based on these abnormalities, combinations of appropriate markers can be selected. Because of the relatively frequent phenotypic changes associated with long-term clonal evolution in AML, the use of such combinations is more suitable for detecting residual disease at the end of induction or consolidation therapy rather than later relapses.

FCM IMMUNOPHENOTYPING AND DNA ANALYSIS

37

2.9 FCM immunophenotyping data representation Irrespective of the number of colors used in FCM testing, the standard format for displaying FCM data still consists of two-dimensional (2D) graphics with x- and y-axes. Based on the current commercially available software for FCM data analysis, the most common approach for displaying immunophenotyping data derived from an antibody panel is the automatic tubeby-tube approach. The parameters available from one tube are displayed, followed by those from the next tube, and so on in a sequential manner. The parameters are shown in various permutations for the x- and y-axes, that is, FL1 versus FL2, FL1 versus FL3, FL2 versus FL3, light scatter (FSC or SSC) versus antibody fluorescence, and single-parameter fluorescence histograms. This approach to data display is rather inefficient because not every permutation is informative and there are a high number of graphics being generated. With some effort and careful planning, however, the FCM software currently available from most manufacturers can be used to create analysis panels based on a medical rationale, which provides a more logical approach to analyzing immunophenotyping data than the tube-by-tube approach. It may be judicious to install the analysis panels on workstation(s) connected via a network, so that the FCM data from any given case can be analyzed at any workstation rather than solely at the computer associated with a particular flow cytometer.

2.9.1 Analysis panels In the authors’ laboratory, one analysis panel has been created for each of the standard antibody panels (BBS, TF). The organization of the FCM displays in the BBS analysis panel follows the rationale that in the BBS compartment, acute leukemias (lymphoid and myeloid) and B-cell NHL/LPD occur more frequently than T-cell disorders. The TF analysis panel reflects a similar reasoning (i.e., that mature B-cell malignancies are more frequent than their T-cell counterparts). Because acute leukemias are considered life threatening, the corresponding displays are among the first to appear on the analysis panels, so as to facilitate the technologist in flagging the case and starting any necessary “add-on” antibody testing based on the above-mentioned “automated” decision-making guidelines. For instance, the BBS analysis panel begins with the SSC/CD45 dot plot, followed by the displays (gated on mononuclear cells) of CD117/CD34, CD13/CD34, CD33/CD34, CD10/CD19, CD10/CD20, CD10/CD58, CD38/CD45 and CD56/CD38. This grouping of dot plots helps one to screen quickly for involvement by myeloid malignancies, neoplastic precursor B-cells or plasma cell dyscrasia. The preferred format for data representation adopted by the authors is the dot plot whereby each dot corresponds with one event. The printout of the dot plots, especially if done in one color (e.g., in black and white), is best done at the 25% to 50% level of the actual cells acquired to facilitate the visual resolution of closely placed cell clusters. The analysis panels should be designed to include, in addition to the familiar dual fluorescence dot plots, displays that correlate the cell size (FSC) and antibody fluorescence data. The information derived from these two types of dot plots complement each other. The dot plots for FSC versus antibody fluorescence can be used for displaying ungated cell populations (i.e., in the case of the BBS analysis panel, granulocytes can be included). The dual fluorescence dot plots correlating antibody expressions can be focused on mononuclear cells, in order to identify and characterize the critical cells. Where appropriate, a gate can be drawn around the neoplastic cells (Figure 2.12) so as to estimate their relative proportion. In addition, the BBS analysis panel also includes dot plots gated solely on the granulocytic clusters to evaluate the expression of myeloid antigens (namely CD13, CD16, and CD11b) on the maturing myeloid precursors. The observed pattern of antigenic coexpression is often altered in myeloid

38

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

Figure 2.12 Bone marrow with residual AML. (a) The neoplastic cells form a distinct cluster in the blast region (arrow) coexpressing CD33 and CD34. (b) The blast content can be derived from a gate (R1) drawn around the CD33+ CD34+ cluster.

processes, such as some myelodysplastic syndromes and myeloproliferative disorders (see Sections 4.2.2.2 and 4.5.2).

2.9.2 Color display The immunophenotyping data can be displayed in color (e.g., each cell cluster in an SSC/CD45 display can be assigned a color), which may facilitate the identification of cell populations in other 2D projections of the data. Once a given population is depicted with a particular color on the anchor SSC/CD45 dot plot, it will appear in the same color in other dot plots. It is then possible to follow that population from one tube to the next throughout the panel. To use color displays effectively requires that the antibody panel be built around a single anchor-marker such as CD45–PerCP. In other words, the antibody panel cannot have two primary anchor-markers, such as CD45–PerCP in some “tubes” and CD20–PerCP in other “tubes.” The TF antibody panel is not based on any single primary anchor-marker. The graphics of the TF analysis panel are therefore displayed in one color. Where applicable, colors can be assigned to different populations using FSC as the anchor “marker.” In the correlated dual fluorescence data displays (e.g., CD71/CD20) (Figure 2.13), the color of the population of interest provides information about its cell size (i.e., small cells vs. large cells). In such instances, the use of color provides a third parameter (FSC) to the 2D fluorescence data displays. This does not preclude displaying the fluorescence data in correlation with the cell size, however. In most laboratories, the emphasis has been on dual fluorescence 2D graphics (e.g., FL1 vs. FL2) and little attention has been given to the correlated FSC and fluorescence data displays. The valuable information (sometimes subtle), which can be derived from the pattern and the relationship of the clusters present on the correlated FSC and antibody fluorescence dot plots, has therefore been overlooked (Figure 2.14). The main utility of displaying the FCM graphics in color is to reduce the number of FCM graphics to be printed for diagnostic review. The use of color also helps to identify different cell populations, especially those sharing similar reactivity for a given marker. A typical example is the expression of activation markers such as HLA-DR or CD38 by cells of various lineages. For instance, on the dot plot FSC/HLA-DR or HLA-DR/CD34 from a bone marrow

FCM IMMUNOPHENOTYPING AND DNA ANALYSIS

a

b

c

d

39

Figure 2.13 Lymph node with high-grade B-NHL. (a) Two distinct cell populations differing in cell size (FSC), for which the use of colors can be applied. (b–d) Cells with high FSC (black) are neoplastic B-cells monoclonal for kappa and expressing high levels of CD71. Cells with low FSC (gray) are benign T-cells negative for the markers displayed.

specimen, there is an apparent continuous population with heterogeneous HLA-DR intensity, from negative to extremely bright. On color displays, it can be more easily appreciated that the apparent continuous population is actually composed of several cell populations of approximately the same cell size merging into each other (Plate 1). In other instances, what appears to be two adjacent cell clusters is actually one cell population with bimodal reactivity for a given marker. With appropriate color displays, potential confusion of one versus two populations may be avoided (Plate 2). Many commercially available FCM software packages assign a hierarchy to the colors selected by the user, the color selected first being on the top of the hierarchy. Where two cell clusters overlap with each other, the one with the higher-ranked color will obscure the other, partly or completely, depending on the degree of overlap. For example, on the SSC/CD45 dot plot, the colors red and light green are assigned to the monocytic and myeloid populations, respectively, with the latter being the higher-ranked color. On another dot plot (e.g.,

40

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 2.14 Utility of FSC versus fluorescence displays. The lymphoid cells are of similar cell size (a, b) despite their antigenic heterogeneity. (b–d) The benign and neoplastic B-cells share similar intensities of CD20 and CD19 (not shown). It is thus difficult to evaluate monoclonality and quantify the malignant cells on the dual fluorescence dot plots. (e, f) The monoclonal cell cluster (arrow) is more clear-cut on the FSC/kappa (evaluated together with FSC/lambda) display.

FCM IMMUNOPHENOTYPING AND DNA ANALYSIS

41

CD33/CD34), these two populations largely overlap each other. Despite the fact that red is a bold and darker color than light green, very few, if any, red signals can be seen through the lighter green cluster (Plate 3). The effect is equivalent to that caused by an opaque light green color. In such instances, there is no advantage compared with a monochrome (black-and-white) display. This hierarchical “opacity” is the main limitation to the usefulness of the color display. In some cases, the use of colors may be distracting, causing subtle findings to be paradoxically missed. In the authors’ opinion, it is important to be versed in inspecting the immunophenotyping data irrespective of whether the graphics are in color or black-and-white.

2.10 Approach to DNA data analysis The purpose of DNA analysis is to determine the DNA content (i.e., the ploidy level) of the cells of interest and their growth rate (i.e., the relative proportion of cells in each phase of the cell cycle). The cell cycle can be compartmentalized based on the amount of DNA in the nucleus at a given time in the cycle. Cells with the 2N amount of DNA are either noncycling (G0 phase) or in the presynthetic growth (G1) phase of the cell cycle, during which RNA and some proteins are accumulated. This is followed by the synthetic (S) phase during which DNA is being replicated. Cells in the S-phase have an intermediate amount of DNA between 2N and 4N. After DNA duplication, the cells enter the postsynthetic (G2) growth phase and, finally, undergo mitosis (M). By FCM, cells in the G2- and M-phases are considered together, as they both have a 4N amount of DNA. DNA fluorescence has a relatively limited biological dynamic range (typically, fourfold to eightfold). The measurements are therefore made using linear amplification instead of the logarithmic scale used for antigenic fluorescence. The approach to DNA analysis is to first inspect the dot plot, FSC versus fluorescence of the DNA dye, correlating the total DNA content and cell cycle measurements with the relative cell size (Figure 2.15a). The FSC/DNA fluorescence dot plot is also useful for detecting small aneuploid population(s), which may be missed on the single parameter DNA histogram, and evaluating the presence of debris and

a

b

Figure 2.15 Lymph node with high-grade B-NHL. Cells were fixed in ethanol, treated with RNAse and stained with PI. Doublets were excluded. (a) The aneuploid cells are larger than the diploid cells. Most of the S-phase signals are associated with the aneuploid cells. (b) The same data shown on the DNA content histogram. Using a diploid control (not shown) it was established that the left peak represents diploid cells. The tumor is aneuploid (DI: 1.16) and highly proliferative (S% 23.9).

42

FLOW CYTOMETRY IN HEMATOPATHOLOGY

aggregates. The conventional approaches to correct for doublets and aggregates have been based primarily on the altered pulse shape produced by the doublets or clumps when illuminated by a focused laser beam. For instance, using the gating strategy of integrated versus peak DNA fluorescence, doublets of G0/G1 cells can be eliminated because they produce an integrated fluorescence equivalent to a G2/M cell, but a peak fluorescence equal to a G0/G1 cell. This exclusion technique does not eliminate the type of doublet caused by two G0/G1 cells passing through the laser side-by-side, however. An alternative approach to this electronic exclusion is to use commercially available DNA analysis software whereby mathematical algorithms can be applied to subtract doublets based solely on the DNA content distribution.

2.10.1 DNA ploidy The ploidy status of the cells of interest is determined by the position of its G0/G1 peak relative to the G0/G1 peak of diploid cells. By convention, the aneuploid peak must be separable from the diploid peak (i.e., the histogram is bimodal, with a trough between the diploid and aneuploid populations) (Figure 2.15). The DNA index, which reflects the DNA content in the aneuploid population, is the ratio of the peak fluorescence of the aneuploid G0/G1 peak to that of the diploid G0/G1 cells. The resolution of two cell populations depends on the CV of the DNA analysis and the percentage of aneuploid cells in the sample. The closer to diploid the aneuploid cells are and the lower the proportion of aneuploid cells, the lower the CV must be to achieve separation. Most lymphomas are diploid or near diploid; frank aneuploidy and polyploidy are less frequent (Figure 2.16). Therefore, if the specimen contains a large nonneoplastic population, this diploid component may obscure the near-diploid neoplastic cells on the single parameter DNA histogram and interfere with calculations of the tumor cell S-phase. Prior to the replacement of PI by DRAQ5, the solution to this problem was to perform DNA testing paired with an appropriate FITC-conjugated antibody (e.g., CD20) that would help to delineate the abnormal population (Figure 2.17). With the routine combination of DRAQ5 and two other dyes for surface antigens, measurements of the DNA content can be directly obtained on the gated tumor cells (Figure 2.18) or, where applicable, on a subpopulation thereof. In addition, the implementation of DRAQ5 has also eliminated the technical limitations associated with PI, including the artifact-induced aneuploidy (e.g., ethanol-fixed, PI-stained whole blood or bone marrow may produce extra peaks on the DNA histograms because of the presence of granulocytes). In ALL, aneuploidy (especially in combination with TdT) is a useful marker for detecting low-level involvement of a sanctuary site (e.g., CSF), as well as for monitoring residual disease in a posttreatment hypocellular marrow or impending relapse in a regenerative marrow (Figure 2.19). In addition, DNA ploidy may be of prognostic significance in ALL. Data on pediatric ALL have revealed that a DNA index (DI) greater than 1.16, which corresponds with the presence of more than 53 chromosomes, is associated with longer remission. Although there is a good correlation between FCM aneuploidy and cytogenetic aneuploidy, karyotypic abnormalities (e.g., small deletions, balanced translocations) that result in a pseudodiploid karyotype are not detectable by flow cytometry.

2.10.2 S-phase Calculation of the cell cycle phases requires the deconvolution of the DNA histogram into three areas (G0/G1, S, and G2/M) and integration of each individual area. The main difficulty in this process is to separate cells in early S-phase from those in G0/G1, and cells in late S-phase from those in G2/M. Various computing methods have been applied to calculate the frac-

FCM IMMUNOPHENOTYPING AND DNA ANALYSIS

a

b

c

d

43

Figure 2.16 Examples of DNA histograms on different lymphomas. (a) FCC II lymphoma: Diploid and low S-phase fraction (S% 2). (b) PTCL: Diploid, S-phase fraction in the intermediate range (S% 10). (c) PTCL: The tumor (arrow) is near tetraploid (DI: 1.84) with an S-phase fraction of 6%. (d) FCC II lymphoma: The tumor (arrows) is polyploid and the S-phase low (DI: 1.1 and 2.1, S% 2).

tion of cells in each phase of the cell cycle. Commercially available programs can be applied to complex histograms (e.g., the presence of debris or overlapping peaks) or those with very high proliferative fractions. For discrete populations with a rather evenly distributed S-phase, low content of debris, and low to intermediate S-phase fractions (<15%), a simple rectangle method for S-phase calculation is suitable to most cases of malignant lymphoma. With PI-based DNA analysis (i.e., prior to the DRAQ5 era), the presence of normal cells in the sample can affect the cell cycle phase calculations on the neoplastic cells, and therefore the calculations need to be corrected by subtracting the normal cells from the G0 /G1 peak. If the abnormal cells cannot be identified by a specific feature such as higher FSC, one can assume that normal cells, in specimens other than the bone marrow, are mostly quiescent and do not contribute significantly to the cycling pool. This assumption can be applied irrespective of how close the DNA peak of the neoplastic cells is to that of the normal cells.

44

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 2.17 Combined analysis of light scatter, CD20 and DNA content to detect and characterize the presence of high-grade B-cell NHL in the peripheral blood. (a) The neoplastic B-cells (R1) are of heterogeneous cell size and comprise 16% of the total cell population. (b) DNA analysis on all cells in the sample. The S-phase signals are associated with large cells. The near diploid component is more discernible on the FSC/DNA dot plot than on the single parameter DNA histogram (c). (d) DNA analysis on CD20+ cells only (R1 gate): The tumor is near diploid (DI: 1.1) and extremely proliferative. The tumor-associated S-phase fraction (S% 54) is much higher than that determined on the entire sample (c).

The preferred approach, especially if the normal and neoplastic populations closely overlap, is multiparameter DNA–antigen analysis whereby an uncontaminated tumor cell cycle fraction can be calculated (Figures 2.17, 2.18, and 2.20). This is the method of choice when the aneuploid cells are the minor component and/or the normal cells are bone marrow hematopoietic precursors, whereby the S and G2/M signals from the tumor cells on the single parameter DNA histogram are obscured by the “noise” from the larger normal component. In lymphomas, the S-phase fraction has been proven to correlate well with the biological grade (i.e., clinical course) of the disease. Highly proliferative lymphomas confer a shorter remission duration and poorer overall survival than lymphomas with a low S-phase. On the other hand, the incidence of clinical “complete remission” is higher in lymphomas with a high S-phase fraction because highly proliferative tumors are more amenable to aggressive therapy. Thus, the S-phase fraction is extremely useful for grading lymphomas and assessing progression/transformation from low-grade to aggressive lymphoma, irrespective of the histological classification scheme. From studies correlating the S-phase fraction in lymphomas with the clinical

FCM IMMUNOPHENOTYPING AND DNA ANALYSIS

a

b

c

d

e

f

45

Figure 2.18 Lymph node with follicular lymphoma. The tumor cells are CD10+ (b) and monoclonal for lambda (c, d) with a bimodal cell size distribution (a). (e) DRAQ5 tube: Gated on the CD10+ CD20+ cells for DNA analysis. From this tube, the intensities of the surface antigens (CD10, CD20) are invariably artifactually reduced. (f) The lymphoma is near tetraploid (DI: 1.9) with a low S-phase (1%).

46

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 2.19 Combined analysis of TdT and DNA content applied to two different cases of precursor B-ALL with appropriate controls (a, c) included. (b) Blasts in the first case are TdT positive and diploid. The proliferative fraction is high. (d) Blasts in the second case are TdT positive and hypodiploid (DI: 0.77). The S-phase fraction is low.

FCM IMMUNOPHENOTYPING AND DNA ANALYSIS

a

b

c

d

e

f

47

Figure 2.20 B-cell lymphoma, intermediate grade. (a–c) The neoplastic cells display distinct expression of CD19, CD20 and kappa. The cell size ranges from small to medium. Higher FSC signals (beyond 500) are artifacts. (d–f) DRAQ5 tube: DNA analysis is gated on CD19/CD20 cells (d). Doublets are excluded (e). The tumor is diploid (DI: 1) with an intermediate S-phase of 8% (f).

48

FLOW CYTOMETRY IN HEMATOPATHOLOGY

survival and/or relapse-free interval, an S-phase threshold can be derived to establish the grading and prognostic categories. The accepted S-phase threshold for lymphoma grading is 5%. Because FCM parameters, similar to any other laboratory data, contain some margin of error, it is preferable to create three rather than two prognostic groups. Low-grade lymphomas invariably demonstrate proliferative fractions well below the 5% level. The S-phase fraction in more aggressive disease is scattered over a wide range, however. In this group, high-grade lymphomas such as Burkitt lymphoma or B-cell lymphoma with “plasmablastic” differentiation (also referred to as immunoblastic lymphoma) have S-phases in the range of 20% and beyond. In between the two extremes of low- and high-grade is the category of intermediategrade lymphomas in which the S-phase fraction usually falls between 5% (or the vicinity thereof) and 15%.

CHAPTER

3

FCM data analysis on nearly homogeneous samples

During data acquisition, the flow cytometer measures the light-scattering properties and fluorescence characteristics of each cell in sequence. The main purpose of the next step, data analysis, is to identify the cells of interest. This process involves distinguishing any abnormal cell population present from normal cell types and, when an abnormal population is found, determining its lineage, maturity, and other characteristics (cell size, antigens expressed, fluorescence pattern of the antigenic expression). The antigenic profile of the abnormal population may lead to a specific diagnosis. Once a neoplastic population is characterized, the multiparameter data can be used to quantify its proportion relative to other cells in the sample. In leukemia–lymphoma immunophenotyping, the percentage of neoplastic cells is a less important feature than the fluorescence intensity of specific antigens on their surface, especially because the percentage of abnormal cells is easily altered by sampling-related factors such as severe hemodilution in bone marrow or fine-needle aspirates. Furthermore, tumor cells are often more fragile than normal cells and therefore preferentially lost during transport, storage, or specimen processing. The ability to distinguish the cells of interest from other normal cell types depends on (1) the relative frequency of the neoplastic cells in the final FCM sample (i.e., after processing) and (2) how much they differ from the normal cells. In most instances, clinical samples consist of heterogeneous cell populations and the critical cells are often not the predominant component. In some cases, however, the sample is overrun by a nearly homogeneous neoplastic cell population, producing a cell culture-like picture such that even the crude and suboptimal technique of single-color FCM can determine the phenotype of the malignant cells. The FCM dot plots produced by heterogeneous samples are inherently complex (with multiple cell clusters) and therefore potentially difficult to evaluate, especially by untrained observers. In contrast, samples with very high numbers of neoplastic cells generate more straightforward-appearing FCM graphics, in which the abnormal cell cluster is virtually the sole component. Several hematopoietic neoplasms can produce this type of FCM picture, including (a) most acute leukemias and (b) various LPD/NHL with extensive involvement at the time of presentation, such as CLL/SLL, prolymphocytic leukemia (PLL), mantle cell lymphoma (MCL), and Burkitt lymphoma. In this chapter, the focus is on nearly homogeneous specimens, because they serve as a useful starting point for inexperienced laboratory staff to familiarize themselves with the visual approach to FCM data analysis. Because the tumor population is already obvious, the task of data analysis in homogeneous specimens is primarily concerned with characterizing the phenotype of the abnormal population. It is easy to appreciate the light scatter properties and antigenic characteristics of a homogeneous abnormal cell population, as reflected by its location and shape on the various dot plots. Because the graphics in homogeneous specimens are simplified, the logical sequence that should be applied to the review of all FCM cases, both simple and more complex, can be easily understood. It should be noted that a small population of normal cells (mostly residual benign lymphocytes) is almost always present, even in specimens with extensive neoplastic involvement. This minor population serves as an internal control. Its location and shape on the dot plots should not be overlooked.

50

FLOW CYTOMETRY IN HEMATOPATHOLOGY

Lineage assignment of leukemias and lymphomas is facilitated by the fact that the neoplastic cells conserve many antigens present on normal lymphoid or myeloid cells. There have been several attempts to subclassify lymphoid malignancies according to the different stages of Bcell and T-cell ontogeny, based on the speculation that the phenotype of the neoplastic cells represents an arrest at a given stage of B-cell or T-cell differentiation. The classification of ALL in earlier studies has been based on such assumptions. However, when applied to lymphoid neoplasms beyond the simple classification of early and mature stages, these assumptions are of limited practical value in the diagnosis (or prognosis) of leukemia and malignant lymphoma. Furthermore, despite antigenic similarities between malignant and benign lymphoid cells, many antigens are expressed on neoplastic cells differently from their expression on normal lymphocytes (e.g., weaker fluorescence intensity).

3.1 FCM parameters The physical characteristics and antigenic features of a cell population are determined based on the light scatter and fluorescence parameters.

3.1.1 Forward scatter The light scattered by a cell in the forward direction near the axis of the incident beam is a complex function of refractive index and cell diameter and consists mostly of diffracted light. The signal collected by the FSC detectors is approximately proportional to cell size. In clinical samples, this parameter has proved to be a good discriminator in mononuclear cells and is especially useful in the characterization of mature lymphoid neoplasms. Lymphomas composed of cells with high FSC signals tend to be more aggressive and more proliferative than those composed of small cells (low FSC). In the analysis panels used by the authors, FSC is represented on the y-axis of the FSC versus SSC and FSC versus fluorescence dot plots. A linear scale of 0, 200, 400, 600, 800, and 1000 is employed, and instrument settings are established such that signals falling within the 200 to 400 range are considered low FSC, 400 to 600 is considered intermediate, and from around 600 upward is considered high. These, in turn, correspond with small, medium (Figure 3.1), and large cell size (Figure 3.2) respectively. For example, normal small lymphoid cells appear as a compact cluster with low FSC, whereas monocytes display medium FSC (Figure 3.3). Debris and platelet clumps, when present, generate FSC signals below 200. Within any cell population, irrespective of its benign or malignant nature, there is a certain degree of variation in cell size. As a result, the lower-range FSC signals of a cell cluster with medium cell size can overlap slightly with the upper-range FSC signals of a cluster with small cell size. Note that the range limits provided here for the FSC parameter are only rough estimates. In the visual approach to FCM data analysis, the relative position of various cell clusters and the resulting pattern are of greater significance than the absolute numerical values of any of the light scatter or fluorescence signals. Assignment of cell size for a given cell population is based on the FSC signals generated by the majority of the cells in the cluster. For example, in many cases of low-grade B-cell NHL, some of the FSC signals fall into the 400 to 410 range (i.e., medium cell size), but the bulk of the cell population falls within the small size range (Figure 3.4). The shape of the cluster distribution along the FSC axis also provides information on the degree of cell size heterogeneity within the population. A stretched-out cell cluster, with FSC

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

51

Figure 3.1 (a, b) Lymph node with FCC I lymphoma. The cell size (FSC) of the neoplastic cells (CD20+) is small, similar to that of residual T-cells (CD5+). (c, d) Lymph node with Burkitt lymphoma. The malignant B-cells are larger (medium FSC) than the residual T-cells.

signals spanning from the lower limit of the low FSC to the upper limit of the medium FSC range and beyond, is an indication that the cell population is of variable size (Figure 3.5). A microscopic review of the corresponding cytospin will confirm that the population in question is composed of cells ranging from small to large. The picture of variable FSC is typically seen in follicular lymphomas containing a substantial number of large cells (i.e., FCC II and FCC III lymphomas) as well as in those CLLs with an increased number of activated lymphoid cells. Occasionally, the neoplastic population is composed of two nearly distinct clusters with different FSC signals. The cell size distribution is then referred to as bimodal (Figure 3.6). This is not a common phenomenon, but can be seen in some cases of CLL/PL (also referred to as “CLL in prolymphocytic transformation” because of the increased number of prolymphocytes).

52

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 3.2 AML involving the kidney. (a) A prominent cluster (R1) in the blast region. (b–d) Blasts display large cell size, CD33 expression and lack of CD34. The expression of CD33 is brighter than that on granulocytes (arrow).

In general, there is a good correlation between the cell size as determined by FSC and that estimated by light microscopy. Light scatter determinations are derived from live cells in suspension, close to their native state. The morphologic material for light microscopy, on the other hand, can be laden with undesirable artifacts, either from suboptimal fixation/processing of tissue sections or slow drying of smear/imprints, which can easily lead to cell shrinkage. Optimal quality morphologic material is therefore required for correlation. Cells that display medium FSC signals often appear large on air-dried smears and imprints. A typical example is monocytes. This phenomenon is not a discrepancy; it is simply due to the fact that on cytologic preparations these cells, basically spherical structures, have been flattened into circles with an apparent greater surface area. The feature of FSC as a good discriminator of cell size in mononuclear cells can be exploited by displaying correlated dual-parameter FSC/fluorescence plots for each of the antibodies tested. Often, the cells of interest can be followed from one display to another by virtue of their location on the FSC axis (i.e., cell size) (Figure 3.7). Furthermore, the location, size, and

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

53

b

Figure 3.3 Normal peripheral blood. (a) Three distinct populations differing in cell size and granularity: lymphocytes (low FSC, low SSC), monocytes (medium FSC, low SSC) and granulocytes (medium FSC, high SSC). (b) Gated on MNCs (i.e., no granulocytes): B-cells, T-cells and monocytes (M). CD4 expression on monocytes is less intense than that on T-cells.

shape of the neoplastic cluster and its relationship to the background normal cell population(s) seen on the various FSC versus antibody fluorescence dot plots can often be specific for a given type of leukemia or lymphoma.

3.1.2 Side scatter Side-scatter signals are collected by optical detectors located perpendicular to the excitation beam. The SSC light consists mostly of refracted and reflected light, which result from changes in refractive index associated with internal structures such as cytoplasmic organelles. The SSC measurements give an indication of the internal complexity of the cell, which may be loosely referred to as “granularity.” In the authors’ laboratory, the preferred method to collect and display SSC data uses a linear scale from 0 to 1024.

a

b

Figure 3.4 Lymph node with FCC lymphoma. (a, b) The tumor population (CD20+ CD10+) is of small cell size. Only a small number of the FSC signals fall in the medium cell size range.

54

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 3.5 (a, b) Extranodal PTCL. The neoplastic cells (CD3+, CD5−) display variable FSC, from low to high. (c, d) Peripheral blood with activated CLL. The tumor population (CD20+ CD5+) is of small to medium cell size. Because of the increased proportion of activated cells, the intensity of CD20 is brighter than that seen in typical CLL.

With the exception of SSC versus CD45, dot plots of SSC versus antibody fluorescence have no applicability in FCM data analysis because SSC is a poor discriminator of mononuclear cells. Therefore, the analysis panels only include two dot plots displaying the SSC parameter: FSC/SSC and SSC/CD45–PerCP (Figure 3.8). These two graphics are useful for separating the mononuclear cell fraction from the myeloid cells. Granulocytes have much higher SSC signals, usually 300 and above. In addition, the SSC/CD45 dot plot is excellent for defining various subpopulations in the bone marrow or peripheral blood (see Section 4.1.2).

3.1.3 Fluorescence Assessing fluorescence intensity is an important step in characterizing and classifying hematopoietic malignancies. This is even more critical when dealing with closely related disorders that share the expression of several antigens (e.g., CLL/SLL and MCL). The level of

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

55

Figure 3.6 Peripheral blood with CLL/PL (morphology shown in Plate 14). (a–d) The leukemic population demonstrates bimodal FSC. The phenotypic profile (downregulated CD20, coexpression of CD5 and CD23) of typical CLL is still retained, although CD20 and CD5 are brighter in the more activated component (larger cells). CD45 is also downregulated in comparison with the small population of normal lymphocytes (arrow).

fluorescence detected is dependent on (1) the fluorochrome employed, (2) the binding capacity of the particular fluorochrome–antibody conjugate, and (3) the number of antigenic epitopes on the cell surface. To this end, efforts have been spent on quantitative FCM, with an aim to standardize and measure absolute fluorescence intensity. It is still not clear that this approach is widely applicable, however, or whether absolute fluorescence intensity measurements are criteria sufficient to characterize abnormal populations. In practice, a cell population is recognized as abnormal because of the pattern of one or more aberrancies (e.g., downregulated expression or overexpression) in comparison with normal cells, rather than absolute fluorescence measurements. Therefore, for diagnostic purposes, fluorescence intensity determinations utilize relative units rather than absolute ones. The relative intensities are usually estimated based on the difference between the mean, median, or modal channel (channel number with

56

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.7 Lymph node involvement by mycosis fungoides. (a–f) Using the FSC parameter, the neoplastic cell population (arrow) can be followed from one display to the next, to evaluate its antigenic expression and relationship to other cell clusters, in particular the residual normal T-cells. The tumor cells demonstrate downregulated CD3 (a), upregulated CD5 (d), positivity for CD4 (e) and lack of CD7 (b), CD2 (c) and CD8 (e) expression.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

57

b

Figure 3.8 Peripheral blood with 1% circulating blasts. (a) The various cell populations are well separated on the SSC/CD45 dot plot. B, blasts; G, granulocytes; L, lymphocytes; M, monocytes. (b) Only lymphocytes, monocytes and granulocytes can be seen on the FSC/SSC display.

the most events) of normal and neoplastic cells, and the results are reported in a semiquantitative/qualitative manner (e.g., negative, positive, dim, and bright). Because of the wide range of fluorescence intensities spanning over four decades (i.e., 10,000 : 1), flow cytometers employ logarithmic amplifier systems for compressing the data over 1024 channels into a log scale of four decades of 256 channels each. On the dot plots, fluorescence intensities are often displayed on a scale of 100, 101, 102, 103, and 104. In general, the first decalog (between 100 and 101) represents negative fluorescence, and signals falling in the second, third, and fourth decalogs represent weak (+), moderate (++), and strong (+++) fluorescence intensity, respectively. The evaluation of fluorescence intensities takes into account the isotype negative controls and the presence of normal cells in the sample. The intensity of a given antigen on the cells of interest is determined by visually comparing the pattern produced by the critical cells reacting with the corresponding antibody to that obtained with the isotype-matched control. Placing of the cursor is not useful because, in clinical samples, there is often an overlap between the negative and positive cells. It is important not to rely solely on isotype controls, however, as they may not necessarily match the biochemical properties of every antibody in the panel, despite an identical concentration and protein:fluorochrome ratio. The background normal cells in the sample are important internal controls. In a large panel, there will always be cells or antibodies serving as negative controls. Monocytic cells tend to have nonspecific staining of immunoglobulins, however (Figures 3.9 and 3.10). It is also important not to be too rigid about the above-mentioned scale division, especially when the cell population exhibits a significant degree of autofluorescence. This feature is typical of leukemic blasts in AML-M3 (Figure 3.11). In straightforward situations, the fluorescence intensity of a given marker on the negative population falls within the first decalog, and the cells of interest form a clear-cut cluster in the positive region. This is most easily seen when the sample is composed almost entirely of abnormal cells. The lack of staining of this population with a given antibody serves as its own negative control when compared with the positive staining with another antibody (Figure 3.12). In many instances, however, an increase in the cell size of the malignant cells is accompanied by a higher degree of background staining, as the large cells tend to be more “sticky”

58

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.9 AML with monocytic differentiation in the peripheral blood. (a) A sizable monocytic cluster (arrow) and a smaller cluster in the blast region (thin arrow). Granulocytes (G) are few. The nonspecific reactivity for kappa and lambda (c, d) compared with the isotype control (b) is typical of monocytic cells. They express dim CD4 (f) but not other T-cell antigens, for example, CD5 (e).

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

59

Figure 3.10 AML-M5 (continuation of Figure 3.9). (a, b, d) The abnormal monocytic cells (arrow) lack CD13 but express CD33 and CD64 at higher levels than granulocytes (G). (c) The heterogeneous CD14 expression indicates that the population consists of monocytic cells at various stages of maturation.

(presumably from increased Fc receptors) or have more autofluorescence. The fluorescence signals of the negative isotype control may then extend well into the region usually considered to be positive staining (i.e., into the second decalog and occasionally beyond). A mental correction of this right shift is necessary to assign the appropriate fluorescence intensity to the positive population (Figure 3.11). For example, if the expression of a given antigen on the abnormal cells falls within the third decalog (moderate) but the negative population is rightshifted well into the second decalog (Figure 3.13), the fluorescence intensity on the positive population should be appropriately reported as weak instead of moderate. It is important not to confuse this phenomenon with poor instrument calibration. When the background shift is caused by increased “stickiness” or autofluorescence, it should not be present in other specimens (e.g., neoplasms with small cell size, or normal specimens) run the same day. The reverse phenomenon can also be seen; that is, the negative population does not fill the entire first decalog but is significantly shifted close to 100. This situation is relatively infrequent, however. It may be seen in samples composed of a nearly pure population of small malignant lymphoid cells (Figure 3.14).

60

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.11 AML-M3. The leukemic blasts display the characteristic high SSC (a) and increased autofluorescence (b). (c–f) Taking this feature into account, it can be determined that blasts are negative for CD34, HLA-DR and CD11b, but express CD13, CD33 and weak CD15.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

61

Figure 3.12 Lymph node with large cell lymphoma. The neoplastic cells comprise 81% of the cells in the sample. (b, d) The fluorescence signals of the negative markers, CD10 and lambda, fall in the first decalog. These serve as negative internal controls. The tumor cells are of medium cell size and react with CD20 (a), CD25 (e), kappa (c) and CD71 (f).

62

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 3.13 Bone marrow with AML. (a) Isotype negative control with increased baseline fluorescence extending into the second decalog. Other isotype-matched negative controls (not shown) yielded the same picture. (b–d) This increase is taken into account when judging the fluorescence intensity of the various markers on the leukemic blasts. Blasts (B) are thus CD34+, CD13++ and CD33−. Residual granulocytes (G) and monocytes (M) are present.

A common phenomenon associated with large B-cell lymphomas with intense expression of CD20 is the shedding of CD20, which then coats the residual T-cells. As a result, the T-cell cluster appears to be CD20-positive when viewed on any of the dot plots that feature CD20 (e.g., FSC/CD20, CD5/CD20). This phenomenon is also observed in FCC lymphomas (see Section 3.6.3.1). A difficult but sometimes important task in assessing fluorescence intensity is to distinguish very weak (dim) positive staining from negative cells. This occurs frequently when evaluating surface light chain expression, especially in CLL. With regard to light chains, this difficulty can be resolved in most instances by including two different sets of kappa and lambda reagents in the FCM panel (Figure 3.15). Irrespective of the methods applied to describe fluorescence intensity, either quantitative or qualitative, the distinction between dim positive and negative staining ultimately involves arbitrary criteria. In general, the population is considered positive for an antigen if its histogram is shifted by a certain number of fluorescence channels

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

63

Figure 3.14 Peripheral blood with T-ALL. (a) Isotype-matched negative controls: The fluorescence signals are “shifted” close to 100. (b) Blasts comprise more than 90% of the cells in the sample. (c–f) The fluorescence of the isotype controls is taken into consideration for the grading of the intensities of the markers expressed on the blasts. Blasts are thus CD7+++, CD3+++, CD5++, CD4+ and CD8++. CD2 is present only on a small subset of the blasts. The intensity of CD34 is extremely heterogeneous, ranging from negative to bright.

64

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 3.15 Peripheral blood with CLL. Two sets of kappa/lambda reagents (from different manufacturers) used for evaluating monoclonality. (a, b) Weak lambda light chain detectable on the leukemic cells. (c, d) With the second set of reagents, no surface light chain is detected. Other typical features of CLL (downregulated CD20, CD20 weaker than CD19) can also be appreciated.

in comparison with an internal negative control. Regardless of the exact number of channels chosen to define the minimum shift arbitrarily, it is important that this shift can be clearly visualized when the histograms or the cluster displays are superimposed on one another (Figure 3.16). 3.1.3.1 Heterogeneous fluorescence intensity (bimodal, variable) In a benign cell population, the fluorescence intensity of most cell surface antigens is relatively uniform from one cell to another, thus yielding a rather compact cell cluster on the FCM dot plots (Figure 3.17). In a neoplastic proliferation, however, the quantitative and qualitative

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

65

Figure 3.16 Bone marrow with CLL/SLL. (a–d) The intensity of the positive light chain (kappa) is extremely weak, but it is still readily visualized especially on the overlay kappa/lambda histograms (gated on R1). Downregulation of CD20, variable fluorescence distribution of CD20 (starting in the negative region), and the small cell size are other features typical of CLL.

a

b

Figure 3.17 Benign lymph node. (a, b) Normal T- and B-cell populations with relatively homogeneous fluorescence distribution for CD2 and CD20, respectively. The benign lymphoid cells are of small cell size.

66

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

Figure 3.18 Peripheral blood with CLL. (a, b) The leukemic cells coexpress CD20 and CD5. CD20 expression is of variable distribution and the resulting cell cluster stretches over more than 1 decalog. A minute population of residual T-cells (CD2+) is present.

expression of certain antigens can vary widely from cell to cell. Because of this heterogeneous staining intensity, the malignant cell cluster may appear stretched out, emulating the shape of an ellipse or a “figure of eight” corresponding with variable (Figures 3.16c and 3.18) and bimodal (Figure 3.19) distributions, respectively. Such patterns of antigenic expression may serve as a diagnostic clue or fingerprint for follow-up purposes. It can be difficult to convey this useful, but complex graphical data to the clinical staff, who for the most part, are unfa-

a

b

Figure 3.19 An unusual case of FCC II lymphoma. (a, b) The tumor cell population is mostly of small cell size. Most unusual is the bimodal expression of CD20; it is markedly downregulated to absent in about half of the neoplastic cells, but well expressed in the other half. The latter display slightly dimmer and more variable CD10 expression.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

67

miliar with the complexities of FCM data and are mainly interested in the “bottom line.” The extent of the heterogeneity in the fluorescence staining, especially if bimodal, also makes it problematic how to report the fluorescence intensity in semiquantitative terms (e.g., weak, moderate, strong). In the authors’ experience, the brighter half of the bimodal distribution provides the more clinically relevant information in most instances, as shown in the following example. The fluorescence intensity reported is that of the brighter half, with the added term bimodal as a modifier. Example A case of precursor B-ALL demonstrated a bimodal distribution for CD34 (Figure 3.20), with the fluorescence intensity starting from the negative region (first decalog) and extending well into the positive weak region (second decalog). In this case, the clinically relevant information concerned the maturity status of the abnormal population. Therefore, CD34 was reported as “weak, bimodal.” Positive staining for TdT further confirmed that the tumor cells were blasts. Alternatively, the intensity of CD34 may be reported as “weak, hemipositive,” although this terminology does not clearly indicate whether the fluorescence intensity is bimodal or variable. It is important not to describe the data simply as “+/−” however, because this format has also been used to mean equivocal results.

Another aspect of heterogeneity is the situation where one or more antigens are found only on a subset of neoplastic cells (Figure 3.21). This subset may or may not be present in subsequent follow-up specimens. Until now, there has been no uniform system for describing or reporting fluorescence distributions. This, together with the lack of standard antibody panels across institutions, has made interlaboratory comparison of FCM data difficult. For clinical purposes (e.g., to monitor patients on whom FCM assays have been performed by more than one institution), review and comparison of the FCM list mode data is currently the most appropriate method for long-term follow-up. It is important to be aware that the appearance of a bimodal distribution in fluorescence intensity may represent two distinctive but closely related populations (e.g., some cases of AML with monocytic differentiation), instead of one single population, especially if the “figure of eight” has a narrow waist. This can be suspected when the same bimodal distribution is seen across several antigens or when the two subpopulations are clearly separate based on the coexpression of certain key markers (Figure 3.22). In such instances, it is more correct to characterize the overlapping halves of the bimodal distribution as two separate clusters: the population of critical cells (i.e., blasts, in the case of AML with monocytic differentiation) and a “related” population (i.e., maturing monocytic cells, which include promonocytes and monocytes). Using the approach of reporting the average fluorescence intensity based on the mean fluorescence peak channel, it can be difficult to achieve an adequate description of the fluorescence intensity when faced with a complex distribution. In such instances, the authors have sometimes employed the term variable as a modifier to the semiquantitative results to denote the heterogeneous staining pattern. Depending on the type of cells and the antigen expressed, the fluorescence signals may begin in the negative region (and/or span over two decalogs or more) (Figure 3.14c). A specific example is the pattern of CD20 expression in CLL/SLL. The appearance of the neoplastic cell cluster on the FSC/CD20 dot plot is virtually pathognomonic for this disorder (Figure 3.16c). In such cases, the intensity of CD20 is best described as “weak, variable.” Note that this pattern of CD20 expression can be altered if the analysis is performed after anti-CD20 therapy or if the manufacturers introduce changes (quantitatively or qualitatively) to the composition of any of their reagent cocktails containing CD20 antibody.

68

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.20 Bone marrow with precursor B-ALL. (a) Isotype-matched negative controls. (b) The blast population accounts for 90% of the cells in the sample. (c–e) Blasts are CD19++, CD10++, CD117− and CD34+ (bimodal). (f) The tumor cells are diploid and express TdT.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

69

Figure 3.21 Bone marrow from a 4-year-old child, virtually replaced by an AML with t(9;11)(p22;q23). The dot plots in (b–f) are gated on the blast region shown in (a). Leukemic cells coexpress CD13 and CD33. CD11b, CD14 and CD16 are absent. CD34 and CD64 are partly expressed in a bimodal distribution. CD15 is present only on a minor subset of the blasts.

70

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.22 AML with monocytic differentiation. (a) A large cluster with bimodal CD45 in the blast region. (b–f) The leukemic population is actually composed of two related components: Blasts and monocytic elements differing in the expression of CD34 (positive in blasts), CD64 (brighter in monocytes), CD14 (negative in blasts), CD71 (dimmer in blasts) and HLA-DR (brighter in blasts). The two subpopulations overlap in cell size and CD13/CD33 expression. CD117 (not shown) is also bimodal.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

71

3.2 Fluorescence dynamic range As alluded to in Chapter 2, the graphical pattern of reactivity to certain antibodies or antibody combinations is extremely useful for discriminating cell populations with similar antigenic profiles. In certain diseases, the fluorescence intensity on the malignant cells is different from that on the normal counterparts, or the distribution of the antigen density (and thereby the fluorescence intensity) within the abnormal population is more heterogeneous than the normal cells, resulting in a distinctively shaped cluster. To take advantage of such useful features, it is important to know the range of intensity of a given marker on different hematopoietic elements. In general, a marker with a wide dynamic range (e.g., CD45) offers good discrimination between different cell populations and therefore is most useful diagnostically, especially if it is also lineage-associated (e.g., CD20). For example, CD45 together with the SSC parameter permits the separation of the major cell subpopulations in the bone marrow. Therefore, CD45 serves as the anchor-marker in the BBS panel. An advantageous feature of CD20 is that its expression on malignant B-cells is usually distinctive from that on the benign counterparts, whereas the density of CD19 on benign and neoplastic B-cells is similar. On dot plots correlating CD19 with other parameters such as FSC, benign and malignant B-cells may overlap and form a single cell cluster. In contrast, with CD20, one often sees the benign and malignant B-cell populations as separate clusters, thus facilitating gating procedures for the identification of other antigens (especially surface light chains) on the cells of interest. Because of this characteristic, CD20 is the marker of choice for pairing with other antibodies to further characterize B-cell neoplasms. In a small number of cases, the discriminatory function is demonstrated by CD19 instead. Therefore, one approach to take advantage of this feature is to use two different sets of light chain reagents: one set paired with CD20, the other with CD19 (Figure 3.23). This approach has proved to be extremely helpful in diagnosing and classifying mature B-cell neoplasms.

3.3 Strategy to the visual review of FCM immunophenotyping data Visual FCM data analysis can be performed most efficiently if the inspection of the FCM graphics follows a sequence based on medical logic. The organization of the dot plots in the FCM analysis panels (see Section 2.9.1) impacts on the efficiency of this process. For samples derived from the blood, bone marrow, or spleen (i.e., the BBS panel), the following strategy has proved to be helpful to the authors. Visual inspection begins with the SSC/CD45 dot plot to determine if an overt blast cluster is present and to decide whether or not the sample is composed of multiple cell clusters or overrun by a single, and thereby abnormal, population. The following discussion pertains to the latter situation, in which the dot plots are easier to evaluate than their complex counterparts in specimens harboring neoplastic cells admixed with several other cell types (see Chapter 4). If the shape and location of the abnormal cell cluster on the SSC/CD45 display suggest that it is a blast population (Figure 3.24a), then the visual evaluation is focused on the dot plots that display CD34, CD117, myeloid antigens (namely CD13 and CD33), HLA-DR, CD19, CD10, CD58, T-cell antigens, and if necessary, TdT. TdT is mainly evaluated in T-cell lymphoblastic lymphoma–leukemia because T-lymphoblasts express CD34 at a much lower frequency than their precursor-B counterparts.

72

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 3.23 Coexisting benign and neoplastic B-cells in a case of FCC lymphoma. (a, b) These two populations are well separated based on CD20 expression. The population with weaker CD20 expression is polyclonal (thin arrow) whereas the one with brighter CD20 is monoclonal for kappa (arrow). (c, d) The benign and neoplastic B-cells cannot be separated from each other based on CD19 expression.

If the appearance of the abnormal cell cluster suggests a mature lymphoid disorder (Figure 3.24b), then the cell size (FSC) should be evaluated next. This parameter is of greater importance in mature lymphoid malignancies than in acute leukemia because of its usefulness in subclassifying LPD/NHL. The evaluation continues with the inspection of the light chain expression, along with that of CD19, CD20, and other key antigens (e.g., CD5, CD10, CD23, CD103) necessary for characterizing the tumor. If the light chain expression is polyclonal, then the T-cell and NK-cell markers are evaluated instead. For solid tissue or body fluid samples (i.e., the TF panel), the visual analysis relies more on the FSC/SSC than the SSC/CD45 dot plot because, in most instances, the cellular elements are composed primarily of mature lymphoid cells, irrespective of whether the population is homogeneous or heterogeneous. The visual inspection follows the sequence described above for LPD/NHL.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

73

b

Figure 3.24 (a) Peripheral blood with acute leukemia. The prominent cell cluster in the blast region accounts for 80% of the cells in the sample. (b) Peripheral blood with a B-cell LPD. The specimen is composed almost entirely of lymphoid cells.

3.4 Common SSC/CD45 patterns The SSC/CD45 dot plot is most useful for identifying blast populations. Although some of the patterns are virtually pathognomonic for certain acute leukemias, further corroborating evidence from the remaining FCM data is necessary for the final interpretation. Normal lymphocytes, which serve as an internal control, are invariably present as a cluster with very low SSC and bright CD45.

3.4.1 Assessment of the blast population

Case study 1

The blast population most often occupies the position shown in Figures 3.14b, 3.24a, 3.25a and 3.26a with SSC in the low to moderate range, and CD45 of weak to moderate intensity. CD45 expression on blasts is invariably one decalog or more weaker than that in normal lymphocytes. There are rare instances of precursor T-cell ALL in which the leukemic blasts express bright CD45 at a level similar to that of lymphocytes. The resulting SSC/CD45 pattern with a predominant cluster located in the lymphocyte region thus mimics that seen in mature lymphoid malignancies. Conversely, in some high-grade LPD/NHL or plasma cell neoplasms, the intensity of CD45 is similar to that seen in acute leukemia, so as to produce a cell cluster occupying the blast region (see Section 4.4.3). Because the SSC signals correlate with internal cytoplasmic complexities, it can be inferred that a blast cluster with low SSC corresponds morphologically with blast cells with very scant cytoplasm (Plate 10). In most instances, blasts with very low SSC prove phenotypically to be ALL (Figures 3.25 and 3.26). Some AMLs may display an SSC/CD45 similar to that associated with ALL (Figure 3.27). A substantial increase in normal B-cell progenitors is a common occurrence in regenerative bone marrow, especially that from pediatric subjects. The expansion of these cells does not overrun other normal hematopoietic elements, however. Therefore, the possibility that normal precursor B-cells could manifest as the only cell cluster seen on the SSC/CD45 display is essentially nonexistent. The differential diagnosis of normal B-cell progenitors versus leukemic blasts may need to be considered, however, if the various bone marrow cell populations

74

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.25 Bone marrow with precursor B-ALL. (a) The blast cluster accounts for 92% of the cells in the sample. (b–e) Blasts display medium cell size, weak CD34 with a bimodal distribution, and bright coexpression of CD10 and CD19. CD33 is negative. (f) TdT is also expressed.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

75

Figure 3.26 Precursor B-ALL. (a) Blasts are CD45+. A small subset shows downregulated CD45. (b–f) Gated on blasts: CD19, CD10 and CD58 are brightly coexpressed. CD13 and CD117 are negative. A subset of blasts express CD20 and CD34. CD33 is “aberrantly” expressed in a heterogeneous distribution.

76

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.27 Peripheral blood with AML-M0. (a) The blast population displays very low SSC and heterogeneous CD45 expression; the SSC/CD45 pattern is similar to that associated with ALL. (b–f) Ungated data to include the small populations of residual monocytes (M) and lymphocytes (L). Blasts comprise 85% of the cells in the sample. CD13, CD33, CD34, CD117, HLA-DR and CD7 are well expressed.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

Case study 2

Case studies 3 and 4

77

are well represented on the SSC/CD45 dot plot and a small but conspicuous cluster with very low SSC is seen in the blast location. When evaluated in the context of positive myeloid antigen expression, the shape of the blast cluster on the SSC/CD45 dot plot can be an important clue to subclassifying AML. For this purpose, the bone marrow sample must be representative (i.e., minimally contaminated with granulocytes and monocytes from the peripheral blood). The finding of a prominent blast cluster (expressing myeloid antigens) and virtually no other cell populations on SSC/CD45 indicates AML with minimal evidence of maturation, which can be either AML-M0 or AML-M1 depending on the results of myeloperoxidase cytochemistry (Plate 11). A virtually single major population with weak to moderate CD45 and SSC signals encompassing a wide range (180 to 800) from moderate to high is strongly suggestive of AML-M3 (Figures 3.11 and 3.28). The bone marrow in two other conditions may occasionally produce a similar appearance on the SSC/CD45 dot plot (Figure 3.29): (1) CML or CML-like MPD and (2) a vigorous response to granulocyte colony-stimulating factor (G-CSF) therapy (see Section 4.5). These can be easily distinguished from AML-M3 based on the remaining FCM graphics such as the pattern of the AML-M3 cell cluster on the CD13/CD16 (or CD13/CD11b) and CD11b/CD16 dot plots (Figures 3.11, 3.28 and 3.29). The pattern of CD15/CD34 expression, in which the neoplastic cells lose CD34 prior to acquiring low levels of CD15, has also been reported as another characteristic of AML-M3. Other additional corroborating evidence for promyelocytic leukemia includes the absence of reactivity for HLA-DR, coexpression of CD13 and CD33, and heterogeneous CD13 (as compared to CD33) intensity (Figures 3.11 and 3.28). The constellation of these features correlates highly with t(15;17) and the PMLRARα gene rearrangement, as well as the morphology of AML-M3 blasts. In AML-M3v, the shape of the cluster on the SSC/CD45 dot plot depends on the degree of hypogranularity (i.e., reduced SSC signals). The morphology in most cases displays a variable degree of hypogranularity along with a small proportion of typical M3 blasts and possibly some evidence of maturation to the later myeloid stages. The resulting shape of the blast population on the SSC/CD45 dot plot can be that of a triangle or a teardrop whereby the bulk of the cell cluster represents hypogranulated blasts (Figure 3.30). Typical AML-M3, when relapsed after ATRA (all-trans-retinoic acid) therapy, may also manifest as M3v. In unusual cases of M3v composed of markedly hypogranular to agranular blasts, the appearance of the blast cluster can be indistinguishable from that seen in AML with minimal maturation. In addition, CD34 and HLA-DR may be expressed, but the expression of these antigens is of weak intensity and variable distribution (Figure 3.31). Other “aberrancies” such as weak/partial expression of CD11b, CD117 or CD2 occur much more rarely. In AML with monocytic differentiation (i.e., AML-M4 and M5), the shape and location of the blast cluster on the SSC/CD45 dot plot depend on the degree of monocytic differentiation and the relative content of the mature monocytic elements. The type of specimen (i.e., bone marrow vs. peripheral blood) also needs to be taken into account because the neoplastic population in the blood may be seen as two distinct clusters (blasts plus monocytes), whereas the corresponding bone marrow yields a single merging cluster. The appearance is usually that of a large cluster shifted upward and to the right (i.e., toward the monocytic region) (Figure 3.32b). Alternatively, the leukemic cells can form a widened cluster extending from the blast region into the monocytic region, thus displaying a variable CD45 expression (Figure 3.32c). If the analysis is on a representative bone marrow sample and the neoplastic cluster displays bimodal CD45 (i.e., merging of the cellular events from the blast region and monocytic regions), then it is important to ensure that the tumor is not actually composed of two separate populations (see Section 3.1.3.1). Occasionally, the internal complexities of the malignant cells in AML-M5 are sufficient to generate high SSC signals, in which case the shape and location of the tumor cell cluster can

78

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.28 Phenotypic characteristics of AML-M3. (a) Isotype-matched negative controls showing increased baseline fluorescence. (b–f) The sample is composed of virtually a single predominant blast cluster, with high SSC and lack of both CD34 and HLA-DR. CD11b, CD16 and CD15 (not shown) are also negative. The intensity of CD13 (c) is relatively more heterogeneous than that of CD33 (d).

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

79

Figure 3.29 Conditions that may produce a bone marrow SSC/CD45 pattern similar to that of AML-M3. (a, b) G-CSF effect; (c, d) CML-like MPD; (e, f) AML-M3 (see also Figure 3.28). The single predominant cluster of myeloid precursors (a, c) mimic the appearance of the AML-M3 blast cluster. There is reactivity for CD16 and CD11b (b, d). In contrast, CD16 and CD11b are negative in AML-M3 (f).

80

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

Figure 3.30 Bone marrow with AML-M3v from two different patients. (a, b) A predominant blast cluster with moderate SSC.

a

b

c

d

Figure 3.31 Peripheral blood with agranular AML-M3v (morphology shown in Plate 12). This case displays several unusual features: a relatively low SSC (a), expression of CD34 with a variable distribution best appreciated in (d) and variable expression of HLA-DR in a small subset of the leukemic cells. Other typical features of AML-M3, namely heterogeneous CD13 (b) and homogeneous CD33 (c) are retained, however. The t(15;17) and PML-RARα gene rearrangement were demonstrated in this case.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

81

mimic that observed in AML-M3 (Figures 3.32d and 3.33). The remaining FCM data, however, especially the reactivity of the malignant cells to HLA-DR, CD11b, CD14, CD64, and CD56, provides additional evidence of monocytic differentiation (see Section 3.5.1). Cytochemical staining for nonspecific esterase (NSE) may also be added to the immunophenotyping results if there is a need for subclassifying an AML with monocytic differentiation into either AML-M4 or M5 according to the previously published criteria.

3.4.2 Immature neoplastic cells with downregulated CD45 CD45 is markedly downregulated or negative in a substantial number of acute leukemias, many of which are precursor B-ALL. On the SSC/CD45 dot plot, the blast population is seen

a

b

c

d

Figure 3.32 Various SSC/CD45 patterns in AML with monocytic differentiation from four different patients. (a) Peripheral blood: The neoplastic population with bimodal CD45 is formed by the merging and overlapping of the blast and monocytic clusters. (b) Bone marrow: The blast cluster is shifted toward the monocytic region. (c) Bone marrow: The leukemic population spans from the blast region into the monocytic region. (d) Bone marrow: The appearance of the blast cluster mimics that seen in AML-M3v.

82

FLOW CYTOMETRY IN HEMATOPATHOLOGY

Case study 5

in the far-left position (Figure 3.34). Other cell types can also appear in this location, however, including erythroid precursors, plasma cells (Figure 3.35), and nonhematopoietic metastases. Occasional high-grade mature B-cell neoplasms (especially those with plasmablastic differentiation) or rare mature T-cell LPD/NHL may also exhibit downregulated CD45. Large numbers of unlysed red cells manifest as a prominent cluster in the CD45-negative region; these should have been excluded from the data acquisition based on a quick screening of the cytospins prior to running the samples on the flow cytometer, however. A prominent cluster with low SSC and a bimodal CD45 distribution spanning over two decalogs or more from the negative into the positive range is an uncommon finding. This may represent one of the following: • The merging of two distinctive populations. The cluster on the left with markedly downregulated CD45 may be composed of erythroid precursors, and the cluster on the right may be composed of

a

b

c

d

Figure 3.33 Two unusual cases of AML with monocytic differentiation. The SSC is unusually high (a, c), resulting in a picture reminiscent of AML-M3. CD11b and CD15 are brightly expressed in the first case (b), whereas CD56 is positive in the second (d).

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

83

Figure 3.34 Bone marrow with CD45-negative precursor B-ALL. Blasts are of medium cell size (d) and display markedly downregulated CD45 (a), and bright coexpression of CD19 and CD10 (c). CD34 is weak and of variable distribution (b). An unusual feature is the bright expression of CD20 (d).

Case study 6

blasts. Alternatively, they may represent two different clones of blasts (e.g., ALL and AML) with similar SSC characteristics. • One single neoplastic population with bimodal CD45. An occasional ALL or AML may display this pattern (Figure 3.36).

3.4.3 SSC/CD45 in mature lymphoid disorders The SSC/CD45 dot plot is of limited usefulness in mature lymphoid disorders because different types of NHL/LPD generate closely similar SSC signals and CD45 intensities. In specimens with extensive involvement by any NHL/LPD, the usual finding on the SSC/CD45 dot plot is a prominent cluster with low SSC and moderate to strong CD45 expression (Figure 3.24b). Although the intensity of CD45 on the neoplastic cells is closely similar to that of normal lymphoid cells, it may be slightly downregulated and the distribution more heterogeneous. This feature is most commonly seen in CLL, with the neoplastic cells forming an elongated cluster occupying the region of the bone marrow B-cell progenitors (Figure 3.37).

84

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 3.35 Plasma cell leukemia (morphology shown in Plate 48). (a) A predominant neoplastic cluster in the CD45-negative region. (b–d) The tumor is positive for CD56 and cytoplasmic kappa.

a

b

Figure 3.36 Precursor B-ALL with bimodal CD45. (a, b) Two subpopulations of blasts: The major one is CD45-negative, the minor one CD45-positive. Some normal B-cells (arrow) and T-cells (open arrow) are present.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

85

Figure 3.37 Bone marrow with CLL. (a) A predominant cluster in the hematogone region, with CD45 less intense than that on normal lymphocytes. (b–f) The leukemic cells display the typical phenotypic features of CLL: (1) small cell size; (2) weak and heterogeneous CD20 expression; (3) CD20 less intense than CD19; (4) dim lambda-positive surface light chain (f); and (5) reactivity for CD5 and CD23. In this case CD23 is downregulated and of variable distribution. HLA-DR is expressed.

86

FLOW CYTOMETRY IN HEMATOPATHOLOGY

The light scatter characteristics and CD45 intensity of hairy cells, as well as neoplastic cells in a substantial number of large cell lymphomas, are similar to those of monocytes (see Sections 3.6.3.2 and 3.6.3.4). Thus, in extensive involvement of the bone marrow or blood, the resulting cell cluster seen on the SSC/CD45 dot plot may simulate the SSC/CD45 picture of a neoplastic monocytic disorder. A small number of aggressive LPD/NHL may lack detectable CD45 (see Section 3.6.3.4).

3.5 Other dot plot patterns useful in acute leukemia diagnosis The correlated displays CD13/CD34, CD33/CD34, or CD117/CD34 are useful for quantifying the proportion of leukemic myeloblasts (Figures 3.22 and 3.27). Similarly, the proportion of precursor B-ALL blasts can be derived from the CD19/CD34 dot plot (Figures 3.20, 3.25, and 3.34), CD10/CD58 (Figure 3.26), or TdT/CD19, especially if the blasts are CD34-negative. In most T-ALLs, the blast proportion can be determined based on TdT/CD7 coexpression or, alternatively, CD4/CD8 coexpression (Figure 3.14). In a substantial number of AMLs, especially those with monocytic differentiation, the blasts are negative for CD34. Many of these demonstrate reactivity for CD117, however. In contrast, the blasts in ALL typically lack CD117.

3.5.1 Useful antigenic features in AML In the context of a prominent blast population with myeloid antigen expression, the results of other markers such as HLA-DR, CD14 (when correlated with CD64), CD16, or CD56 can be helpful for subclassifying AML. Although the blast population may demonstrate a bimodal or variable distribution for CD34, CD13, or CD33, this feature does not add much to the subclassification of AML. The diagnostic importance of HLA-DR lies in its lack of expression, typically seen in AML-M3 blasts. In contrast, blasts in non-M3 AML express moderate to very bright levels of HLA-DR. Two potential pitfalls are to be avoided, however:

Case study 7

1. Absence of HLA-DR can occur in some cases of AML-M1 (Figures 3.38 and 3.39). The phenotypic profile is similar to that of AML-M3 in that the neoplastic cells are negative for both HLA-DR and CD34. However, the shape of the blast cluster on the SSC/CD45 dot plot is not that associated with AML-M3 (Figures 3.11, 3.28, 3.30, and 3.31). The relationship between the CD13 and CD33 reactivities (both markers being conjugated to same type of fluorochrome) is also different, whereby CD13 and CD33 are either expressed at similar levels, or CD13 is at lower intensity (i.e., downregulated to absent) than CD33. This pattern is essentially the reverse of that observed in AML-M3, where the expression of CD13 is characteristically brighter and more heterogeneous than CD33. Furthermore, these unusual cases of AML-M1 also lack the cytogenetic abnormalities of AML-M3 and do not respond to ATRA therapy. 2. HLA-DR expression may occur in AML-M3. These are often AML-M3v, in which the blasts are severely hypogranular or agranular (Plate 12) and, therefore, may be morphologically mistaken for non-M3 blasts or lymphoma cells. However, only a subset of the M3v cells express HLA-DR, and the intensity of HLA-DR is distinctively weaker (Figure 3.31) than that seen in non-M3 blasts. Other phenotypic features such as a lack of expression of CD117, CD11b, and CD16, plus the cytogenetic abnormalities and biological behavior are otherwise identical to typical AML-M3.

The CD14/CD64 dot plot is useful for determining monocytic differentiation. There exist noticeable differences in reactivity between different clones of anti-CD14 antibodies. The observations described here pertain primarily to the Leu-M3 clone of anti-CD14, which labels predominantly mature monocytes. Some of the other clones of anti-CD14 also label myeloid cells, but at a lesser level of reactivity than that seen in monocytes. Similarly, CD64 expression

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

87

Figure 3.38 An unusual case of AML-M1. (a) A single predominant cluster in the blast region with low SSC. (b–d) Blasts are CD13++, CD33++ and negative for CD34. HLA-DR is minimally present on a tiny subset of the blasts. Cytogenetics and molecular genetics did not reveal a PML-RARα rearrangement.

is less intense on granulocytes than monocytes. On the correlated dot plot, mature monocytes appear as a homogeneous cluster with bright CD14 and CD64 (Figure 3.40). In some cases of chronic myelomonocytic leukemia (CMMoL), the circulating abnormal monocytic cells overrun the granulocytic elements such that the resulting SSC/CD45 dot plot of the peripheral blood displays a single major cluster in the monocytic position (Figure 3.41) (i.e., a picture similar to that of AML with monocytic differentiation [Figures 3.32 and 3.42]). Conversely, residual/relapsed AML with monocytic differentiation may contain a substantial number of myeloid cells, and the resulting SSC/CD45 of the bone marrow can thus mimic that seen in CMMoL. In such instances, the pattern of CD14/CD64 reactivity is a clue to distinguish these two different disorders. The neoplastic population in CMMoL is composed mostly of mature elements coexpressing bright CD14 and CD64 (see Section 4.5.1.1). In contrast, the coexpression of CD14 and CD64 on neoplastic cells in AML with monocytic differentiation often manifests as a diffuse “trailing” cluster (Figure 3.42). This vertical “trail” pattern reflects the marked heterogeneity in CD14 expression (ranging from negative to bright)

88

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.39 AML-M1. Blasts display low SSC (a) and no increased autofluorescence (b). (c–f) The phenotypic profile, with coexpression of CD13, CD33 and CD117, but lack of CD34, HLA-DR, CD11b, CD16 and CD15, mimics that of AML-M3. Note that CD13 is less bright than CD33, whereas the reverse is observed in M3. Molecular and cytogenetics showed no PML-RARα translocation.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

89

Figure 3.40 Coexpression of CD14 (Leu-M3) and CD64 on monocytes. (a, b) Normal bone marrow with a small population of monocytes forming a relatively compact cluster positive for CD64 and CD14 (thin arrow). Myeloid precursors (open arrow) are CD14− and CD64+. There is increased baseline fluorescence (not shown) in this case. (c, d) Peripheral blood with AML-M4. Distinct blast (arrow) and monocytic clusters. (d) Mature monocytes (thin arrow) are prominent and coexpress bright CD14 and CD64. A small subset of the monocytic elements display variable CD14 expression, seen as a short illdefined trail attached to the CD14/CD64 positive cluster. Blasts are CD14− and CD64−.

and thereby the variable degree of monocytic differentiation/maturation among the neoplastic cells. This, in turn, translates morphologically as a spectrum of blasts, promonocytes, and monocytes (Plate 13) in which the more immature elements predominate. The most immature cells often produce a loosely formed cell cluster located at the lower end (i.e., CD64 bright, CD14 negative or markedly downregulated) of the vertical trail (Figure 3.42d). In some cases, the entire tumor population lacks CD14 expression (Figure 3.43). The morphologic picture in the latter instance is usually that of AML-M5a. In others, a third pattern is observed whereby the blast and monocytic components (although seen together as one cell population on the SSC/CD45 dot plot) separate from each other into (1) a blast cluster negative for CD14 and CD64 and (2) a CD14/CD64-positive “trail” of monocytic elements (Figure 3.42a,b). Such

90

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 3.41 Peripheral blood with CMMoL. (a) A single predominant cluster in the monocytic region. Granulocytes (G) are markedly reduced. (b) Brighter CD33 on monocytes (M) than on granulocytes. (c, d) Gated on the monocytic cluster: the neoplastic monocytes are CD64++, CD14+++ and CD56++. An unusual finding is the presence of a small subpopulation with weak CD16 reactivity, as well as lower levels of CD14 and CD64.

Case studies 8 and 9

cases may be referred to as AML with a monocytic component. The terminology “AML with monocytic differentiation” and “AML with a monocytic component” can be used interchangeably, however, since these two subgroups share a similar biological behavior. The FCM diagnosis of AML with monocytic differentiation should be based on a representative bone marrow aspirate rather than a blood specimen because (a) the proportion of blasts in the peripheral blood is often much lower than the arbitrary level currently accepted for the diagnosis of acute leukemia, and (b) the circulating monocytic elements are more mature than those residing in the marrow. In such cases, the peripheral blood picture may be misinterpreted as “advanced” CMMoL (Figures 3.44 to 3.46). CD56 expression may occur in acute leukemias, either of lymphoid or myeloid lineage. In myeloid leukemias, CD56 reactivity tends to be associated with monocytic differentiation (Figure 3.43f). This feature is also present in other abnormal monocytic proliferations such (text continues on page 95)

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

91

Figure 3.42 AML with monocytic differentiation. Case 1: (a) Merging blast (B) and monocytic (M) clusters. (b) Blasts are CD14− and CD64−. Monocytic cells display variable CD14 expression (CD14/CD64 trail). Case 2: (c) A prominent cluster in the monocytic region. Some residual granulocytes (G) are present. (d) CD14/CD64 trail. The majority of the monocytic cells are CD14− and CD64++, which correspond with the most immature cells. Case 3: (e) A predominant cluster extending from the blast to the monocytic region. (f) The blast component is partially CD64 positive and lacks CD14. There is a large proportion of more mature elements in the monocytic component.

92

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.43 AML with monocytic differentiation (morphology shown in Plate 27). The SSC/CD45 picture (a) is similar to that of CMMoL as the blast cells occupy the monocytic region and display moderate SSC. (b–f) Blasts are negative for CD34, CD117, CD14 and CD16. CD4, CD13, CD64 and CD56 are expressed. Residual granulocytes are CD13++ and CD16++.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

93

Figure 3.44 AML-M5. The leukemic cells produce a “CMMoL-like” SSC/CD45 pattern in the peripheral blood (a) with a prominent monocytic cluster and decreased granulocytes. In the bone marrow (b) of the same patient, the SSC/D45 is more diagnostic of AML-M5. CD15 is more heterogeneous in the peripheral blood (e). The leukemic cells from both sites display similar reactivities for CD4 (c), CD56 (d) and CD33 (e, f).

94

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.45 AML-M5 (continuation of Figure 3.44). The expression of CD13, CD11b and CD14 on the leukemic cells in the bone marrow (b, d, f) is much more heterogeneous than those in the peripheral blood (a, c, e). The intensities of CD13, CD11b and CD14 correlate inversely with the maturation of the monocytic cells. The peripheral blood contains a lower proportion of the immature elements (e).

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

95

Figure 3.46 AML-M5 (continuation of Figure 3.45). Bone marrow DNA analysis with DRAQ5, gated on CD14+ CD45+ cells (a). (b) High proliferative activity from the more immature monocytic cells with lower CD14 expression. (c) Gated on R3 (more mature cells): the S-phase is 3%. (d) Gated on R2: the S-phase is 19%.

Case study 10

as CMMoL (Figure 3.41). Because these phenotypic abnormalities are by themselves not diagnostic of abnormal monocytic proliferations and may be seen in other non-M3 AMLs (Figure 3.47), the data need to be evaluated in the context of other immunophenotypic findings. CD16 and CD11b are normally present on late myeloid cells and may be expressed in AML. Although it is not possible to correlate the presence or absence of CD11b and/or CD16 with any particular subtype of AML, it is of interest to note that both CD16 and CD11b are absent in AML-M3 (Figures 3.28 and 3.29e,f).

3.5.1.1 Myeloid phenotypic abnormalities and MRD detection In comparison with ALL, the evaluation of MRD in AML is more challenging and requires a larger panel of antibodies. More than one clone of AML cells may be present at the time of diagnosis, and the residual/relapsed disease can consist of any of the original clones. Changes in the antigenic expression, associated with long-term clonal evolution, may also complicate the detection of early relapse.

96

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 3.47 AML-M1 with CD56 expression. (a) Virtually a single predominant cluster with low SSC in the blast region. (b–d) Blasts are CD34−, CD13++, CD33++, CD16− and CD56++. The distribution of CD56 is bimodal. The results of other markers (not shown) include reactivity for HLA-DR and CD117, but lack of CD14 and CD64.

A substantial number of AMLs exhibit antigenic abnormalities at the time of diagnosis, based on which custom-built panels can be designed for the analysis of follow-up specimens. The frequently encountered abnormalities include the following:

Case study 11

• Reactivity for a B- or T-cell associated marker. • Expression of a late myeloid antigen (e.g., CD11b, CD14, or CD15). When coexisting with CD34, such antigenic expression is considered asynchronous. • Lack of either CD13 or CD33 (Figures 3.13 and 3.48). • Absent HLA-DR (Figures 3.38, 3.39 and 3.48). • Markedly downregulated to absent CD45. • Coexpression of CD56. A rare case of AML with monocytic differentiation may lack both CD13 and CD33 and demonstrate reactivity for only CD14 and CD56. • Abnormal light scatter characteristics, such as low FSC and SSC seen in the rare AML with basophilic differentiation.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

97

Figure 3.48 AML with minimal differentiation. Neoplastic myeloblasts display low SSC (a), heterogeneous CD117 (b) and several antigenic abnormalities, including downregulated CD13 (c) and lack of HLA-DR (f). CD33 is well expressed (d). Dim CD11b is present on a subset of blasts (f). CD34, CD16 and CD14 are absent. The phenotypic profile of HLA-DR−, CD34− superficially resembles that of AML-M3.

98

FLOW CYTOMETRY IN HEMATOPATHOLOGY

3.5.2 Precursor B-ALL versus bone marrow B-cell progenitors Normal B-cell progenitors (morphologically referred to as hematogones) in the bone marrow and precursor B-ALL blasts share similar phenotypic features, namely TdT, CD19, and CD10 as well as CD58. At the time of diagnosis, most cases of precursor B-ALL present with an overwhelming proportion of immature B-cells with the resulting suppression of other hematopoietic elements in the bone marrow. This feature is, by itself, sufficient to indicate that the proliferation is neoplastic. In the follow-up bone marrow specimens, however, the distinction between residual/relapsed precursor B-ALL and normal B-cell progenitors may not be straightforward, especially because these two populations may coexist. On the other hand, the presence of precursor B-cells in the peripheral blood or extramedullary sites is virtually an indication of a malignant process. Evaluation of one or more of the following parameters may help to distinguish precursor B-ALL from a proliferation of normal B-cell precursors.

Case study 12

Case studies 13 and 14

1. DNA aneuploidy. This parameter, when present, indicates the neoplastic nature of the cell population. The analysis can be done in conjunction with TdT (Figure 2.19d). 2. The pattern of coexpression of TdT and CD19 (Figure 3.49). On the correlated TdT/CD19 dot plot, ALL cells form a single compact cluster. In contrast, benign precursor B-cells manifest as two separate clusters, one expressing TdT (and also CD34) and the other lacking TdT. 3. The pattern of CD10 and CD20 coexpression (Figure 3.49). The levels of CD10 on the leukemic blasts of most precursor B-ALLs are higher than those on bone marrow B-cell progenitors. The latter also display markedly heterogeneous CD20 intensity, ranging from negative to bright, thus producing a “trail” pattern on the CD10/CD20 dot plot. The CD10/CD20 trail is J-shaped (Figure 3.49b) whereby the less immature B-cell progenitors have brighter CD20 but weaker CD10 than the more immature ones. In contrast, the majority of precursor B-ALLs lack CD20 expression. Among those expressing CD20, the distribution of CD20 on the leukemic blasts is less heterogeneous than that seen on hematogones (Figure 3.34). A sporadic case may display a CD10/CD20 pattern nearly mimicking that of normal B-cell precursors, however (Figure 3.50; Plate 4). 4. Light scatter characteristics. Bone marrow B-cell progenitors produce lower SSC and FSC signals than ALL cells (see Section 4.1.2.2). 5. CD38 expression (Figure 3.51). The intensity of CD38 on normal B-cell precursors is one of the brightest among hematopoietic precursors. By comparison, CD38, although bright, is less intense in most precursor B-ALLs. 6. The patterns of CD58/CD19, and CD58/CD10 coexpression. Leukemic blasts demonstrate much higher levels of CD58 than bone marrow B-cell progenitors (Figure 3.52). Because precursor B-cell ALLs also display brighter CD10, the combination of CD10 and CD58 is a useful discriminating feature for the detection of MRD, particularly in those instances where benign and neoplastic precursor B-cells coexist (see Section 4.3). 7. Lack of CD10 expression. This feature is seen in a small number of precursor B-ALL, most often in infants. There is a high association with the (4;11) translocation (Figure 3.53). In some patients, the expression of CD10 on the leukemic blasts in the initial sample is bimodal/heterogeneous, ranging from negative to bright. At relapse, however, the leukemic population is often more homogeneous, either with distinct CD10 expression, or lacking CD10 completely. This apparent “phenotypic shift” is not a common phenomenon. 8. Aberrant expression of other lineage-associated markers. Expression of myeloid antigens, usually in a subset of the leukemic blasts, is a common finding in precursor B-ALL.

In some cases, the distinction between malignant and normal precursor B-cell populations may not be possible, such as in a vigorously regenerative marrow where the residual/relapsed blast cells are extremely rare (at the 10−4 limit of FCM detection) and masked by an overwhelming number of normal B-cell progenitors. Such instances pose a serious dilemma that can only be resolved by close clinical follow-up and repeat analysis.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

99

Figure 3.49 Bone marrow B-cell precursors versus precursor B-ALL. (a, b) Gated on the hematogone and lymphocyte regions of the SSC/CD45 dot plot (not shown) of a normal pediatric bone marrow. Precursor B-cells (thin arrows) separate into two subpopulations: one positive for TdT and the other negative. CD20 expression is markedly heterogeneous. The brightest CD20 and dimmer CD10 are seen on the least immature B-cell progenitors. (c, d) Precursor B-ALL (ungated data). The leukemic cells (arrow) coexpress CD19 and TdT. CD20 is negative. A minute population of hematogones (thin arrow) is present.

3.5.3 Useful antigenic features in precursor T-lymphoma/leukemia In general, malignant immature T-cells attempt to replicate the phenotypes seen on thymocytes in their different stages of maturation and differentiation. Therefore, in most precursor T-cell ALL, the blasts either lack CD4 and CD8 or coexpress both of these antigens. They also lack surface CD3 but contain cytoplasmic CD3. The expression of CD2 and CD5 varies from case to case. Antigenic aberrancy such as coexpression of a myeloid antigen (e.g., CD13 or CD33) is not uncommon (Figures 3.54 and 3.55). Occasionally, the phenotypic profile may be indistinguishable from that of a post-thymic T-cell neoplasm, in that CD3 is well expressed

100

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.50 Precursor B-ALL in a 75-year-old patient. A prominent blast cluster with heterogeneous CD45 (a), and coexpression of CD22 and TdT (b). The FSC is heterogeneous and reveals two subpopulations of blasts (c, d). Blasts with the highest FSC are also brightest for CD45. (e) Gated on R3: heterogeneous CD20 on the smaller blasts. (f) Gated on R2: similar CD20 staining pattern on the larger blasts.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

101

b

Figure 3.51 Precursor B-ALL (a) versus bone marrow B-cell progenitors (b). (a) The ALL cells are CD45− and CD38++. (b) CD38 expression on bone marrow B-cell precursors (R2) is brighter than that on ALL cells. Note: The data in (b) is gated on MNCs.

Case study 15

(Figure 3.56), or the blasts are either CD4+ CD8− or CD4− CD8+. Furthermore, CD10 and CD56 can be found in both immature and mature T-cell malignancies. The key to distinguishing precursor T-cell from post-thymic malignancy relies on the expression of CD1, CD34 and/or TdT.

a

b

Figure 3.52 Precursor B-ALL (a) versus bone marrow B-cell progenitors (b). (a) The leukemic blasts (R1) coexpress bright CD19 and CD58. (b) Normal pediatric bone marrow (ungated data) with a conspicuous population of B-cell progenitors (R2) expressing lower levels of CD58.

102

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.53 Precursor B-ALL in a 5-month-old infant. (a) Leukemic blasts with low SSC. (b–e) Gated on the blast region: Leukemic cells lack CD20 and CD10, but coexpress bright CD58, CD19, CD34 and TdT. CD15 is heterogeneously expressed in a subset. (f) The tumor is diploid with a low S-phase.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

103

Figure 3.54 Precursor T-ALL. A prominent blast cluster with low SSC (a) and heterogeneous CD34 expression (e). (b–f) Leukemic blasts lack CD3, CD5, CD4 and CD8. CD2 and CD7 are present. CD13 is “aberrantly” coexpressed. CD33 and HLA-DR are negative.

104

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.55 Precursor T-ALL (continuation of Figure 3.54). Leukemic blasts display aberrant CD117 expression (b), but are negative for CD10 (a) and CD56 (c). (d–f) Reactivity for cCD3 and TdT, and lack of MPO confirm that the leukemic cells are precursor T-cells. Cytogenetics revealed del(6)(q23q25) and del(12)(p11.2p12).

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

105

Figure 3.56 Bone marrow with precursor T-ALL. (a, b) Leukemic cells with low SSC lacking CD34 and HLA-DR. The intensities of CD2 and CD5 are similar to that on mature T-cells (c). CD7 is upregulated and CD3 is slightly less intense than that on mature T-cells (d). CD4 is heterogeneous and coexpressed with CD8 (e). Reactivity for TdT confirms that the leukemic cells are precursor T-cells (f).

106

FLOW CYTOMETRY IN HEMATOPATHOLOGY

3.6 Evaluation of mature lymphoid malignancies Mature lymphoid neoplasms (LPD/NHL) can be categorized on the basis of the cell lineage (B-cell vs. T-cell vs. NK cell) and biological grade (low-grade vs. aggressive). In addition to the S-phase fraction, the FSC parameter and the expression of the transferrin receptor (CD71, a marker of cellular activation/proliferation), contribute to the grading of mature lymphoid malignancies. In general, a higher tumor grade is associated with a higher S-phase, increased cell size, or bright CD71 (Figure 3.57).

3.6.1 Assessment of surface light chain expression Visual evaluation of surface light chain expression is required to establish the monoclonality of a B-cell proliferation. The optimal approach is to analyze the surface light chain data gated on B-cells. Other methods, such as determining the percentage of kappa-positive and lambdapositive cells with a calculation of the kappa:lambda ratio on all cells, can easily result in a misinterpretation of polyclonality. In heterogeneous specimens where the B-cell subpopulation consists nearly entirely of neoplastic cells or in homogeneous samples composed virtually of a single population of abnormal B-cells (e.g., CLL with marked lymphocytosis), the B-cell gate can encompass the entire CD20 or CD19 population (Figure 3.58). If the B-cell population is heterogeneous, consisting of two or more clusters of different cell sizes or CD20 intensities, then several selective gates may be necessary to identify and quantify the monoclonal population (see Section 4.4.1). In determining the fluorescence intensity of the expressed light chain, it is helpful to take into account the following: • The position of the peak of the positive light chain on the fluorescence scale. • The background intensity of the negative light chain and the degree of overlap, if present, between the kappa and lambda fluorescence clusters. • The extent of the difference between the mean peak fluorescence channels of kappa and lambda.

Using this approach, the intensity of the positive light chain is determined relative to that of the negative one, not just based on where it falls on the fluorescence scale. This is especially applicable when there is an increase of background fluorescence on the negative cluster (Figure 3.58d), either due to the increased nonspecific binding of immunoglobulins by the neoplastic cells (usually associated with increased cell size) or a high level of serum immunoglobulins. In general, if there is obvious overlap in the fluorescence intensity distributions of the light chains, then the intensity of the positive light chain is weak (Figures 3.15, 3.16, 3.37, 3.57, and 3.58). This usually corresponds with a difference of less than 1 decalog between the mean peak fluorescence channels of the negative and positive histograms. The above-described technique is best applied when there are an adequate number of B-cell events. If the malignant B-cell population is minimal (1% or less), additional data collection of B-cells may be necessary. A mature B-cell population negative for both light chains is, by definition, abnormal, except when the specimen analyzed is a solid lymphoid organ (tonsils, lymph nodes) with florid reactive follicular hyperplasia (FRFH). In such cases, there may be a subpopulation of activated benign B-cells lacking expression of either light chain (i.e., germinal center cells). Inspection of other dot plots will reveal diagnostic clues for FRFH (see Section 4.1.1.1). Aside from this exception, a mature B-cell population negative for both light chains can be proven to be neoplastic by other corroborating parameters, such as aneuploidy or an overwhelming proportion of the abnormal cells in the sample analyzed. Lack of surface light chain expression has been found in a number of NHL/LPD, including CLL/SLL, diffuse large-cell lymphoma (DLCL), and

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

107

Figure 3.57 High-grade B-cell NHL. The neoplastic cells display the characteristics of an aggressive lymphoma, including increased FSC (a), bright CD71 expression (b) and an elevated S-phase fraction (S% 20) with aneuploidy (DI: 1.1) (f). CD25 is also well expressed (e). Surface light chain (kappa) is very weakly expressed (c, d).

108

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 3.58 Lymph node (FNA) with diffuse large cell lymphoma. (a–d) The tumor cells are of variable cell size (medium to large) and comprise 90% of the cells in the sample. CD19 is well expressed. The background fluorescence of the negative light chain (lambda) is increased. The neoplastic cells are thus kappa+ and not kappa++, as they would be if the results were simply read off the fluorescence scale (kappa/lambda overlay gated on R2).

FCC (Figure 3.59) lymphomas. Potential diagnostic confusion between FRFH and FCC lymphomas can be avoided by comparing the respective patterns of CD10/CD20 or bcl-2/CD20 coexpression (see Section 4.1.1.1).

3.6.2 Assessment of pan B-cell antigens The evaluation of pan B-cell antigens takes advantage of the wide dynamic range of CD20 and is most useful in mature B-cell neoplasms with low FSC. A comparison of the relative intensities of CD19 and CD20 (both conjugated to same type of fluorochrome) can provide an additional clue for subclassifying small B-cell LPD/NHL. The following possibilities can be encountered, based on the relationship of CD20 and CD19 fluorescence intensities: 1. CD20 intensity < CD19 (Figure 3.37). This pattern, together with small cell size, is virtually pathognomonic of CLL, in which CD20 is downregulated and displays a heterogeneous fluores-

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

109

Figure 3.59 FCC lymphoma in the thyroid. The neoplastic cells (arrow) are slightly larger than benign T-cells (a), coexpress CD10 and CD20 (b), but lack detectable surface light chain (c, d). A minute population of polyclonal B-cells (thin arrow) and residual T-cells (T) are present. Note the shifting of the T-cell cluster into the CD20-positive region; this is due to CD20 coating, commonly seen in FCC lymphoma.

cence distribution. The fluorescence signals for CD20 often begin in the negative region (Figure 3.16). CD19 is usually of moderate intensity. It is still important, however, to evaluate this pattern in the context of other features (i.e., CD5 and CD23 expression) as well as the intensity of the monoclonal light chain. Other closely related entities, such as CLL/PL, as well as some cases of LPC lymphoma–leukemia, often exhibit a phenotypic profile similar to CLL, with subtle differences in the intensity of CD20 and/or surface light chains. These CLL-related disorders contain a variable increase in activated cells. When significant, the activated cells are reflected as an increased FSC, with either a bimodal (Figure 3.6) or variable distribution (Figure 3.5c,d). Morphologically, the tumor demonstrates a spectrum of small lymphocytes, slightly larger lymphoid cells with more cytoplasm, plasmacytoid lymphocytes, and/or prolymphocytes (Plate 14). A dimorphic picture may result if small lymphocytes and prolymphocytes predominate. Because of the increase in activated cells, the intensity of CD20 or the surface light chain can become relatively brighter than that seen in CLL (Figures 3.5c/d, 3.60 and 3.61). The pattern of CD20 < CD19 and the variable distribution of CD20 density are often retained, however. CD38 becomes brighter, whereas CD5 or CD23 may be downregulated. As a result, the antigenic profile of these CLL-related disorders varies from case

110

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.60 Peripheral blood with activated CLL. (a, b) Leukemic cells with slightly decreased CD45 and mildly increased FSC. (c–e) Gated on MNCs: Coexpression of CD5, CD23 and CD20. CD20 is brighter than usual, but still of variable distribution and less intense than CD19. CD38 is expressed. (f) Gated on B cells: The difference between the mean peak fluorescence channels of the negative (lambda) and positive (kappa) histograms is greater than 1 decalog (i.e., kappa intensity is not weak).

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

111

Figure 3.61 Peripheral blood with CLL/PL. (a, b) A predominant lymphoid cluster with SSC shifting toward the monocytic region, bimodal FSC and bimodal CD20. (c–f) Gated on MNCs: The expression of CD5, CD23 and kappa is also bimodal, that is, more intense on the activated lymphoid cells with higher FSC and brighter CD20.

112

FLOW CYTOMETRY IN HEMATOPATHOLOGY

Case studies 16 and 17

Case study 18

to case, ranging from a CLL-like profile (CD5+ CD23+) to that which mimics MCL (CD5+ CD23−), or the phenotype can be nondescript (CD5− CD23−). The terms mixed CLL, atypical CLL, or CLL in prolymphocytic transformation have been used to designate some of these more activated forms of CLL. Correlation of the morphologic findings from the different sites of involvement (e.g., peripheral blood and lymph node), together with the review of the corresponding FCM studies and serum protein data, will provide a better understanding of CLL and its related disorders. In patients with multiple FCM studies, it is not unusual to observe a progression from CLL to the more activated form over the course of time. Furthermore, the FCM picture of the lymph node may reveal a larger activated component than that seen in the blood or bone marrow. This apparent discrepancy is translated morphologically as the presence of prominent proliferation centers in the lymph node, whereas the blood or bone marrow is involved mostly by small lymphoid cells. 2. The fluorescence intensities of CD19 and CD20 are closely similar (CD19 = CD20) and usually in the moderate range. This pattern is the least informative, being present in various low-grade B-cell LPD/NHLs, such as FCC lymphoma, MCL, mucosa-associated lymphoid tissue (MALT) lymphoma, and some of the above-mentioned CLL-related disorders (Figures 3.62 and 3.63). 3. CD19 intensity < CD20. This feature is frequently encountered in FCC and mantle cell lymphoma. Its main utility is primarily in the context of a CD5+ CD23− small B-cell LPD/NHL (Figure 3.64). Because the combination CD5+ CD23− is found not only in MCL but also in a number of CLL related malignancies and LPC disorders, the CD19 intensity < CD20 pattern is a helpful clue for recognizing MCL. Unlike the fluorescence signals of downregulated CD20 in CLL, the downregulated CD19 usually falls within the second or third decalog, rather than the first decalog.

Downregulated expression or loss of CD19 and/or CD20 (Figure 3.65) can also occur in large, mature B-cell neoplasms, especially those with a very high S-phase. Many of these demonstrate morphologic features reminiscent of large nucleolated neoplastic plasma cells (i.e., a “plasmablastic” or “immunoblastic” cytology).

3.6.3 Useful antigenic features in mature B-cell malignancies The differential diagnosis of B-cell LPD/NHL can be further narrowed on the basis of the expression of certain antigens or variations in density thereof, in addition to the relationship of CD20 and CD19 intensities described earlier. A substantial number of small B-cell LPD/ NHL bear no distinctive antigenic characteristic, however, (i.e., a nondescript B-cell phenotype) and are thus referred to as B-cell LPD/NHL, NOS (not otherwise specified). The useful antigenic features are described next. 3.6.3.1 CD10 expression: Follicular center cell lymphomas

Case studies 19 and 20

In the context of a mature B-cell malignancy, coexpression of CD10 and a pan B-cell antigen infers follicular center cell differentiation. CD10 intensity varies from case to case (Figures 3.66 to 3.68), and may be downregulated to absent (Figure 3.68b). According to the literature, approximately 20% to 25% of well-documented FCC lymphomas lack any detectable CD10 expression. CD20 is well expressed unless the specimen analyzed is obtained after anti-CD20 immunotherapy. Occasionally, the neoplastic population may display marked heterogeneity in the expression of CD20 (e.g., a bimodal distribution) (Figure 3.19). On FCM dot plots derived from lymph node specimens, the tumor cell cluster is usually obvious (Figure 3.66a,b) whereas in the blood or bone marrow, the lymphomatous population can be inconspicuous because the involvement can be as low as 1% (or less). Lack of a detectable light chain is observed in a small number of FCC lymphomas (Figures 3.59 and 3.68). The FSC parameter in FCC lymphomas varies considerably from case to case (depending on the relative preponderance of small and large follicular center cells) ranging from small (Figure 3.66), through variable (Figure 3.67), to large. Discrepancy between the FSC results

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

113

Figure 3.62 Peripheral blood with activated CLL (WBC: 113 × 109/L). (a–f) Leukemic population with coexpression of CD5 and CD23 and slightly decreased CD45. The FSC is slightly increased. A small subset of larger cells is present. CD20 and CD19 (both conjugated to the same fluorochrome) are well expressed. The intensity of the monoclonal kappa light chain is not weak.

114

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.63 Activated CLL (continuation of Figure 3.62). (a, c) Normal blood as a control: T-cells are Zap-70 positive (arrow). (b, d) Patient’s sample: CLL cells are CD38+. Zap-70 expression appears indeterminate. (e, f) The overlay Zap-70 histograms comparing normal blood (gray) and the patient’s blood, either ungated (e) or gated on R2 (f), demonstrate that the patient’s CLL cells do not express Zap-70.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

115

Figure 3.64 Peripheral blood involvement by MCL (morphology shown in Plate 41). (a–f) A predominant lymphoid cluster with CD45 downregulated into the hematogone region. The cell size is small. The leukemic cells are CD19++, CD20+++, CD5+, CD23− and lambda+ +. CD19 is less intense than CD20. A small number of T-cells and granulocytes are present.

116

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 3.65 Pelvic mass with plasmablastic large cell lymphoma. (a–d) Neoplastic cells are large. CD20 is not detectable. Lambda is extremely dimly expressed. CD10 is present.

Case study 21

and the morphology may occur as a result of tissue sampling, when the sample allocated for FCM analysis is taken from areas containing a preponderance of small cells while the remaining tissue contains foci with high numbers of large cells. Otherwise, there is generally a good correlation between the cell size and each subtype of FCC lymphoma (small, mixed, and large [i.e., FCC I, II, and III, respectively]). In the natural course of the disease, low-grade FCC lymphoma ultimately transforms into either FCC III or diffuse large cell lymphoma (DLCL), both of which are traditionally considered to be aggressive lymphomas. Although the S-phase fraction in these transformed FCC lymphomas is higher than that observed in FCC I and FCC II, it is much lower than that observed in de novo non-FCC large cell lymphoma. In a substantial number of FCC III, the S-phase fraction is either low (Figure 3.68) and similar to that of low-grade LPD/NHL, or may only slightly exceed the typical 5% threshold. Concurrently, CD71 is either absent or minimally expressed. Furthermore, the waxing and waning clinical course of transformed FCC lymphoma contrasts with the rapidly progressive behavior of de novo LCL. CD10 is also present in Burkitt lymphoma. Although the FCM immunophenotyping data in Burkitt lymphoma and CD10-positive DLCL (many of which represent transformation from

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

117

Figure 3.66 Variations in CD10 intensity. Two different cases of FCC I lymphoma. (a, b) Case 1: The tumor is of small cell size, coexpressing bright CD20 and CD10. (c, d) Case 2: The neoplastic cells (arrow) are small. The intensity of CD10 is weak and variable.

low-grade FCC lymphoma) may be indistinguishable (Figures 3.69 and 3.70), the extremely high S-phase (usually exceeding 20%) is an important clue for recognizing Burkitt lymphoma by FCM analysis (Figure 3.69e,f). Occasional high-grade B-cell lymphomas (e.g., those with “plasmablastic” morphology) may exhibit CD10 expression and a high proliferative fraction (Figures 3.71 and 3.72) similar to that seen in Burkitt lymphoma. A subtle differentiating clue is downregulation of one of the pan B-cell antigens, usually CD20, however. Furthermore, they also differ from Burkitt lymphoma by their cytologic and genotypic (lack of c-myc translocation) characteristics. 3.6.3.2 Pattern of CD20 and CD11c coexpression

Case study 22

CD11c is expressed in various B-cell neoplasms, including hairy cell leukemia. CD25, typically found in HCL, may be found in non-HCL B-cell LPD/NHL (Figures 3.12, 3.57 and 3.73), in particular CLL and CLL-related disorders (Figures 3.74 and 3.75). In HCL, however,

118

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 3.67 Two different cases of FCC lymphoma with an increased number of large FCC cells. (a, b) Case 1: The tumor is of variable cell size, ranging from small to medium. CD10 and CD20 are well expressed. Note the unusual variable distribution in CD20 intensity. (c, d) Case 2: The tumor is of bimodal cell size with coexpression of CD10 and CD20. A small but conspicuous population of benign CD10− B-cells is present (arrow) in addition to the residual T-cells (T).

both CD20 and CD11c are intensely expressed, resulting in a coexpression pattern distinctive from that in other diseases (Figure 3.76). Furthermore, the cell size of hairy cells is larger than lymphocytes and the SSC similar to that of monocytes. The constellation of these findings, together with the coexpression of CD25 and CD103, is essentially pathognomonic for HCL and serves as a fingerprint for the detection of a very low level of hairy cells in the peripheral blood. Rarely, CD10 may be also found on hairy cells. In most of the non-HCL B-cell LPD/ NHL with CD11c expression, the intensity for CD11c is heterogeneous, thus producing a “trail” pattern on the CD11c/CD20 dot plot (Figure 3.77). In the authors’ experience, cases of LPD that may be misinterpreted as HCL are those in which the blood or bone marrow smears display lymphoid cells with hairy or villous projections. Slow drying of the air-dried smears can easily induce such cytologic features, however, and the artifacts are most visible in cells with ample cytoplasm and a relatively low content of organelles/granularity. Lymphoid cells such as plasmacytoid lymphocytes, activated lymphocytes, or prolymphocytes fall into this category. The pattern of CD11c/CD20 coexpression

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

119

Figure 3.68 Follicular lymphoma with a preponderance of large cells (FCC III). (a) The tumor displays bimodal FSC and bimodal CD20 with brighter CD20 on the larger cells. A minute residual T-cell population (T) is present. (b–d) Weak and downregulated CD10 expression on the brighter CD20 cells. The smaller cells with dimmer CD20 are negative for CD10. No surface light chain is detectable. (e, f) Dim CD71 and low S-phase (0.8%) indicate a low proliferative activity. The DNA index is 1.9.

120

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.69 Bone marrow with Burkitt lymphoma–leukemia. (a) Virtually a single predominant cluster in the monocytic region with moderate SSC. (b–c) The neoplastic cells are of variable (medium to large) cell size, coexpressing CD10, CD19, CD20 (CD20 brighter than CD19). (d) Gated on the tumor cells: Lambda light chain is well expressed. (e, f) The tumor is diploid with a very high proliferative fraction (S% 51.3).

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

121

Figure 3.70 Lymph node with CD10-positive diffuse large cell lymphoma. (a–f) The neoplastic cells display increased FSC. They are CD20++, CD10+ and kappa++. The tumor is near diploid (more discernible on the FSC/DNA content dot plot than on the single parameter DNA histogram) with an S-phase fraction of 7%.

122

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 3.71 Extranodal high-grade B-cell lymphoma. (a–d) A single cluster of large tumor cells positive for CD19 and CD10. CD20 is markedly downregulated. There is no surface light chain expression.

is one of several clues, which if evaluated carefully, can prevent potential misinterpretation. This is especially true because some of these LPDs with “hairy/villous” cytology (see Section 5.2.1.7) may express weak CD25 (Figure 3.77) or less commonly, CD103, especially the B-ly7 clone (Figure 3.78). The distribution of CD103 intensity, because of its limited dynamic range, is similar between HCL and the occasional case of CD103+ non-HCL LPD. The extremely bright CD20 and CD11c coexpression in HCL is a differentiating characteristic, however. Other additional features, such as an increased WBC count with abundant circulating abnormal cells, and a cellular bone marrow aspirate, also permit one to recognize that the LPD with “hairy” cytology is not HCL. This can be further confirmed by the patient’s lack of response to the HCL therapeutic regimen. It cannot be stressed enough that the reactivity for CD25 and CD103 and the very intense coexpression of CD11c/CD20 need to be evaluated together because one or two of these three features may be present in non-HCL B-cell LPD (Figures 3.77 to 3.79).

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

123

Figure 3.72 Extranodal high-grade B-cell lymphoma (continuation of Figure 3.71). (a–c) The lack of TdT and CD34 confirms that the neoplastic B-cells are mature B-cells. (d) DNA analysis with DRAQ5: The tumor is aneuploid (DI: 1.19) with a very high S-phase of 40%.

3.6.3.3 CD5 expression Normally present on T-cells, the presence of CD5 in B-cell tumors is the most wellrecognized “aberrancy.” Among the CD5+ B-cell malignancies, the intensity of CD5 varies from case to case (Figure 3.80). It is often weaker than that in normal T-cells. Distinguishing the various entities within this subgroup of CD5+ B-cell LPD/NHL, which include CLL/SLL, CLL-related disorders, MCL, and CD5+ large cell lymphoma, requires the evaluation of other parameters, including cell size, the presence or absence of CD23, the intensity of CD20 and surface light chain, and the relationship between CD19 and CD20 intensities (see Sections 3.6.2 and 4.4.1.1) when conjugated to the same fluorochrome. The differential diagnosis of CD5+ B-cell disorders is presented in Chapter 5.

124

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.73 FNA of a retroperitoneal mass. Intermediate-grade B-cell NHL. (a) The malignant cells (arrow) display high FSC and a slightly decreased CD45. (b–d) The tumor coexpresses CD19, CD20 and monoclonal kappa light chain. CD25 is also present. (e, f) DNA analysis with DRAQ5: The tumor is diploid (DI: 1) with an intermediate S-phase of 10.7%.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

125

Figure 3.74 Lymph node with CLL. (a–f) The tumor cells are of small cell size and coexpress CD19, CD5, CD23, IgD and IgM. The intensity of CD20 is weak. CD10, IgA and IgG are absent.

126

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 3.75 Lymph node with CLL (continuation of Figure 3.74). (a, b) The tumor cells are monoclonal for lambda, but expressed at low levels. (c) The difference between the mean peak fluorescence channels of the negative (kappa) and positive (lambda) histograms is less than 1 decalog. (d) CD25 is unusually brightly expressed. The phenotype is otherwise that of a typical CLL.

3.6.3.4 Aberrant B-cell profile Compared with T-cell malignancies, B-cell disorders have a lower frequency of aberrancies in pan B-cell antigens or surface immunoglobulin expression. The antigenic abnormalities include the absence of CD19 or CD20 (Figure 3.65) and lack of surface light chains despite detectable surface heavy chain expression (Figure 3.81). In the latter scenario, it is judicious to ensure that both markers of immaturity (CD34 and TdT) are negative because the combination of positive surface heavy chain (IgM) and absent surface light chain has been reported as a characteristic of transitional ALL, a rare subtype of ALL. These phenotypic abnormalities tend to occur in CLL, lymphoplasmacytic neoplasms, and high-grade B-cell LPD/NHL. The

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

127

Figure 3.76 Bone marrow core biopsy with hairy cell leukemia. (a) The neoplastic cells (arrow) are seen in the monocytic region. (b–f) CD25 and CD103 are well expressed on hairy cells. There is intense coexpression of CD20 and CD11c. The tumor is monoclonal for kappa. Some residual T-cells (with evidence of CD20 coating) are present.

128

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.77 Peripheral blood with CD25-positive B-cell LPD NOS, morphologically LPC leukemia. (a–f) The leukemic cells (arrow) are small to medium size, with CD20, CD19, CD25 and kappa reactivity, but lack CD103. CD5 and CD10 are negative (not shown). CD11c is heterogeneous, from negative to positive (a CD11c/CD20 staining pattern different from that in HCL). There is indirect evidence of hypergammaglobulinemia (e, f).

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

129

Figure 3.78 Peripheral blood with CD103-positive B-cell LPD NOS, morphologically LPC leukemia. (a–e) A prominent lymphoid cluster composed of normal T-cells and neoplastic B-cells. The tumor cells are small and express bright CD19 and CD20. CD25 is absent. CD103 is present and of variable distribution (from negative to positive). CD11c is markedly heterogeneous. (f) Gated on B-cells: The surface light chain (lambda) is weak.

130

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.79 Peripheral blood with B-cell LPD NOS, morphologically PLL (shown in Plate 38). (a, b) Leukemic cells (arrow) with medium FSC occupying the monocytic region. (c–e) Gated on MNCs: weak CD25, minimal CD103 on a subset. CD5 and CD10 (not shown) are negative. The intense coexpression of CD20 and CD11c mimics that of HCL. (f) Gated on CD20-positive cells: Lambda is positive (weak). The background fluorescence of the negative kappa light chain is increased.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

131

Figure 3.80 Variations in CD5 expression. Four different lymph node samples. (a) CLL/SLL: CD5 as bright as that on T-cells. (b) CLL/SLL: CD5 less intense than that on T-cells. (c) LPC lymphoma: two subpopulations, one CD5-negative and the other with downregulated CD5. (d) MCL: bimodal CD5 expression (an unusual finding).

Case studies 23 to 25

latter is characterized by large cell size, high S-phase with or without aneuploidy (Figures 3.71 and 3.72), and frequently, a “plasmablastic” cytology. These features are often observed in acquired immunodeficiency syndrome (AIDS)-related NHL, such as primary effusion lymphomas. CD45 expression in these high-grade B-cell lymphomas is often weak or absent (Figure 3.82). Other sporadic aberrancies include the expression of a myeloid (Figure 3.83) or a T-cell-associated antigen (Figure 3.84) other than CD5. When evaluating follow-up specimens in patients with B-cell LPD/NHL, especially the lowgrade lymphomas, it is important to be aware that an apparent lack of CD20 expression may have been induced by anti-CD20 (rituximab) therapy (see Section 4.4.1.2). Absent CD20 may also be observed in some cases of lymphoplasmacytic lymphoma–leukemia (Figures 3.85 and 3.86).

3.6.4 Identification of abnormal mature T-cells The absence of one or more T-cell antigens in otherwise mature T-cells has been regarded as one of the most useful criteria in the diagnosis of post-thymic T-cell LPD/peripheral T-cell (text continues on page 138)

132

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.81 Bone marrow with high-grade B-cell NHL. (a, b) The tumor cells are of medium cell size and form a cluster in the monocytic region. (c–f) Gated on MNCs: The neoplastic cells (CD20+ + +, CD19+ +) express IgG only. IgM, IgD, IgA (not shown) and both light chains are negative.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

133

Figure 3.82 Lymph node with CD45-negative high-grade B-cell lymphoma. (a–f) The tumor displays bimodal FSC and no detectable CD45. Residual T-cells (CD45+ +) are present. The neoplastic cells are positive for CD20, CD10 (not shown) and CD19. CD34, TdT (not shown) and both light chains are negative. The tumor is aneuploid (DI: 1.77) and highly proliferative, with bright CD71 and a high S-phase fraction (S% 34.7).

134

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.83 Bone marrow with B-cell LPD NOS. (a) The tumor cluster (gray) occupies the lymphocyte region. (b–e) The neoplastic cells are of variable cell size, coexpressing CD19, CD20 and bright monoclonal kappa light chain. CD5 and CD10 (not shown) are absent. (f) Gated on CD19/CD20 cells: The malignant cells also display aberrant and heterogeneous CD13 expression.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

135

Figure 3.84 Low-grade B-cell NHL NOS with aberrant CD8 reactivity. (a–f) The neoplastic population (arrow) is composed mostly of small cells. The cells are CD20++, CD19++ and kappa+. An unusual feature is the weak and variable expression of CD8. Small populations of benign B-cells (thin arrow) and normal CD4 and CD8 (open arrow) T-cells are present.

136

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.85 Pleural fluid with lymphoplasmacytic lymphoma–leukemia. (a–f) The neoplastic cells (arrow) are of small cell size, with coexpression of CD19, CD5, CD25 and CD38. CD10 and CD23 are negative. The profile CD5+ CD23− is similar to that of MCL. Weak surface light chain and downregulated CD20 (see Figure 3.86) are not features of MCL, however.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

137

Figure 3.86 Lymphoplasmacytic lymphoma–leukemia (continuation of Figure 3.85). (a–d) The tumor cells (arrow) coexpress IgM, IgD and weak monoclonal lambda light chain. CD20 is virtually absent. (e, f) DRAQ5 tube: DNA analysis is gated on CD19+ CD5+ B-cells (R2). The tumor is diploid with an S-phase of 0.4%.

138

FLOW CYTOMETRY IN HEMATOPATHOLOGY

lymphomas (PTCLs). Also important is the quantitative variation in antigenic density that often occurs in lymphoid tumors. It is usually difficult to appreciate these alterations as well as the abnormal coexpression of antigens on immunohistologic preparations, however. Because there exist no T-cell antigens equivalent to the surface light chains, identifying the malignant nature of a T-cell proliferation can be problematic. Often the diagnostic workup requires molecular studies, namely the PCR analysis of the TCR-γ genes. This technique is more widely applied than the more labor-intensive Southern blot analysis of TCR-β genes because the rearrangement repertoire of TCR-γ is far less complex than that of the β chain genes. However, the limited complexity of TCR-γ is also a limiting factor in the specificity of the PCR assay, whereby benign conditions such as aging, autoimmune disorders, or viral infections may produce banding patterns similar to those seen in T-cell malignancies. The recent commercial availability of antibodies recognizing 70% of the TCR-Vβ repertoire has facilitated the study of T-cell clonality by FCM. Abnormalities in the distribution of the TCR-Vβ antigens are found in both immature T-cell malignancies and post-thymic T-cell neoplasms. Furthermore, other antigenic alterations such as the lack of CD3, coexpression of CD4 and CD8, or the aberrant expression of CD10 can be present in either disorder. Lack of HLA-DR is typically seen in T-ALL. This feature is not a reliable indicator of immature T-cells, however, because it can be found in a number of mature T-cell malignancies (e.g., Sézary syndrome). Reactivity for TdT or CD34 is therefore the most reliable feature to separate these two groups of disorders. Another helpful clue is the location of the neoplastic cell cluster on the SSC/CD45 dot plot. In precursor T-cell ALL, the tumor cells most often form a cluster in the blast region. Malignant lymphoid cells in T-cell LPD/NHL, on the other hand, display brighter CD45 and therefore would be found in the lymphocyte region or vicinity thereof. Occasionally, CD45 expression on mature neoplastic T-cells is either lost or quite downregulated (Figure 3.87), resulting in a blast-like cluster on the SSC/CD45 dot plot. The identification of a post-thymic T-cell process relies on the combination of several abnormalities including: 1. Aberrant antigenic profile. The determination of clonality in suspected T-cell LPD/NHL is mainly based on the detection of an abnormal phenotype (i.e., phenotypes not observed in normal mature Tcells). The main abnormality concerns the lack of expression of one or more pan T-cell antigens, most often CD3, CD5, or CD7 (Figure 3.88). In many mature T-cell disorders, the aberrancy can be less overt. Instead of an obvious loss of antigenic expression, the abnormality can manifest as a quantitative variation or a heterogeneous distribution in antigen density of one or more T-cell markers. Such subtle changes usually escape detection on immunohistochemistry preparations. On the FCM dot plots, the neoplastic T-cells can be distinctively separated from normal T-cells on the basis of the altered fluorescence intensity, either downregulated or upregulated, of one or more pan T-cell antigens (Figures 3.7, 3.89 and 3.90). The sole finding of downregulated CD7 does not imply a malignant process, however, especially when the specimen analyzed is peripheral blood from children or young adults. In acute infection with Epstein-Barr virus (EBV), it is not unusual to find a conspicuous CD8+ T-cell population with increased FSC and dim CD7 expression (see Section 4.1.2.3). A rare aberrancy observed in post-thymic T-cell neoplasms is the expression of CD1. This marker is normally present on thymocytes and is expressed in many T-ALLs. Another rare but important abnormality is the coexpression of markers strongly associated with another lineage, e.g., CD20 (Figure 3.91). 2. CD4/CD8 abnormalities. The CD4 : CD8 ratio does not indicate clonality. The ratio was initially established in the peripheral blood mainly for the purpose of evaluating immunocompromised patients. In general, values within the 0.5 to 4 range are considered “normal.” The traditional thinking is that a marked increase or decrease (not associated with immunodeficiency) in the CD4 : CD8 ratio, namely > 10 : 1 or < 1 : 10, would suggest a malignant T-cell proliferation (Figures 3.89 and 3.92). The usefulness of this feature pertains mainly to the peripheral blood and bone marrow, but it still requires other corroborating evidence such as increased FSC, persistent absolute lymphocy-

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

139

Figure 3.87 Aggressive NK-like T-cell lymphoma–leukemia in the CSF. (a–f) Neoplastic cells (arrow) with high FSC and markedly downregulated CD45, thus producing a “blast-like” SSC/CD45 pattern. The tumor displays downregulated CD3, heterogeneous CD7 and CD56, and bright CD8. Benign CD4 and CD8 lymphocytes are present in the specimen.

140

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 3.88 Lymph node (FNA) with peripheral T-cell lymphoma. (a–d) The neoplastic cells display an aberrant T-cell phenotype with lack of CD3 and CD7. CD5, CD4 and CD2 (not shown) are well expressed. CD10 is dim and of variable distribution. NK markers (not shown) are negative.

tosis, and no evidence of active viral infection (e.g., IM) or immunodeficiency. In solid lymphoid organs, the finding of an abnormal CD4 : CD8 ratio (in particular, elevated CD4 : CD8) by itself does not necessarily imply a neoplastic T-cell process (Figures 3.93 to 3.95). There is also considerable overlap in the CD4 : CD8 ratios between different groups of disorders including NHLs, Hodgkin disease, and reactive lymphadenopathies. The ratios are spread over a much wider range, and a CD4 : CD8 ratio > 10 : 1 can be encountered in conditions other than T-cell NHLs. A wellknown example is the marked increase in CD4+ cells that may occur in Hodgkin disease as a response to cytokines released by Reed-Sternberg cells (Figure 3.96). An occasional case of large B-cell NHL may present with such a marked preponderance of CD8 T-cells as to obscure the rare neoplastic B-cells, and the resulting picture may be mistaken for a CD8+ small T-cell LPD/NHL (Figures 3.97 to 3.99). More reliable than the CD4 : CD8 ratio in the diagnosis of mature T-cell malignancies is either coexpression of CD4 and CD8 (Figure 3.91) or lack of both CD4 and CD8 (Figure 3.90) in a sizable lymphoid population that is negative for both CD34 and TdT. In normal tissues outside the thymus, CD4+ CD8+ and CD4− CD8− cells are present in insignificant numbers. Either feature is more common in T-cell lymphoblastic leukemia–lymphoma than in post-thymic T-cell disorders. Coexpression of CD4 and CD8 is a normal finding if the specimen analyzed is the

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

141

Figure 3.89 Peripheral T-cell lymphoma in leukemic phase. Subtle antigenic aberrancies. (a) Downregulated CD45. The tumor (arrow) occupies the hematogone region. (b) Slightly downregulated CD3 compared with normal T-cells (thin arrow). (c, d) Slightly downregulated CD4 compared with normal T-cells, which can be better appreciated on the FSC/CD4 than the CD4/CD8 dot plot. The CD4 : CD8 ratio is markedly increased. The tumor cells are of medium cell size.

Case studies 26 to 28

thymus (i.e., thymoma), however, because this is the expected phenotype of thymocytes (Figure 3.100). Furthermore, caution should be taken when analyzing peripheral blood specimens from pediatric patients. In certain congenital immunodeficiencies such as severe combined immunodeficiency (SCID), the lymphocytes may lack CD3 expression or be negative for both CD4 and CD8. The presence of lymphopenia and the clinical setting are important clues to the underlying immunodeficiency. 3. TCR-αβ/TCR-γδ abnormalities. Most normal T-cells express TCR-αβ (TCR-αβ+/TCR-γδ−) as a result of the normal maturation and differentiation process in thymocytes. Therefore, the finding of TCR-αβ−/TCR-γδ+ or TCR-αβ−/TCR-γδ− in a prominent lymphoid population implies that the proliferation is neoplastic. 4. The proportion of the cells of interest. The expression of T-cell antigens in some mature T-cell malignancies may not display any aberrancy (i.e., the tumor cells may have a profile identical to either normal CD4 T-cells or normal CD8 T-cells). In the blood or bone marrow, the process can still be recognized as neoplastic if the population is present in overwhelming numbers. A specific example is T-PLL, in which the peripheral blood is essentially composed of a single population of (text continues on page 147)

142

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.90 Lymph node with peripheral T-cell lymphoma. The neoplastic cells (arrow) are of medium cell size and display an aberrant T-cell phenotype with: (1) lack of CD3, CD4 and CD8 (a, b); (2) upregulated CD5 (c) compared with the normal Tcells (thin arrow); and (3) CD10 expression (e). The intensity of CD2 and CD7 is similar to that on normal T-cells. NK markers (not shown) are negative. (f) The tumor is diploid with a low S-phase fraction (S% 3.8).

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

143

Figure 3.91 Peripheral blood with T-PLL (morphology shown in Plate 50). (a) Virtually a single predominant cluster in the lymphoid region. (b–f) CD3, CD7, CD2 and CD4 are well expressed on the leukemic cells. The leukemic cells also display weak CD8 and weak, aberrant CD20 expression. CD25 is negative. HLA-DR is absent. TdT (not shown) and CD34 are negative.

144

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.92 Lymph node involvement by mycosis fungoides. (a–d) The tumor cells are medium size, coexpressing CD2, CD3, CD7 and CD4 at the same level as normal T-cells. The CD4 : CD8 ratio is markedly increased. HLA-DR is weakly expressed. (e) The only antigenic aberrancy is the loss of CD5. (f) The tumor is diploid with a moderate S-phase fraction (S% 9.5).

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

145

Figure 3.93 Reactive lymphocytosis in the pleural fluid of a patient with lung carcinoma. (a–f) Most of the lymphocytes are T-cells, with a normal pattern of reactivity for CD3, CD5, CD2 (not shown) and CD7. CD25 and HLA-DR are weakly expressed. CD56 is negative. The CD4 : CD8 ratio is abnormally increased to 13.7 : 1.

146

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.94 Reactive CD4 T-cell lymphocytosis in the pleural fluid (continuation of Figure 3.93). (a) The cell size of CD4 T-cells is slightly larger than that of other lymphocytes. (b–f) The analysis of the TCR-Vβ repertoire is gated on the CD5+ CD4+ T-cells (R2). A polyclonal pattern of reactivity is seen in tubes 1-4.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

147

Figure 3.95 Reactive CD4 T-cell lymphocytosis in the pleural fluid (continuation of Figure 3.94). (a–d) CD4 T-cells produce a polyclonal pattern of reactivity to antibodies contained in TCR-Vβ tubes 5 through 8.

phenotypically “normal” CD4 T-cells. The increased cell size (FSC in the moderate range) is another clue that these are abnormal T-cells (Figure 3.101). 5. Abnormal distribution of TCR-Vβ antigens. Using the current commercially available panel of TCRVβ antibodies, this abnormality can manifest in one of two ways: (a) restricted reactivity to a single Vβ family (Figures 3.102 to 3.104), which is direct evidence that the T-cell proliferation is monoclonal, or (b) the neoplastic T-cell population is TCR-αβ+, but demonstrates no reactivity with any of the Vβ antibodies in the kit (Figures 3.105 to 3.107). This complete lack of reactivity is accepted as indirect evidence of clonality. The explanation for this apparent abnormality is that the tumor cells express either a Vβ with an altered epitope (and therefore not recognized by the antibody to its Vβ family), or a Vβ family for which no antibody is yet available. In those cases of T-cell LPD/NHL with expansion of TCR-Vβ, the expansion most commonly involves a single Vβ family. Occasionally, one or two additional smaller Vβ expansions may be also present. 6. DNA and cell cycle parameters. Aneuploidy and/or an increased S-phase are additional parameters supportive of a malignant process.

148

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.96 Lymph node with Hodgkin disease. (a–f) Most of the cells analyzed are T-cells with normal expression of CD3, CD7, CD5 and CD2 (not shown). CD25 and HLA-DR are heterogeneously expressed. CD56 is negative. The CD4 : CD8 ratio is increased to 8.3 : 1.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

149

Figure 3.97 Gastric biopsy with reactive CD8 T-cell lymphocytosis in a patient with a history of FCC lymphoma. (a–e) The lymphoid cells in the sample are nearly all T-cells, with normal expression of CD3, CD7, CD5 and CD2 (not shown). CD56 is negative. The CD4 : CD8 ratio is reversed to 0.2 : 1. CD8 T-cells are of small cell size. (f) Gating on CD5+ CD8+ T-cells (R3) for the analysis of the TCR-Vβ repertoire.

150

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 3.98 Gastric biopsy with reactive CD8 T-cell lymphocytosis (continuation of Figure 3.97). (a–d) CD8 T-cells produce a polyclonal pattern of reactivity to antibodies contained in TCR-Vβ tubes 1 through 4.

3.6.5 Useful antigenic features in mature T-cell malignancies Once a mature T-cell population is identified as neoplastic, the presence of certain additional phenotypic features may permit one to establish the specific subtype of T-cell LPD/NHL. Antigens useful for this purpose include CD25, CD30 (Figures 3.108 and 3.109) and ALK-1, and the family of NK markers. The current WHO classification of mature T-cell malignancies is complex. Some of the neoplasms, especially the extranodal T/NK tumors, share closely similar phenotypic features, aggressive biological behavior (and thereby similar treatments), and a frequent occurrence in the context of altered immunity. Despite these similarities, these tumors are categorized as separate entities on the basis of the initial/primary site of involvement (e.g., nasal, intestinal, cutaneous) and some other secondary features. From a practical diagnostic viewpoint, it is much less confusing to subdivide the mature T-cell LPD/NHL according to their biology based on (1) the expression of

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

151

Figure 3.99 Gastric biopsy with reactive CD8 T-cell lymphocytosis (continuation of Figure 3.98). (a–d) CD8 T-cells produce a polyclonal pattern of reactivity to antibodies contained in TCR-Vβ tubes 5 through 8.

Case studies 29 and 30

several key antigens, especially CD3, CD4, CD8, TCR-αβ, TCR-γδ, and NK markers, (2) biological behavior, and (3) where appropriate, certain very specific laboratory or clinical features (e.g., the typical morphologic features seen in mycosis fungoides/Sézary syndrome). In some of the post-thymic T-cell malignancies, such as T-PLL, aggressive true NKleukemia, the NK-like T-cell leukemias, and many cases of adult T-cell leukemia– lymphoma (ATLL), the peripheral blood involvement at the time of diagnosis is so extensive as to resemble the picture of a cell culture. Expression of CD25, together with the absence of one or more pan T-cell antigens, is nearly pathognomonic for ATLL (Figure 3.110). Although most cases of ATLL are CD4+, a small number coexpress CD4 and CD8. Rarely, CD25 (dim) may be found in a T-cell LPD/NHL other than ATLL. Laboratory evidence of HTLV-I infection is therefore necessary to confirm the diagnosis. Other helpful features include hypercalcemia and the typical blood smear morphology, showing a spectrum of small to large lymphoid cells with the “flower cell” cytology.

152

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 3.100 Thymocytes, from a case of thymoma. (a–c) The degree of heterogeneity in the expression of T-cell antigens (including CD1) varies from one marker to another. This reflects the presence of different subpopulations of thymocytes at various stages of maturation and differentiation. (d) The great majority coexpress bright CD4 and CD8. Similar to CD4 and CD8, the intensity of CD2 (not shown) is bright and homogeneous.

Case studies 31 and 32

T-prolymphocytic leukemia is recognized by its extremely high WBC count and an apparent normal helper T-cell phenotype (Figure 3.101). The overwhelming number of neoplastic cells results in a markedly elevated CD4 : CD8 ratio. This feature and the expansion of a single Vβ family are the main phenotypic abnormalities in T-PLL. The blood and bone marrow manifestations of aggressive NK disorders, which include true NK and NK-like T-cell malignancies, may be confused with an acute leukemia morphologically. The FCM characteristics are clear-cut, however, and indicate a mature population with medium to high FSC and coexpression of several T-cell markers along with one or more NK markers (CD16, CD56 or CD57). The results for CD16 may vary because of the wide difference among various anti-CD16 antibodies. In general, the aggressive NK or NK-like T-cell malignancies are usually associated with CD56 expression, whereas CD57 is more common in the low-grade counterparts (see Sections 4.1.2.3 and 4.4.2). A small number of aggressive cases may express CD57 instead of CD56, however. The increased cell size, high proliferative fraction, and mul(text continues on page 163)

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

153

Figure 3.101 Peripheral blood with T-PLL. (a) A predominant lymphoid cluster. (b–f) The leukemic cells are of medium cell size, but smaller than monocytes (arrow). CD3, CD7, CD5, CD2 and CD4 are of similar intensities to those on normal T-cells. HLA-DR is of weak and variable intensity.

154

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.102 Lymph node with PTCL. (a–f) A conspicuous cluster of abnormal CD4 T-cells with downregulated CD3 (arrow). The abnormal cells are better separated from the normal T-cells on the CD4/CD3 dot plot than on other graphics. CD2 and CD7 are coexpressed. CD56 is absent. CD5 is very slightly upregulated. HLA-DR is higher than on normal T-cells. The CD4 : CD8 ratio (2.86 : 1) is not overtly abnormal.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

155

Figure 3.103 Lymph node with CD4+ PTCL (continuation of Figure 3.102). (a, b) B-cells are virtually absent in the lymph node. The abnormal CD4 T-cells are larger than normal lymphoid cells. CD10 is “aberrantly” expressed. (c–f) The analysis of the TCR-Vβ repertoire is gated on the HLA-DR+ CD4+ population, which includes a minority of normal polyclonal CD4 T-cells. The tumor demonstrates clonal expansion of Vβ17.

156

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.104 Lymph node with CD4+ PTCL (continuation of Figure 3.103). (a–e) Residual normal CD4 T-cells demonstrate a polyclonal pattern of reactivity to antibodies contained in TCR-Vβ tubes 4 through 8. (f) DNA analysis gated on the tumor cells: The PTCL is diploid with a low S-phase.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

157

Figure 3.105 Tonsil with CD8+ peripheral T-cell lymphoma (PTCL). (a–f) A sizable cluster of abnormal lymphoid cells (gray) with markedly downregulated CD2 and CD5. CD3 is heterogeneous. CD7 is well expressed. The tumor cells are CD8+, resulting in a reversed CD4 : CD8 ratio. CD25, CD16 and CD57 are absent. CD56 is present in a subset.

158

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.106 Tonsil with CD8+ PTCL (continuation of Figure 3.105). (a–c) Lymphoma cells (gray) are of large cell size, coexpressing TCR-αβ and CD7. TCR-γδ is negative. (d–f) The analysis of the TCR-Vβ repertoire is gated on the lymphoma cluster (gray) with downregulated CD45 and heterogeneous CD8 expression. The tumor displays no reactivity to antibodies contained in TCR-Vβ tubes 1 and 2.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

159

Figure 3.107 Tonsil with CD8+ PTCL (continuation of Figure 3.106). (a–f) The abnormal CD8 T-cells display no reactivity to antibodies contained in TCR-Vβ tubes 3 through 8. A complete lack of reactivity to all of the currently available TCR-Vβ antibodies is indirect evidence of clonality.

160

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.108 CD30+ T-cell NHL in an elderly patient. (a–f) The tumor cells (gray) are of medium to large cell size and display no reactivity for CD2, CD5, CD7, CD4 or CD8. CD3 is markedly downregulated (nearly negative). CD56 and the myeloid markers (CD13 and CD33) are not expressed. At this point, the tumor appears to be of a so-called null phenotype.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

161

Figure 3.109 CD30+ T-cell NHL in an elderly patient (continuation of Figure 3.108). (a–d) The tumor cells (gray) display reactivity for HLA-DR, cCD3 and CD30. CD15 is also partly expressed. (e, f) The tumor cells are aneuploid (DI: 1.32) with a high S-phase of 20%.

162

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.110 Peripheral blood with ATLL (morphology shown in Plate 49). (a–f) The leukemic cells are of medium cell size, but smaller than monocytes. CD2, CD3, CD4, CD5 and CD25 are well expressed. CD7 is absent. HLA-DR is negative.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

163

Figure 3.111 NK-like T-cell leukemia. (a–d) The leukemic cells (arrow) display low FSC and weaker CD5 than normal T-cells (thin arrow). CD3, CD7 and CD8 are expressed at the same level as normal Tcells. CD56 is positive. Despite the small cell size, this case was biologically aggressive with marked leukocytosis and a short clinical course. There is a minute CD4++ CD8++ T-cell population (open arrow).

Case studies 33 to 35

tiorgan involvement are clear-cut indicators of the aggressive biology of this group of lymphoma–leukemias. Because of the infrequent occurrence of these diseases, not much is yet known regarding the expression of the more recently available NK markers, namely the family of KIR antigens. Furthermore, the number of commercially produced KIR antibodies is still too few for an adequate assessment of the highly polymorphic KIR repertoire. CD8 and CD5, when expressed, are often downregulated compared with normal suppressor T-cells (Figure 3.111). True NK cells differ from NK-like T-cells by the lack of a CD3/TCR complex and the germ-line configuration of the TCR genes. In other words, true NK cells do not express TCR-αβ, TCR-γδ, or surface CD3 (Figures 3.112 and 3.113). The truncated mRNA for the ε-chain of CD3 is present in true NK cells, however, thus producing isolated CD3 ε-chains that react with the polyclonal anti-CD3 antibodies used in immunohistochemistry. As a result, the distinction between true NK and NK-like T-cell lymphoma may not be possible with paraffin immunostaining.

164

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 3.112 Aggressive true NK-cell leukemia. (a–d) The neoplastic cells are negative for CD3, TCRαβ and TCR-γδ. CD7 is expressed but of heterogeneous distribution. This case has two unusual features: (1) coexpression of CD4 and CD8; and (2) reactivity for CD57 instead of CD56. CD34 and TdT are negative (not shown). Residual normal T-cells (positive for CD3, CD7 and TCR-αβ) are present.

The following is a summary of the phenotypes often encountered in the aggressive NK disorders:

Case study 36

Case study 37

• Positive CD56/CD2/CD7, downregulated or absent CD8/CD5, absent CD3/TCR-αβ/TCR-γδ/CD4. These features indicate a true NK cell proliferation. • Positive CD56/CD2/CD7/CD3/TCR-γδ, downregulated or absent CD5, absent TCR-αβ/CD4/CD8 (Figure 3.114). This NK-like T-cell, TCR-γδ phenotype occurs mainly in the rare γδ-T-cell lymphoma–leukemia (also known as hepatosplenic lymphoma). TCR-γδ T-cell neoplasms at other sites have also been reported (e.g., as subcutaneous panniculitis T-cell lymphoma). • Positive CD56/CD2/CD7/CD3/TCR-αβ, downregulated CD5 and CD8, absent TCR-γδ/CD4. This NK-like T-cell, TCR-αβ phenotype (Figures 3.105 to 3.107, 3.111) is found in the majority of aggressive NK-like T-cell malignancies. A rare case may express CD57 instead of CD56.

Aggressive mature T-cell LPD/NHLs with a pure suppressor phenotype (i.e., bright CD8, NK markers absent) are encountered infrequently (Figures 3.115 to 3.117).

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

165

Figure 3.113 Liver biopsy with aggressive NK-cell lymphoma. (a–f) The abnormal lymphoid cells in the sample are of medium cell size, coexpressing bright CD2, CD7 and CD56. CD3, CD5, CD4 and CD8 are absent. TCR-αβ and TCR-γδ are also negative (not shown). The phenotypic profile is most consistent with that of a true NK cell.

166

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 3.114 Bone marrow with γδ-T cell lymphoma–leukemia (morphology shown in Plate 56). (a, b) The neoplastic cells (arrow) coexpress CD7 and CD3, but lack both CD4 and CD8. There is increased background fluorescence. Monocytes (thin arrow) are CD4+ and CD8−. (c, d) The tumor displays bright CD2, CD56 and TCR-γδ expression.

3.7 Assessing the biological behavior of mature lymphoid neoplasms The biological behavior of any given lymphoma affects the choice of treatment modalities as well as the patient’s overall prognosis. The grading can be derived from several FCM parameters including cell size (FSC), CD71 levels, and tumor S-phase as determined by DNA analysis (see Sections 1.3 and 2.10). Increased FSC, bright CD71 and high S-phase are features seen in high-grade LPD/NHL, irrespective of whether the tumor cells are of B- or T-cell lineage. The S-fraction and CD71 are especially useful in separating large cell LPD/NHL with a high degree of aggressiveness (Figure 3.118) from those cases with a more waxing and waning clinical course (Figure 3.73). In patients with long-term follow-up (which may span over decades), these two parameters also provide information on any change, either progressive or sudden, in the biological behavior of the tumor. (text continues on page 172)

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

167

Figure 3.115 Peripheral blood with CD8+ T-cell LPD. (a, b) A prominent lymphoid cluster composed virtually entirely of T-cells of small cell size. (c–f) The abnormal cells display bright coexpression of CD3, CD2 and CD5. CD7 is heterogeneous. CD8 is slightly weaker than that on the few residual normal CD8 T-cells.

168

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.116 CD8+ T-cell LPD (continuation of Figure 3.115). (a–d) A subset of the abnormal CD8 T-cells displays aberrant coexpression of CD20 and CD10. CD25, CD103 (not shown) and NK markers (CD16, CD56 and CD57) are not expressed. (e, f) The tumor cells show no reactivity to the antibodies contained in TCR-Vβ tubes 1 and 2.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

169

Figure 3.117 CD8+ T-cell LPD (continuation of Figure 3.116). (a–f) The tumor cells display no reactivity to antibodies contained in TCR-Vβ tubes 3 to 8. The results of the TCR-Vβ analysis constitute indirect evidence that the abnormal CD8 T-cells are clonal.

170

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 3.118 Lymph node with high-grade B-cell NHL NOS. (a–c) The neoplastic cells display high FSC, bright coexpression of CD19, CD20 and monoclonal kappa light chain. (d–f) DRAQ5 tube: DNA analysis gated on CD19+ CD20+ B-cells. The tumor is aneuploid (DI: 1.18) with a high S-phase (32.5%).

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

e

f

171

Figure 3.119 Bone marrow with small cell carcinoma. (a) Abnormal FSC/SSC picture. (b–f) The tumor cells form a prominent cluster in the CD45-negative region with variable FSC signals from low to high. CD10, CD19 and all of the remaining lymphoid and myeloid markers (not shown) are negative, except for CD117 and CD56.

172

FLOW CYTOMETRY IN HEMATOPATHOLOGY

3.8 Dot plot patterns in histiocytic proliferations and nonhematopoietic malignancies The main nonhematopoietic neoplasm that passes through the flow cytometer is small cell carcinoma (neuroendocrine neoplasms) because the tumor cells are more likely to occur as individual cells and the cell size is in the range of that seen in aggressive lymphoid malignancies. Small cell carcinoma (Figure 3.119) is often accompanied by extensive necrosis. The resulting FCM dot plots may show an uninterpretable and disarrayed distribution. In many cases, the tumor cells can be recognized by their variable and extremely high FSC signals, along with CD56 expression. CD117 may also be present (Figure 3.119c). Specimens harboring a proliferation of histiocytes/macrophages are very rarely encountered in the FCM laboratory. The graphical FCM data, when viewed in the context of myeloid/monocytic antigenic expression, can provide helpful clues for recognizing these unusual situations, including the following:

a

b

c

d

Figure 3.120 Macrophages in a “nadir” (after chemotherapy) bone marrow. (a) Only lymphocytes and a population with very high SSC located in the monocytic region. (b) Isotype-matched negative controls: High level of background fluorescence. (c, d) Only CD33 is expressed. CD34, CD13, CD16 and all of the remaining lymphoid and myeloid markers (not shown) are negative.

FCM DATA ANALYSIS ON NEARLY HOMOGENEOUS SAMPLES

a

b

c

d

173

Figure 3.121 Langerhans cell histiocytosis (a) Isotype-matched negative controls: Slightly increased background fluorescence. (b) A monocytic-like cell cluster with high SSC (arrow). (c, d) The cells of interest display very intense CD1 expression. CD4 is positive and CD8 essentially negative.

Case study 38

• A markedly high level of autofluorescence (Figure 3.120). • The pattern of FSC, SSC, and CD45. On the SSC/CD45 dot plot, the histiocytic population occupies the position of monocytes, but with high SSC similar to that produced by granulocytes (Figures 3.120 and 3.121). The FSC signals are much higher than that of monocytes or granulocytes. • CD1, when present, points toward the diagnosis of Langerhans cell histiocytosis, also known as histiocytosis X (Figure 3.121).

CHAPTER

4

FCM data analysis on heterogeneous specimens

Most clinical specimens, whether normal/reactive or harboring neoplastic cells, are heterogeneous. The highest degree of heterogeneity is seen in bone marrow samples. Solid lymphoid tissues such as tonsils and lymph nodes are less heterogeneous because the granulocyte/monocyte component is usually insignificant. The benign nonpurulent effusion, when not overloaded with a large number of macrophages or mesothelial cells, is the least complicated type of specimen consisting of B- and T-cells in a similar proportion to that seen in the blood. The FCM graphics in heterogeneous samples are inherently complex, displaying multiple clusters, some of which may be overlapping. The task in analyzing these dot plots is twofold: first, to determine whether the sample is benign/reactive or contains a neoplastic population (which may or may not be evident at first glance) and, second, to characterize the tumor cells, if present. In some malignant conditions such as a myelodysplastic syndrome (MDS), CML, or CMMoL, an overt neoplastic cluster (i.e., increased blasts) may not be present. The FCM graphics often demonstrate several features useful for recognizing these disorders however, such as an altered proportion of the granulocytic or monocytic component and qualitative antigenic abnormalities in the myeloid (e.g., altered maturation curves), monocytic, or erythroid elements. The visual FCM data can become highly complex if the specimen harbors more than one malignant process (e.g., high-grade MDS and multiple myeloma [MM], or CLL and HCL). In order not to overlook potentially subtle abnormalities, it is necessary to be familiar with the appearance of the various cell clusters (shape, size, density, relative position) on certain key dot plots in different types of normal or benign reactive specimens. Visual FCM data analysis in heterogeneous samples follows a similar sequence to that applied to the nearly homogeneous specimens (see Chapter 3). Visual inspection begins with either the SSC/CD45 dot plot if the specimen is blood or bone marrow (BBS panel), or the FSC/SSC dot plot in the case of solid tissue or body fluids (TF panel). For blood and bone marrow samples, data analysis also includes an evaluation of the maturing granulocytes (granularity and antigenic maturation) or other elements (e.g., B-cells, erythroid precursors) in addition to identifying any neoplastic population present.

4.1 Identifying normal FCM samples A normal FCM sample is one in which no detectable clonal/neoplastic proliferation exists as determined by the light scatter and fluorescence parameters.

4.1.1 Benign/reactive solid lymphoid tissue (e.g., lymph nodes, tonsils) In general, the FCM dot plots derived from reactive lymph nodes display findings similar from case to case, irrespective of the specific morphologic subtype of the reactive process. Polyclonality and the absence of phenotypic aberrancies characterize a benign sample. It is important to be aware, however, that the absence of abnormalities by FCM analysis does not

176

FLOW CYTOMETRY IN HEMATOPATHOLOGY

necessarily exclude the existence of lymphoma (or other tumors that may not be recovered in the cell suspension). If the lymphomatous involvement is focal, either sampling error or selective loss of neoplastic cells (especially when present in low numbers) during tissue preparation can easily result in a false picture of “negative” dot plots. Furthermore, lymph nodes involved by Hodgkin disease also yield “negative” FCM data because the malignant cells are usually not present in the final cell yield. In Hodgkin disease, the composition of the background lymphocytes and their expression of lymphoid surface antigens are indistinguishable from that seen in reactive lymph nodes. In benign/reactive lymph nodes, the relative proportion of mature B- and T-cells varies widely from case to case. On the FSC/SSC or FSC/CD45 dot plots, both B- and T-cell populations are merged into a single cluster of variable cell size (Figure 4.1a,b). On the dot plots showing reactivity to B- and T-cell antibodies, the B- and T-cell clusters are well segregated based on the clear-cut expression of their respective lineage markers (Figures 4.1 and 4.2). The expression of HLA-DR (Figure 4.2f) and/or CD38 (Figure 4.1e) reflects evidence of activation in a proportion of lymphocytes. CD25 expression may be present on a small number of reactive T-cells (Figure 4.2b), and CD103 (B-ly7) can be detected in a subset of T-cells (Figure 4.2c) of the suppressor family. A ratio of helper to suppressor T-cells can be derived from the CD4/CD8 dot plot (Figure 4.2b). In the lymph node, this ratio can be higher than that in the peripheral blood, and may occasionally exceed 10 : 1. Such an increase in CD4 : CD8 ratio is frequently observed in Hodgkin disease (Figure 3.96). Other reactive and neoplastic conditions may also give rise to similar CD4 : CD8 ratios, however (Figure 4.3). In general, the CD4 : CD8 ratio in solid lymphoid organs is of limited diagnostic usefulness (see Section 3.6.4). The assessment of clonality in B-cells can be best performed by comparing the kappa– FITC/CD20 (or CD19) and lambda–FITC/CD20 (or CD19) dot plots side by side. In the authors’ experience, this approach is preferable to that using the kappa–FITC/lambda–PE dot plot or expressing the results as a kappa:lambda ratio, as it prevents potential misinterpretation of reactive B-cells as being monoclonal (Figures 4.4 and 4.5), especially in some cases of marked florid reactive follicular hyperplasia (FRFH), and facilitates the detection of small populations of malignant B-cells (Plate 5). With the exception of activated germinal center cells (GCCs), polyclonal B-cells demonstrate the normal bimodal distribution of kappa and lambda as viewed on the kappa/CD20 and lambda/CD20 dot plots (Figure 4.6). The bimodal distribution is most evident as a “double camel hump” (Figure 4.6d) seen on the corresponding single parameter histograms for kappa and lambda when gated on B-cells. In patients with marked polyclonal hypergammaglobulinemia, immunoglobulin coating of Bcells may obscure the normal bimodal distribution and produce a broad distribution with a single peak instead. A small number of benign B-cells positive for CD5 (with a weaker intensity than that on T-cells) may be seen in some reactive lymphadenopathies (Figure 4.7). A more frequent finding is the expression of CD10 in a subset of reactive B-cells. The corresponding lymph node sections invariably demonstrate follicular hyperplasia (Plate 15). These CD10+ reactive B-cells are larger and display the brightest CD20 intensity (Figure 4.8a). They correspond to the activated larger GCCs. In marked FRFH, the increased number of these cells is seen as a sizable cluster with coexpression of CD10, intense CD38, and CD20. This may lead to potential confusion with FCC lymphoma, especially because the larger cells in most cases of FRFH lack surface light-chain expression (Figures 4.8 and 4.9d) and conversely, a small number of FCC lymphomas have no detectable surface kappa or lambda (see Section 3.6.3.1; Figure 3.59). Characteristically, benign GCCs do not contain intracellular bcl-2 protein, however.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

177

Figure 4.1 Benign lymph node. (a, b) A single lymphoid cluster, mostly in the low FSC range, with bright CD45. (c, d) Normal B- and T-cells. (e, f) Variable expression of CD38 in B- and T-cells. CD71 is minimally expressed, indicative of low proliferative activity.

178

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.2 Benign lymph node (continuation of Figure 4.1). (a) B- and T-cells, mostly of small cell size. (b) CD4 : CD8 ratio is within the normal range. (c, d) Dim CD25 on a subset of T-cells. Lymphoid cells with NK markers are virtually absent. (e, f) A small subset of T-cells express CD103 (B-ly7). HLA-DR is intense on the B-cells, but more heterogeneously distributed on the T-cells.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

179

Figure 4.3 Increased CD4 : CD8 ratio in lymph nodes. (a, b) Case 1: FRFH. T-cells with normal expression of CD3 and CD7. CD2 and CD5 are also well expressed (not shown). The CD4 : CD8 ratio is 5.7 : 1. (c, d) Case 2: Castleman’s disease. Normal T-cells coexpressing CD3 and CD7. The CD4 : CD8 ratio is 11.8 : 1.

4.1.1.1 Pattern of CD10/CD20 coexpression: Distinction between FRFH and FCC lymphoma Close evaluation of the CD10/CD20 dot plot (CD10 on the y-axis and CD20 on the x-axis) will reveal the subtle differences between FRFH and the majority of FCC lymphomas (Figures 3.59, 3.66, and 3.67). In this regard, the choice of fluorochromes (see Section 2.6.2) to be conjugated with CD10 and CD20 is important in bringing out these differences. The reactivities to CD20 and CD10 by the various lymphoid cell populations in FRFH produce a pattern reminiscent of a “horizontal hockey stick” whereby the cluster of the larger reactive B-cells with coexpression of CD10 and CD20 is the foot of the “hockey stick” (Figures 4.8 and 4.10). The “horizontal hockey stick” is formed by several closely adjoined cell clusters arranged in the following order: T-cells, B-cells without CD10, and B-cells coexpressing CD10 and brightest CD20. In contrast, the CD10+ CD20+ neoplastic cell cluster in FCC lymphoma is well (text continues on page 185)

180

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.4 Tonsil with marked FRFH. (a–f) The two subsets of B-cells are better visualized with CD20 than CD19. The cells with lower CD20 intensity display smaller cell size and polyclonality for kappa and lambda. The brighter CD20 cells are larger, with no detectable light chain expression. Kappa (2) and lambda (2) are both conjugated to FITC. Both B-cell subsets lack IgA and IgG (data not shown).

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

181

Figure 4.5 Marked FRFH (continuation of Figure 4.4). (a–c) The brighter CD20 cells are CD10+ germinal center cells, heterogeneous for IgD and IgM. The dimmer CD20 cells are CD10−, homogeneous for IgD and IgM. (d) Gated on CD19+ B-cells: Bimodal distribution of kappa and lambda. (e, f) Confusing results from the reagent combination kappa (1)–FITC and lambda (1)–PE; a reversed kappa : lambda ratio (1 : 2.2) but apparent monoclonality for kappa.

182

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.6 Benign lymph node (same sample as that in Figure 4.1). (a, b) Polyclonal B-cells, easily appreciated by a sideby-side comparison of the kappa/CD19 and lambda/CD19 dot plots. (c, d) Overlay kappa/lambda histograms gated on B-cells (R2) only, showing the double camel hump pattern of polyclonal light chain expression. (e, f) B-cells are CD23 positive and CD10 negative.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

183

Figure 4.7 Reactive lymph node with CD5-positive B-cells. (a) Two subpopulations (no. 1 and no. 2) of B-cells differing in CD20 intensity. (b) A small population of lymphocytes with dim CD5 (arrow). (c–f) B-cells are polyclonal: The subpopulation with dimmer CD20 is CD10 negative; a subset thereof is CD5 positive, comprising 21% of the cells in the sample. The subpopulation with brighter CD20 expresses weak CD10 (d).

184

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.8 Lymph node with FRFH. (a) Two subpopulations of B-cells: one with smaller cell size and dimmer CD20 (R2), and the other composed of larger cells with brighter CD20 (R3). (b) CD10/CD20 “hockey stick” pattern. The brighter CD20 cells coexpress CD10. (c, d) The dimmer CD20 cells are polyclonal. The brighter CD20 cells have no detectable light chain. (e, f) Using CD19, it is not possible to discriminate between the two subpopulations.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

185

Figure 4.9 Lymph node with FRFH (continuation of Figure 4.8). (a) Dim CD71 expressed on the larger cells. (b) A subset of the dimmer CD20 cells is positive for CD5. (c) Gated on R2 (from Figure 4.8): Overlay kappa/lambda histograms showing the double camel hump pattern of polyclonal light chain expression. (d) Gated on R3 (from Figure 4.8): A single peak in the negative region (this region extends beyond the first decalog).

Case studies 39 and 40

separated from the other lymphoid cell populations (Figures 3.59b, 3.66d, and 3.67), and the intensity of CD10 on malignant FCCs is, in most instances, stronger than that on the reactive counterparts. CD20 levels on benign GCCs overlap with that on neoplastic FCCs, however. The “horizontal hockey stick” pattern so typical of FRFH cannot be appreciated if the fluorochrome chosen for CD10 is PE instead of FITC. The high quantum yield of PE blurs the subtle difference in CD10 intensities between reactive GCCs and neoplastic FCCs (Figures 2.7 to 2.9), resulting in similar CD10–PE/CD20 patterns, which would then require unnecessary additional testing for bcl-2. There are two infrequent scenarios that require caution in the interpretation of the CD10– FITC/CD20 pattern. The first scenario is florid hyperplasia, in which the subpopulation of CD10+ reactive B-cells, which lacks surface immunoglobulins, can predominate over the remaining T- and B-cells, producing a CD10/CD20 picture simulating that seen in FCC lymphoma. Testing for bcl-2 is useful in such instances, because the patterns of the combined reactivities to bcl-2 and CD20 demonstrate a clear-cut difference between FRFH and FCC lymphoma (Figures 2.9 and 4.11). The second scenario concerns those cases of FCC lymphoma in which the tumor cells express CD10 poorly, at levels either lower than (Figures 3.68 and 4.12) or overlapping

186

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 4.10 CD10/CD20 reactivity in four different lymph nodes with FRFH. (a–d) The proportion of the brighter CD20 cells with CD10 expression varies with the degree of follicular hyperplasia in the sample.

Case study 41

with that on reactive GCCs. Furthermore, the lymph node, especially if it is only partially involved by lymphoma, may still contain a predominant component of residual benign B- and T-cells. Such combination of poor CD10 expression and partial nodal involvement would produce a CD10–FITC/CD20 pattern simulating that seen in FRFH (Figure 4.13). However, the finding of a monoclonal surface light chain, DNA aneuploidy (Figure 4.12), or bcl-2 expression on the brightest CD20 cells provides the evidence that the follicular process is neoplastic. In very rare instances of FRFH, a small clonal population of CD10+ B-cells may be detected by FCM and its presence confirmed by molecular assay for IgH rearrangement. Such clonal proliferations may represent an exaggerated response of GCCs to antigenic stimulation. The affected patients may or may not have overt underlying immune disorders or active viral infections, however. There are helpful clues to avoid misinterpreting such cases as FCC lymphoma, including (1) the CD10 level is that expected on reactive GCCs, (2) bcl-2 expression is absent, and (3) t(14;18) is not detectable by molecular analysis. Furthermore, follow-up of the patients proves that the process is not malignant.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

187

Figure 4.11 FRFH versus FCC lymphoma. bcl-2 reactivity. (a, b) FRFH: Two subpopulations of B-cells differing in FSC signals and CD20 intensity. The brighter CD20 subpopulation (arrow) is bcl-2 negative. The other benign B- and T-cells are weakly bcl-2 positive. (c, d) FCC lymphoma: The neoplastic cells (thin arrow) are of small cell size, with bright bcl-2 expression.

4.1.2 Normal peripheral blood and normal bone marrow The peripheral blood and bone marrow are discussed together because marrow aspirates contain all of the blood elements. The following discussion does not pertain to specimens sent in for the evaluation of minimal residual disease because a standard data acquisition of 20,000 cells may not detect rare residual leukemic cells, especially when they are present at about the 10−4 level. A normal-appearing specimen by FCM analysis is defined by the absence of abnormalities. The following should not be present in normal specimens: (1) an increased number of blasts; (2) a marked preponderance of a specific cell population (lymphocytes, monocytes, granulocytes, red cell precursors, or eosinophils); (3) abnormal B-cells, T-cells, or plasma cells; (4) abnormal “antigenic” maturation in the granulocytic elements; and (5) nonhematopoietic elements. The size of each cell cluster only reflects its relative proportion. Therefore, the

188

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.12 Follicular lymphoma (FCC II) with dim CD10. (a–d) The tumor cells (R2) are larger than normal lymphocytes, coexpressing weak CD10 (the reagent used is CD10–PE) and weak monoclonal lambda light chain. (e) The staining results are less clear-cut with the reagent combination kappa (1)– FITC and lambda (1)–PE. (f) The tumor is near tetraploid (DI: 1.94) with a low S-phase of 2.3%.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

189

Figure 4.13 FCC lymphoma with a CD10/CD20 pattern simulating FRFH. (a, b) Two populations of B-cells, differing in cell size and CD20 intensity. The larger cells coexpress brighter CD20 and dim CD10. (c, d) The dimmer CD20 cells are polyclonal. The brighter CD20 cells are neoplastic, monoclonal for lambda. In this case, the poorly expressed CD10 on the tumor cells, combined with the substantial number of residual benign B-cells, result in a CD10/CD20 pattern mimicking that of FRFH.

evaluation needs to take into account the CBC and bone marrow cellularity, if the specimen is marrow aspirate. For peripheral blood specimens, the cell populations are assessed in light of the total WBC count. The bone marrow FCM picture in some neoplastic disorders such as low-grade MDS (e.g., refractory anemia) or myeloproliferative disorders (MPDs) (e.g., polycythemia rubra vera [PRV] or essential thrombocythemia [ET]) can be indistinguishable from that in a normal subject. The different cell clusters in the blood or bone marrow can be best appreciated on the SSC/CD45 dot plot where granulocytes, lymphocytes, and monocytes are well separated based on their respective granularity and CD45 intensity (Figure 4.14). Lymphocytes and monocytes demonstrate high levels of CD45, whereas the intensity of CD45 on granulocytes is lower by 1 decalog. In a representative normal bone marrow specimen, the monocytic component is not conspicuous. The appearance of the granulocytic cluster in the bone marrow reflects the

190

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

Figure 4.14 (a) Normal bone marrow composed predominantly of granulocytic (G) precursors. The monocytic (M) and lymphoid (L) cell clusters are small. The blast level is less than 1%. A conspicuous erythroid cluster is not evident because most of the nucleated red cells (13% to 15% on the aspirate smears) have been lysed. (b) Normal peripheral blood with granulocytes, monocytes and lymphocytes.

heterogeneity of the myeloid precursors, which consist of both mature and immature elements ranging from promyelocytes to neutrophils. The bone marrow granulocytic cluster is actually composed of two merging clusters with a variable degree of overlap, whereby the more granular elements (with higher SSC signals) display lower CD45 (Figure 4.14a). Eosinophils, when present in sufficient numbers, manifest as a cluster with extremely high SSC and a CD45 intensity slightly brighter than that on myeloid cells (Figure 4.15). The cell size of the various elements in the blood or bone marrow can be appreciated on ungated FCM graphics correlating FSC and antigenic fluorescence (e.g., FSC/CD45). An adequate bone marrow sample without gross abnormalities typically produces a “tunnel-like” pattern on the FSC/CD45 dot plot (Figure 4.16). Based on the SSC/CD45 dot plot, the cellular events to be evaluated in the subsequent graphics can be separated into two broad groups: granulocytes and mononuclear cells. The latter group includes events from the regions associated with blasts, bone marrow B-cell precursors, lymphocytes, and monocytes, as well as signals from the CD45-negative to the borderline region which may represent blasts (see Section 3.4.2), erythroid precursors, plasma cells, platelets, or nonhematopoietic elements. 4.1.2.1 Blast region In normal specimens, the blast region is essentially empty, with only a few scattered events present (Figure 4.14a). Enumeration of the blast events can be achieved by drawing a gate around this region. A better approach, if blasts express CD34 (the majority of leukemic blasts do), is to determine the blast content from the CD117/CD34, CD33/CD34 (or CD13/CD34), or CD19/CD34 dot plots (Figure 4.17). Note that CD117 may occasionally be found on cells other than blasts, such as plasma cells, abnormal maturing granulocytes, and small cell carcinoma (see Section 3.8). In the blood, the presence of a discrete cluster of blasts, however tiny it may be, is a significant finding (Figure 4.18). In the bone marrow, the threshold generally employed is that which has been traditionally accepted for clinical remission of acute leukemias (i.e., less than

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

191

Figure 4.15 Bone marrow eosinophilia. (a, b) The eosinophilic precursors (arrow) display slightly brighter CD45 than the granulocytes, along with the characteristic combination of low FSC and high SSC. (c, d) Eosinophils are CD16 negative. CD13 is expressed. CD15 is dimmer than that on myeloid cells. The intensity of several of the myeloid antigens on eosinophils overlaps with that on granulocytes. Eosinophils are thus best evaluated on the FSC versus fluorescence dot plots.

5% myeloblasts). However, because of the manner in which the bone marrow is collected in most institutions, the aspirate sample allocated for FCM analysis is often much more hemodilute than the sample spread on the smears. Therefore, a bone marrow blast level of less than 5% in the FCM sample does not necessarily exclude a neoplastic condition. Such hemodilution-related artifacts can be circumvented by performing FCM analysis on either (a) essentially “blood free” concentrated marrow spicules obtained from bone marrow aspirate collected in EDTA tubes (see Section 2.1.1) or (b) cells extracted from bone marrow core biopsies. The presence of a discrete blast cluster in the bone marrow, even without phenotypic aberrancies (see Section 3.5.1.1) and at a level considered within the normal range, can still be cause for concern. Knowledge of the antecedent clinical history and close follow-up are warranted in such cases to exclude or confirm either preneoplastic conditions (MDS) or minimal residual AML.

192

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

Figure 4.16 (a) Normal bone marrow with erythroid precursors (E), granulocytes (G), lymphocytes (L) and monocytes (M). (b) Their relative cell size can be appreciated on the FSC/CD45 dot plot.

4.1.2.2 Bone marrow B-cell precursors

Case studies 14 and 42 to 44

In close proximity to the blast region on the SSC/CD45 dot plot is the location for bone marrow B-cell precursors (morphologically known as hematogones). These cells can be easily recognized by the extremely low SSC and heterogeneous CD45 intensity intermediate between that of lymphocytes and blasts (Figure 4.19 and Plate 6). This characteristic appearance, along with the low FSC, distinguishes benign B-cell precursors, especially when present in significant numbers, from ALL cells. Increased numbers of normal B-cell progenitors, though most common in the pediatric age group (Plate 16), can be seen at any age. Their presence is especially conspicuous in bone marrows affected with solid tumors (e.g., neuroblastoma) or nonneoplastic hematologic conditions, such as idiopathic thrombocytopenic purpura (ITP) or maturation arrest of hematopoietic precursors, as well as in regenerative marrows after chemotherapy/transplantation for acute leukemia. It is not unusual to find leukemic blasts and normal B-cell precursors coexisting in the same sample, especially in residual or early relapsed disease. The distinction of B-cell progenitors from residual AML cells is straightforward because their antigenic characteristics do not overlap (Figure 4.20). In contrast, in the setting of residual/relapsed ALL, it is often necessary to use other approaches (see Section 3.5.2) to differentiate benign B-cell progenitors from residual ALL cells, especially because both components may coexist in a regenerative marrow after induction chemotherapy (see Section 4.3).

4.1.2.3 Lymphocytes The evaluation of B- and T-cells in the blood and bone marrow follows the same steps applied to the lymph node, using the appropriate B- and T-cell dot plots gated on mononuclear cells. The B-cell component is minor and simpler than that in the lymph node, because normal blood and bone marrow contain no follicular mature B-cells with CD10/CD20 coexpression. The variations in the relative proportion of T-cells and B-cells in the peripheral blood from one person to the next fall within a narrower range than that observed in lymph nodes. Similarly, determination of the CD4 : CD8 ratio is more useful in the blood than in solid lymphoid organs because the values are scattered over a more limited range, and the evaluation takes

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

193

Figure 4.17 Normal bone marrow (same sample as in Figure 4.14a). (a) Lymphocytes (L), monocytes (M) and myeloid (G) precursors form distinct clusters on the FSC/CD33 dot plot. (b–d) CD34-positive events are minimal. The blast level is 0.5%.

into account the relative and absolute lymphocyte count. Not infrequently, peripheral blood from normal individuals may contain a very small subset of suppressor T-cells with downregulated CD7, forming a “CD7 tail” on the CD3/CD7 dot plot (Figures 4.21 and 4.22). The T-cell population in normal peripheral blood is heterogeneous. It consists of (1) a major subpopulation of CD3+ TCR-αβ T-cells segregated into CD4+ CD8− and CD4− CD8+ cells, and (2) a minor but highly variable subpopulation of CD3+ TCR-γδ T-cells, some of which express CD8 while most lack both CD4 and CD8. Clinical studies have been focused on the former subgroup. Using the current eight-tube kit of TCR-Vβ antibodies, which detects 70% of the known Vβ repertoire, it has been found that the subsets CD4+ TCR-αβ and CD8+ TCR-αβ display similar distributions of Vβ antigens. However, some of the Vβ families tend to be preferentially expressed by either subset, for instance: (a) Vβ2, Vβ5.1, and Vβ6.7 by CD4+ T-cells, and (b) Vβ7.1 by CD8+ TCR-αβ T-cells. The pattern of the Vβ repertoire remains relatively stable with aging. Small expansions of one or more of the Vβ families may be encountered in the elderly, however, are most commonly associated with expanded CD8+ T-cells (Figures

194

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 4.18 Peripheral blood with 1% circulating blasts (arrow). (a) A tiny cluster in the blast region. (b–d) Blasts display an abnormal phenotype: CD34++, CD13++, CD33−. CD19 is negative. CD117 (not shown) is expressed.

4.23 to 4.25). Such expansions may be regarded as the T-cell equivalent of the low levels of monoclonal B-cells of undetermined significance, also often seen in the older age group. Non-neoplastic alterations affecting blood or bone marrow lymphocytes involve primarily the T-cell component, most commonly manifesting as an inverted CD4 : CD8 ratio and/or activation changes. This can be observed in immunosuppressive states, including those secondary to chemotherapy or underlying malignancies. Decreased numbers of CD4 cells and/or increased numbers of CD8 cells, when seen in advanced stages of CLL and other LPD/NHL, may imply a poor prognosis. Evaluation of the TCR-Vβ repertoire in some of these conditions may reveal small polyclonal or oligoclonal expansions, most often found in the CD8+ population (Figures 4.26 to 4.28). Several studies have reported such expansions in conditions marked by chronic antigenic stimulation, including autoimmune disorders, underlying malignancies, and allogeneic BMT. In viral processes, such as EBV-infectious mononucleosis or cytomegalovirus (CMV) infection, the reactive proliferation of suppressor T-cells also produces an inverted CD4 : CD8 ratio

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

195

Figure 4.19 Pediatric bone marrow with 29% B-cell progenitors (BCPs). (a–e) The BCPs are variable in cell size. They are also heterogeneous in terms of CD45, CD19, CD20 and CD10 expression. This reflects the presence of different subpopulations of BCPs at various stages of maturation and differentiation. (f) TdT is present in a subpopulation.

196

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.20 Coexisting leukemic blasts and B-cell progenitors. (a–f) Blasts (arrow), hematogones (thin arrow) and erythroid precursors (open arrow) are present in addition to the lymphocytes and granulocytes. Blasts, coexpressing CD34, CD13 and CD33, comprise 18% of the cells in the sample; CD19 is negative on the blasts. CD19 and CD10 are present on hematogones. Erythroid precursors display intense CD71; most are of small cell size.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

197

Figure 4.21 (a, b) Peripheral blood with benign relative lymphocytosis, composed mostly of T-cells. (c–e) Gated on the small subset with markedly downregulated CD7 (arrow): The cells coexpress normal levels of CD8, CD2 and CD5. (f) A small number of true NK cells (R2), lacking CD3 and coexpressing CD56 and bright CD7 are present. NK-like T-cells (thin arrow) also express CD56.

198

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 4.22 Peripheral blood with benign relative lymphocytosis (continuation of Figure 4.21). (a–d) NK cells (R2) are CD5−, CD4− and CD2++. CD8 is dim and partly expressed. A small number of NK-like T-cells (thin arrow) expressing heterogeneous CD8 are present. They are positive for CD3, CD2, CD5, CD7 and CD56 (see Figure 4.21).

Case studies 45 and 46

(Figure 4.29). Samples from the sites involved (peripheral blood, lymph node) by viral infections are rarely submitted for FCM analysis except when the clinical picture is unusual (e.g., older age group) and the patient is suspected of harboring a malignant process instead. The peripheral blood (or bone marrow) cytology and lymph node morphology in EBV infection can simulate that of a large cell lymphoma (Plate 17). Furthermore, the FCM data often reveals a conspicuous T-cell population (CD4− CD8+) with decreased CD7 intensity (Figures 4.29 and 4.30). In most instances, this is the only antigenic “aberrancy” observed. Rarely, CD5 may also be downregulated. The positive serological data and the absence of clonality based on the evaluation of the TCR-Vβ repertoire are helpful clues to avoid misdiagnosing EBV infection as malignant lymphoma, however. Ultimately, the benign nature of the EBV induced CD8 T-cell proliferation is confirmed by the self-limited clinical course of the infection. In HIV-infected subjects, an inverted CD4 : CD8 ratio is the result of the destruction of CD4 cells and, in the early stages, a compensatory increase in CD8 cells. Other abnormalities

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

199

Figure 4.23 Oligoclonal CD8 T-cell expansion in an 80-year-old patient with relative lymphocytosis. (a–f) Gated on peripheral blood lymphocytes. A cluster of abnormal T-cells of small cell size (arrow) with (1) normal CD3 levels; (2) lack of CD2; and (3) slight downregulation of CD5 and CD7. The CD4 : CD8 ratio is reversed (1 : 2). The abnormal T-cells are CD8++.

200

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.24 Oligoclonal CD8 T-cell expansion in an 80-year-old patient with relative lymphocytosis (continuation of Figure 4.23). (a–f) Analysis of the TCR-Vβ repertoire gated on CD8+, CD2− abnormal T-cells (gate R2 from Figure 4.23f). There are small expansions of the Vβ17 and Vβ21.3 families (tubes 2 and 6, respectively).

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

201

Figure 4.25 Oligoclonal CD8 T-cell expansion (continuation of Figure 4.24). (a, b) Gated on CD8+, CD2− abnormal T-cells (gate R2 from Figure 4.23f): No abnormality detected in TCR-Vβ tubes 7 and 8. (c, d) Gated on CD8+, CD2+ (gate R3 from Figure 4.23f): Expansion of the Vβ17 family. No abnormality detected in tube 3 and the other TCR-Vβ tubes (not shown). Polyclonal TCR-Vβ in the CD8−, CD2+ population (data not shown). PCR studies were negative for T-cell gene rearrangement.

Case studies 47 to 49

include perturbations in the Vβ repertoire of both CD4+ and CD8+ subsets, and “clonal” expansions of CD4+ T-cells. In some cases, indirect evidence of a marked polyclonal hypergammaglobulinemia may be seen on the FCM graphics (Figures 4.31 and 4.32). The coating of excess immunoglobulins on non-B-cells results in an apparent “positive” reactivity for all heavy and light chains. The loss of the suppressor subset of CD4 T-cells, which normally regulates B-cell production of immunoglobulins, most likely accounts for this abnormality. With additional washing steps during cell processing and staining, it is possible to eliminate this “undesirable” finding from the FCM data. Lymphocytes expressing NK markers (see Section 1.4) are more easily detectable in the blood than in the bone marrow or lymph nodes because of their distinctive cytology, that of

202

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.26 Oligoclonal CD8 T-cell expansion in a 68-year-old patient with an absolute reactive lymphocytosis. (a–f) Gated on peripheral blood lymphocytes: a prominent population of “abnormal” T-cells (arrow) with downregulated CD7 and slightly decreased CD5. The cells are CD8+ without CD56 expression. A small population of NK cells (thin arrow) with the profile CD3−, CD5−, CD2++, CD7++, CD56++ and dim CD8, is present.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

203

Figure 4.27 Oligoclonal CD8 T-cell expansion in a 68-year-old patient with an absolute reactive lymphocytosis (continuation of Figure 4.26). (a, b) The “abnormal” T-cells (arrow) display low FSC. The CD4 : CD8 ratio is reversed. (c–f) Analysis of the TCR-Vβ repertoire gated on the “abnormal” T-cells: There are small expansions of the Vβ3 and Vβ13.1 families (tubes 1 and 4, respectively).

204

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 4.28 Oligoclonal CD8 T-cell expansion in a 68-year-old patient with an absolute reactive lymphocytosis (continuation of Figure 4.27). (a–d) No abnormality detected in TCR-Vβ tubes 5 through 8.

large granular lymphocytes (LGLs). In addition to the presence of one or more NK antigens, this group can also be recognized by weak or absent CD8 expression (Figures 4.21, 4.22, 4.26 and 4.33). LGLs are phenotypically heterogeneous, however, broadly classified into two major groups according to the expression of both CD3 and TCR complex: (1) true NK cells, lacking both CD3 and TCR (Figure 4.34), and (2) NK-like T-cells (CD3+ and TCR+), with the majority expressing the αβ heterodimer. In normal healthy individuals, NK-like T-cells express heterogeneous CD57. Other NK antigens, namely CD16, CD94 and the KIR antigens, are infrequently present (Figure 4.35). In the peripheral blood, true NK cells, CD3− CD56+, account for about 15% of the circulating lymphocytes. The NK-cell population is further classified into two different subgroups based on the expression of CD56. The subpopulation with lower levels of CD56 predominates in the blood while that with higher CD56 reactivity is found in the tissues and organs. The

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

205

Figure 4.29 Peripheral blood with EBV viral lymphocytosis in a young adult male. (a) A prominent lymphoid cluster. The SSC/CD45 picture is reminiscent of that seen in an LPD. (b–f) The lymphoid cells are of variable cell size, coexpressing CD3, CD7, CD2, CD5 (not shown) and CD8. The expression of CD7 and CD8 is heterogeneous. CD56, CD57 and CD16 (not shown) are not expressed. EBV serology was positive in this case.

206

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.30 Infectious mononucleosis in an 8-year-old child. (a–f) Gated on peripheral blood lymphocytes: a prominent population of “abnormal” T-cells (arrow), with downregulated CD7, slightly increased FSC and bright CD8 expression. CD5 is slightly decreased and more heterogeneous than on normal T-cells, whereas CD2 and CD45 are slightly brighter. The diagnosis of EBV infection was confirmed serologically.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

207

Figure 4.31 Peripheral blood from an HIV-positive patient. (a–d) Gated on MNCs: normal T-cells with a reversed CD4 : CD8 ratio (0.47 : 1). There is also indirect evidence of marked polyclonal hypergammaglobulinemia (the CD20-negative cells are coated with immunoglobulins and appear to be positive for kappa and lambda). A small number of polyclonal B-cells are present.

two subsets also differ morphologically, functionally, and phenotypically in regard to the pattern of expression of several antigens such as CD5, CD62L (a homing antigen), CD16 and the KIR antigens. Phenotypically, the CD56+high subset is homogeneously positive for CD94 and CD62L, but lacks CD5, CD57, CD158a and CD158e. CD16 reactivity is absent or downregulated. CD56+high NK cells produce abundant cytokines but have little cytotoxic activity, and show no cytoplasmic granularity by light microscopy. In contrast, CD56+low NK cells contain cytoplasmic granules and express CD16, but lack CD62L. A significant fraction of these cells show positivity for CD57. The expression of CD94 is more heterogeneous. The proportion of cells expressing CD5 (dimmer than T-cells) and the KIR antigens varies highly from one individual to another. Increased LGLs can be observed in altered immune conditions, including autoimmune disorders, chemotherapy, and underlying malignancy. Reactive LGL proliferations can persist over long periods of time. Because of overlapping clinical and laboratory features, the

208

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 4.32 Peripheral blood from an HIV-positive patient (continuation of Figure 4.31). (a–d) Gated on MNCs: marked hypergammaglobulinemia. B-cells are IgM and IgD positive. IgA and IgG are negative.

distinction between reactive and neoplastic LGL proliferations has been problematic, unless the number of LGLs is so overwhelming as to suggest malignancy or the LGLs are composed of NK-like T-cells (i.e., CD3+ TCR-αβ+), in which case rearrangements of the TCR-β gene can be investigated by Southern blot analysis. For NK LGL proliferations, it is more difficult to prove clonality due to the lack of rearrangement of any of the TCR genes. If the patient is female, however, then clonal determination may be obtained through X-linked polymorphism gene analysis. More recently, the evaluation of the TCR-Vβ repertoire using antibodies against several Vβ families has made it possible to assess T-cell clonality in persistent NK-like T-cell LGL proliferations by FCM analysis. 4.1.2.4 Monocytes In addition to its specific location on the SSC/CD45 dot plot, the mature monocytic population also displays the highest levels of CD14 (Leu-M3) and CD64 (Figures 3.40 and 4.36).

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

209

Figure 4.33 NK cells in the peripheral blood. (a–d) Gated on MNCs (which account for 50% of the total cell population). NK cells (R2) are negative for CD3, but coexpress CD7, CD5, CD2 (not shown), CD8, CD16 and CD56. CD5 (arrow), as well as CD2 (not shown), are slightly weaker than on T-cells. CD8 is heterogeneous and dimmer than that on suppressor T-cells. In this sample, NK cells comprise 6% of the cells analyzed.

The CD14/CD64 dot plot is useful for evaluating the maturation of the monocytic elements, especially in neoplastic monocytic proliferations (see Section 3.5.1) in which the more immature elements demonstrate lower CD14 reactivity. Other myeloid antigens such as CD13, CD33, and CD11b are also present on monocytes. The intensities of some of these antigens differ between monocytes and myeloid cells (e.g., brighter CD33 on monocytes than on granulocytes), thus facilitating the distinction between these cell types on the ungated dot plots correlating FSC with the respective myeloid antigens (Figures 4.17a and 4.36b). 4.1.2.5 Plasma cells In the bone marrow, the brightest CD38+ cells are plasma cells. Under normal conditions, the proportion of plasma cells in the FCM sample rarely exceeds 1% for several reasons,

210

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.34 Reactive lymphocytosis (WBCs: 13.9 × 109/L, 50% lymphocytes). (a–e) Twenty-five percent of the lymphoid cells are true NK cells (gate NK). The cells display coexpression of CD7, CD2 and CD56. CD8 is heterogeneous and less intense than that on CD8 T-cells. CD3 and CD5 are absent. NK-like T-cells (arrow) are also present. (f) Gated on NK cells: TCR-αβ and TCR-γδ are not expressed.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

211

Figure 4.35 Reactive lymphocytosis (continuation of Figure 4.34). (a) CD56 present on NK and NK-like T-cells (gate T-NK). (b–d) Gated on NK-like T-cells: heterogeneous expression of CD7 and CD94. CD158a, CD158b and CD158e1 (NKB1) are absent. CD2 and CD5 are homogeneous (data not shown). (e, f) Gated on NK cells: CD94 is expressed, CD158a dim to absent, CD158b bimodal, CD158e1 present on a small subset.

212

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 4.36 Normal bone marrow. Monocytes (M) can be recognized by their location on the SSC/CD45 dot plot (a). The cell size is in the medium range (b). CD33 is brighter than that on granulocytes (G), whereas CD15 is dimmer (c). CD14 is intense (d).

Case study 48

including (1) the aspirates allocated for FCM analysis are often hemodilute, (2) the distribution of plasma cells in the bone marrow is focal, (3) plasma cells tend to adhere to the spicules and, therefore, are not well released into the cell suspension, and (4) plasma cells may be relatively fragile and therefore easily lost during cell processing. A 5- to 20-fold discrepancy in the proportion of plasma cells between the FCM sample and the aspirate smear is a common occurrence. Therefore, the finding of a plasma cell population at levels >1% as determined from the CD138/CD38 or CD45/CD38 dot plot (Figure 4.37) would suggest some degree of plasmacytosis. Evaluation of other surface antigens or cytoplasmic light chains will determine whether the plasma cell population is reactive or neoplastic. Polyclonal plasmacytosis in solid lymphoid organs is an uncommon occurrence, and may be associated with altered immune conditions. In addition to the very intense CD38 expression, benign plasma cells also display the following immunophenotypic characteristics: • Polyclonal cytoplasmic light chain expression. • Lack of surface immunoglobulins.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

213

Figure 4.37 Normal bone marrow with mild plasmacytosis. (a, b) A small population (R3) with the brightest CD38 expression and FSC in the medium range. CD45 is weak. (c, d) Gated on the brightest CD38 cells: Both cytoplasmic kappa and lambda are detected. In this sample, plasma cells comprise 1% of the cells analyzed. The corresponding fresh bone marrow aspirate smears contain 10% plasma cells.

• Downregulated to absent CD45. • Well-expressed CD138. • The combination of CD19 reactivity and a lack of CD56 (as well as the lower FSC and SSC values) differ from that seen on neoplastic plasma cells.

4.1.2.6 Erythroid precursors By FCM analysis, the erythroid component in normal bone marrow specimens often appears insignificant because many of the late erythroid precursors (i.e., polychromatophilic and later) are eliminated during the red cell lysis step. This effect can be appreciated by comparing the marrow aspirate smear to the cytospin made after sample preparation. There is variability in the degree of red cell lysis across laboratories, however. On the SSC/CD45 dot plot, the erythroid events loosely form a cluster with very low SSC and negative to borderline CD45 intensity (Figure 4.38).

214

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 4.38 Erythroid precursors in a normal bone marrow. (a–d) The erythroid population forms a cluster in the CD45-negative to borderline region. Most of the precursors are of small cell size and display the brightest coexpression of CD71 and glycophorin A.

The few antibodies currently available for evaluating erythroid cells include CD71 (transferrin receptor) and anti-glycophorin A (Gly-A). The maturation process from the early erythroid precursor to the erythrocyte stage is accompanied by a loss of CD45 while the glycophorin A level remains unchanged after reaching its peak at the basophilic erythroblast stage. In a normal bone marrow, early erythroid precursors (i.e., basophilic erythroblasts and earlier) are few. The erythroid population consists predominantly of late erythroid precursors, identifiable on the FCM dot plots by their low FSC and bright coexpression of CD71, CD36 (a platelet glycoprotein), and Gly-A (Figures 4.38 and 4.39). In benign conditions with erythroid hyperplasia such as B12/folate deficiency or red cell hemolysis, the erythroid cluster seen on the SSC/CD45 dot plot becomes more conspicuous. The FCM graphics correlating FSC and erythroid markers (Figures 4.39 and 4.40) help to determine whether or not the hyperplasia also contains a preponderance of early (and thus larger) erythroid precursors. Other preneoplastic or neoplastic conditions with erythroid hyperplasia, such as low-grade MDS or PRV, in which most abnormalities are in the erythroid series, can yield a similar picture. These disorders may display no other FCM abnormalities because

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

215

Figure 4.39 (a–d) Bone marrow with erythroid hyperplasia (M : E ratio 1 : 2 by morphology). Erythroid precursors (gray) comprise 41% of the cells in the FCM sample and display normal coexpression of CD71, CD36 and Gly-A. There is a relative increase in erythroid precursors with higher FSC (i.e., more immature). (e, f) Unlysed RBCs (gray), CD71−, Gly-A+, in the peripheral blood of a patient with hemoglobinopathy.

216

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 4.40 Normal bone marrow with a mildly increased number of erythroid precursors and slight eosinophilia. (a, b) A conspicuous erythroid cluster (arrow) present in the CD45-negative region. A small ill-defined cluster of eosinophils (thin arrow) is also visible, best seen on the FSC/SSC dot plot. (c, d) Gated on MNCs: The erythroid precursors range from small to large (most are small) and display the brightest CD71 expression. In this sample, erythroid cells comprise 12% of the cells analyzed. The fresh bone marrow smears contain 33% erythroid cells (M : E ratio 1.5 : 1) and 7% eosinophils.

the blast level is usually within the normal range and myeloid antigenic maturation is often devoid of overt abnormalities. The presence of either unlysed red cells, recognizable as a CD71− Gly-A+ population (Figure 4.39e, f), or pronounced thrombocytosis with platelet clumping can also produce a substantial cluster located in the CD45-negative region of the SSC/CD45 dot plot. Reactivity for CD41 and a cell size much smaller than lymphoid cells, as well as the tendency to pick up PE nonspecifically, are clues to confirm the identity of the platelet population (Figures 4.41 and 4.42). 4.1.2.7 Maturing myeloid cells Flow cytometry data analysis on maturing myeloid cells consists of assessing the granularity and antigenic maturation. Overt hypogranularity (an infrequent finding, Figure 4.41a) or

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

217

Figure 4.41 Peripheral blood with neoplastic thrombocytosis. (a–c) Giant platelets and platelet clumps form a conspicuous cluster in the CD45-negative to borderline region (arrow), with low FSC and brightest reactivity for CD41. CD71 is negative. A tiny blast cluster is also visible (thin arrow). The granulocytes display low SSC (see Figures 4.14b and 4.18a for comparison) and are not well separated from the monocytes. (d) Myeloblasts (R1) comprise 2% of the total cell population.

hypergranularity (e.g., secondary to G-CSF; see Section 4.5.1) can be detected on the SSC/ CD45 dot plot. The maturation process is best evaluated on the CD13/CD16, CD11b/CD13 and CD11b/CD16 displays, with the data gated on the granulocyte cluster only. On these three graphics, the myeloid component in the peripheral blood, being composed mostly of mature elements, forms a single well-defined cluster with bright coexpression of CD13, CD11b, and CD16 (Figure 4.43b). This appearance is altered if the blood contains a substantial number of circulating intermediate myeloid precursors (e.g., associated with G-CSF effect), in which case the patterns on the CD13/CD16 and CD11b/CD16 dot plots resemble those seen in an aspicular hemodilute bone marrow. In contrast with peripheral blood granulocytes, the bone marrow myeloid precursors generate a continuous curvilinear population with heterogeneous coexpression of CD13, CD16, and CD11b, thus reflecting the antigenic maturation of the myeloid elements (Figure 4.43). The normal CD13/CD16 maturation curve is reminiscent of a “wide open checkmark,” whereas the normal CD11b/CD16 curve appears more angular and conical-like

218

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 4.42 Bone marrow with marked thrombocytosis (platelet count 902 × 106/L). (a–d) The prominent cluster with low SSC and downregulated CD45 (gray) are platelet clumps positive for CD41 and CD36. CD71 and Gly-A are negative. Erythroid cells are few. CD71 is abnormally downregulated in some of the erythroid precursors (arrow).

to horseshoe-like. The uneven “U” shape of the normal CD11b/CD13 maturation curve reflects the bimodal distribution of CD13 (present at the highest levels at the beginning and end of the granulocytic maturation process) and the high levels of CD11b in the intermediate and late myeloid precursors. Alternatively, the shape of the single parameter CD16 and CD11b histograms (gated on granulocytes) can be used to assess myeloid maturation. In the authors’ experience, there exists a good correlation between the antigenic maturation and the morphologic maturation seen on the bone marrow aspirate smears. Abnormalities in the maturation curves can be observed in both neoplastic (e.g., AML, MDS) and benign conditions (e.g., G-CSF therapy). Another useful marker to study myeloid cells is CD10. CD10 is well expressed on mature granulocytes, but at lower levels than on normal B-cell progenitors (Figure 4.43e,f).

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

219

Figure 4.43 Antigenic features of normal granulocytes in the bone marrow and peripheral blood. (b, f) Peripheral blood: The mature granulocytes form a single cluster with bright CD13, CD16, CD11b (not shown) and clear-cut CD10 coexpression. (a, c–e) Bone marrow: CD13/CD16, CD11b/CD16 and CD11b/CD13 maturation curves reflecting the orderly granulocytic maturation. CD10 is positive on the more mature granulocytes.

220

FLOW CYTOMETRY IN HEMATOPATHOLOGY

4.2

Abnormal heterogeneous samples with a detectable immature neoplastic population In the blood and lymphoid organs, the finding of any cellular events in the blast region (normally “empty”) on the SSC/CD45 dot plot is synonymous with a malignant process. In the bone marrow, however, the process is recognized as neoplastic only when either the increased number of blasts exceeds the currently accepted threshold (Figure 4.44), or their phenotype is not that expected of normal myeloblasts. Several diagnostic possibilities come into consideration, including acute leukemia (de novo, relapsed, or residual), a myeloproliferative disorder (MPD) with increased blasts, or a high-grade myelodysplastic syndrome (MDS). The expression of immaturity markers (CD34, CD117, or TdT), myeloid antigens, or lymphoid antigens is determined from the appropriate FCM graphics to establish the phenotype of the blasts. To facilitate this visual review, it is usually preferable to limit the data on these dot

a

b

c

d

Figure 4.44 Bone marrow with high-grade MDS. (a) A small distinct blast cluster (arrow) is present. (b–d) Ungated data: Blasts are distinct from the myeloid precursors (thin arrow), coexpressing CD34, CD13 and CD33. The intensity of CD13 and CD33 on the blasts is relatively heterogeneous. CD19 is negative. Blasts comprise 12% of the cells analyzed.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

221

Figure 4.45 Bone marrow with high-grade MDS (continuation of Figure 4.44). (a–c) Gated on granulocytes: Altered myeloid maturation curves with downregulated CD16. There is also loss of CD10 expression. (d) Data ungated: No abnormalities noted in the CD14/CD64 expression on granulocytes (thin arrow).

plots to either mononuclear cells (i.e., granulocytes excluded) or the blast cluster only by appropriate gating on the SSC/CD45 dot plot. In addition to evaluating the blast population, the myeloid maturation curves and CD10 expression on granulocytes also need to be assessed, especially if the blast cluster is of myeloid lineage. In myeloid disorders, abnormalities in the CD13/CD16, CD13/CD11b and/or CD11b/CD16 maturation curves, or altered CD10 expression are common (Figure 4.45).

4.2.1 Blasts of lymphoid lineage If the blast population is of lymphoid origin, either precursor T-cell or precursor B-cell, then the diagnosis is ALL/lymphoblastic lymphoma irrespective of the body site and the relative proportion of tumor cells. One notable exception is the thymus, as encountered in thymic hyperplasia or thymoma, in which the overall phenotype of normal thymocytes is superficially

222

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 4.46 Precursor B-ALL. (a, b) Blasts (R1) comprise 20% of the cells analyzed and coexpress bright CD34 and CD19 (CD19/CD34 dot plot gated on R1). (c) Blasts coexpress bright CD10 and CD19. Granulocytes (R2) display weaker CD10 expression and are negative for CD19 (there is higher background fluorescence associated with PE). (d) Gated on R2: essentially normal CD13/CD16 myeloid maturation curve.

Case studies 13, 15, 28, 50 and 51

similar to that of T-lymphoblasts. The graphical FCM data are quite dissimilar between thymocytes and T-ALL cells (Figures 3.14 and 3.100), however, with the latter being a more homogeneous population than the former. In those cases of ALL where the involved bone marrow still contains a substantial number of myeloid precursors, the granulocytic maturation curves appear essentially normal (Figure 4.46), unless the ALL originates from an underlying CML (Figure 4.47). In lymphoid organs, granulocytes are not normally present. Therefore, the FCM findings of lymphoblasts accompanied by significant numbers of granulocytes give a straightforward diagnosis of a CML or MPD in lymphoid blast crisis.

4.2.2 Blasts of myeloid lineage The differential diagnosis is more complex if the blast population is of myeloid lineage, requiring morphologic correlation with the hemogram, blood, and bone marrow smears to

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

223

Figure 4.47 CML with lymphoid blast crisis. (a–c) Blasts (R1) comprise 35% of the cells analyzed and coexpress bright CD19, CD34 and CD10. Weak CD10 is present on a proportion of the myeloid precursors. (d) Gated on R2: abnormal myeloid maturation curve similar to that in Figure 4.45a.

subclassify the disorder (AML vs. high-grade MDS vs. MPD with increased blasts) according to the currently accepted criteria. This step is especially important if the FCM sample is the bone marrow aspirate, where, as already mentioned, hemodilution can artifactually lower the blast proportion. The bone marrow is generally the specimen of choice for analysis in neoplastic myeloid disorders, especially if a further classification of AML into the different subtypes is deemed necessary. Otherwise, if the blast level in the peripheral blood is that of an overt leukemia (the current criteria fluctuates between 20% and 30%), then FCM analysis of a blood sample is sufficient for establishing the diagnosis of non-M3 AML. A more precise interpretation of AML with monocytic differentiation (see Section 3.5.1) can be achieved if the monocytic cluster is also prominent (Figure 3.42a,b). However, if both the monocytic and blast clusters in the blood are conspicuous and the proportion of blasts is below the threshold for acute leukemia but higher than that accepted for CMMoL (e.g., in the 10% to 15% range), then the distinction between AML with monocytic differentiation and the so-called CMMoL “accelerated phase” cannot be made with certainty in the peripheral blood (Figures 3.44 to

224

FLOW CYTOMETRY IN HEMATOPATHOLOGY

3.46). The difficulty also lies in the fact that in AML with monocytic differentiation, especially the subtype M5b, the circulating population consists of a much larger proportion of more mature elements than the tumor population in the bone marrow. 4.2.2.1 AML

Case study 52

Case studies 53 to 55

Case study 56

A heterogeneous bone marrow picture, in which blasts are admixed with other maturing hematopoietic elements at the time of the initial presentation, is seen primarily in AMLs with evidence of maturation. The main subtypes of AML that fall into this category include AML with maturation (AML-M2), AML with erythroid hyperplasia (AML-M6), and the rare acute megakaryoblastic leukemia (AML-M7) (see Section 5.1.2.6). Some cases of AML with monocytic differentiation (or “AML with a monocytic component”), namely AML-M4 and M4E may also present a similar heterogeneous picture. In AML-M6, the blast population is accompanied by a conspicuous erythroid cluster (Figure 4.48). The relative proportion of erythroid cells by FCM is often not as dramatic as the degree of erythroid hyperplasia seen on the aspirate smears, as many of the late precursors are likely to be eliminated at the red cell lysis stage during specimen preparation. Inspection of the CD71/Gly-A or CD71/CD45 dot plots may disclose evidence of abnormal antigenic maturation in the erythroid precursors. The appearance of the FCM dot plots in AML-M4 and M4E varies widely from case to case depending on the relative proportion of blasts, granulocytic and monocytic components, and whether or not the blast and monocytic cell clusters merge with each other. In many cases, the number of myeloid precursors is minimal, and the blasts and monocytic elements produce a single large merging cluster, as seen on the SSC/CD45 dot plot (see Section 3.4.1). Cases of AML-M4 with a substantial number of granulocytic elements may appear morphologically similar to AML-M2, especially if the monocytic component is not cytologically obvious. Furthermore, the NSE stain in AML-M4E is often negative and the abnormal eosinophils (Plate 18) may not be conspicuous. The FCM graphics in such instances are most helpful, as the monocytic cluster on the SSC/CD45 dot plot is insignificant in AML-M2, but more prominent in M4 and M4E (Figures 4.49 and 4.50). Occasionally, the SSC/CD45 display in AML-M2 may mimic that seen in monocytic disorders, however, as a result of the unusual location and shape of the blast cluster, shifting toward the monocytic region (Figure 4.51). Another useful feature pointing toward AML-M4 is the “trail” pattern on the CD14/CD64 dot plot, which corresponds with the spectrum of immature and mature monocytes (Figure 4.49) showing heterogeneous CD14 expression. The blast population itself, especially in AML-M4E, may not express any monocytic associated antigen, however (see Section 3.5.1). The eosinophilic component in AML-M2Eo (Figure 4.52) or M4E may be visible on the SSC/CD45 and FSC/SSC dot plots, as well as on the bivariate displays of FSC versus myeloid antigens. The phenotypic profile of the blast population in non-M3 AML varies widely from case to case. Other myeloid antigens besides CD13 or CD33 may be expressed. In some cases, a lymphoid-associated antigen or CD56 may also be present. As a result, a given combination of antigenic features present at the time of diagnosis can serve as a “fingerprint” for the FCM analysis of subsequent bone marrow samples to assess the patient’s disease status and response to therapy (see Section 3.5.1.1). However, because of clonal evolution, the antigenic profile of leukemic myeloblasts at relapse may not be entirely identical to that established at the time of the initial diagnosis. In any of the above-described non-M3 AMLs, the maturing myeloid elements may also display phenotypic abnormalities such as hypogranularity (decreased SSC), downregulated CD16, CD11b (Figures 4.48 and 4.51), or CD10 expression (Figure 4.52), especially when the leukemia is preceded by a myelodysplastic process.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

225

Figure 4.48 AML with erythroid hyperplasia. (a) Conspicuous erythroid cluster. The blast cluster (arrow) is poorly separated from the granulocytic cluster. (b) Gated on MNCs: Blasts (R2) are weakly positive for CD117; a subset expresses dim CD34. (c) Gated on blasts: CD33 and CD13 (not shown) are expressed. Blasts comprise 15% of the cells analyzed. (d) Increased erythroid precursors: Intense CD71 expression. (e, f) Gated on granulocytes: Abnormal maturation curves with downregulated CD16 and CD11b.

226

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 4.49 AML with a monocytic component. (a) Merging blast and monocytic clusters. (b, c) CD34positive blasts (B) comprise 20% of the cells analyzed. CD13 and CD33 are expressed. (d) Gated on R1: CD14/CD64 trail pattern reflecting the heterogeneity of monocytic cells at various stages of maturation. When compared with the isotype-matched negative control (not shown), blasts are CD14 negative. CD64 is partially expressed.

4.2.2.2 High-grade MDS and MPD with increased blasts The graphical FCM data in high-grade MDS (Figures 4.44 and 4.53), where the blast count is, by current criteria, between the threshold accepted for clinical remission and that for overt acute leukemia, are closely similar to that observed in AML. The blast cluster is smaller, however. Abnormalities in the granulocytic maturation curves are common. CD10 on the granulocytes can also be downregulated. The myeloid cluster may display abnormally low SSC signals (Figure 4.53). Occasionally, the appearance of the markedly hypogranular myeloid population on the SSC/CD45 dot plot may mimic that of a prominent monocytic cluster (Figure 4.54). Paralleling the FCM abnormalities, the aspirate smears in high-grade MDS demonstrate abnormal morphology with an excess of blasts and left-shifted myeloid maturation along with qualitative abnormalities such as hyposegmentation and hypogranulation. Because both the

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

227

Figure 4.50 AML-M4E (morphology shown in Plate 18). (a, b) Merging blast and monocytic clusters (R1). Eosinophils (R2) are conspicuous, best seen on the FSC/SSC dot plot by their low FSC. (c, d) Blasts (R3) comprise 49% of the cells analyzed and express CD34, CD13 and CD33 (not shown). CD13 is of variable intensity. CD19 is negative. (e) Blasts and eosinophils (arrow) are negative for CD14 and CD64 (not shown). (f) Gated on R2: CD16 is absent on eosinophils.

228

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.51 AML with maturation. (a) The SSC/CD45 picture is reminiscent of a monocytic disorder because of the shape of the blast cluster (R2). (b–e) Blasts comprise 24% of the cells analyzed and are positive for CD34, CD13, CD33 and CD7. CD14 and CD64 are not expressed. (f) Gated on granulocytes: altered CD13/CD16 myeloid maturation curve with downregulated CD16.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

229

Figure 4.52 AML-M2Eo (morphology shown in Plate 26). (a) A conspicuous blast cluster (R1) and an ill-defined population of eosinophils (arrow). (b–d) Gated on R1: Blasts comprise 30% of the cells analyzed and coexpress CD34, CD13, CD33 and CD19. A small portion of the blasts lack CD34. (e, f) Gated on granulocytes: Myeloid precursors are antigenically abnormal, with loss of CD10 and downregulated CD13, CD16 and CD11b (not shown). Cytogenetics revealed t(8;21).

230

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 4.53 Bone marrow with high-grade MDS. (a, b) Decreased SSC on granulocytes. There is poor separation of the granulocytic, monocytic and blast clusters on the SSC/CD45 dot plot. (c) Blasts (R2) comprise 8% of the cells analyzed and express CD34, CD33 and CD13 (not shown). (d) Gated on granulocytes (using the FSC/SSC dot plot): altered myeloid maturation curve with downregulated CD16.

Case study 57

Case studies 58 and 59

FCM and morphologic pictures of the bone marrow in high-grade MDS can closely resemble that of residual disease/impending relapse of AML, it is critical to be aware of the antecedent clinical history or to review any earlier FCM and bone marrow data. An overt blast cluster of myeloid lineage, along with abnormal myeloid maturation curves similar to that seen in AML and MDS may also be observed in certain MPDs with increased blasts such as CML in accelerated phase. The bone marrow morphology may not be distinguishable from that of high-grade MDS. Knowledge of the peripheral blood data becomes important to discriminate between these two groups of disorders. The diagnosis of CML with increased blasts is made straightforward by the conspicuous presence of basophils on the SSC/CD45 dot plot, a finding more easily appreciated in the blood than in the bone marrow (Figure 4.55). The manifestation of MPD/CML in the peripheral blood, with abundant circulat-

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

231

Figure 4.54 High-grade MDS (8% blasts). (a–f) A small blast cluster (arrow) and hypogranular granulocytes (gray) producing a prominent monocytic-like cluster. The coexpression of CD15, CD16 and CD10 confirms the granulocytic identity of this cluster. CD16 is downregulated (e). Graphics (a), (c) and (e) are backgated from (b), (d) and (f), respectively. Cytogenetics revealed del 5q, del 7q, −7, −8, +13, −14, and + markers.

232

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.55 Peripheral blood: CML with 14% blasts. (a–e) Blasts (thin arrow) express CD34, CD13 and CD33. Platelet clumps/giant platelets (arrow) display dim CD45, but intense CD41 (P). Basophils (open arrow), located between the lymphoid and hematogone regions on the SSC/CD45 dot plot, express CD33 (and other myeloid markers) similar to granulocytes, but display lower FSC (Baso). (f) Gated on granulocytes: CD16 is downregulated on the circulating intermediate myeloid precursors (IMP).

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

233

ing immature myeloid precursors, may produce an FCM and morphologic picture similar to that of the bone marrow.

4.3 Minimal residual disease

Case study 14

The study of MRD has been focused on disorders with blastic proliferations, namely acute leukemias. The evaluation of MRD by FCM analysis requires (1) knowledge of the detailed antigenic features present on the leukemic cells at the time of the initial diagnosis and (2) data acquisition on a large number of events (in the range of 1 million) using a limited antibody panel tailored to the phenotype sought after. FCM data analysis, in turn, is centered on those dot plots and histograms most useful for showing the known antigenic characteristics of the leukemic blasts and, where applicable, separating these from the normal cells with a similar phenotype. A typical example is the evaluation of MRD in precursor B-ALL, whereby the dot plots TdT/DNA content, CD10/CD20, and CD58/CD19 are some of the most useful graphics (Figures 2.19, 3.49, 3.52 and 4.56) for differentiating leukemic blasts from bone marrow B-cell precursors (see Section 3.5.2). In AML, the detection of residual disease relies on the presence of phenotypic aberrancies identified at the time of diagnosis (see Section 3.5.1.1) such as a lack of either CD13 or CD33, or markedly downregulated/negative CD45 (Figure 4.57). When such aberrancies are not observed, molecular genetics may be helpful if the AML is associated with certain specific genetic abnormalities for which molecular probes are currently available.

4.4 Abnormal heterogeneous samples with detectable mature neoplastic populations The finding of an abnormal mature B- or T-cell population is virtually synonymous with involvement by an LPD/NHL. In patients with overt systemic involvement by LPD/NHL, the disease in the lymph node may progress to a higher-grade large cell lymphoma while the bone marrow involvement remains unchanged and composed of small neoplastic cells. This asynchronous picture is a well-known phenomenon in FCC lymphomas. Similarly, in CLL/SLL, the relative proportion of activated cells in the lymph node is higher than that present in the blood or bone marrow. In the bone marrow or blood involved by LPD/NHL, the SSC/CD45 dot plot can appear deceptively normal, especially if the tumor cells are small in number or in cell size (Figures 4.58 and 4.59). In disorders such as HCL, or low involvement by small cell lymphoma (e.g., Sézary syndrome, and the early stages of CLL or LGL leukemia), benign cells in the background may obscure a small population of tumor cells. If the neoplastic cluster is relatively conspicuous, subtle abnormalities such as slightly downregulated CD45 expression in the lymphoid cells can be appreciated (see Section 3.4.3). In such instances, the lymphoid cluster may extend toward the left into the hematogone region. In involvement by large cell LPD/ NHL, the neoplastic cluster is usually more easily discerned on the FSC/SSC or FSC versus fluorescence displays (Figures 2.13 and 3.90).

(text continues on page 238)

234

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.56 MRD, day 30 after allo-PBSCT in a patient with t(9,22) precursor B-ALL. (a) Increased signals in the blast region. (b–f) Gated on MNCs: Myeloblasts (CD117+, CD34+) are below 0.5%. The critical cells (gray) are residual ALL cells (1%), coexpressing CD34, CD58, CD19 and CD10. CD117 and CD20 are negative. Benign B-cell progenitors (5%) are also present (arrow).

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

235

Figure 4.57 MRD, day 4 after FLAG induction in a patient with AML. (a) G-CSF effect, producing an SSC/CD45 pattern reminiscent of AML-M3. Note the scattered signals in the CD45-negative to borderline region. (b–d) There are 0.6% myeloblasts (B) coexpressing CD34, CD117, CD13 and CD33 (not shown). (e) Backgated from the “B” gate: The abnormal myeloblasts are CD45-negative to dim. (f) Granulocytes are abnormal with downregulated CD16.

236

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.58 Bone marrow with 2% residual CLL. (a) Essentially normal SSC/CD45 picture. (b–f) Gated on MNCs: The small population of small B-cells with homogeneous CD19 expression is actually composed of two populations differing in CD20 intensity. The population with dimmer CD20 (R4) coexpresses CD5 and CD23, but has no detectable light chain. CD20 intensity is also less than CD19. The population with brighter CD20 is polyclonal.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

237

Figure 4.59 Bone marrow with minimal involvement by MCL (negative morphology on the corresponding aspirate smears and core biopsy). (a) Essentially normal SSC/CD45 picture. The small population of B-cells with homogeneous CD19 expression (b) consists of two populations differing in CD20 intensity. (c–f) Gated on MNCs: The brighter CD20 cells (R4) are CD5+ CD23−. Lambda is well expressed and CD20 slightly brighter than CD19. Cells with weaker CD20 are hematogones (arrow).

238

FLOW CYTOMETRY IN HEMATOPATHOLOGY

4.4.1 Abnormal mature B-cells The most efficient way to detect abnormal B-cells in a heterogeneous sample is by the combined staining of surface light chains and pan B-cell antigens (see Section 2.6.1.1). Heterogeneous specimens, especially those derived from solid tissue, usually contain a substantial number of benign reactive B-cells in addition to malignant B-cells. In some cases, the benign component predominates. In the authors’ experience, CD20 has proved to be most useful in separating the malignant and benign B-cell populations from each other. Depending on the type of lymphoma, the neoplastic cells often display either brighter or weaker CD20 than the benign cells (Figures 1.1, 3.23, 4.13, 4.58, and 4.60). On the dot plots that correlate CD20 and surface light chains, this differential in CD20 intensity permits the identification of the monoclonal cluster. In most cases, the picture is less clear-cut on the light chain/CD19 dot plots, as both the malignant and benign populations often display similar CD19 intensities (Figures 3.23, 4.58, and 4.60). The sensitivity of the combined staining of the light chains with CD20 is sufficient for detecting very low numbers of abnormal B-cells, such as in HCL, “T-cell rich B-cell” lymphoma, or minimal involvement by FCC lymphoma in which the corresponding cytologic or histologic preparations may be negative by light microscopy. The heterogeneity of the B-cell component in B-cell LPD/NHL may also be appreciated on the FSC/CD20 or FSC/CD19 dot plots, which often show several B-cell clusters differing by cell size and CD20 (or CD19) intensity (Figures 1.1 and 4.60). By evaluating the FSC/CD20 and light chain/CD20 dot plots together, it is possible to deduce which B-cell cluster(s) is neoplastic and calculate its relative proportion. Alternatively, a gate is drawn systematically around each of the B-cell clusters present on the FSC/CD20 (and/or FSC/CD19) dot plot, and the neoplastic cluster(s) is identified based on the pattern of light chain distribution seen on the single parameter histograms (Figures 4.60 and 4.61). The finding of superimposed unimodal distributions of kappa and lambda in the negative region indicates no detectable light chain expression. With the exception of benign CD10+ B-cells in marked FRFH (see Section 4.1.1), the absence of both light chains on mature CD20+ or CD19+ cells implies a neoplastic B-cell process. Once the presence of a monoclonal B-cell population or a B-cell population negative for light chains is established, then graphics displaying the key antigens CD10, CD11c, CD25, and CD103 (see Section 3.6.3), along with CD5 and CD23 are evaluated (Figure 4.62). This permits the characterization of the B-cell LPD/NHL immunophenotypically, while the cell size and S-phase fraction are assessed for grading the tumor. 4.4.1.1 Evaluation of CD5 and CD23

Case study 60

The presence or absence of CD23 in monoclonal B-cells is only meaningful when CD5 is expressed, especially if the neoplastic population is of small cell size. In this context, the fluorescence intensity of the surface light chain, the pattern of CD20 intensity within the tumor population, and the relationship of CD20 and CD19 also need to be evaluated carefully (see Section 3.6.2). The various combinations of findings frequently encountered are described next. Six combined features represent the FCM “fingerprint” of CLL/SLL (Figures 3.16, 3.18, 3.37, and 4.58). These are (1) downregulated CD20, (2) CD20 intensity < CD19 (provided they are conjugated to the same fluorochrome, or to fluorochromes of comparable quantum yield), (3) heterogeneous CD20 expression within the tumor population, (4) weak to absent surface light chain (i.e., the difference between the mean fluorescence intensity for kappa and lambda does not exceed 1 decalog), (5) CD20 (or CD19), CD5, and CD23

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

239

Figure 4.60 Lymph node with low-grade B-cell NHL NOS, morphologically monocytoid B-cell lymphoma (morphology shown in Plates 45 and 46). (a–f) Two populations of B-cells (R2 and R3) differing in cell size and CD20 expression, but with overlapping CD19 intensity. The larger cells (medium size) with brighter CD20 are monoclonal for lambda. The smaller cells with dimmer CD20 are polyclonal. Monoclonality is better appreciated in (c, d) than (e, f).

240

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.61 Lymph node with low-grade B-cell NHL NOS (continuation of Figure 4.60). (a) Gated on R3 (of Figure 4.60): Monoclonal lambda is well expressed. (b) Gated on R2 (of Figure 4.60): A polyclonal pattern. (c–f) The tumor is negative for CD5, CD10 (not shown) and CD103, but positive for CD25. CD25 is also present on some T-cells. Interestingly, the CD4 : CD8 ratio is markedly increased (no evidence of TCR gene rearrangement or aberrant T-cell phenotype).

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

241

Figure 4.62 Peripheral blood with less than 1% hairy cells. (a–f) Gated on MNCs: two B-cell populations differing in cell size. The larger cells (R3) display medium FSC, brighter CD19 and CD20, distinct expression of CD25 and CD103, intense CD11c, and monoclonal kappa light chain (arrow). The smaller CD20-positive cells (R2) are polyclonal.

242

FLOW CYTOMETRY IN HEMATOPATHOLOGY

coexpression, and (6) small cell size. This FCM “fingerprint,” being highly specific, is quite useful for detecting MRD in CLL, especially because more aggressive therapy is now being applied to this group of patients. Recent studies have shown that in this setting, FCM analysis is much more sensitive for the detection of minimal disease than PCR-based methodologies. When the pattern of CD20 or surface light chain expression departs from the above description (i.e., either becomes brighter), a close inspection of the FSC parameter may reveal an increase in larger and more activated cells (see Section 3.6.2). Cases of CLL with an increased number of activated cells usually present with an overwhelming number of neoplastic cells in the blood or bone marrow and suppression of other hematopoietic elements. The resulting picture on the SSC/CD45 dot plot is that of a nearly homogeneous specimen, in which a predominant cluster occupies the lymphoid region (Figures 3.60 and 3.61). Six FCM features in combination, (1) bright CD20, (2) CD20 intensity > CD19, (3) CD5 positive, (4) CD23 absent, (5) bright monoclonal surface light chain expression, and (6) small cell size (in most instances), are a virtual fingerprint of MCL (see Figures 3.64 and 4.59). However, if CD20 and CD19 are similar in intensity (both bright), other LPD/NHLs with antigenic features similar to MCL need to be considered in the differential diagnosis (see Section 5.2.1.4). Additional testing for cyclin D1 expression either by FCM or immunohistochemistry, and correlation with the morphologic preparations, serum immunoelectrophoresis results, and bcl-1 rearrangement or the (11;14) translocation, are warranted in such cases. The disorder is unlikely to be MCL if the surface light-chain expression is weak or CD20 is dim and less intense than CD19 (Figure 4.63). The latter two features favor a CLL-related malignancy instead, such as lymphoplasmacytoid/lymphoplasmacytic (LPC) lymphoma/leukemia. The combination of medium/high FSC, CD5+, and CD23− occurs infrequently (Figure 4.64). The diagnostic possibilities include CD5+ high-grade B-cell lymphoma, PLL, and, less likely, the “blastic” variant of MCL. CD20 and a monoclonal surface light chain are well expressed in the latter, but may be variable in the other two disorders.

4.4.1.2 FCM features suggestive of anti-CD20 therapy While evaluating the expression of CD20, it is possible to suspect the effect of anti-CD20 therapy if the pattern of CD20 reactivity is incongruous with the remaining FCM data (Figure 4.65). Although a history of anti-CD20 therapy should be provided when submitting the specimen to the FCM laboratory, it is unfortunately not always available. After anti-CD20 administration, the content of neoplastic cells can be dramatically reduced, especially if the B-cell lymphoma is one with bright CD20 expression. In the follow-up specimens, the interaction between the therapeutic chimeric anti-CD20 antibody and the residual B-cells masks the detection of CD20, resulting in an apparent decrease to absence of CD20 expression. This finding, when seen in the context of an abnormal B-cell population with either (1) CD10 expression, (2) small cell size, CD5+, CD23−, and bright surface light chain, or (3) small cell size, CD5−, CD23−, CD10− and bright surface light chain, should raise the suspicion of anti-CD20 therapy. If the neoplastic B-cells are large, however, then it is unclear whether the downregulated CD20 is inherent or secondary to anti-CD20 therapy, unless a prior FCM study is available for comparison. When CD20 is markedly downregulated, such as in some cases of CLL/SLL or after anti-CD20 therapy, the proportion of neoplastic cells may not be easily determined based on CD20, surface light chains, or FSC, especially if the tumor cells are of small cell size. Because

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

243

Figure 4.63 Bone marrow with low involvement by CD5+ CD23− B-cell LPD, morphologically LPC lymphoma–leukemia. (a) A conspicuous lymphoid cluster. (b–f) Gated on MNCs: The lymphocytes are of small cell size and consist mostly of T-cells. B-cells are CD5+ (arrow), CD23− and express weak monoclonal kappa. CD20 is less intense than CD19. The tumor comprises 4% of the cells analyzed.

244

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.64 Lymph node FNA with CD5-positive LCL. (a–e) The neoplastic B-cells (R1) are large and comprise 18% of the cells analyzed. The tumor expresses CD5 and weak kappa light chain. A small number of granulocytes (arrow) represent blood contamination. (f) Gated on R2: The population of small B-cells (R2) is polyclonal for kappa and lambda. The tumor has a DI of 1.07 and an S-phase fraction of 21% (data not shown).

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

245

Figure 4.65 Lymph node with residual FCC lymphoma after anti-CD20 therapy. (a–d) The neoplastic cells are of variable cell size. CD20 (which in FCC lymphoma is normally more intense than CD19) is downregulated and dimmer than CD19. CD10 is well expressed. The tumor displays no detectable light chain.

benign and malignant B-cells often share similar CD19 intensities, examination of CD19 may not be helpful in this regard either. In such instances, the problem can often be resolved by analyzing the surface light chain expression (FITC) in conjunction with CD5 (PE) or CD10 (PE).

4.4.2 Abnormal mature T-cells and NK cells A small population of abnormal mature T-cells can be detected based on one or more of the findings mentioned in Section 3.6.4. Most useful is the absence/downregulation or upregulation (however slight) of one or more T-cell markers, which causes the neoplastic T-cells to be visibly separated from the normal T-lymphocytes (Figures 3.89, 3.90, and 4.66). Such phenotypic abnormalities form the basis for identifying mature T-cell LPD/NHLs, the majority of which are CD4+. The T-cell marker with altered expression can then serve as an anchor marker

246

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 4.66 Peripheral blood involvement by PTCL. (a) The tumor (arrow) is part of the lymphoid cluster, but displays slightly increased SSC. (b, c) The neoplastic cells comprise 15% of the cells analyzed and display variable cell size, upregulated CD5 and downregulated CD7. A tiny population of residual small benign T-cells (thin arrow) is present. (d) The CD4 : CD8 ratio is 2.7 : 1. Monocytes (open arrow) display dim CD4.

Case studies 61 and 62

in the subsequent analysis of the TCR-Vβ repertoire (Figures 4.67 to 4.69). Other antigenic features, for example, expression of CD30 and ALK-1 (see Section 5.2.3.3), and biological parameters, including cell size and proliferative fraction, help to further characterize the mature T-cell neoplasm (Figures 3.108 and 3.109). In the early stages of mycosis fungoides (MF)/Sézary syndrome, the fluorescence pattern of T-cell markers on the tumor cells in the peripheral blood varies from case to case (Figure 4.70) and may be identical to that of normal T-cells. On the FCM displays, the malignant and benign cells are merged into one single cluster (Figure 4.70a). The CD4 : CD8 ratio may be markedly increased, however, with the Sézary cells contributing to the excess of CD4 cells (Figure 4.70b). This is a rare instance where the “percent-positives” approach (otherwise applicable only for the enumeration of B- and T-cell subsets) may be useful for the initial identification of a mature T-cell malignancy. Further testing for the TCR-Vβ repertoire on the CD4 population should be performed to obtain a more definite proof of clonality, however. Other types of cutaneous T-cell lymphomas may also have low-level involvement of the bone

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

247

Figure 4.67 Bone marrow (hemodilute) with low involvement by PTCL. (a, b) Abnormal T-cells (R2) comprise 1% of the cells analyzed. They display loss of CD7 and coexpression of CD3 and CD4. (c–f) Gated on R2: The abnormal cells display bright CD2 and CD5. There is expansion of the Vβ13.6 family. The other TCR-Vβ tubes (e.g., tubes 3 and 6) revealed no abnormality.

248

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.68 Indolent NK-like T-cell leukemia. (a) Hemodilute bone marrow with a conspicuous lymphoid cluster. (b–f) Gated on lymphocytes: Most of the T-cells are abnormal (arrow), with downregulated CD5 and slightly decreased CD2 and CD7. CD8 is brightly expressed and CD56 is absent. The malignant cells are small.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

249

Figure 4.69 Indolent NK-like T-cell leukemia (continuation of Figure 4.68). (a, b) The malignant cells display abnormal coexpression of CD16 and CD57. (c–f) Analysis of the TCR-Vβ repertoire gated on the CD5−, CD8+ population (arrow). There is clonal expansion of the Vβ2 family. No abnormality noted in the other TCR-Vβ tubes (e.g., tubes 1 and 7).

250

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 4.70 Peripheral blood with Sézary cells. Gated on MNCs. (a, b) Case 1: Sézary cells are not well separated from normal T-cells and are located at the lower end (i.e., with dimmer CD7) of the CD3/CD7 cluster. The CD4 : CD8 ratio is 7.3 : 1. The level of Sézary cells is about 5%. (c, d) Case 2: Sézary cells comprise 10% of the cells analyzed and are distinct (R2) from normal T-cells with loss of CD7 and downregulated CD3 expression. CD4 intensity is similar to that on normal T-cells. The CD4 : CD8 ratio is 7.5 : 1. Monocytes (M) are CD4 positive.

Case study 63

marrow/peripheral blood (Figure 4.71). Therefore, the diagnosis of Sézary syndrome requires confirmatory clinical and laboratory data concerning the patient’s skin lesion(s). More difficult to assess are the early stages of low-grade true NK and NK-like T-cell malignancies (Figures 4.68, 4.69, 4.72 and 4.73), the majority of which consist of bland-appearing LGLs. These two groups of disorders are often designated together under the term LGLleukemia. The latter group occurs more frequently than the former. Phenotypically, the FSC signals, as well as the fluorescence intensity patterns of CD8, NK-cell markers, and other T-cell antigens are essentially identical between the reactive and low-grade malignant LGL proliferations (see Section 4.1.2.3). Furthermore, the blood and bone marrow involvement by indolent LGL-leukemia is usually not overwhelming, in contrast to the aggressive T/NK disorders (see Section 3.6.5). The most commonly encountered phenotypic combination is that of an NK-like T-cell with a TCR-αβ phenotype (i.e., positive CD3/TCR-αβ, CD57+, down-

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

251

Figure 4.71 Bone marrow from a patient with multiple skin plaques but no erythroderma. (a) The bone marrow contains 7% eosinophils (arrow) and 5% lymphocytes. (b–e) Gated on lymphocytes: Half of the cells (R2) are abnormal T-cells with loss of CD7, slightly downregulated CD3 and coexpression of CD56. CD5 and CD4 are coexpressed. CD2 is slightly upregulated. There is clonal expansion of the TCR-Vβ13.1 family (f). PCR revealed a rearranged T-γ gene.

252

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 4.72 Indolent NK-cell leukemia. (a) Hemodilute bone marrow with a prominent lymphoid cluster. (b–d) Gated on lymphocytes: NK cells (gray) account for 78% of the lymphoid cells. They are CD3−, CD7++ and lack both CD4 and CD8. B-cells are virtually absent.

regulated CD8 and CD5, TCR-γδ−, and CD4−) (Figure 4.74). Other antigenic combinations such as CD3+, TCR-γδ+, CD4−/CD8−/TCR-αβ− occur much less frequently. Two phenotypic features serve as diagnostic clues to separate low-grade malignant CD3+ TCR-αβ+ NK-like T-cell proliferations from their reactive counterparts: • Coexpression of CD16 and CD57 (Figure 4.69). This is considered an aberrancy because in normal individuals, CD16 is rarely present on CD57+ CD8+ T-cells. • Significant clonal expansion of a TCR-Vβ family (Figures 4.68 and 4.69), especially when the restricted expression involves a family not normally preferentially expressed by cytotoxic T-cells. This finding is analogous to the rearrangement of the β-chain gene when assessed by molecular analysis. Alternatively, the neoplastic T-cell LGLs lack reactivity against all of the currently available Vβ antibodies. This finding indirectly implies that the proliferation is clonal, expressing a Vβ determinant for which no antibody is yet available.

Clonal expansions of T-cells, especially within the CD8+ population, also occur with aging and in conditions with chronic antigenic stimulation (see Section 4.1.2.3). Non-neoplastic expansions of other T-cell subsets occur much less frequently. In the evaluation of the TCR-

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

253

Figure 4.73 Indolent NK-cell leukemia (continuation of Figure 4.72). (a–f) Gated on lymphoid cells: The leukemic NK cells (gray) lack both CD5 and CD3. They display weak CD11b, bright CD16, and coexpression of CD56 and CD57.

254

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.74 Bone marrow with indolent NK-like T-cell leukemia. (a) A conspicuous lymphoid cluster. (b–f) The neoplastic cells comprise 23% of the cells analyzed and display small cell size and normal expression of CD2, CD3 and CD7, but weaker CD5 (arrow) than normal T-cells. CD8 and CD57 are expressed. The intensity of CD8 is dimmer than that on suppressor T-cells.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

255

Vβ repertoire, the following clues are helpful for distinguishing a monoclonal/neoplastic T-cell proliferation from a polyclonal or oligoclonal process: • The number of expanded Vβ families. The great majority of malignant proliferations demonstrate a single Vβ expansion. In contrast, reactive conditions are usually associated with expansion of more than one Vβ family (Figures 4.23 to 4.28). • The degree of the expansion, that is, the proportion of T-cells under consideration expressing a particular Vβ family compared with that observed in the blood of healthy individuals. This is an important factor to be considered because there exist benign cases with a single Vβ expansion; the expansion is much less significant than that seen in malignant proliferations, however. On the other hand, some cases of LGL leukemia display more than one expanded Vβ family. One of the Vβ expansions is the dominant one, involving a significant number of T-cells. The other expansions are much smaller, thus suggesting that the tumor might have been preceded by an oligoclonal process. • The expansion involves a Vβ family that is not normally preferentially expressed by CD8 T-cells in healthy individuals. This abnormality is highly suggestive of malignancy even when the size of the expansion is relatively small.

Case studies 64 and 65

Indolent CD3-negative LGL (i.e., true NK) leukemias are rare (Figure 4.75). To determine the clonal nature of these proliferations is a challenging task because NK cells do not have rearrangements of the TCR genes. In normal individuals, the great majority of the circulating NK cells are CD16+ CD56+, and a significant number of these (about 66%) express CD8. Therefore, lack of CD56 or CD8, or coexpression of CD57, would suggest that the NK proliferation is abnormal (Figures 4.72 and 4.73). In addition, abnormalities in the expression of the KIR antigens have been reported in NK and NK-like T-cell indolent LGL-leukemias. The restricted expression of a single KIR antigen on NK cells may be accepted as putative evidence of clonality. On the other hand, because the number of currently commercially available KIR antibodies is disproportionately small compared with the large KIR repertoire of antigens, complete lack of reactivity to the KIR antibodies should not be assumed to be indirect evidence of clonality.

4.4.3 Abnormal plasma cells present Abnormal plasma cells are defined by the expression of a monoclonal cytoplasmic light chain, which, in most instances, can be evaluated by a side-by-side comparison of the cKappa/ CD38 and cLambda/CD38 dot plots (Figure 4.76). Because of the high correlation between bright CD56 expression (Figure 4.76b) and clonal/neoplastic plasma cell proliferations, testing for cytoplasmic light chains can be omitted in the great majority of cases. Other antigenic features include intense CD38 expression and downregulation or absence of CD45, pan B-cell antigens, and HLA-DR. As in normal plasma cells, CD138 is usually well expressed. Plasma cells, benign or malignant, may also express a myeloid antigen, usually CD14 or CD13. The immunophenotypic characteristics of neoplastic plasma cells can be best contrasted with that of their benign counterparts in conditions where the two coexist, namely in patients with monoclonal gammopathy of undetermined significance (MGUS). Clonal plasma cells often display one or more of the following abnormalities: 1. Higher FSC and SSC than that produced by reactive plasma cells. In many cases of plasma cell dyscrasia, the heterogeneous mixture of small to large neoplastic cells imparts a variable distribution to the FSC parameter. 2. Lack of CD19 reactivity. 3. The expression of CD138 and CD56 varies from case to case. Although intense, the expression of CD38 on clonal plasma cells is often lower than that on normal counterparts (Figure 4.77). Occasionally, CD38 may even be absent. In some cases, the tumor may have no detectable cytoplasmic light chain although the cytoplasmic heavy chain (IgG/IgA) is present. Some plasma cell tumors lack all antigenic reactivity except for CD38 and CD56.

256

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.75 Indolent NK cell leukemia. (a) Hemodilute bone marrow with a sizable lymphoid cluster (R1). (b) Gated on MNCs: The neoplastic cells have identical CD2 intensity to normal T-cells. (c–f) Gated on R1: Neoplastic cells (14% of the cells analyzed) lack CD5, CD3, CD4 and CD8, but express CD7, CD16 and CD56 (arrow).

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

257

Figure 4.76 Bone marrow with plasma cell dyscrasia. (a–d) The intensity of CD45 on the neoplastic plasma cells (arrow) ranges from negative to weak. The tumor cells (15% of the cells analyzed) coexpress intense CD38 and CD56. Monoclonal cytoplasmic kappa is present. There is indirect evident of kappa hypergammaglobulinemia.

Case studies 66 to 68

4. Variable reactivity for markers normally not present on plasma cells such as CD117, CD20, and monoclonal surface light chain (of the same clone as the cytoplasmic light chain) (Figures 4.78 and 4.79). CD45 may also be brightly expressed. As a result, the neoplastic plasma cell cluster may be seen occupying the blast or monocytic region of the SSC/CD45 dot plot (Figure 4.77). Bright CD45 and high FSC tend to correlate with an “immature” plasma cell morphology (i.e., plasma cells with a prominent nucleolus and pale blue instead of deep blue cytoplasm).

The finding of coexisting residual polyclonal and neoplastic cells has been reported to be a strong discriminating feature to separate MGUS from the early stages of multiple myeloma. On the SSC/CD45 dot plot, the plasma cell cluster and erythroid precursors share the same CD45-negative to borderline region. The SSC signals from plasma cells are slightly

258

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.77 Bone marrow with plasma cell dyscrasia. (a) The neoplastic plasma cells (arrow) comprise 17% of the cells analyzed and display bright CD45, merging in with monocytes. (b–e) Gated on MNCs: The tumor cells display variable FSC and express CD56. CD38 is downregulated compared with normal plasma cells. There is indirect evidence of kappa hypergammaglobulinemia. (f) Gated on the brightest CD38 cells: Cytoplasmic kappa is expressed.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

259

Figure 4.78 Bone marrow with plasma cell dyscrasia. (a, b) The tumor cells (arrow) comprise 17% of the cells analyzed and display dim CD45, intense CD38 and reactivity for CD138 and CD56 (not shown). (c, d) CD20 and surface light chain (kappa) are unusually well expressed. A minute population of polyclonal B-cells is present. (e, f) Gated on the brightest CD38 cells: The tumor cells contain cytoplasmic kappa.

260

FLOW CYTOMETRY IN HEMATOPATHOLOGY

higher and more variable than that from erythroid cells, however. In addition, plasma cells can be separated from erythroid elements on the FSC/CD45 dot plot because plasma cells are about the size of monocytes or even larger. Plasma cells can also be spotted on the FSC/SSC dot plot, seen as a cluster with medium/high FSC and with lower SSC than monocytes. This cluster location is not unique to plasma cells, as it can be observed in any large lymphoid malignancy (which may also lack CD45) involving the blood or bone marrow. Because of hypergammaglobulinemia in multiple myeloma, the nonspecific coating of M-protein to cell surfaces is often not completely washed off during specimen processing. This nonspecific binding is most pronounced on monocytes, but it can also be seen on other cells (e.g., activated T-cells, NK cells). As a result, the monocytic cluster appears as an artifactual monoclonal B-cell population, which can be best appreciated when the kappa/CD20 and lambda/CD20 (or kappa/CD19 and lambda/CD19) dot plots are inspected side by side (Figures 4.77 and 4.79). The presence of monoclonal surface light chains on cells other than B-cells is therefore highly suggestive of a plasma cell dyscrasia.

4.5 Abnormal blood or bone marrow samples with no detectable neoplastic cells Samples with no detectable neoplastic cells as referred to in this section are those in which (1) the percentage of myeloblasts is within normal limits, (2) blasts have no detectable phenotypic aberrancies, and (3) no abnormal (or suspicious) mature lymphoid cells or plasma cells are present. However, the relative cellular composition (mainly granulocytes or monocytes) of the blood or bone marrow is altered. The antigenic maturation of the hematopoietic elements may also be abnormal. Neoplastic disorders with such phenotypic abnormalities are predominantly MPDs in the chronic phase (in particular, CML and CMMoL) or low-grade MDS. Certain benign conditions such as a vigorous response to G-CSF therapy or myeloid maturation arrest from different underlying causes can also demonstrate similar findings. Most clinical laboratories are not concerned about the FCM data from the abovementioned conditions. This is especially true in those laboratories where morphology screening is employed as a method for selecting antibodies to be tested on any given sample. Initially, FCM testing was not recommended in MPD or MDS unless there was a suspicion of progression to an acute leukemia. More recently, however, the usefulness of FCM analysis in the evaluation of MDS is being recognized, especially in those cases where other parameters (e.g., morphology, cytogenetics) are indeterminate. The FCM data also contribute to the prognostic assessment of MDS patients. In addition, if the bone marrow sample is representative, a more objective determination of the blast percentage can be obtained by FCM studies (in lieu of a manual differential count), which facilitates the distinctions between lowgrade and high-grade MDS, and between the chronic phase and the accelerated phase of MPD. Because the abnormalities are found mainly in the myeloid or monocytic population, the SSC/CD45, CD14/CD64, CD13/CD16, CD13/CD11b and CD11b/CD16 dot plots (the latter three gated on granulocytes) are the main FCM graphics for assessing the altered cellular composition and abnormal myeloid maturation. The FCM graphics displaying CD56, CD10, CD117 and HLA-DR, as well as the dot plot CD71/CD235a (for those cases with increased erythroid elements) also need to be evaluated.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

261

Figure 4.79 Bone marrow with plasma cell dyscrasia. (a) Neoplastic plasma cells (10% of the cells analyzed) form a distinct cluster (gray) in the CD45-negative region. (b–f) The tumor cells display CD38 and intense CD117. There are several unusual features: expression of HLA-DR and surface light chain (kappa), and lack of CD56. CD56 is present on a subset of granulocytes (the patient was on G-CSF).

262

FLOW CYTOMETRY IN HEMATOPATHOLOGY

4.5.1 Altered cellular composition and abnormal SSC Altered cellular composition can result from an increase in the proportion of any of the normal elements in the peripheral blood or bone marrow. Most often this affects the monocytic lineage. Eosinophilia or basophilia is encountered less frequently. 4.5.1.1 Increased monocytic elements

Case study 69

The most obvious finding in specimens with an altered cellular composition is the increased prominence of the monocytic cluster observed on the SSC/CD45 dot plot, usually accompanied by a corresponding decrease in the size of the granulocytic cluster. This picture, when seen in the bone marrow, is quite suggestive of CMMoL. In the peripheral blood, however, this interpretation cannot be made with certainty because reactive or compensatory monocytosis may yield a similar picture. The presence of an elevated WBC count (via correlation with the hemogram data) or a small blast cluster (blast level well below 10%) on the SSC/CD45 display will point toward CMMoL. Persistent CD56 expression on a monocytic population is another helpful clue for confirming CMMoL (Figure 4.80). The finding of CD56 as the sole abnormality on monocytes must be interpreted with caution, however. In reactive conditions (e.g., with relative monocytosis or mild absolute monocytosis in the peripheral blood), CD56 may be transiently expressed on monocytes, usually at low levels or in a small subset of the cells. In some cases of CMMoL, the bone marrow granulocytic component is markedly diminished (Figure 3.41). The resulting pattern of the cell clusters on the SSC/CD45 dot plot resembles that observed in AML with monocytic differentiation (Figure 3.42). These are cases in which the distinction between an AML (usually M5b) and CMMoL based on the bone marrow morphology is also problematic. Helpful clues for differentiating these two disorders include CD117 expression in AML and the difference in the patterns of CD14/CD64 coexpression (Figures 3.41c, 3.42, and 4.81), which reflect the difference in the relative distribution of blasts and maturing monocytic elements (i.e., promonocytes and monocytes) between the two diseases. Either disorder can produce a vertical “trail” pattern on the CD14/CD64 dot plot. This pattern reflects the heterogeneous and downregulated CD14 expression on the abnormal monocytic population. CMMoL, in contrast with AML with monocytic differentiation (see Section 3.5.1), has a much higher proportion of the more mature elements, however. The mature elements are seen as a cluster expressing bright CD14 and CD64, located at the upper end of the vertical “trail” (Figure 4.81). The lower part of the “trail” is attenuated to insignificant, reflecting the smaller proportion of less mature monocytic cells in CMMoL.

4.5.1.2 Increased eosinophils An extremely prominent eosinophil population is a rare occurrence (Figure 4.82). On the SSC/CD45 dot plot, eosinophils form a cluster with high SSC located immediately to the right of the myeloid cluster (i.e., expressing slightly brighter CD45 than granulocytes). Because of its low FSC, the cluster of eosinophils seen on the FSC/SSC display may be mistaken for degenerated granulocytes (Figure 4.15b). The phenotype of eosinophils is best derived from the dot plots correlating the FSC parameter with the expression of various myeloid antigens. On these graphics, the eosinophil population stands out because of its characteristic low FSC signals (Figures 4.15d and 4.82b,c). The levels of CD13, CD33, and CD11b on eosinophils overlap with that on myeloid cells. The expression of CD15 is, however, distinctly less intense than that on the most mature granulocytes. CD16 is not expressed. The lack of

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

263

Figure 4.80 Peripheral blood with CMMoL. (a–e) The monocytic cluster is slightly enlarged (21% monocytes) and displays brighter CD33 than granulocytes. Monocytes (M) express CD13, CD11b, CD64 and intense CD14. Circulating myeloid precursors (arrow) are evident. The more mature granulocytes have a higher level of CD14 than usual (open arrow). (f) Gated on MNCs: The monocytic population is abnormal by its reactivity for CD56. A tiny cluster of CD16/CD56 positive NK cells is present.

264

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.81 Bone marrow with CMMoL. (a, b) There are 2% blasts and 31% monocytes. Blasts (R3) are CD34++ and CD33++. (c) Gated on MNCs: CD14/CD64 trail pattern reflecting the heterogeneity of monocytic cells at various stages of maturation; the more mature cells predominate. (d) The more mature granulocytes (thin arrow) display a higher level of CD14 than usual. (e) Monocytes are CD56 positive (arrow) and thus neoplastic. (f) Normal pattern of CD10 expression on myeloid precursors.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

Case study 70

265

CD16 expression is, in addition to the low FSC, a useful identifying feature especially when the eosinophilia is composed of hypogranular eosinophils. Hypogranularity in eosinophils, though infrequent, may occur in non-neoplastic conditions including altered immunity and HIV infection. Because of hypogranularity, the eosinophil population is not obvious on the SSC/CD45 display. It remains detectable on the CD16/CD45 dot plot as well as the FCM graphics correlating the expression of myeloid antigens (Figure 4.83), however. Massive eosinophilia of the blood and bone marrow strongly suggests the hypereosinophilic syndrome. Because this MPD is a diagnosis of exclusion, correlation with the clinical picture and cytogenetics is necessary. The diagnosis becomes straightforward if basophilia, however mild, is also present.

4.5.1.3 Conspicuous basophils or mast cells

Case study 71

In contrast to eosinophilia, the finding of basophilia in the blood or bone marrow implies a myeloid malignancy, namely a myeloproliferative process. Because the level of basophilia is much more modest than that seen in eosinophilia, the cell cluster may be overlooked during the evaluation of the FCM graphics. On the SSC/CD45 dot plot, basophils form a small but distinct cell cluster “sandwiched” between the blast region, the monocytic, and the lymphoid clusters. The expressions of CD13, CD33, and CD11b on basophils overlap with that of granulocytes. Because of its distinctive FSC (“sandwiched” between that of lymphocytes and monocytes/granulocytes), the population can be identified on the FCM graphics correlating the cell size and myeloid antigens (Figure 4.84). It can also be recognized on the CD13/CD16 and CD11b/CD16 dot plots as a homogeneous cell cluster lacking CD16. Basophils also lack CD15 and HLA-DR, but may express CD117 in a heterogeneous pattern. The finding of mast cells in a bone marrow aspirate submitted for FCM analysis is extremely rare. Neoplastic mast cell proliferations are uncommon, and the associated marked fibrosis prevents the mast cells from being released into the bone marrow aspirate. The identification of a mast cell population by FCM is straightforward based on several key phenotypic features, namely (1) moderate to high SSC, (2) bright CD45, and (3) intense CD117 (Figure 4.85). Several myeloid antigens such as CD13 and CD33 are also expressed. The presence of a cell cluster with the usual combination of high SSC and bright CD45 yields an SSC/CD45 picture nearly pathognomonic of mast cell disease. Another rare disorder, Langerhans histiocytosis, may produce the same SSC/CD45 pattern (Figure 3.121). Its phenotypic profile does not overlap with that of mast cell disease, however. According to recent reports, neoplastic mast cells also express CD2 and CD25. These two antigens are not expressed in benign mast cell proliferations occurring in reactive marrow or in association with other hematologic malignancies.

4.5.1.4 Abnormal SSC in granulocytes

Case study 47

On the SSC/CD45 dot plot, abnormalities of the myeloid cluster are usually not easily discernible, apart from hypogranularity or hypergranularity (Figures 4.86 and 4.87). For the observation to be valid, the instrument gain for the SSC parameter should be appropriately set (i.e., not excessively high). Low SSC signals can be observed in some cases of MDS in which the corresponding morphology displays extensive hypogranulation among the myeloid precursors, irrespective of whether the MDS is low- or high-grade. When present as the sole abnormality, low SSC is not a sufficient feature for identifying MDS, however, as hypogranularity can be encountered in other conditions such as HIV infection or some cases of MPD (Figure 4.41).

266

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.82 Bone marrow eosinophilia in a 41-year-old patient, 10 months after matched unrelated donor bone marrow transplant (MUD-BMT) for T-ALL. (a–f) The bone marrow contains 37% eosinophils (gray), identifiable by their high SSC, low FSC and lack of CD16. CD45 levels are slightly higher and CD15 levels slightly lower than those on granulocytes. CD13, CD33 and CD11b are present. There are no detectable T-ALL cells (data not shown). Cytogenetics: Normal karyotype.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

267

Figure 4.83 Bone marrow with hypogranular eosinophils. Graphic (a) is backgated from graphic (d) where the eosinophil cluster (gray) is most visible. (a) The cluster of hypogranular eosinophils (16% of the cells analyzed) can be misidentified as a subset of the granulocytic population. (b–f) Other phenotypic features are retained, including low FSC, lack of CD16, CD45 brighter and CD15 weaker than on granulocytes.

268

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.84 Bone marrow with basophilia in CML-like MPD. (a) A conspicuous basophil (gray) cluster (6%). (b–f) Basophils are identified by their FSC, which is intermediate between that of lymphocytes and granulocytes. The intensities of CD33, CD13 (not shown) and CD11b overlap with those on granulocytes. CD16 and CD117 are absent (arrow). Granulocytes are abnormal, with downregulated CD16 and expression of CD117 in a subset.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

269

Figure 4.85 Mast cell disease in the bone marrow. (a) Mast cells (MC) produce an additional cell cluster on the SSC/CD45 dot plot with characteristic high SSC and bright CD45 (at higher levels than that on granulocytes). (b, c) CD33, CD13 (not shown) and CD11b are expressed; CD15 is absent. (d) Intense CD117 is a key characteristic of mast cells.

Increased granularity may be observed in an intense bone marrow response to granulocyte colony stimulating factor (G-CSF) therapy (Figure 4.87), as well as some cases of CML or CML-like MPD (i.e., those MPDs with a blood and bone marrow picture similar to CML but different genotypic findings). The hypergranular myeloid cluster in these conditions occasionally simulates the SSC/CD45 picture of hypergranular AML-M3 (Figure 3.29). The findings on other FCM dot plots, for example, CD13/CD16 and CD11b/CD16, as well as the HLA-DR and CD15 results, differ from those observed in AML-M3 (see Section 3.4.1), however. In recent years, G-CSF has become part of the treatment regimen for acute leukemia, in conjunction with fludarabine and Ara-C in the FLAG protocol. Because G-CSF is administered simultaneously with other chemotherapeutic agents, the bone marrow picture during the early follow-up period of a treated AML (2 to 6 weeks after the start of induction) may be difficult to interpret. Under the effect of G-CSF, a large number of blasts can simultaneously undergo some attempt at (incomplete) maturation. As a result, the treated bone marrow contains a large cohort of intermediate myeloid precursors with a promyelocyte-like appearance due to the

270

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.86 Bone marrow with low-grade MDS. (a, b) Hypogranular myeloid precursors (arrow) and a small number of eosinophils (thin arrow). (c) The blast content is 1% (R2). (d) Loss of CD10 expression on granulocytes. Eosinophils are CD10-negative. (e) Gated on granulocytes: Abnormal CD13/CD16 maturation curve. (f) Ungated: Monocytes (open arrow) and eosinophils (thin arrow) are CD16 negative.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

271

Figure 4.87 Bone marrow with G-CSF effect. (a) Myeloid precursors with increased SSC (hypergranulation). (b–d) Myeloid cells display increased levels of CD64 (as bright as that on monocytes) and CD14. The resulting CD14/CD64 picture nearly simulates the CD14/CD64 trail pattern seen in neoplastic monocytic proliferations. There are only 4% monocytes (M) in the sample analyzed. The myeloid maturation curves are nearly normal.

Case study 72

intense granulation. Although the morphology may simulate AML-M3 (Plate 19), these abnormal myeloid precursors differ from AML-M3 cells phenotypically (Figure 3.28). In rare cases, however, the large cohort of abnormal myeloid cells closely mimics AML-M3 antigenically (e.g., lacking both CD11b and CD16 expression) (Figure 4.88). The presence of distinct CD15 expression, and the lack of the PML-RARα rearrangement are clues pointing away from a possible misinterpretation. A repeat bone marrow with FCM analysis within 4 to 6 weeks should also clear up the diagnostic dilemma, as the picture would be that of a regenerative bone marrow if the patient responded to therapy and the cohort of abnormal cells had (presumably) undergone apoptosis. Caution is also warranted in the interpretation of minimal residual AML in the context of G-CSF therapy, especially if the specimen analyzed is peripheral blood. Even in non-neoplastic conditions, rare circulating blasts may be encountered as part of a vigorous response to G-CSF.

272

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.88 Bone marrow with residual AML and G-CSF effect (G-CSF was part of the induction chemotherapy regimen). (a) Isotype-matched negative controls. (b) Myeloid cells with increased SSC and 9% residual blasts (AML-M1 at presentation). (c–f) Gated on granulocytes: The phenotypic features simulate that seen in AML-M3. Granulocytes are CD13 positive and CD33 positive (not shown) and lack CD11b, CD16 and HLA-DR. Differing from M3, however, is the distinct expression of CD15.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

273

The diagnosis is more straightforward, however, when the myeloblasts, however few, exhibit distinct phenotypic aberrancies (Figure 4.57).

4.5.2 Abnormal antigenic maturation in myeloid or erythroid precursors The diagnostic utility of phenotypic abnormalities on myeloid and erythroid precursors in the study of MDS has been demonstrated in the literature. Some institutions have also devised a scoring of the FCM abnormalities observed on myeloid and monocytic lineages, which can be applied in determining the prognosis of patients with MDS undergoing allogeneic stem cell transplant. It is important, however, that any myeloid and erythroid antigenic abnormalities be evaluated in conjunction with the bone marrow and peripheral blood data, as well as any relevant clinical information. 4.5.2.1 Antigenic abnormalities in myeloid precursors

Case studies 44, 49 and 73

Case study 74

In MDS, the phenotypic abnormalities on maturing myeloid precursors can be best appreciated from the abnormal patterns of the maturation curves CD13/CD16, CD11b/CD16 or CD13/CD11b (Plate 7). The abnormal patterns result from the fact that CD16 or CD11b is downregulated on a large proportion of granulocytes while the expression of CD13 is either at normal levels or upregulated. The differences between normal and abnormal CD13/CD16 maturation curves are usually more obvious than that between normal and abnormal CD11b/ CD16 or CD13/CD11b curves. The antigenic maturational abnormalities on granulocytes are not pathognomonic of MDS, however, because altered myelopoiesis can also be encountered in some MPDs (especially CML and CML-like MPD) as well as certain non-neoplastic conditions, such as HIV infection, severe myeloid maturation arrest of various etiologies (e.g., drug-induced agranulocytosis) or intense G-CSF effect (Figures 4.89 to 4.91). Furthermore, the lack of CD16 expression on granulocytes is also seen in the rare congenital absence of CD16, as well as in patients with unsuspected PNH. In the latter case, CD14 is also absent on monocytes. To confirm the diagnosis of PNH requires further FCM analysis of CD55 and CD59 on erythrocytes. In all of the above-mentioned conditions featuring abnormal myelopoiesis, the common morphologic feature that appears to correlate with the abnormal antigenic maturation curves is a substantial degree of left-shifted myeloid maturation whereby the intermediate myeloid precursors exceed the mature elements. When the myeloid series is normal-appearing or just slightly left-shifted (e.g., low-grade MDS or MPD in chronic phase), the antigenic maturation curves are often essentially normal. In addition to the altered maturation curves, other antigenic aberrancies such as a loss of CD10 or CD64 may also be present on maturing myeloid precursors. Furthermore, a substantial subpopulation may display (a) retention of HLA-DR or CD117 (Figures 4.89 and 4.90, and Plate 7), or (b) expression of nonmyeloid markers, namely CD56 or CD7. Other reported abnormalities include lack of CD13 or CD33, downregulated CD45, and retention of dim CD34 on a subset of granulocytes. These findings are a manifestation of abnormal myelopoiesis, thus pointing toward a stem cell disorder. Their presence is not unique to MDS, however. For instance, granulocytes with loss of CD10, retention of HLA-DR or CD117, or aberrant CD56 expression can be also encountered in CML, CML-like MPD, or G-CSF therapy (Figure 4.90). Increased levels of CD14 and/or CD64 on granulocytes (Figures 4.87, 4.90 and 4.92) is a helpful feature for identifying G-CSF effect. The increased expression is similar to that seen in the neutrophilic response to infectious or inflammatory processes, which, in turn, suggests that the effect of G-CSF on myeloid cells is mediated by the tumor necrosis factor. Increased CD14 or CD64 on myeloid precursors may sporadically occur in MPDs, including CMMoL

274

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.89 CML-like MPD in an elderly patient with leukocytosis and splenomegaly. (a–c) The bone marrow consists mostly of myeloid precursors. Blasts (R2) comprise less than 1% of the cells analyzed. (d–f) Gated on granulocytes (R1): Myeloid cells are abnormal with HLA-DR expression and altered CD11b/CD13 and CD16/CD13 maturation curves. CD117 is expressed in a subset.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

275

Figure 4.90 G-CSF effect on the bone marrow of a patient with lupus. (a) The bone marrow consists mostly of myeloid precursors. (b–f) Gated on granulocytes (R2): Myeloid cells are abnormal with coexpression of HLA-DR and CD56, upregulated CD14, downregulated CD16 and an altered CD11b/CD13 maturation curve.

276

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.91 MPD NOS. (a, b) Bone marrow with 1% myeloblasts (arrow) and a small cluster of basophils (夹). (c–f) CD10 is expressed on granulocytes. A significant proportion of myeloid cells are abnormal with upregulated CD14 and downregulated CD16. Monocytes (gray) are also abnormal with upregulated CD13. CD56 is aberrantly present on both granulocytes and monocytes.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

277

Figure 4.92 Peripheral blood with G-CSF effect. (a) The granulocytic cluster displays increased SSC. A tiny blast cluster (arrow) is present. (b) Blasts (R2), coexpressing CD34 and CD117, comprise less than 1% of the cells analyzed. CD13 and CD33 (data not shown) are expressed. (c) CD14 is increased on most of the granulocytes (thin arrow). There are 4% monocytes. (d) Gated on granulocytes: The presence of circulating intermediate myeloid precursors yields a maturation curve similar to that seen in a hemodilute normal bone marrow (i.e., with an excess of mature granulocytes).

Case study 75

(Figures 4.80e and 4.81d), CML (Figure 4.93), or CML-like MPDs (Figure 4.94), however. The degree of G-CSF-associated phenotypic abnormalities (hypergranularity, altered maturation, upregulated CD14, downregulated CD10) on myeloid precursors varies from case to case depending on the extent of G-CSF therapy and when the bone marrow specimen is obtained. A vigorous response to G-CSF may yield a peripheral blood picture (with a low level of circulating blasts, Figure 4.92) mimicking that of CML. CML usually contains a conspicuous population of basophils (Figure 4.93), however. Among the myelodysplastic syndromes, the phenotypic abnormalities on the maturing myeloid population are less frequently observed in the low-grade than in high-grade MDS. When present in low-grade MDS (Figure 4.86), however, they tend to correlate with abnormal cytogenetic findings, which, in turn, accounts for a more rapid progression to high-grade MDS and eventually acute leukemia.

278

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.93 Peripheral blood with CML. (a) There are 4% basophils (arrow) and 1% monocytes (thin arrow). (b, c) Blasts (B), coexpressing CD34, CD13 and CD33, comprise 0.5% of the cells analyzed. (d) Upregulated CD14 expression on the more mature granulocytes (open arrow). (e, f) Basophils (arrow) express CD13 and CD11b, but lack CD16. Circulating myeloid precursors (see Figure 4.43b for comparison) are present, with an abnormal CD13/CD16 maturation curve.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

279

Figure 4.94 Bone marrow with a CML-like MPD (monosomy 7 syndrome). (a–c) Blasts (R2), which comprise less than 2% of the cells analyzed and coexpress CD34, CD13 and CD33, are not easily seen on the SSC/CD45 dot plot. (d) The myeloid precursors display increased CD64; CD14 is upregulated on the more mature cells (arrow). There are less than 8% monocytes (M) in the sample. (e, f) Gated on granulocytes: Abnormal myeloid maturation curves with downregulated CD16 and CD11b.

280

FLOW CYTOMETRY IN HEMATOPATHOLOGY

4.5.2.2 Antigenic abnormalities in erythroid precursors

Case study 49

The evaluation of phenotypic abnormalities on erythroid precursors is more difficult than that on the myeloid series because (1) a substantial number of cells are eliminated during the red cell lysis step of processing the bone marrow sample, and (2) a very limited number of antibodies are available for studying the erythroid series. The altered antigenic maturation in erythroid precursors is therefore best appreciated in the context of a prominent erythroid component, as seen on the SSC/CD45 dot plot. The most frequent abnormality is downregulated CD71, its expression asynchronous with that of CD45 and CD235a (Gly-A). As a result, the distribution of CD71 on the abnormal erythroid precursors becomes heterogeneous, ranging from very dim to bright (Figure 4.95), in contrast with the homogeneously intense CD71 expression seen on normal erythroid elements. A CD71 abnormality on erythroid precursors is not unique to MDS, however. It has been described in cases of aplastic anemia with morphologic evidence of dyserythropoiesis. The altered erythropoiesis in HIV-infected patients may also demonstrate downregulated CD71. In regenerative marrow rebound erythroid hyperplasia after chemotherapy for acute leukemia, decreased levels of CD71 may be present (transiently) in a subset of erythroid precursors (Figure 4.96). It still remains unclear if benign disorders with erythroid hyperplasia and dyserythropoiesis (namely B12/folate deficiency) also produce similar abnormal CD71 expression.

4.5.2.3 Identifying low-grade MDS

Case study 76

Case studies 74 and 77

The diagnostic evaluation of low-grade MDS can be a difficult process, especially when clear-cut cytologic (e.g., ring sideroblasts) or cytogenetic abnormalities are not present. Furthermore, the patient may be treated with growth factors (namely G-CSF) to stimulate hematopoiesis. In such instances, it may not be possible to determine if the phenotypic abnormalities on the granulocytes are inherent to the MDS itself or secondary to G-CSF therapy. The phenotypic (as well as cytologic) findings can be normal in a substantial number of patients with low-grade MDS, whereas in others, several aberrant antigenic features are present and correlate with abnormal karyotypes. In HIV infection, peripheral cytopenias as well as the cytologic and antigenic abnormalities in erythroid and myeloid cells, may closely mimic those seen in low-grade MDS. The combination of several of the following FCM abnormalities, which should be evaluated in the context of adequate clinical and laboratory data, including the peripheral blood and bone marrow findings, can facilitate the diagnosis of low-grade MDS: 1. Maturing myeloid population with low SSC (hypogranularity). 2. Abnormal antigenic maturation, including (a) altered myeloid maturation curves, (b) aberrant expression of HLA-DR, CD117, CD34, or (c) loss of CD13, CD33, CD64 or CD10 (Figure 4.86 and Plate 7). In the absence of HIV infection, the presence of myeloid cells with abnormal antigenic maturation is most useful for distinguishing hypocellular MDS from aplastic anemia, both of which manifest with peripheral cytopenias and bone marrow hypocellularity. 3. Expression of nonmyeloid antigens (CD56, CD7) on granulocytes. 4. A discrete blast cluster on the SSC/CD45 dot plot, but at a level not exceeding the currently accepted normal range. The presence of any phenotypic aberrancy on the blast population is the most useful diagnostic feature (Figure 4.97). 5. Abnormal CD71 expression on erythroid precursors (Figure 4.95). The proportion of erythroid cells is usually increased.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

281

Figure 4.95 (a, c, d) Low-grade MDS. Bone marrow with hypogranular granulocytes (G), 3% myeloblasts (B) and 53% erythroid precursors (E1). Monocytes (M) are not increased. A small subset of the erythroid cells is abnormal with decreased CD71 and CD36 levels. (b, e, f) Peripheral blood in a patient with agnogenic myelofibrosis: 28% circulating abnormal erythroid precursors (E2) with downregulated CD71 and CD36.

282

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.96 Erythroid hyperplasia in the regenerative marrow of a child with ALL. (a, b) The sample contains hypergranular myeloid cells and 50% erythroid precursors (gray). (c, d) Gated on erythroid cells: A very small subset displays downregulated CD71 and CD36 (these abnormalities disappeared in a subsequent bone marrow). (e, f) Gated on granulocytes: Abnormal maturation curves secondary to G-CSF effect.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

e

f

283

Figure 4.97 Low-grade MDS with abnormal myeloblasts. (a–c) Bone marrow with hypogranular myeloid precursors and 1% blasts (R2) coexpressing CD117, CD34 and CD13. (d) Half of the myeloblasts are abnormal (arrow), lacking CD33. (e, f) Myeloid cells are abnormal with downregulated CD11b and CD16. A subset displays CD117 expression (b).

284

FLOW CYTOMETRY IN HEMATOPATHOLOGY

The abnormal myeloid maturation curves can also be useful when evaluating the follow-up bone marrow specimens of AML patients in clinical remission. In the early posttreatment phase, it is unclear whether the maturational abnormalities, both antigenically and morphologically, reflect residual disease or are secondary to chemotherapy/G-CSF effect. If the patient remains in stable remission, the maturation curves will become essentially normal. When evaluating later follow-up bone marrow samples, it is helpful to compare the maturation curves from one specimen to the next. Via this exercise, it is possible to detect a trend toward an impending relapse even though the blast proportion still remains below the threshold level for clinical remission. During the 3 to 6 months preceding the relapse, the maturation curves may change from normal to abnormal (Figure 4.98).

a

b

c

d

Figure 4.98 CD13/CD16 myeloid maturation curves from remission to relapse in a patient with AML. (a) In remission, 10 months prior to relapse: Normal CD13/CD16 maturation curve. (b) Still in clinical remission (bone marrow blasts <5%), 4 months prior to relapse. Abnormal CD13/CD16 curve with downregulated CD16. (c) Impending relapse (6% bone marrow blasts), 1 month prior to full blown relapse: The abnormal CD13/CD16 curve is essentially identical to that in the diagnostic sample (not shown) and the relapse sample (d), which contains 23% blasts.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

285

4.6 Coexisting malignancies The occurrence of two (or sometimes more) coexisting malignancies, although infrequent, is not rare. One of the neoplastic populations may comprise a very small component in the specimen analyzed and, thus, may escape FCM detection if a large antibody panel was not applied initially. The systematic approach to FCM data analysis, as thus far described, is applied to establish the phenotype of each individual neoplastic process. The various combinations of coexisting malignancies encountered in the FCM–hematopathology laboratory can be broadly grouped as follows:

Case studies 78 and 79

Case study 80

Case studies 81 and 82

1. A mature B-cell neoplasm, usually CLL or plasma cell dyscrasia, coexisting with a myeloid disorder (Figure 4.99) that can be an MPD, high-grade MDS, or AML. In the context of a low-grade LPD/NHL, the superimposed high-grade MDS or AML is likely to be associated with a history of aggressive chemotherapy. 2. Two or more coexisting LPDs/NHLs of different subtypes (e.g., CLL and HCL), which may or may not share the same light chain isotype. A high-grade B-cell malignancy occurring in the context of CLL suggests Richter syndrome. 3. Two different clones within the same type of LPD/NHL (e.g., biclonal CLL). This is a very rare occurrence. 4. Biclonal acute leukemia (i.e., coexisting AML and ALL) (Figure 4.100). 5. Hematopoietic neoplasms coexisting with a nonhematopoietic malignancy. The metastatic tumor cells, individually and in clusters, may be too large to pass through the flow cytometer and may be seen only on the morphologic preparations.

286

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 4.99 Bone marrow with AML and plasma cell dyscrasia. (a) A prominent blast cluster (arrow) and an ill-defined plasma cell cluster (thin arrow). (b–d) Gated on MNCs: Blasts, coexpressing CD34, CD13 and CD33, comprise 25% of the cells analyzed. There are 16% plasma cells with dim CD45 and bright CD38. (e, f) Monoclonal cytoplasmic kappa is expressed. There is kappa hypergammaglobulinemia.

FCM DATA ANALYSIS ON HETEROGENEOUS SPECIMENS

a

b

c

d

287

Figure 4.100 Biclonal acute leukemia (AML + ALL). (a–d) The two leukemic processes are best appreciated on the FSC versus fluorescence correlated displays. The major blast population (arrow) demonstrates low FSC and extremely low SSC, with coexpression of CD34, CD19 and CD33. The minor blast population (thin arrow) is of higher SSC and FSC, coexpressing CD34 and CD33. CD19 is negative.

CHAPTER

5

FCM interpretation and reporting

The previous chapters focus on analyzing FCM data by evaluating the location, shape, density, and size of the cell clusters observed on the various FCM graphics produced by light scatter signals and the antigenic expression of the cells analyzed. In this chapter, the focus is on the last step of FCM analysis (i.e., FCM data interpretation and result reporting), which requires the integration of FCM results and other relevant laboratory and clinical information, including the correlation of the FCM and morphologic data. Similar to data analysis, FCM interpretation is best carried out using a pattern approach. Each of the major immunophenotypic patterns (presented in detail in Diagnostic Hematology: A Pattern Approach) corresponds with a cell lineage and maturation, as inferred from the combination of relevant results derived from the FCM graphics. The antigenic patterns, similar to the leukemias and lymphomas they represent, can be segregated into two broad categories according to the maturity status of the critical cells. The mature group includes neoplasms with a mature B-cell, mature T-cell, NK cell, or plasma cell phenotype. The immature group consists of blastic proliferations, which can be of myeloid, immature B-cell, or immature T-cell differentiation, or, less commonly, a bilineage myeloid-lymphoid differentiation. A leukemic blast population with no evidence of lineage commitment (i.e., a stem cell phenotype) is an extremely rare occurrence. Correlation of the FCM results and morphologic material is greatly facilitated in institutions where the FCM and hematopathology laboratories are integrated together as one single diagnostic service. For this exercise to be worthwhile, the air-dried preparations and tissue sections should be of optimal quality. The traditional approach in the literature has been to describe the immunophenotypes of hematologic neoplasms according to the accepted morphologic categories used in the classification schemes of that era. In the earlier schemes, the categorization of lymphomas and leukemias relied heavily on the site of involvement (namely, solid lymphoid organs vs. bone marrow/peripheral blood) and, particularly, on morphology as the “gold standard.” There are many flaws associated with a morphologic gold standard, however. Morphologically based categories suffer from serious subjectivity and reproducibility drawbacks, and may not all represent distinct disease entities. The WHO scheme has brought much improvement in this regard by incorporating newer analytical techniques such as immunology, cytogenetics, and molecular genetics into its classification scheme. Even with these additions, however, morphologic interpretations still remain subjective and dependent on optimally prepared histologic sections or cytologic smears. Therefore, in this chapter, the FCM/morphologic correlation departs from the traditional approach. The conventional terminology based on the current WHO classification is employed when there is reasonable certainty that a given disorder is distinctive enough to be identified in a reproducible manner.

5.1 Immature hematopoietic malignancies The immature status of the critical cells can be identified by the downregulation of CD45 and the presence of early markers, such as CD34, CD117, TdT, and/or CD1a. The expression of early antigens is most important in the lymphoid lineage because of overlapping cytologic

290

FLOW CYTOMETRY IN HEMATOPATHOLOGY

features between mature and immature lymphoid cells in terms of cell size, nuclear chromatin, nuclear contour and nuclear/cytoplasmic ratio. In contrast, myeloid blasts are cytologically different from maturing myeloid precursors and, therefore, can be more easily recognized as immature elements. On air-dried preparations such as peripheral blood or bone marrow, the only lineage-specific cytologic feature is the presence of Auer rods. In the majority of cases, the leukemic blasts display a high nuclear/cytoplasmic (N/C) ratio, scant to moderately abundant pale blue cytoplasm, fine nuclear chromatin, a visible nucleolus, and a variable degree of heterogeneity in cell size. These morphologic features are nondescript, however, and give no hint to the associated phenotype. Furthermore, some cases of T-ALL demonstrate conspicuous azurophilic granules, simulating the appearance of myeloblasts (Plate 20). Occasionally, the appearance of lymphoblasts closely mimics that of a mature lymphoid disorder (Plate 21) even in optimally prepared smears. Conversely, large lymphoma cells in the blood or bone marrow may be cytologically misidentified as blasts. Biopsies of lymph node specimens involved by any leukemic process usually reveal a diffuse pattern of infiltration. A less frequent morphologic manifestation is the interfollicular pattern of involvement, in which residual follicles are widely separated by an expanded interfollicular zone containing the neoplastic infiltration. On routine tissue sections, the morphology of AMLs can be easily mistaken for that of large cell lymphoma (Plate 22), whereas the cytology of lymphoblasts, with their medium-sized nuclei with a stippled (“salt and pepper”) chromatin, may be simulated by the “blastic” variant of MCL (Plate 23). In the authors’ experience, such diagnostic confusion is not an uncommon occurrence in institutions where FCM immunophenotyping or extensive immunohistochemistry studies are not applied to the workup of solid tissue specimens with suspected hematopoietic malignancies.

5.1.1 ALL/lymphoblastic lymphoma

Case studies 1, 10 and 12 to 14

Case study 15

Case study 83

The immature lymphoid malignancies are best designated according to their immunophenotypes; that is, as precursor T-ALL/LL (lymphoblastic lymphoma) or precursor B-ALL/LL (see Sections 3.5.2, 3.5.3 and 4.2.1). In some institutions, the precursor-B leukemias are further subclassified, into “common” ALL (CD10+) and “null” ALL (CD10−) for prognostic purposes, especially in the pediatric age group. There is a higher frequency of CD10− precursor B-ALL among infants, whereby the absence of CD10 is highly associated with adverse cytogenetic abnormalities, especially those involving chromosome 11 (Figure 3.53). The usefulness of identifying a third subgroup, pre-B-ALL, by testing for cytoplasmic IgM (cmu) for prognostic purposes is questionable, because a substantial number of pre-B leukemias are not associated with specific chromosomal abnormalities. On the other hand, hyperploidy in childhood ALL, as evaluated by DNA analysis or by cytogenetics, is a well-established good prognostic indicator (Figure 5.1). Among the precursor-T malignancies, the phenotypic variations (e.g., CD3+ vs. CD3−, or CD4+ CD8+ vs. CD4− CD8− vs. CD4+ CD8−) have no demonstrable prognostic significance. Therefore, it is not necessary to subdivide precursor T-ALL/LL according to the stages of thymocyte differentiation. CD10 expression, present in a substantial number of precursor T-ALL/LL, does not affect the prognosis either. Rare cases of ALL/LL may display a very complex phenotype whereby antigens of several lineages are coexpressed. The clinical history or genotypic studies may disclose an underlying CML or an association with the (9;22) translocation.

FCM INTERPRETATION AND REPORTING

a

b

c

d

e

f

291

Figure 5.1 Precursor B-ALL in a 2-year-old child. (a–e) The leukemic blasts (gray) display: (1) low SSC; (2) heterogeneous and downregulated CD45; and (3) bright CD10 and CD58. CD20 expression is heterogeneous, in a pattern reminiscent of that seen on benign B-cell progenitors. CD34 and TdT are positive. (f) The leukemia is hyperdiploid (DI: 1.17) with a low S-phase of 2.9%.

292

FLOW CYTOMETRY IN HEMATOPATHOLOGY

5.1.2 Myeloid malignancies In the myeloid lineage, the most useful antigens for recognizing myeloblasts include CD13 and CD33, CD117, and MPO. Other antigens, such as CD11b and CD16, are less useful because their expressions vary highly among AMLs and they may appear or disappear at relapse (and, therefore, may not be useful for follow-up purposes). Using a combined FCM and morphologic evaluation of the bone marrow, several subtypes of AML can be identified. 5.1.2.1 AML-M3

Case studies 2 to 4

The characteristic finding of bundles of Auer rods or dense granulation facilitates the morphologic recognition of typical AML-M3. The cytology of M3v may simulate that of other hematologic malignancies (Plates 12 and 24), however. In either instance, inspection of the FCM dot plots reveals the characteristic pattern of antigenic expression associated with AML-M3 (see Sections 3.4.1 and 3.5.1). CD13 and CD33 are well expressed and other antigens, namely CD34, CD117, HLA-DR, CD11b, CD15, and CD16, are absent (Figure 3.28). CD15 may be poorly expressed in some cases. As shown in Section 3.4.1, weak HLA-DR and/or weak CD34 (bimodal) may be observed in the severely hypogranular/agranular variant of AML-M3 (Figure 3.31).

5.1.2.2 AML with minimal maturation

Case studies 6, 7 and 84

In AML with minimal maturation, the neoplastic population is composed almost solely of myeloblasts. Further subclassification into AML-M0 and M1, if deemed necessary, can be made based on either the presence of Auer rods (Plate 25) or a positive myeloperoxidase cytochemistry (Plate 11) in AML-M1. The leukemic blasts often share the phenotype of normal myeloblasts (i.e., CD13, CD33, and HLA-DR are expressed) with or without CD34. CD7 may be present. Whether this finding should be considered aberrant has been the subject of controversy. When expressed, CD7 may be used as a “fingerprint” to detect early relapse or residual disease. Other phenotypic aberrancies (see Section 3.5.1.1), when present, are also useful for follow-up purposes. A frequent aberrancy encountered in AML with minimal maturation is CD13−, CD33+. Some of the cases with this aberrant myeloid pattern also lack detectable CD34 and HLA-DR (Figure 5.2), which then may lead to a misinterpretation as AML-M3. AML with positive MPO antigen (or MPO cytochemistry) but lacking both CD13 and CD33 is a very rare occurrence.

5.1.2.3 AML with maturation

Case studies 54 to 56

In AML with maturation, the neoplastic clone includes a substantial proportion of maturing myeloid precursors in addition to the blasts. Most of the AMLs preceded by MDS fall into this category, which corresponds with AML-M2. The phenotype of the blasts is similar to that mentioned in Section 5.1.2.2. When present, the aberrant coexpression of CD19 (Figure 4.52) is highly associated with the translocation (8;21), a favorable prognostic indicator (Plate 8). CD56 is also often expressed. The corresponding bone marrow often demonstrates some increase in eosinophils (i.e., AML-M2Eo) (Plate 26). The latter is not to be confused with AML-M4E (Plate 18), itself characterized by the presence of abnormal eosinophils and abnormalities of chromosome 16. In the classification schemes, M4E has been regarded as a subtype of AML-M4. The bone marrow morphology in M4E more often resembles that in AML-M2 than M4, however. Phenotypically, the leukemic blasts in AML-M4E do not show evidence of monocytic differentiation either. Although the blast and monocytic components may merge together as one cell cluster on the SSC/ CD45 dot plot (Figure 4.50), they are seen as two distinct populations on the CD14/CD64 display

FCM INTERPRETATION AND REPORTING

a

b

c

d

e

f

293

Figure 5.2 AML with minimal maturation (AML-M1). (a) Isotype-matched negative controls. (b–f) The sample is composed virtually entirely of blasts. The leukemic cells are CD33++, CD117++, CD34−, CD13−, CD11b− and CD16−. HLA-DR is essentially negative. The phenotype superficially mimics that of AML-M3. However, the shape of the cluster on the SSC/CD45 dot plot, bright CD117 expression and an aberrant CD13/CD33 pattern are features not supportive of the diagnosis of AML-M3.

294

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 5.3 Bone marrow with AML-M4E. (a–d) The blast (arrow) and monocytic clusters are distinct from each other on the SSC/CD45 dot plot. Eosinophils (thin arrow) display the characteristic low FSC. Blasts coexpress CD34 and dim CD117, but lack both CD14 and CD64. The monocytic component is composed mostly of mature elements with bright CD14/CD64 coexpression. The CD14/CD64 trail is relatively attenuated, reflecting the lower number of immature monocytes.

because blasts in AML-M4E usually demonstrate no reactivity for the monocytic associated markers CD14 and CD64 (Figure 5.3). Furthermore, NSE staining is essentially negative in AMLM4E, and peripheral monocytosis is not always present. 5.1.2.4 AML with monocytic differentiation The AMLs with monocytic differentiation can be identified by FCM based on the shape and location of the blast cluster on the SSC/CD45 dot plot and the pattern of CD14 and CD64 coexpression (see Section 3.5.1). Cases composed predominantly of the most immature cells (so-called monoblasts), that is, AML-M5a (Plate 27), can also be distinguished from those containing maturing monocytic elements according to the pattern of CD14/CD64 reactivity (Figures 3.42 and 3.43). This approach is more reproducible than the morphologic enumeration of monoblasts, promonocytes, “atypical” monocytes, and monocytes. The most immature cells (i.e., cells with lower levels of CD14) also demonstrate high proliferative activity (Figures

FCM INTERPRETATION AND REPORTING

Case studies 8, 9, 11 and 53

295

3.44 to 3.46). Blasts in this group of AMLs may not express CD34. Further subclassification into AML-M4 and M5, if deemed necessary, is based on the results of the NSE cytochemistry using the recommended thresholds of 20% and 80%, respectively. The sensitivity of NSE staining varies with the type of substrate used, however. The preferred substrate is α-naphthyl butyrate. Furthermore, the relative proportions of the granulocytic and monocytic components fluctuate during the course of the disease. For instance, the bone marrow picture may be that of AML-M5 at diagnosis, but residual/relapsed disease may resemble AML-M4. Irrespective of the subtype M4 or M5, it is important to separate AML with monocytic differentiation from other subtypes of AML because of its propensity to involve extramedullary sites.

5.1.2.5 AML with erythroid hyperplasia

Case study 52

The diagnosis of acute myeloid leukemia with erythroid hyperplasia (AML-M6) requires optimal bone marrow smear preparations. In AML-M6, the erythroid component exceeds the combined population of blasts and myeloid precursors (Plate 28). A small number of MDS (including high-grade MDS) also demonstrate erythroid preponderance/hyperplasia, however. Despite the red cell lysis step, erythroid hyperplasia remains evident on the SSC/CD45 dot plot and those dot plots featuring CD71, such as FSC/CD71 (Figure 4.48) or CD71/CD235a. The antigenic profile of the blasts in AML-M6 is similar to that of typical myeloblasts. Antigenic alterations on the erythroid and myeloid precursors (see Section 4.5.2), as well as their abnormal cytology, can be present in both AML-M6 and the group of MDS with erythroid preponderance/hyperplasia. Because of these similarities, the distinction is based on the blast proportion relative to nonerythroid cells as determined from the bone marrow smears. In many instances, AML-M6 is a transient manifestation of either AML-M2 or AML with minimal maturation. The diagnosis of “pure erythroleukemia” requires extreme caution. An occasional reactive process (e.g., viral-associated hemophagocytic syndrome, B12/folate deficiency) may display such striking erythroid proliferation as to give a picture of “pure erythroleukemia” (Figure 5.4). With the exception of severe hemophagocytic syndrome, which leads to a fulminant course and early demise, the patient should be followed-up closely and a subsequent bone marrow evaluation within a month may reveal a more clear-cut clinical picture.

5.1.2.6 AML with megakaryocytic differentiation

Case studies 85 and 86

The identification of the rare AML-M7 (AML with megakaryocytic differentiation, Figure 5.5) is based on megakaryocyte-associated antigen expression (e.g., CD41, CD61) and occasionally may require platelet peroxidase ultracytochemistry. Either CD13 or CD33 may also be expressed. The marrow aspirate is often aspicular because of severe reticulin fibrosis, thus necessitating examination of the bone marrow biopsy and corresponding imprints instead. Contrary to popular belief, the cytologic finding of blasts with cytoplasmic blebs does not necessarily imply megakaryocytic differentiation. Slow drying of the blood or bone marrow smears can induce bleb formation on any cell with ample cytoplasm and minimal cytoplasmic granularity.

5.1.2.7 MPD and MDS For myeloid disorders in which the blast component is below the currently accepted threshold for AML, a correlation of the blood and bone marrow findings is necessary to separate MPD from MDS. It is also necessary to distinguish the latter from residual/early-relapsed AML based on the appropriate clinical history. As a rule of thumb, hematopoiesis is ineffective in MDS, but excessive in MPD. This in turn requires two different therapeutic approaches, one

296

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 5.4 Bone marrow with marked erythroid hyperplasia in a patient with fulminant viral-associated hemophagocytic syndrome. (a, c–f) The specimen contained 70% erythroid cells. There is marked left shift with a preponderance of larger erythroid precursors (e). CD71 and Gly-A are coexpressed. HLA-DR is negative. (b, d) Blasts are virtually nonexistent. The patient expired within 2 weeks after the bone marrow procedure.

FCM INTERPRETATION AND REPORTING

a

b

c

d

297

Figure 5.5 AML with megakaryocytic differentiation. (a) A prominent cluster (R1) with variable CD45 expression, ranging from negative to weak. (b–d) Gated on R1: Blasts are of medium to large cell size. Megakaryocytic markers (CD41, CD42b and CD61) are well expressed. CD117 and CD34 (not shown) are present in a minute subset of the blasts.

Case study 57

to stimulate hematopoiesis, the other to control it. In many patients with MPD, the peripheral blood demonstrates an elevated WBC count and circulating immature elements, a picture mirroring the bone marrow proliferative process. In most patients with MDS, however, the increased cellularity in the bone marrow stands in sharp contrast with the progressive cytopenias in the peripheral blood. This is the key difference between MPD and MDS, more so than the presence of any particular cytologic feature. An exception to the above rule of thumb is the occasional patient with hypoplastic MDS, in whom the clinical history often reveals previous exposure to intense chemotherapy and thereby severe damage of the bone marrow microenvironment. Adequate cytologic evaluation is much hampered by the scantiness of the bone marrow aspirate material. Hypoplastic MDS can be morphologically indistinguishable from aplastic anemia (Plate 29). Although certain FCM abnormalities such as low SSC in granulocytes or decreased CD71 expression on erythroid precursors have been reported in aplastic anemia, the presence of several abnormal FCM findings, including altered myeloid maturation curves or a slight excess in blast content, are

298

FLOW CYTOMETRY IN HEMATOPATHOLOGY

helpful clues for identifying hypoplastic MDS. The diagnosis can be further confirmed by the presence of cytogenetic abnormalities. Follow-up bone marrow studies would also reveal a subsequent rapid progression to a typical hypercellular, high-grade MDS. The role of FCM study is limited in MPD, particularly in those cases with no apparent pathology affecting the myeloid lineage (e.g., polycythemia vera, essential thrombocythemia). In CML and CML-like MPD, the myeloid precursors may or may not demonstrate abnormal phenotypic features (Figures 5.6 and 5.7). Some of the antigenic abnormalities overlap with those seen in MDS or G-CSF therapy (see Section 4.5.2.1), and cytologic features considered specific to MDS (e.g., hypogranulation) may also occur in MPD. The distinction between MPD, MDS and reactive myeloid proliferations is usually not problematic, however, if all the clinical, peripheral blood, and bone marrow data are evaluated together. The one MPD in which the FCM picture is relatively specific is CMMoL, such that a diagnosis can be made when the dot plots are examined in the context of a peripheral leukocytosis

a

b

c

d

Figure 5.6 Bone marrow of a patient with polycythemia vera. (a) Bone marrow with 3% blasts (B), 50% granulocytes (G) and 28% erythroid cells (E). Monocytes (M) are few. (b–d) Gated on G: No overt abnormal expression of CD117. Myeloid maturation curves are nearly normal with a mild left shift.

FCM INTERPRETATION AND REPORTING

a

b

c

d

e

f

299

Figure 5.7 Polycythemia vera (continuation of Figure 5.6). (a) Upregulated CD14 in a subset of granulocytes (G). (b–d) Gated on blasts: Coexpression of CD117, CD34, CD13 and CD33. HLA-DR is negative (the intensity of the isotype controls extends far into the second decalog). (e, f) Gated on erythroid cells: Essentially normal coexpression of CD71, CD36 and Gly-A. Some unlysed red cells (CD71−, Gly-A+) are present (arrow).

300

FLOW CYTOMETRY IN HEMATOPATHOLOGY

Case study 69

or hyperplastic marrow (see Section 4.5.1.1). Evaluation of the pattern of CD14/CD64 coexpression is also helpful in those instances where the bone marrow morphology poses a diagnostic dilemma between CMMoL and AML-M5b (see Section 3.5.1).

5.1.3 Acute leukemias with a multilineage antigenic profile

Case studies 87 to 89

Expression of a myeloid antigen can be seen in some ALLs, especially those of precursor-B lineage. Conversely, some AMLs may express a lymphoid-associated marker (Figures 3.27, 4.51, and 4.52). Because of the variable criteria and lack of a standard uniform panel between different laboratories, there has been no clear definition regarding what constitutes aberrant antigenic expression in acute leukemia. In the authors’ opinion, the label “aberrant myeloid antigen expression” or “aberrant lymphoid antigen expression” should take into account the relative specificity of the marker in question. For instance, CD13, CD33, CD19, and CD20 are relatively more specific than CD7, CD2, CD14, CD15, or CD56. Similar stringent criteria should be also applied in defining biphenotypic leukemia (Figure 5.8), a process characterized by the presence of blasts coexpressing several lymphoid and myeloid surface antigens. It is best that the diagnosis be established based on the coexpression of the cytoplasmic antigens, that is, MPO and cCD22, or MPO and cCD3. These three markers basically define their respective cell lineages. In the literature, the frequency of this type of leukemia has varied widely, ranging from less than 1% to greater than 20% of all acute leukemias. As a result of the lack of consistent criteria, it has been difficult to assess the clinical significance of the so-called biphenotypic leukemias. These leukemias may represent the neoplastic transformation of a stem cell at a stage prior to lineage commitment. Some studies have shown a high association with the presence of the Philadelphia chromosome. Biphenotypic leukemias should not be confused with biclonal (or multiclonal) acute leukemia (Figure 4.100), defined as the coexistence of two or more distinct leukemic populations with different phenotypes. This very rare process may also be associated with the (9;22) translocation.

5.2 Mature lymphoid malignancies The mature lymphoid malignancies consist of mature B-cell, T-cell, and NK leukemias and lymphomas, as well as plasma cell dyscrasias. Most of the leukemias are chronic lymphoproliferative disorders and follow an indolent disease course. However, some of these mature neoplasias, such as ATLL, certain aggressive NK leukemias, and Burkitt lymphoma–leukemia run a rapidly progressive course. The diagnosis and classification of malignant lymphomas, which traditionally have relied mostly on morphology, have in recent years taken into account the results of immunologic studies. In most instances, the antigenic profiles of these neoplasms have been derived from immunohistochemistry, however, rather than from FCM immunophenotyping, despite the fact that FCM provides a much better delineation of antigen expression, as well as a better appreciation of antigen density and antigen coexpression. One of the main reasons is that in many institutions, including those with FCM laboratories (especially outside the United States), lymph nodes and other tissue specimens with suspected lymphoma are rarely submitted for FCM analysis. As a result, large amounts of valuable information contained in the rich data derived from the FCM analysis are not being taking into consideration in the current lymphoma classification schemes.

FCM INTERPRETATION AND REPORTING

a

b

c

d

e

f

301

Figure 5.8 Biphenotypic acute leukemia. (a–f) The sample is essentially a pure cell culture of blasts. Blasts are of medium cell size and lack CD45. There is coexpression of CD34, CD13, CD33, CD10 and CD19. CD34 is weak and of variable distribution. Cytogenetics revealed t(9;22).

302

FLOW CYTOMETRY IN HEMATOPATHOLOGY

The authors have chosen to use the FCM immunophenotyping features, along with DNA and cell cycle analysis (see Section 2.10), as the basis for the discussion of the categorization of mature lymphoid malignancies. The WHO nomenclature is applied where applicable.

5.2.1 B-cell LPD/NHL In the Western world, the bulk of mature lymphoid malignancies passing through the FCM– hematopathology laboratory are of B-cell lineage. Several major categories can be identified with a high degree of accuracy and reproducibility based on their FCM patterns of antigenic expression and S-phase fraction. 5.2.1.1 CD10 expression

Case studies 19 to 21

The correlation of CD10 expression with FCC lymphomas is straightforward (see Sections 3.6.3.1 and 4.1.1.1). On histologic sections, the follicular pattern (however vague it may be) with back-to-back follicles in the lymph node and the paratrabecular pattern in a bone marrow biopsy are useful morphologic features for recognizing FCC lymphoma. Bone marrow involvement by other disorders (e.g., PTCL and mast cell disease) may also demonstrate a similar picture of paratrabecular infiltration, however (Plate 30). Occasionally, the follicular pattern is vague and focal, and therefore not easily appreciated. By dimming the light of the microscope, the follicular architecture becomes more evident. Alternatively, it can be highlighted by immunostaining for follicular dendritic cells. In some cases, the lymph node morphology is less clear-cut, as the neoplastic follicles are farther apart and still surrounded by an adequate mantle cuff. Occasionally, an FCC lymphoma adopts a so-called floral pattern on histologic sections (Plate 31). On air-dried preparations, the presence of deep nuclear indentations (e.g., “buttock” cells) facilitates the identification of the neoplastic cells cytologically. Nuclear indentation is a cytologic feature common to lymphoma cells of various types and therefore does not imply an FCC origin, however. In addition, it is important to be aware that nuclear irregularities can be artifactually induced by prolonged storage (Plates 32 and 33). The WHO classification further subdivides follicular center cell lymphomas into FCC I, II, and III (applicable to solid lymphoid tissue only). The subclassification would be more reproducible by taking into account the FSC parameter as an indicator of cell size (see Section 3.6.3.1) instead of relying solely on the manual counting of the number of large cells per microscopic high-power field (Figure 5.9). The clinical significance of the subclassification has remained a matter of controversy, however. CD10 expression, common in FCC lymphomas, is also found in some DLCLs (Figure 3.70). Knowledge of the clinical history, along with the availability of previous lymph node biopsies and corresponding FCM analyses are helpful to determine if the CD10+ DLCL is a de novo process or a progression of a previous FCC lymphoma. The finding of a tiny focus with neoplastic follicles is also sufficient to confirm that the DLCL originates from an underlying FCC lymphoma. During the clinical course of patients with FCC lymphoma, the biological behavior of the tumor invariably worsens over time as its proliferative fraction increases. This is best documented with DNA analysis performed on the sequential lymph node biopsies (Figures 5.10 and 5.11). The combined findings of a very high S-phase and CD10 expression in a mature B-cell proliferation is essentially pathognomonic of Burkitt lymphoma (Figure 3.69). This diagnostic combination is most useful when analyzing fine-needle aspirate specimens from deep-seated lesions where an adequate morphologic assessment is not possible. Exceptionally, a markedly reactive lymph node may contain a minute subpopulation of monoclonal B-cells with dim CD10 expression (see Section 4.1.1.1). Such cases should not

FCM INTERPRETATION AND REPORTING

a

b

c

d

e

f

303

Figure 5.9 Lymph node with FCC III lymphoma. (a–d) The lymphoma cells (arrow) are of large cell size, coexpressing CD10, CD19 and monoclonal kappa light chain. A few residual polyclonal benign B-cells (thin arrow) are present. (e) Total DNA content: The tumor is aneuploid (DI: 1.21); benign T-cells and residual B-cells are diploid. (f) Gated on CD19/CD10 cells: The tumor has an intermediate S-phase of 7.3%.

304

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 5.10 Disease progression in a patient with FCC lymphoma. The time interval between the initial diagnosis of FCC I (a, b) and progression to FCC III (c, d) was 10 years. Transformation to an aggressive lymphoma (e, f) occurred 18 months later. During disease progression, there was an increase in cell size (a, c, e) and the S-phase fraction increased from 1% to 8% and then 21% (b, d, f). The tumor remained diploid throughout.

FCM INTERPRETATION AND REPORTING

a

b

c

d

e

f

305

Figure 5.11 Disease progression in a patient with FCC lymphoma (continuation of Figure 5.10). The intensity of CD10 is retained between the time of initial diagnosis (a) and the transformation to an aggressive lymphoma (b). Although the brightness of lambda is similar between the FCC I at diagnosis (not shown) and the FCC III (c, d), the intensity becomes comparatively reduced with transformation (e, f).

306

FLOW CYTOMETRY IN HEMATOPATHOLOGY

be diagnosed as “in situ FCC lymphoma” especially because other laboratory and clinical findings, such as the age of the patient and the lack of t(14;18), do not support the diagnosis of FCC lymphoma. 5.2.1.2 Coexpression of CD11c, CD25 and CD103

Case study 22

When evaluated properly (see Section 3.6.3.2), the coexpression of CD11c, CD25 and CD103 is the antigenic fingerprint of HCL. Hairy cells can be identified by diffuse tartrateresistant acid phosphatase (TRAP) positivity, but this approach may not be useful because the number of hairy cells on the blood film or aspicular bone marrow smear is often very low. On optimal air-dried preparations, hairy cells demonstrate a characteristic bland ovoid/reniform nucleus and a voluminous amount of pale cytoplasm (Plate 34). A bone marrow procedure is rarely necessary nowadays, given the excellent sensitivity of FCM to detect hairy cells in the peripheral blood (Figure 4.62). The diagnosis of HCL is further confirmed by its excellent response to 2-chlorodeoxyadenosine (2-CdA) therapy.

5.2.1.3 Coexpression of CD5 and CD23 Coexpression of CD5 and CD23 is associated with CLL/SLL and disorders closely related to CLL. Biologically the same disorder, SLL only differs from CLL by the presence of some adhesion molecules (CD11a, CD18). In most cases of CLL/SLL, the neoplastic population is composed mostly of small lymphoid cells; the larger and more activated cells, namely plasmacytoid lymphocytes and prolymphocytes, are essentially insignificant. On lymph node sections, the activated cells are loosely packed into scattered pseudofollicles (i.e., proliferation centers). Microscopic appreciation of pseudofollicles is much enhanced by dimming the light of the microscope. Occasionally, the activated cells may form a ring of pale cells around residual germinal centers, thus producing a marginal zone pattern (see Section 5.2.1.7). Rarely, proliferation centers may be seen on the bone marrow biopsy. The typical morphology of CLL/SLL correlates with the classic antigenic profile of weak CD20, a weak monoclonal surface light chain, and coexpression of CD5 and CD23 (Figures 3.16, 3.18, and 3.37). In the disorders closely related to CLL, the number of activated lymphoid cells is increased and occasional nuclear irregularities may be present (Plate 35), especially in long-standing disease resistant to chemotherapy. The cytology of the heterogeneous-appearing neoplastic cells, ranging from small lymphoid cells to prolymphocytes can be better appreciated on the blood film (Plate 14) than on bone marrow smears, where the higher density of nucleated cells invariably renders the cytoplasmic features of lymphoid cells less discernible. Various terminologies have been used in the literature to designate these activated CLLs (e.g., “atypical” CLL or CLL/PL [Plate 14; Figures 3.6 and 3.61]). Traditionally, the arbitrary thresholds of 10% and 50% prolymphocytes in the peripheral blood have been used to define CLL, CLL/PL, and B-PLL. The solid tissue equivalent of “activated” CLL manifests as prominent proliferation centers showing a variable degree of merging with each other (Plate 36). The morphologic similarities to the blood picture are best appreciated on lymph node imprints (Plate 37). Longterm follow-up usually reveals a slow steady increase in the number of activated cells in the blood of CLL patients, along with other manifestations of increased tumor burden. The rate of disease progression varies highly from patient to patient, however. Several unfavorable prognostic features have been identified, including (1) increased CD38 expression on CLL cells, (2) high serum levels of thymidine kinase and β2-microglobulin, and (3) genomic abnormalities, such as trisomy 12, deletions of chromosomes 11q22–23 or 17p13, or an absent/ altered p53 gene. Because several of these parameters only appear during disease progression, there has been a continuing search for prognostic markers that can be detected in the early

FCM INTERPRETATION AND REPORTING

Case study 16

Case study 90

Case study 80

307

stages and remain stable during the course of the disease. This would help to identify the unfavorable group of patients who may benefit from early therapeutic intervention. To this end, the most recently identified prognostic marker is the expression of cytoplasmic Zap-70. In normal peripheral blood, Zap-70 protein is present abundantly in T-cells but not in B-cells or monocytes. Several studies have demonstrated that the presence of Zap-70 in CLL correlates highly with the absence of somatic mutations in IgVH genes, and thereby identifies patients with progressive disease and shorter survival. It also appears that, based on a combined evaluation of Zap-70 and CD38 expressions, CLL patients can be stratified into different risk groups as follows: (1) good prognosis, with lack of both Zap-70 and CD38 (Figures 5.12 and 5.13), (2) poor prognosis, with both markers present (Figures 5.14 and 5.15), and (3) intermediate prognosis, with only one of the two expressed (Figures 3.62 and 3.63). The aforementioned unfavorable factors are more frequently observed in the poor prognostic group. In most instances, the coexpression of CD5 and CD23 is retained in the activated variants of CLL/SLL, although CD20 and/or the surface light chain tend to be brighter and the FSC slightly higher (Figure 3.60) than that observed in classical CLL/SLL. Occasionally, CD23 may also be downregulated. Although most cases of LPC lymphoma–leukemia (Figures 5.16 and 5.17) express no specific markers (i.e., a nondescript B-cell phenotype), a small number display a phenotypic profile similar to that of either activated CLL or MCL (CD5+ CD23−). Because the terminology lymphoplasmacytoid or lymphoplasmacytic (designated together as LPC) is based mostly on morphology, other laboratory data, especially serum immunoglobulins, are helpful to confirm the diagnosis. B-prolymphocytic leukemia occurs much less frequently than activated CLL. In most instances the leukemic population is CD5− and devoid of any specific antigenic characteristics (Figure 3.79); that is, a nondescript mature B-cell profile. An occasional case may be either CD5+ CD23+ or CD5+ CD23−. Because of the distinctive higher FSC of the prolymphocytic proliferation, there is no confusion with either CLL or MCL, however. CD20 is brighter in B-PLL, and the proliferative fraction, though variable, is also higher. The clinical presentation and the blood picture, with its markedly high WBC count composed predominantly of prolymphocytes (Plate 38), are also useful features for identifying B-PLL. Lymph node biopsies have only been infrequently studied in PLL. The morphologic equivalent of PLL in the lymph node is known as paraimmunoblastic lymphoma (Plate 39). Richter syndrome, which refers to the transformation of CLL into a high-grade large cell lymphoma, is a rare occurrence and thus has been rather poorly defined. Therefore, it is unclear if a distinction between paraimmunoblastic lymphoma and Richter syndrome can be made with certainty. In the former, the rather low number of mitotic figures is asynchronous with the large cell morphology. The large-cell lymphoma in Richter syndrome, on the other hand, is expected to demonstrate a high mitotic rate and a corresponding elevated S-phase fraction similar to that seen in aggressive B-cell NHLs.

5.2.1.4 CD5+ CD23− B-cell neoplasms

Case study 91

The antigenic profile of CD5+, CD23−, bright CD20, and bright surface light chain is very suggestive of MCL. However, some cases of activated CLL or LPC leukemia–lymphoma may exhibit a similar profile. The intensity of CD19 relative to that of CD20 (i.e., CD19 < CD20) (Figure 3.64) is an important clue for identifying MCL (see Section 3.6.2). Other findings useful for distinguishing MCL from other CD5+, CD23− B-cell LPD/NHLs include (1) absence of serum monoclonal protein (IgM, IgA), (2) evidence of bcl-1 rearrangement, that is, the (11;14) translocation (this abnormality may occur in a small number of lymphoplasmacytoid lymphomas and B-PLL, however), (3) a mantle zone or a nodular growth pattern (even if focal) on lymph node/adenoid tissue sections (Plate 40), (4) lymphoid cells with cytologic

308

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 5.12 Peripheral blood with CLL. (a–f) The neoplastic cells (gray) display the typical antigenic features of CLL including: (1) small cell size; (2) slightly downregulated CD45; (3) weak and heterogeneous CD20; and (4) dim surface light chain expression (monoclonal for lambda). CD20 is less intense than CD19 (both markers conjugated with the same type of fluorochrome). The tumor cells are CD5++ and CD10−.

FCM INTERPRETATION AND REPORTING

a

b

c

d

309

Figure 5.13 Peripheral blood with CLL (continuation of Figure 5.12). (a) CLL cells (gray) with bright CD23 expression. (b–d) The intensity of Zap-70 in the CLL cells is less than that in the (internal control) T-cells. The lack of both CD38 and Zap-70 is a favorable prognostic finding.

Case study 92

features (e.g., indented nuclei) suggestive of lymphoma cells on air-dried smears, and/or (5) expression of cyclin D1. The nuclear indentations of the lymphoma cells can be better appreciated in uncrowded areas of the smears. The neoplastic cells in some cases may not display any nuclear indentations, however (Plate 41). The finding of nuclear irregularities, unless marked, is less meaningful on histologic sections because this feature may be artifactually induced by tissue fixation and processing. Mantle cell lymphoma is considered to be a more aggressive disease than other small B-cell LPD/NHLs. In the authors’ experience, the S-phase fraction varies widely between cases (Figure 5.18), and not all patients have a limited survival. In some cases, the S-phase fraction is similar to that of low-grade lymphomas; in others, it is in the range seen in intermediategrade lymphomas. The “blastic” variant of MCL (Plate 23b) is considered to represent a rapidly progressive stage of the disease. In laboratories that do not routinely perform FCM studies on solid tissue, this diagnosis is based mainly on the nuclear chromatin, which, in blastic MCL, appears finely stippled, similar to the chromatin of the lymphoblasts in ALL/LL. Blastic MCL can be recognized more reliably when the diagnostic evaluation takes into account the

310

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 5.14 Peripheral blood with CLL. (a–d) Neoplastic B-cells (gray) monoclonal for kappa. Surface light chain expression is dim. CD5 is present. Note that CD20 intensity is brighter and less heterogeneous than that expected in CLL.

Case study 93

S-fraction, CD71 reactivity, and FSC parameter instead of being based solely on morphology, especially because the appearance of the nuclear chromatin can be easily affected by tissue processing and staining. The immunophenotype CD5+ CD23− may be found in a small number of large B-cell malignancies (Figure 4.64). Although some of these may be “blastic” MCL or PLL, most are high-grade B-cell lymphomas (Plate 42), which morphologically and genotypically bear no relationship to either PLL or MCL. Some may display an immunoblastic/plasmablastic appearance. Knowledge of the antecedent clinical history, or review of any available diagnostic material, may help to differentiate de novo CD5+ large B-cell lymphoma from the other groups of disorders.

5.2.1.5 CD45 and/or pan B-cell antigens markedly downregulated In addition to the decreased levels of CD45 or B-cell antigens (Figure 3.65), surface light chains may also be downregulated (Figure 3.82) in B-cell NHLs. These antigenic features are

FCM INTERPRETATION AND REPORTING

a

b

c

d

311

Figure 5.15 Peripheral blood with CLL (continuation of Figure 5.14). (a) CLL cells (gray) positive for CD23. (b) The tumor is diploid with a low S-phase. (c, d) The expression of both CD38 and Zap-70 is heterogeneous. Reactivity for both of these markers is an unfavorable prognostic finding.

Case studies 23, 24 and 94

consistently associated with high-grade B-cell lymphoma with a variable degree of plasma cell differentiation. There is a propensity for this type of lymphoma to occur in the context of altered immunity. One example is the uncommon “primary effusion lymphoma” associated with human herpes virus-8 in HIV-infected patients. Careful inspection, at least in some patients, may reveal that the lymphoma is not just effusion-based, but also involves solid lymphoid tissues. CD38 and CD56 may be present and light chain-restricted intracytoplasmic immunoglobulins can be detected in the majority of cases.

5.2.1.6 Nondescript B-cell phenotype and high FSC Most large B-cell lymphomas display no specific antigenic characteristics. The aggressive nature of these lymphomas correlates with an increased S-phase fraction (Figures 3.73 and 3.118), high FSC signals, and bright CD71 expression (Figure 2.13). Occasionally, CD30 may be present. These lymphomas are heterogeneous, both morphologically and genotypically. In

312

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 5.16 Peripheral blood with CD5+, CD23+ B-cell LPD, morphologically LPC lymphoma–leukemia. (a–f) The leukemic cells form a predominant lymphoid cluster with slightly increased FSC (the larger cells have brighter lambda). CD20 is bright (nearly as bright as CD19) and homogeneous. IgM and IgD are expressed. IgG and IgA are negative (not shown). Lambda is brightly expressed.

FCM INTERPRETATION AND REPORTING

a

b

c

d

313

Figure 5.17 Peripheral blood with CD5+, CD23+ B-cell LPD (continuation of Figure 5.16). (a–c) The leukemic cells coexpress CD20, CD5, CD23 and CD38. (d) Gated on B-cells: The difference between the mean peak fluorescence channels of the negative (kappa) and positive (lambda) histograms is greater than 1 decalog (i.e., lambda intensity is not weak).

lymphoid organs, several patterns of infiltration, such as interfollicular, sinusoidal, or diffuse, can be observed. The content of residual benign cells, especially T-cells, is highly variable. Simultaneous biopsies of several lymph nodes or sequential lymph node biopsies over the course of the disease may demonstrate the various morphologic manifestations of the same large cell lymphoma. For instance, the infiltration in one node may consist of scattered large neoplastic cells with a high content of benign T-cells (i.e., the so-called T-cell-rich B-cell lymphoma), whereas in another node, the large cells predominate as sheets and clusters. In the current classification, lymphomas with a nondescript B-cell phenotype and high FSC are subdivided into several categories, based on either the location (e.g., mediastinal B-cell lymphoma) or morphologic findings (e.g., intravascular large B-cell lymphoma). However, the biological behavior of the lymphoma, which dictates the therapeutic approach, is based on its proliferative activity rather than on any of these features.

314

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 5.18 Proliferative fraction in MCL. (a) Case 1: The tumor is diploid; the S-phase fraction is low (S% 2.4). (b) Case 2: The tumor is diploid; the S-phase fraction is in the intermediate range (S% 10). (c) Case 3: The tumor is diploid; the S-phase fraction is high (S% 17). (d) Case 4: The tumor is near tetraploid (DI: 1.97) with a low S-phase fraction (S% 1.4).

In most instances, diffuse large B-cell lymphomas (DLCLs) arise de novo. An occasional DLCL may have evolved from a previous CD10− FCC lymphoma. Recognition of this type of case is based on the knowledge of previous lymph node studies, or on the evidence of bcl-2 rearrangement. 5.2.1.7 Nondescript B-cell phenotype and low FSC Small B-cell lymphomas with a so-called nondescript phenotype include the infrequent FCC lymphomas without detectable CD10 and the heterogeneous group of small B-cell malignancies showing various morphologic manifestations and genotypic abnormalities. Several “entities” have been described, based primarily on morphologic features. The lack of uniform and well-defined morphologic criteria plus the lack of specific antigenic expression have very much hampered the reproducibility and accuracy of diagnosis in these low-grade B-cell neoplasms with a nondescript phenotype, however. Morphologic artifacts and the fact that the same disease may be given a different name depending on whether the specimen is blood/bone marrow or solid lymphoid tissue further compound the diagnostic difficulties. The authors’

FCM INTERPRETATION AND REPORTING

315

approach to the diagnosis of these low-grade disorders is to append the morphologic name to the FCM interpretation if the smears or tissue sections demonstrate the corresponding cytologic or histologic features. For example, if a marginal zone pattern is seen on the tissue section, then the diagnosis can read “low-grade B-cell NHL, morphologically marginal zone.” Similarly, if the neoplastic cells exhibit lymphoplasmacytoid/lymphoplasmacytic cytology on the blood smears, then the diagnosis can be “low-grade B-cell LPD/NHL, morphologically LPC lymphoma/leukemia.” Some of the relatively common low-grade B-cell LPD/NHL with a nondescript phenotype are presented below. Most lymphoplasmacytoid/plasmacytic malignancies have a nondescript antigenic profile (Figures 3.77 and 5.19). There is a high degree of variability in the expression of several antigens, such as the following:

Case study 95

Case studies 17 and 96

1. Surface light chain: From weak to bright, depending on the degree of differentiation toward the plasma cell stage. Cytoplasmic immunoglobulins can also be detectable. The laborious FCM procedure for cytoplasmic heavy chains is rarely called for, as a serum IEP will yield the same information. 2. CD20 may be downregulated. 3. CD5 and CD23: Although most cases do not express CD5, some display the pattern of CD5+ CD23− (Figure 4.63) similar to that associated with MCL. The finding of weak CD20 and/or weak surface light chain expression helps to avoid a potential misinterpretation, however. An occasional case may have the CD5+ CD23+ profile of CLL, but CD20 or a monoclonal surface light chain is brightly expressed (Figure 5.16).

The plasmacytoid cytology of the neoplastic cells in LPC neoplasms is best appreciated on well-prepared blood films (Plate 43a) or in uncrowded areas of any air-dried preparation. Patients in whom the primary manifestation of the disease is in the blood and bone marrow invariably have splenomegaly. Cellular crowding or slow drying of the air-dried smears often results in the formation of hairs and villous projections (Plate 43b). The degree of basophilia varies from case to case; cells with increased basophilia are considered as “lymphoplasmacytic.” A small number of larger lymphoid cells are usually present. Rouleaux is evident if the M-protein (usually IgM) is sufficiently elevated, in which case the disease is then called Waldenstrom macroglobulinemia. In most instances, the M-protein level is much lower than the artificial threshold set for Waldenstrom macroglobulinemia, however. Indirect evidence of monoclonal hypergammaglobulinemia may be visible on the FCM graphics (Figure 5.19). On histologic sections, recognition of the plasmacytoid cytology is often hampered by the many artifacts associated with tissue fixation and processing (Plate 44). Splenic lymphoma with villous lymphocytes (SLVL) has been described in the literature as a distinct entity on the basis of hairy and villous projections on neoplastic lymphoid cells as seen on the blood or bone marrow smears. Another morphologic name for this disorder is splenic marginal zone lymphoma. The disease is in the group of LPC lymphoma–leukemias, however, as a large number of SLVL cases are associated with monoclonal gammopathy. Emphasis in the literature has been on distinguishing this so-called SLVL from HCL because of the “hairy cytology,” the expression of CD11c, and, occasionally, CD25 or CD103 (Figures 3.78 and 5.20) positivity. The distinction is straightforward, however, in view of the characteristic phenotypic fingerprint of HCL (see Section 3.6.3.2), its sensitivity to 2-CdA therapy, and the marked differences in the hematologic pictures between the two diseases. In solid tissue, disorders labeled as monocytoid B-cell lymphoma (MBCL), marginal zone lymphoma, and MALT lymphoma also have a nondescript B-cell phenotype (Figures 4.60 and 4.61). In the lymph node, a serpentine pattern of infiltration of the interfollicular zone by bland-appearing neoplastic cells with ample pale cytoplasm (a hairy cell-like cytology) and some admixed neutrophils are the main features helpful for recognizing MBCL morphologically (Plates 45 and 46). The histology may appear deceptively non-neoplastic, as prominent

316

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 5.19 Peripheral blood with B-cell LPD/NHL NOS, morphologically LPC lymphoma–leukemia. (a–f) The cell size of the neoplastic cells is slightly larger than that of normal T-cells. A small subset displays higher FSC. The tumor cells lack CD5, CD23 and CD103. CD25 and CD10 are negative (not shown). CD11c is present in a small subset. Surface light chain (kappa) is well expressed. There is indirect evidence of hypergammaglobulinemia.

FCM INTERPRETATION AND REPORTING

a

b

c

d

e

f

317

Figure 5.20 Peripheral blood with CD103-positive B-cell LPD NOS, morphologically LPC leukemia. (a–e) Gated on MNCs: Neoplastic cells comprise 20% of the cells analyzed. Most of the cells are of small cell size. CD25 is absent. A subset expresses CD103. CD11c is bright, but of variable distribution. The expression of CD38 is heterogeneous. (f) Gated on B-cells: Kappa is weakly positive.

318

FLOW CYTOMETRY IN HEMATOPATHOLOGY

residual hyperplastic follicles are not uncommon. The tumor cells may also encircle residual follicles, thus forming a marginal zone pattern. The term marginal zone implies that the germinal center and its mantle zone (which may be attenuated) are surrounded by another ring (layer) of cells. This ring pattern stands out on histologic sections because the rings are composed of pale appearing cells. The marginal zone pattern can be produced by various neoplastic proliferations, however, such as monocytoid B-cells, plasmacytoid lymphocytes, as well as the proliferation centers of CLL (Plate 47). Very often the diagnosis of “marginal zone B-cell lymphoma” (MZBL) has been rendered at the time of presentation even when a marginal zone pattern cannot be documented morphologically (e.g., the specimen is peripheral blood). For these reasons, the diagnostic usefulness of the name MZBL is rather questionable. Small B-cell neoplasms in mucosal/glandular sites are referred to as MALT lymphomas. Another name is extranodal MZBL. The minute size of the biopsies from these sites and the frequent dense inflammatory background are some of the factors precluding an optimal histologic assessment of these tumors, which, in turn, accounts for the difficulty of distinguishing MALT lymphoma from a reactive process morphologically. The clinical presentation and the ease of establishing the clonal (and thereby neoplastic) nature of the lymphoid cells by FCM immunophenotyping have made the diagnosis more straightforward, however. 5.2.1.8 Monoclonal B-cells of undetermined significance An important issue in the bone marrow is the finding of a minute population of nondescript monoclonal B-cells in the range of 1% to 3% in elderly subjects. In the authors’ experience, it is not infrequent for older individuals without any evidence of LPD/NHL to harbor a low level of monoclonal B-cells, asymptomatically and stable over the course of many years (10 years or more). The situation is analogous to MGUS. In these instances, knowledge of the clinical history is critical because the differential diagnosis is low level of involvement by a B-cell LPD/NHL. Thus, in the absence of a positive clinical history, it is judicious to refrain from labeling the finding as lymphomatous involvement of the bone marrow. It is important, however, to review both the bone marrow aspirate smears and core biopsy sections to ensure that the low level of monoclonal cells in the FCM specimen is not due to sampling. Periodical clinical follow-up is also helpful.

5.2.2 Plasma cell dyscrasias

Case studies 66 to 68

The combined findings of a monoclonal cytoplasmic light chain and plasma cell morphology are adequate criteria for the diagnosis of a plasma cell dyscrasia. Other phenotypic characteristics include decreased CD45, downregulation or absence of pan B-cell antigens, and the expression of CD38 and CD56 (Figure 4.76). Phenotypic aberrancies (see Section 4.4.3) are not infrequent (Figures 4.77 to 4.79). The clinical diagnosis of multiple myeloma requires other clinical and laboratory data (e.g., the levels of the M-protein in the serum or urine) unless the bone marrow involvement is overt or neoplastic plasma cells are present in the peripheral blood (Plate 48). Malignant plasma cells may be “immature” appearing (i.e., with dispersed chromatin and conspicuous nucleoli) and therefore referred to as plasmablasts (Plate 48). Extramedullary plasmacytomas with a high content of plasmablasts are referred to as anaplastic plasmacytomas. The distinction between these tumors and high-grade B-cell lymphomas with plasmablastic features may not be clear-cut, in view of the cytologic similarities. The lack of B-cell antigens and surface Ig would support the diagnosis of anaplastic plasmacytoma, however (Figures 5.21 and 5.22). Furthermore, if the involved site is bone marrow/peripheral blood, then the distinction can be made based on the presence of a serum M-protein.

FCM INTERPRETATION AND REPORTING

a

b

c

d

e

f

319

Figure 5.21 Extramedullary plasma cell tumor (nasal mass). (a–f) The tumor cells display increased FSC and a high degree of autofluorescence. CD45 is downregulated. HLA-DR, CD20 and surface light chains are not expressed. CD10 is minimally present.

320

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

Figure 5.22 Extramedullary plasma cell tumor (continuation of Figure 5.21). (a–c) The tumor cells display bright CD38 and aberrant expression of myeloid antigens (CD13/CD33 combination). CD56 is absent. (d) The CD4:CD8 ratio is markedly reversed.

Case study 97

Plasmacytoma in the lymph node usually displays a diffuse pattern of involvement. Occasionally, the disease may present with an interfollicular pattern instead. The morphologic picture of residual reactive follicles surrounded by abundant mature-appearing plasma cells in the interfollicular zone may therefore be misinterpreted as the plasma cell variant of Castleman disease. The latter condition occurs infrequently, and as a reactive process, should not be composed of monoclonal plasma cells.

5.2.3 T-cell LPD/NHL The systematic evaluation of T-cell antigens and DNA analysis by FCM can be applied to the diagnosis and grading of most mature T-cell neoplasms (see Section 3.6.4). The antigenic aberrancy can be subtle, such as a slight increase or decrease in the density of a pan T-cell antigen (Figures 3.89, 3.90, 4.66, and 4.70). The great majority of cases are composed of neoplastic TCR-αβ T-cells. Clonality can therefore be determined with analysis of the TCR-Vβ repertoire (Figures 3.102 to 3.107, 3.115 to 3.117, 4.67 to 4.69 and 4.71). This analy-

FCM INTERPRETATION AND REPORTING

321

sis is particularly useful because downregulation of one pan-T antigen may occur in certain reactive conditions, and, conversely, some T-cell LPD/NHL may display no antigenic abnormalities. Neoplastic TCR-γδ T-cell proliferations, on the other hand, are much rarer. It has been thought that the TCR-γδ+ profile is characteristic of hepatosplenic T-cell lymphoma. Cases composed of neoplastic TCR-αβ T-cells with NK antigen expression have been recently described, however. In either case, the disease is not limited to the spleen/liver, but is systemic and rapidly progressive. The bone marrow and peripheral blood involvement are indistinguishable from that of an acute leukemia. Some of the antigenic features (e.g., lack of both CD4 and CD8) are similar to those seen in T-ALL/LL. The FCM diagnostic workup should therefore include testing for the immature markers TdT and CD34 (Figure 5.23 and Plate 9). Mature T-cell malignancies, especially those involving solid tissue, are currently classified into multiple subtypes. The nomenclature and often the criteria for classification are based primarily on morphologic features (e.g., T-cell LGL leukemia, angioimmunoblastic T-cell lymphoma) or derived from the sites of involvement (e.g., enteropathy-type T-cell lymphoma, hepatosplenic T-cell lymphoma, subcutaneous panniculitis-like T-cell lymphoma, extranodal NK/T-cell lymphoma, nasal type, and aggressive NK-cell leukemia). A biological classification of T-cell LPD/NHL, based on the antigenic characteristics and proliferative activity (S-fraction) to indicate the cell lineage and tumor grade, respectively, would be more reproducible and less confusing, especially when the neoplasm under consideration consists of NK or NK-like T-cells. Where appropriate, very specific morphologic or clinical features can be included in the classification. Using this biological approach, several specific entities of mature T-cell LPD/NHL can be recognized under the two broad categories of CD4+ and CD8+/NK malignancies. The majority of post-thymic T-cell neoplasms are CD4+. Many of the CD8+ LPD/NHL also express one or more NK-associated antigens. The proportion of indolent versus aggressive mature T-cell neoplasms varies between different parts of the world. The aggressive disorders are more prevalent in the Far East than in Western countries. 5.2.3.1 CD4+ T-cell LPD/NHL

Case studies 29 to 32

Case studies 61 to 63

Adult T-cell leukemia–lymphoma (ATLL) (Plate 49 and Figure 3.110) and T-prolymphocytic leukemia (Plate 50; Figures 3.91 and 3.101) are the two most easily identifiable CD4+ T-cell disorders based on the combined immunologic, morphologic and clinical findings (see Section 3.6.5). The morphology of either disease is best appreciated in the peripheral blood, especially in ATLL, where the involvement of the bone marrow invariably appears less obvious than that in the blood. In general, T-PLL demonstrates a much higher degree of leukocytosis than ATLL, but the cytologic features are less dramatic, with fewer nuclear clefts and indentations. On optimally prepared blood films, a distinct nucleolus can be appreciated in T-PLL cells; the overall cytology bears no similarity to B-PLL cells (Plate 38), however. In contrast to T-PLL and ATLL, recognition of Sézary syndrome may be less straightforward, especially if the cerebriform cytology is not readily apparent. The abnormal lymphoid cells may be low in numbers and display a normal helper T-cell profile (see Section 4.4.2). On the other hand, cases with a large number of neoplastic cells and a high degree of nuclear irregularities may simulate the picture of ATLL (Plate 51). Circulating lymphoma cells from nodal-based PTCL or other subtypes of cutaneous T-cell NHL (Figure 4.71), although infrequent, can result in a blood picture similar to that in Sézary syndrome. Conversely, the lymph node picture in the advanced stages of mycosis fungoides/Sézary (MF/SS) syndrome is that of a PTCL that cannot be distinguished from other nodal-based PTCLs. The most frequent antigenic abnormality in Sézary syndrome is the loss of CD7 expression. This finding in itself is not pathognomonic of MF/SS, however. Knowledge of other corroborating laboratory/clinical data

322

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 5.23 TCR-γδ T-cell lymphoma–leukemia (same patient as in Plate 9). (a–d) Gated on peripheral blood MNCs: Leukemic cells (gray) are TCR-γδ+, lacking both CD4 and CD8. CD16 and CD56 are not expressed. (e, f) Ungated data: The absence of both CD34 and TdT confirms that the tumor cells are mature γδ-T-cells.

FCM INTERPRETATION AND REPORTING

Case study 98

323

relating to the cutaneous lesions helps to confirm the diagnosis. Skin biopsy features such as a superficial psoriasiform lichenoid lymphocytic infiltrate and Pautrier abscesses (Plate 52) are quite characteristic of mycosis fungoides. In addition to the above-mentioned disorders, the bulk of CD4+ T-cell malignancies consist of nodal-based PTCLs. The variability of several nonspecific features such as prominent vascularity, epithelioid histiocytes, and nuclear pleomorphism accounts for the diverse morphologic manifestations of PTCLs.

5.2.3.2 CD8+ disorders

Case studies 34, 35, 37, 64 and 65

CD8+ disorders can be further divided based on the expression of NK markers and the TCR/CD3 complex, into the subgroups of (1) suppressor T-cell (NK markers absent), (2) NKlike T-cell, and (3) true NK (TCR/CD3 complex absent) disorders. Each subgroup is heterogeneous and includes reactive conditions (see Section 4.1.2.3) as well as indolent and aggressive malignant disorders (see Sections 3.6.5 and 4.4.2). In normal individuals, a small subset of normal suppressor T-cells coexpressing CD103 may be detected in the peripheral blood. Their malignant counterparts occur primarily in the gastrointestinal tract of patients with intractable gluten enteropathy (celiac disease), itself a condition of altered immunity with an increased risk of malignancy (Figures 5.24 and 5.25). The CD8+ indolent neoplasms are found mainly in the blood and bone marrow and are composed of normal-appearing LGLs (Plate 53). The cytologic identification of LGLs may be challenging when the cytoplasmic granules are few and of a size below the threshold detectable by light microscopy. The great majority of indolent LGL neoplasms display an NK-like T-cell, TCR-αβ+ CD57+ phenotype (Figure 4.74). Similar to the reactive LGL proliferations, indolent CD3+ LGL leukemia has a high association with autoimmune disorders such as rheumatoid arthritis. Because of the similarities in phenotype, cytology, and clinical picture, the differential diagnosis between indolent neoplastic and reactive LGL proliferations can be difficult, especially if the number of large granular lymphocytes does not exceed the arbitrary level previously established for the diagnosis of LGL leukemia. In most instances, the evaluation of the TCR-Vβ repertoire helps to resolve the diagnostic dilemma (see Section 4.4.2) and may establish the diagnosis of LGL malignancy even in the absence of overt LGL lymphocytosis (Figures 5.26 to 5.28). Indolent NK-LGL neoplasms (negative for CD3, TCR-αβ and TCR-γδ) occur less frequently (Figures 4.72 and 4.73). The tools available for determining clonality in this group of disorders still remain limited, however (see Section 4.4.2). In general, when the clinical and laboratory data (including molecular genetics) fail to establish the malignant nature of any LGL proliferation, a conservative approach of “watch and wait” with periodic follow-up FCM analysis of the blood or bone marrow, is warranted. As a group, aggressive CD8+ neoplasms are uncommon, although those with NK-associated antigens are found at high frequency in certain regions of the Far East and Central/South America, where the tumors also tend to be associated with EBV. The malignant nature of the CD8+ aggressive neoplasms can be easily recognized based on features such as aneuploidy, high S-phase, large cell size, or diffuse/extensive involvement. The cytology of the tumor cells on air-dried preparations is variable, depending on the quality of the cytoplasm (pale to basophilic) and the number and size of the azurophilic granules (Plate 54). On tissue sections the eye-catching feature is the ample pale cytoplasm (Plate 55), but this is not a pathognomonic finding. The tumor cells may also mimic the appearance of AML (especially AML-M5) when presenting in the blood or bone marrow (Plate 56). The current classification and terminology of aggressive CD8+ neoplasms are confusing, especially for the tumors that express NK antigens. Some of these may occur as a

324

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 5.24 Jejunal biopsy with CD103+ T-cell NHL in a patient with protracted diarrhea. (a–f) The biopsy contains a preponderance of T-cells with low FSC and no overt abnormalities in the expression of CD3, CD5, CD2 and CD7. The CD4:CD8 ratio is markedly reversed (1:12). HLA-DR and CD56 are heterogeneously expressed.

FCM INTERPRETATION AND REPORTING

a

b

c

d

e

f

325

Figure 5.25 CD103+ T-cell NHL (continuation of Figure 5.24). (a, b) The abnormal T-cells display coexpression of CD8 and CD103. Myeloid antigens (CD13/CD33 combination) are present. (c–f) Analysis of the TCR-Vβ repertoire gated on CD3+ CD8+ T-cells: Clonal expansion of the Vβ11 family. There is virtually no reactivity to antibodies in the other TCR-Vβ tubes (e.g., tubes 1 and 3).

326

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 5.26 NK-like T-cell leukemia without overt LGL lymphocytosis. (a) Peripheral blood with platelet clumps (thin arrow) and 70% lymphocytes. (b–f) A conspicuous cluster of abnormal T-cells (arrow) with slightly downregulated CD7, CD2, CD5, CD3 and CD8. CD57 is expressed. CD56 is absent. The abnormal NK-like T-cells comprise 20% of the lymphocytes.

FCM INTERPRETATION AND REPORTING

a

b

c

d

e

f

327

Figure 5.27 NK-like T-cell leukemia without overt LGL lymphocytosis (continuation of Figure 5.26). (a) The abnormal cells (arrow) are of small cell size. (b–f) TCR-αβ is expressed; TCR-γδ is absent. The analysis of the TCR-Vβ repertoire is gated on the T-cells with downregulated CD3 and CD5. The tumor cells display no reactivity to the antibodies contained in TCR-Vβ tubes 1 and 2.

328

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 5.28 NK-like T-cell leukemia without overt LGL lymphocytosis (continuation of Figure 5.27). (a–f) The abnormal NK-like T-cells display no reactivity to the antibodies contained in TCR-Vβ tubes 3 through 8. Lack of reactivity to all currently available antibodies to the Vβ families is indirect evidence of clonality.

FCM INTERPRETATION AND REPORTING

Case studies 33 and 36

329

posttransplant LPD and therefore are classified as such. Many factors have contributed to the confusion, including (1) as a group, these malignancies occur at low frequency; each subcategory is therefore even rarer, (2) the phenotypic characterization of the reported cases was often based on immunohistochemistry instead of FCM, and therefore was made with a more limited panel, and (3) the associated clinical data (e.g., systemic disease involvement at other sites) have not always been included in the published studies. As a result, the true NK and NK-like T-cell malignancies are being referred to by different names (each considered as a distinct entity) depending on whether the disease is found, at least initially, in solid tissue sites (often extranodal) or in the peripheral blood/bone marrow compartment. In many instances, however, the tumor already involves other organs at the time of presentation, or quickly becomes systemic irrespective of its initial site of involvement. Many of these “distinct entities” share closely similar cytologic, phenotypic, biological, and clinical features (e.g., an association with EBV infection). Conversely, some of the “distinct entities” are actually very heterogeneous in terms of phenotypes and biological behavior. For instance, the “entity” subcutaneous panniculitis-like T-cell lymphoma consists of at least two different diseases, one indolent and the other rapidly progressive, often with systemic involvement of liver, spleen, blood and bone marrow at presentation. Furthermore, the phenotype of this “entity” is also heterogeneous whereby the neoplastic population displays either a cytotoxic profile (TCR-αβ+ with or without CD56 expression) or a TCR-γδ+ profile. To circumvent the above confusion, it would be preferable to designate these aggressive lymphomas/leukemias by their phenotypes (see Section 3.6.5) instead of by the organ site or histologic features. A more simplified terminology can be used, such as “CD8+ T-cell lymphoma, NK markers absent,” “true NK lymphoma-leukemia, CD56+,” “NK-like T-cell lymphoma-leukemia, TCR-αβ+, CD56+” or “TCR-γδ lymphoma-leukemia” (Figures 3.111, 3.112 and 3.114).

5.2.3.3 CD30+ lymphoma

Case studies 99 and 100

Although the presence of CD30 (Figures 5.29 and 5.30) in a non-B-cell lymphoma suggests a so-called Ki-1 (CD30) lymphoma, this diagnosis needs to be rendered with caution. In some cases of CD30-lymphoma, the tumor cells have lost all pan T-cell antigens, thus displaying an apparent “null” phenotype. Tumors expressing this antigen are also referred to as anaplastic large cell lymphomas (ALCLs) despite the fact that a substantial number of cases do not demonstrate pleomorphic morphology (Plate 57). Morphologic features such as sinusoidal growth pattern can also be found in other B- or T-cell lymphomas. Because CD30 expression and pleomorphic features may occur in other post-thymic T-cell malignancies (e.g., in the late stages of mycosis fungoides or some NK and NK-like T-cell malignancies), some other more specific criteria should be relied upon to define CD30 lymphoma. It is preferable to restrict the diagnosis to those lymphomas with the translocation (2;5)(p23q35). This translocation juxtaposes the gene encoding nucleophosmin (NPM) on chromosome 5p23 with the ALK (anaplastic lymphoma kinase) gene encoding a tyrosine kinase receptor. The resulting product is the NPM-ALK chimeric protein. The intracytoplasmic ALK portion of the chimeric protein can be recognized by the currently available ALK-1 antibodies, thus facilitating the distinction of CD30+ ALK-1+ NHL from other CD30+ ALCLs, such as the CD30 primary cutaneous ALCL, which affects adults and has a relatively good prognosis if confined to the skin. CD30+ ALK-1+ ALCL, on the other hand, is a rare systemic disease affecting children and adolescents. Disease involvement can be nodal or extranodal. ALK-1 reactivity is also observed in other lymphoid neoplasms, in which there is either expression of the full-length ALK receptor or activation of the ALK gene by a different translocation, for example, t(2;7). The clinical setting (older age group) and other antigenic features (B-cell phenotype or lack of CD30 expression) differ from that of CD30+ ALK-1+ lymphoma, however. Molecular or cytogenetic studies for

330

FLOW CYTOMETRY IN HEMATOPATHOLOGY

a

b

c

d

e

f

Figure 5.29 CD30+ lymphoma in an inguinal lymph node of a 15-year-old patient. The tumor cells (gray) are of variable cell size (a), with increased autofluorescence (e). (a–f) CD3, CD4 and CD8 are negative. CD2, CD7 and CD56 are present in a subset of the tumor cells; very few are CD5+ dim. Residual lymphocytes consist mostly of T-cells, with a small number of CD56+ NK-like T-cells. B-cells are virtually absent.

FCM INTERPRETATION AND REPORTING

a

b

c

d

e

f

331

Figure 5.30 CD30+ lymphoma (continuation of Figure 5.29). (a, b) Neoplastic cells (gray) coexpress CD25 and myeloid markers (CD13/CD33 combination). (c, d) The lack of MPO and the presence of cCD3 (albeit at lower levels than in normal T-cells) confirm that the tumor cells are of T-cell lineage. (e, f) CD30 is expressed. The tumor is near-diploid with a high S-phase of 25%.

332

FLOW CYTOMETRY IN HEMATOPATHOLOGY FLOW CYTOMETRY IMMUNOPHENOTYPING ANALYSIS Patient Name: Date of Birth: 11/14/1947

Patient ID Number: Gender: M

Specimen Date:

Specimen: Bone marrow

Clinical History:

Case Number: Report Date:

Follicular lymphoma treated with Rituxan. ? Bone marrow disease.

ANTIBODIES TESTED Kappa Lambda HLA-DR CD2 CD3 CD4 CD5 CD7 CD8 CD10

Negative + Moderate + Moderate Negative Negative Negative Negative Negative Negative + Dim

Viability: 96%

CD11b CD11c CD13 CD14 CD16 CD19 CD20 CD23 CD25 CD33

Negative Negative Negative Negative Negative + Moderate Negative Negative Negative Negative

CD34 CD38 CD45 CD56 CD64 CD103 CD117 CD138

Negative + Dim + Bright Negative Negative Negative Negative Negative

Cell yield: 20 million

Abnormal cells present by flow cytometry: Yes Percentage of cells with abnormal phenotype: < 1% Cell size: Small

PHENOTYPE:

Monoclonal lambda CD10-positive B-cell population (0.3%).

MORPHOLOGY:

Bone marrow aspirate and biopsy, normocellular with no overt abnormalities in the hematopoietic precursors.

DIAGNOSIS:

A small population of monoclonal lambda B-cells is present in the bone marrow. The antigenic profile is consistent with follicular center cell lymphoma. In the context of the clinical history, the findings indicate minimal residual disease. The apparent "loss" of CD20 is secondary to antiCD20 therapy.

NOTE:

Molecular genetics pending.

Figure 5.31 An example of an FCM report. Information related to the patient has been omitted to preserve patient confidentiality. The clinical history was not provided initially. It was obtained at the time when the FCM data and the bone marrow morphology were reviewed.

FCM INTERPRETATION AND REPORTING

333

the (2;5) translocation provides the ultimate confirmatory evidence of CD30+ ALK-1 lymphoma.

5.3 FCM reporting In the last critical step of the FCM study, the complexities of the FCM data are translated into a form that can be understood by clinicians and pathologists who may not always be familiar with hematopathology or flow cytometry. The goal is to package all of the relevant information in a short and concise report, so as to communicate the FCM data and diagnosis to other physicians most efficiently. The technical aspects (e.g., antibody–fluorochrome conjugation, three- vs. four-color procedure) of FCM analysis as well as the FCM graphics can be omitted from the reports because few clinicians comprehend or care for these details. An example report is shown in Figure 5.31. In addition to the information regarding the patient and the specimen, the FCM report for a hematologic neoplasm should include the following: 1. 2. 3. 4. 5.

The proportion of the abnormal cells. The cell size of the critical cells, especially for LPD/NHL. The ploidy and S-phase fraction (where applicable). The markers evaluated. The antigens expressed by the neoplastic cells, including the respective fluorescence intensities. This may be indicated in the form of a short textual description of the phenotype of the tumor. 6. A short description of the morphologic findings, including cytochemistry or immunohistochemistry results, if applicable. 7. The diagnosis or differential diagnosis based on the available FCM, morphologic, and clinical data. 8. Any additional comment if appropriate, such as in cases with low cell yield, poor viability, or when other additional laboratory tests are recommended.

For specimens without phenotypically abnormal cells, the FCM report can be a summary of the relative proportions of the various cell populations (e.g., granulocytes, monocytes, B-cells, T-cells). If present, qualitative abnormalities on the erythroid and myeloid precursors (e.g., altered myeloid maturation curves) should also be mentioned. In general, the terminology employed for the diagnosis should reflect the lineage and maturation status of the neoplastic cells. In lymphomas, the grade (and thereby the proliferative capacity) of the tumor as inferred from the cell size, CD71 reactivity, or the S-phase fraction should also be included. Where the conventional terminology is unsatisfactory, the authors prefer to use a more conservative and generic terminology instead. For instance, the statement “low-grade B-cell NHL with a nondescript B-cell phenotype” is preferred to “marginal zone lymphoma,” especially if the specimen analyzed is the bone marrow or peripheral blood, in which no appreciation of any marginal zone pattern could be made. This type of statement may not necessarily follow currently accepted (yet predictably volatile) classification schemes, but it reflects objectively the phenotype of the neoplastic proliferation and its biological behavior.

Suggested reading

Almasri NM, Duque RE, Iturraspe J, et al. Reduced expression of CD20 antigen as a characteristic marker for chronic lymphocytic leukemia. Am J Hematol 1992;40:259–263. Almasri NM, Iturraspe JA, Braylan RC. CD10 expression in follicular lymphoma and large cell lymphoma is different from that of reactive lymph node follicles. Arch Pathol Lab Med 1998;122:539–544. Arnulf B, Copie-Bergman C, Delfau-Larue M-H, et al. Nonhepatosplenic γδ T-cell lymphoma: a subset of cytotoxic lymphomas with mucosal or skin localization. Blood 1998;91:1723–1731. Behm FG, Raimondi SC, Schell MJ, et al. Lack of CD45 on blast cells in childhood acute lymphoblastic leukemia is associated with chromosomal hyperdiploidy and other favorable prognostic features. Blood 1992;79:1011–1016. Benharroch D, Meguerian-Bedoyan Z, Lamant L, et al. ALK-positive lymphoma: a single disease with a broad morphologic spectrum. Blood 1998;91:2076–2084. Bennett JM, Catovsky D, Daniel M-T, et al. Proposals for the classification of chronic (mature) B and T lymphoid leukaemias. J Clin Pathol 1989;42:567–584. Blorn B, Verschuren MCM, Heemskerk MHM, et al. TCR gene rearrangements and expression of the pre-T cell receptor complex during human T cell differentiation. Blood 1999;93:3033–3043. Bogen SA, Pelley D, Charif M, et al. Immunophenotypic identification of Sezary cells in the peripheral blood. Am J Clin Pathol 1996;106:739–738. Borowitz MJ, Bray R, Gascoyne R, et al. U.S.–Canadian consensus recommendations on the immunophenotypic analysis of hematologic neoplasia by flow cytometry: data analysis and interpretation. Cytometry (Commun Clin Cytometry) 1997;30:236–244. Borowitz MJ, Guenther KL, Shults KE, et al. Immunophenotyping of acute leukemia by flow cytometric analysis. Use of CD45 and right-angle light scatter to gate on leukemic blasts in three color analysis. Am J Clin Pathol 1993;100:534–540. Borowitz MJ, Shuster JJ, Civin CI, et al. Prognostic significance of CD34 expression in childhood B-precursor acute lymphoblastic leukemia: a Pediatric Oncology Group Study. J Clin Oncol 1990;8:1380–1398. Bouroncle BA. Thirty-five years in the progress of hairy cell leukemia. Leukemia Lymphoma 1994;14 (Suppl 1):1–12. Bowen KL, Davis BH. Abnormal patterns of expression of CD16 (FCRgIII) and CD11b (CRIII) antigens by developing neutrophils in the bone marrow of patients with myelodysplastic syndrome. Lab Hematol 1997;3:292–298. Braylan RC. Flow cytometric DNA analysis in the diagnosis and prognosis of lymphoma. Am J Clin Pathol 1993;99:374–380. Braylan RC. Impact of flow cytometry on the diagnosis and characterization of lymphomas, chronic lymphoproliferative disorders and plasma cell neoplasias. Cytometry Part A 2004;58A:57–61. Braylan RC. Lymphomas. In: Bauer KD, Duque, RE, Shankey TV, eds. Clinical Flow Cytometry: Principles and Application. Philadelphia: Lippincott Williams & Wilkins, 1993:203–234. Braylan RC, Atwater SK, Diamond LW, et al. U.S.–Canadian consensus recommendations on the immunophenotypic analysis of hematologic neoplasia by flow cytometry: data reporting. Cytometry (Commun Clin Cytometry) 1997;30:245–248. Braylan RC, Benson NA, Iturraspe J. Analysis of lymphomas by flow cytometry. Current and emerging strategies. Ann NY Acad Sci 1993;677:364–378. Braylan RC, Benson NA. Flow cytometric analysis of lymphomas. Arch Pathol Lab Med 1989;113:627–633. Braylan RC, Orfao A, Borowitz MJ, et al. Optimal number of reagents required to evaluate hematolymphoid neoplasias: results of an international consensus study. Cytometry (Commun Clin Cytometry) 2001;46:23–27.

336

SUGGESTED READING Burke JS. Are there site-specific differences among the MALT lymphomas—morphologic, clinical? Am J Clin Pathol 1999;111 (Suppl 1):S133–S143. Campana D, Coutan-Smith E. Detection of minimal residual disease in acute leukemia by flow cytometry. Cytometry (Commun Clin Cytometry) 1999;38:139–152. Carulli G, Gianfaldoni ML, Azzara A, et al. FcRIII (CD16) expression on neutrophils from chronic myeloid leukemia: a flow cytometric study. Leuk Res 1992;16:1203–1209. Chen J-S, Coutan-Smith E, Suzuki T, et al. Identification of novel markers for monitoring minimal residual disease in acute lymphoblastic leukemia. Blood 2001;97:2115–2120. Cherian S, Moore J, Bantly A, et al. Peripheral blood MDS score: a new flow cytometric tool for the diagnosis of myelodysplastic syndromes. Cytometry Part B 2005;64B:9–17. Christensson B, Lindemalm C, Johansson B, et al. Flow cytometric DNA analysis: a prognostic tool in non-Hodgkin’s lymphoma. Leuk Res 1989;13:307–314. Colucci F, Caligiuri MA, Di Santo JP. What does it take to make a natural killer? Nat Rev Immunol 2003;3:413–425. Cooke CB, Krenacs L, Tetler-Stevenson M, et al. Hepatosplenic T-cell lymphoma; a distinct clinicopathologic entity of cytotoxic γδ T-cell origin. Blood 1996;88:265–4274. Cornfield DB, Mitchell DM, Almasri NM, et al. Follicular lymphoma can be distinguished from benign follicular hyperplasia by flow cytometry using simultaneous staining of cytoplasmic bcl-2 and cell surface CD20. Am J Clin Pathol 2000;114:258–263. Cornfield DB, Mitchell Nelson DM, Rimsza LM, et al. The diagnosis of hairy cell leukemia can be established by flow cytometric analysis of peripheral blood, even in patients with low levels of circulating malignant cells. Am J Hematol 2001;67:223–226. Coutan-Smith E, Sancho J, Hankcock ML, et al. Clinical importance in minimal residual disease in childhood acute lymphoblastic leukemia. Blood 2000;96:2691–2696. Crespo M, Bosch F, Villamor N, et al. ZAP-70 expression as a surrogate for immunoglobulinvariable-region mutations in chronic lymphocytic leukemia. N Engl J Med 2003;348:1764– 1775. Damle RN, Wasil T, Fais F, et al. IgVH gene mutation status and CD38 expression as novel prognostic indicators in chronic lymphocytic leukemia. Blood 1999;94:1840–1847. Davis BH, Foucar K, Szczarkowski W, et al. U.S.–Canadian consensus recommendations on the immunophenotypic analysis of hematologic neoplasia by flow cytometry: medical indications. Cytometry (Commun Clin Cytometry) 1997;30:249–263. De Rosa SC, Andrus JP, Perfetto SP, et al. Ontogeny of gamma delta T cells in humans. J Immunol 2004;172:1637–1645. Delmer A, Ajchenbaum-Cymbalista F, Tang R, et al. Over-expression of cyclin D1 in chronic B-cell malignancies with abnormality of chromosome 11q13. Br J Haematol 1995;89:798–804. Diamond LW, Bearman RM, Berry PK, et al. Prolymphocytic leukemia: flow microfluorometric, immunologic, and cytogenetic observations. Am J Hematol 1980;9:319–330. Doherty PC. The function of γδ T-cells. Br J Haematol 1992;81:321–324. Duerig J, Nueckel H, Cremer M, et al. Zap-70 expression is a prognostic factor in chronic lymphocytic leukemia. Leukemia 2003;17:2426–2434. Dunphy CH, Dunphy FR, Visconti JL. Flow cytometric immunophenotyping of bone marrow core biopsies. Report of 8 patients with previously undiagnosed hematologic malignancy and failed bone marrow aspiration. Arch Pathol Lab Med 1999;123:206–212. Duque RE, Andreeff M, Braylan RC, et al. Consensus review of the clinical utility of DNA flow cytometry in neoplastic hematopathology. Cytometry 1993;14:492–496. Dworzak MN, Fritsch G, Fleischer C, et al. Multiparameter phenotype mapping of normal and postchemotherapy B lymphopoiesis in pediatric bone marrow. Leukemia 1997;11:1266–1273. Falini B, Pileri S, Zinzani PL, et al. ALK+ lymphomas: clinico-pathological findings and outcome. Blood 1999;93:2697–2706. Gentile TC, Uner AH, Hutchison RE, et al. CD3+, CD56+ aggressive variant of large granular lymphocyte leukemia. Blood 1994;84:2315–2321. Gorczyca W, Weisberger J, Liu Z, et al. An approach to diagnosis of T-cell lymphoproliferative disorders by flow cytometry. Cytometry 2002;50:177–190. Harada H, Kawano MM, Huang N, et al. Phenotypic difference of normal plasma cells from mature myeloma cells. Blood 1993;81:2658–2663. Hernandez O, Oweity T, Ibrahim S. Is an increase in CD4/CD8 T-cell ratio in lymph node fine needle aspiration helpful for diagnosing Hodgkin lymphoma? A study of 85 lymph nodes FNAs with increased CD4/CD8 ratio. Cytojournal 2005;2:14–19.

SUGGESTED READING

337

Huebl W, Iturraspe J, Braylan RC. FMC7 antigen expression on normal and malignant B-cells can be predicted by expression of CD20. Cytometry (Commun Clin Cytometry) 1998;34:71–74. Hurwitz CA, Raimondi SC, Head D, et al. Distinctive immunophenotypic features of t(8,21)(q22;q22) acute myeloblastic leukemia in children. Blood 1992;12:3182–3188. Jaffe ES. Classification of NK cell and NK-like T-cell malignancies. Blood 1996;87:1207–1210. Jennings CD, Foon KA. Recent advances in flow cytometry: application to the diagnosis of hematologic malignancy. Blood 1997;90:2863–2892. Kabutomori O, Iwatani I, Koh T, et al. CD16 antigen density on neutrophils in chronic myeloproliferative disorders. Am J Clin Pathol 1997;107:661–664. Kaleem Z, White G, Zutter MM. Aberrant expression of T-cell-associated antigens on B-cell nonHodgkin lymphomas. Am J Clin Pathol 2001;115:396–403. Karandikar NJ, Hotchkiss EC, McKenna RW, et al. Transient stress lymphocytosis: an immunophenotypic characterization of the most common cause of newly defined adult lymphocytosis in a tertiary hospital. Am J Clin Pathol 2002;117:819–825. Kita K, Nakase K, Miwa H, et al. Phenotypical characteristics of acute myelocytic leukemia associated with the t(8;21)(q22;q22) chromosomal abnormality: frequent expression of immature B-cell antigen CD19 together with stem cell antigen CD34. Blood 1992;80:470–477. Koike T. Megakaryoblastic leukemia: the characterization and identification of megakaryoblasts. Blood 1984;64:683–692. Koo CH, Rappaport H, Sheibani K, et al. Imprint cytology of non-Hodgkin’s lymphomas. Based on a study of 212 immunologically characterized cases. Hum Pathol 1989;20 (Suppl 1):1–138. Krasinskas AM, Wasik MA, Kamoun M, et al. The usefulness of CD64, other monocyte-associated antigens, and CD45 gating in the subclassification of acute myeloid leukemias with monocytic differentiation. Am J Clin Pathol 1998;110:797–805. Kussick SJ, Fromm JR, Rossini A, et al. Four-color flow cytometry shows strong concordance with bone marrow morphology and cytogenetics in the evaluation for myelodysplasia. Am J Clin Pathol 2005;124:170–181. Kussick SJ, Kalnoski M, Braziel RM, et al. Prominent clonal B-cell populations identified by flow cytometry in histologically reactive proliferations. Am J Clin Pathol 2004;121:464–472. Lamy T, Loughran TP Jr. Current concepts: large granular lymphocyte leukemia. Blood Rev 1999;13:230–240. Langerak AW, van den Beemd R, Wolvers-Tettero ILM, et al. Molecular and flow cytometric analysis of the Vβ repertoire for clonality assessment in mature TCRαβ T-cell proliferations. Blood 2001;98:165–173. Lanier LL. NK cell receptors. Annu Rev Immunol 1998;16:359–393. Leclercq G, Plum, J. Thymic and extrathymic T cell development. Leukemia 1996;10:1853– 1859. Lee RV, Braylan RC, Rimsza LM. CD58 decreases as non-malignant B-cells mature in bone marrow and is frequently overexpressed in adult and pediatric precursor B-cell acute lymphoblastic leukemia. Am J Clin Pathol 2005;123:119–124. Li S, Eshleman JR, Borowitz MJ. Lack of surface immunoglobulin light chain expression by flow cytometric immunophenotyping can help diagnose peripheral B-cell lymphoma. Am J Clin Pathol 2002;118:229–234. Liang X, Meech SJ, Odom LF, et al. Assessment of t(2;5)(p23;q35) translocation and variants in pediatric ALK+ anaplastic large cell lymphoma. Am J Clin Pathol 2004;121:496–506. Lima M, Almeida J, Santos AH, et al. Immunophenotypic analysis of the TCR-Vβ repertoire in 98 persistent expansions of CD3+/TCR-αβ+ large granular lymphocytes. Utility in assessing clonality and insights into the pathogenesis of the disease. Am J Pathol 2001;159:1861–1868. Lima M, Almeida J, Teixeira MdA, et al. TCRαβ+/CD4+ large granular lymphocytosis. A new clonal T-cell lymphoproliferative disorder. Am J Pathol 2003:163:763–771. Lima M, Teixeira M, Queiros ML, et al. Immunophenotypic characterization of normal blood CD56+lo versus CD56+hi NK-cell subsets and its impact on the understanding of their tissue distribution and functional properties. Blood Cells Mol Dis 2001;27:731–743. Lin P, Hao S, Handy BC, et al. Lymphoid neoplasms associated with IgM paraprotein: a study of 382 patients. Am J Pathol 2005;123:200–205. Loken MR, Shah VD, Dattilio KL, et al. Flow cytometric analysis of human bone marrow. II. Normal B lymphocyte development. Blood 1987;70:1316–1324. Loken MR, Shah VO, Dattilio KL, et al. Flow cytometric analysis of human bone marrow: I. Normal erythroid development. Blood 1987;69:255–263.

338

SUGGESTED READING Longacre TA, Foucar K, Crago S, et al. Hematogones: a multiparameter analysis of bone marrow precursor cells. Blood 1989;72:543–552. Look AT, Roberson PK, Williams DL, et al. Prognostic importance of blast cell DNA content in childhood acute lymphoblastic leukemia. Blood 1985;65:1079–1086. Loughran TP Jr. Clonal diseases of large granular lymphocytes. Blood 1993;82:1–14. Macon WR, Levy NB, Kurtin PJ, et al. Hepatosplenic α/β T-cell lymphomas: a report of 14 cases and comparison with hepatosplenic γ/δ T-cell lymphomas. Am J Surg Pathol 2001;25:285–296. Macon WR, Williams ME, Greer JP, et al. Natural killer-like T-cell lymphomas: aggressive lymphomas of T-large granular lymphocytes. Blood 1996;87:1474–1483. Malec M, Bjorklund E, Soderhall S, et al. Flow cytometry and allele-specific oligonucleotide PCR are equally effective in detection of minimal residual disease in ALL. Leukemia 2000;15:716–727. Matutes E, Brito-Babapulle V, Swansbury J, et al. Clinical and laboratory features of 78 cases of Tprolymphocytic leukemia. Blood 1991;78:3269–3274. Matutes E, Oscier J, Garcia-Marco J, et al. Trisomy 12 defines a group of CLL with atypical morphology: correlation between cytogenetic, clinical and laboratory features in 544 patients. Br J Haematol 1996;92:382–388. Meyerson HJ, MacLennan G, Husel W, et al. D cyclins in CD5+ B-cell lymphoproliferative disorders: cyclin D1 and cyclin D2 identify diagnostic groups and cyclin D1 correlates with Zap-70 in chronic lymphocytic leukemia. Am J Clin Pathol 2006;125:241–250. Morice WG, Katzmann JA, Pittelkow MR, et al. A comparison of morphologic features, flow cytometry, TCR-Vβ analysis, and TCR-PCR in qualitative and quantitative assessment of peripheral blood involvement by Sezary syndrome. Am J Clin Pathol 2006;125:364–374. Morice WG, Kimlinger T, Katzmann JA, et al. Flow cytometric assessment of TCR-Vβ expression in the evaluation of peripheral blood involvement by T-cell lymphoproliferative disorders: a comparison with conventional T-cell immunophenotyping and molecular genetic techniques. Am J Clin Pathol 2004;121:373–383. Nathwani BN. Diagnostic significance of morphologic patterns of lymphoid proliferations in lymph nodes. In: Knowles DM, ed. Neoplastic Hematopathology. Philadelphia: Lippincott Williams & Wilkins, 2000:507–536. Nguyen D, Moskowitz FB, Diamond LW. Potential diagnostic pitfalls caused by blood film artifacts in prolymphocytic leukaemia. Observations in two cases. Br J Biomed Sci 1994;51:371–374. Nguyen DT, Diamond LW, Schwonzen M, et al. Chronic lymphocytic leukaemia with an interfollicular architecture: avoiding diagnostic confusion with monocytoid B-cell lymphoma. Leuk Lymphoma 1995;18:179–184. Nguyen DT, Diamond LW. Diagnostic Hematology: A Pattern Approach. London: Arnold, 2000. Ocqueteau M, Orfao A, Almeida J, et al. Immunophenotypic characterization of plasma cells from monoclonal gammopathy of undetermined significance patients. Implications for the differential diagnosis between MGUS and multiple myeloma. Am J Pathol 1998;152:1655–1665. Ocqueteau M, Orfao A, Garcia-Sanz R, et al. Expression of the CD117 on normal and myelomatous plasma cells. Br J Haematol 1996;95:489–493. Orchard JA, Ibbotson RE, Davis Z, et al. Zap-70 expression and prognosis in chronic lymphocytic leukaemia. Lancet 2004;363:105–111. Orfao A, Chillon MC, Botolucci AM, et al. The flow cytometric pattern of CD34, CD15 and CD13 expression in acute myeloblastic leukemia is highly characteristic of the presence of PML-RARα gene rearrangements. Hematologica 1999;84:405–412. Plander M, Brockhoff G, Barlage S, et al. Optimization of three- and four-color multiparameter DNA analysis in lymphoma specimens. Cytometry Part A 2003;54A:66–74. Radaev S, Sun PD. Structure and function of natural killer cell surface receptors. Annu Rev Biophys Biomol Struct 2003;32:93–114. Rassenti L, Huynh L, Toy TL, et al. ZAP-70 compared with immunoglobulin heavy-chain gene mutation status as a predictor of disease progression in chronic lymphocytic leukemia. N Engl J Med 2004;351:893–901. Reichard KK, McKenna RW, Kroft SH. Comparative analysis of light chain expression in germinal center cells and mantle cells of reactive lymphoid tissues. A four-color flow cytometric study. Am J Clin Pathol 2003;119:130–136. Richards SJ, Sivakumaran M, Parapia LA, et al. A distinct large granular lymphocyte (LGL)/NKassociated (NKa) abnormality characterized by membrane CD4 and CD8 coexpression. Br J Haematol 1992;82:494–501.

SUGGESTED READING

339

Robertson MJ, Ritz J. Biology and clinical relevance of human natural killer cells. Blood 1990;76: 2421–2438. San Miguel JF, Martinez A, Macedo A, et al. Immunophenotyping investigation of minimal residual disease is a useful approach for predicting relapse in acute myeloid leukemia patients. Blood 1997;90:2465–2470. San Miguel JF, Vidriales MB, Lopez-Berges C, et al. Early immunophenotypical evaluation of minimal residual disease in acute myeloid leukemia identifies different patient risk groups and may contribute to postinduction treatment stratification. Blood 2001;98:1746–1751. Sanchez ML, Almeida J, Vidriales B, et al. Incidence of phenotypic aberrations in a series of 467 patients with B chronic lymphoproliferative disorders: basis for the design of specific four-color stainings to be used for minimal residual disease investigation. Leukemia 2002;16:1460–1469. Sausville JE, Salloum RG, Sorbara L, et al. Minimal residual disease detection in hairy cell leukemia. Comparison of flow cytometric immunophenotyping with clonal analysis using consensus primer polymerase chain reaction for the heavy chain gene. Am J Clin Pathol 2003;119:213–217. Schroers R, Griesinger F, Truemper L, et al. Combined analysis of ZAP-70 and CD38 expression as a predictor of disease progression in B-cell chronic lymphocytic leukemia. Leukemia 2005;19: 750–758. Semenzato G, Zambello R, Starkebaum G, et al. The lymphoproliferative disease of granular lymphocytes: updated criteria for diagnosis. Blood 1997;89:256–260. Shin SS, Sheibani K. Monocytoid B-cell lymphoma. Am J Clin Pathol 1993;99:421–425. Smets LA, Homan-Blok J, Hart A, et al. Prognostic implication of hyperdiploidy as based on DNA flow cytometric measurement in childhood acute lymphocytic leukemia—a multicenter study. Leukemia 1987;1:163–166. Smith PJ, Blunt N, Wiltshire M, et al. Characteristics of a novel deep red/infrared fluorescent cellpermeant DNA probe, DRAQ5, in intact human cells analyzed by flow cytometry, confocal and multiphoton microscopy. Cytometry 2000;40:280–291. Smith PJ, Wiltshire M, Davies S, et al. A novel cell permeant and far red-fluorescing DNA probe, DRAQ5, for blood cell discrimination by flow cytometry. J Immunol Methods 1999;229:131–139. Spiekermann K, Emmendoerffer A, Elsner J, et al. Altered surface marker expression and function of G-CSF-induced neutrophils from test subjects and patients under chemotherapy. Br J Haematol 1996;87:31–38. Spits H, Lanier LL, Phillips JH. Development of human T- and natural-killer cells. Blood 1995;85: 2654–2670. Steltzer GT, Shults KE, Loken MR. CD45 gating for routine flow cytometric analysis of human bone marrow specimens. Ann NY Acad Sci 1993;677:265–279. Stelzer GT, Marti G, Hurley A, et al. U.S.–Canadian consensus recommendations on the immunophenotypic analysis of hematologic neoplasia by flow cytometry: standardization and validation of laboratory procedures. Cytometry (Commun Clin Cytometry) 1997;30:214–230. Stetler-Stevenson M, Arthur DC, Jabbour N, et al. Diagnostic utility of flow cytometric immunophenotyping in myelodysplastic syndrome. Blood 2001;98:979–987. Stewart CC, Behm FG, Carey JL, et al. U.S.–Canadian consensus recommendations on the immunophenotypic analysis of hematologic neoplasia by flow cytometry: selection of antibody combinations. Cytometry (Commun Clin Cytometry) 1997;30:231–235. Tefferi A, Li C-Y, Witzig TE, et al. Chronic natural killer cell lymphocytosis: a descriptive clinical study. Blood 1994;84:2721–2725. Terstappen LWMM, Huang S, Picker LJ. Flow cytometric assessment of human T-cell differentiation in thymus and bone marrow. Blood 1992;79:666–677. Terstappen LWMM, Loken MR. Myeloid cell differentiation in normal bone marrow and acute myeloid leukemia assessed by multi-dimensional flow cytometry. Anal Cell Pathol 1990;2:229–240. Terstappen LWMM, Safford M, Koeemann S, et al. Flow cytometric characterization of acute myeloid leukemia. Part II. Phenotypic heterogeneity at diagnosis. Leukemia 1991;5:757–767. Terstappen LWMM, Safford M, Loken MR. Flow cytometric analysis of human bone marrow III. Neutrophil maturation. Leukemia 1990;4:657–663. The non-Hodgkin’s Lymphoma Classification Project. A clinical evaluation of the International Lymphoma Study Group classification of non-Hodgkin’s lymphoma. Blood 1997;89:3909–3918. Uhrberg M, Valiante NM, Young NT, et al. The repertoire of killer cell Ig-like receptor and CD94: NKG2A receptors in T cells: clones sharing identical αβ TCR rearrangement express highly diverse killer cell Ig-like receptor patterns. J Immunol 2001;166:3923–3932.

340

SUGGESTED READING van den Beemd R, Boor PPC, van Lochem EG, et al. Flow cytometric analysis of the Vβ repertoire in healthy controls. Cytometry 2000;40:336–345. van Lochem EG, van der Velden VHJ, Wind HK, et al. Immunophenotypic differentiation patterns of normal hematopoiesis in human bone marrow: reference patterns for age-related changes and diseaseinduced shifts. Cytometry Part B 2004;60B:1–13. Vandersteenhoven AM, Williams JE, Borowitz MJ. Marrow B-cell precursors are increased in lymphomas or systemic diseases associated with B-cell dysfunction. Am J Clin Pathol 1993;100:60–66. Vilches C, Parham P. KIR: diverse, rapidly evolving receptors of innate and adaptive immunity. Annu Rev Immunol 2002;20:217–251. Wang S, Li N, Heald P, et al. Flow cytometric DNA ploidy analysis of peripheral blood from patients with Sezary syndrome: detection of aneuploid neoplastic T cells in the blood is associated with large cell transformation in tissue. Am J Clin Pathol 2004;122:774–782. Weisberger J, Cornfield D, Gorczyca W, et al. Down-regulation of pan-T-cell antigens, particularly CD7, in acute infectious mononucleosis. Am J Clin Pathol 2003;120:49–55. Wells DA, Benesch M, Loken MR, et al. Myeloid and monocytic dyspoiesis as determined by flow cytometric scoring in myelodysplastic syndrome correlates with the IPSS and with outcome after hematopoietic stem cell transplantation. Blood 2003;102: 394–403. Wells DA, Sale GE, Shulman HM, et al. Multidimensional flow cytometry of marrow can differentiate leukemic from normal lymphoblasts and myeloblasts after chemotherapy and bone marrow transplantation. Am J Clin Pathol 1998;110:84–94. Wiestner A, Rosenwald A, Barry TS, et al. ZAP-70 expression identifies a chronic lymphocytic leukemia subtype with unmutated immunoglobulin genes, inferior clinical outcome and distinct gene expression profile. Blood 2003;101:4944–4951. Wood BL. Flow cytometric diagnosis of myelodysplasia and myeloproliferative disorders. J Biol Regul Homeost Agents 2004;18:141–145. Wu JM, Borowitz MJ, Weir EG. The usefulness of CD71 expression by flow cytometry for differentiating indolent from aggressive CD10+ B-cell lymphomas. Am J Clin Pathol 2006;126:39–45. Yuan C, Douglas-Nikitin VK, Ahrens KP, et al. DRAQ5-based DNA content analysis of hematolymphoid cell subpopulations discriminated by surface antigens and light scatter properties. Cytometry Part B 2004;58B:47–52.

Appendix: Using the case studies CD-ROM

The second edition of Flow Cytometry in Hematopathology: A Visual Approach to Data Analysis and Interpretation is accompanied by a CD-ROM with 100 case studies. The case studies are designed for computers running Windows 32-bit operating systems (Windows 98 or newer). They require a minimum screen resolution of 1024 × 768 and have been optimized for that resolution with 16 million colors (24-bit). Your computer must be capable of displaying at least 65,000 colors (16-bit) at 1024 × 768 in order to view the photomicrographs properly. This edition of the case studies has been designed to run directly from the CD-ROM. No installation is necessary. However, if you choose to transfer the material from the CD-ROM to your hard disk drive, you must copy both folders, “Database” and “FCM Case Study Program” to the same directory with their contents intact for the program to run properly.

Using the case studies The case studies are very simple to use. You will find FCMCaseStudies.exe in the “FCM Case Studies Program” folder on the CD-ROM, and can use “Windows Explorer” or “My Computer” to start the program. Once the program (FCMCaseStudies.exe) is running, simply choose the case study that you wish to open using the drop-down list in the toolbar at the top of the screen. Clicking on the microscope icons in the toolbar allows you to display the photomicrographs associated with the case. The FCM graphics and explanations for each case are presented in a series of tabbed pages with up to six graphics per page. Simply click on the tab (FCM Graphics 1, FCM Graphics 2, etc.) to see the graphics on that page. Troubleshooting: With the latest Windows operating systems, setting the “Keep the taskbar on top of other windows” property can rob you of vertical space, causing the top of the program window to be partially cut off. If you experience this difficulty, simply uncheck “Keep the taskbar on top of other windows” and/or check “Auto-hide the taskbar,” which will restore the full vertical space to your screen.

Limited warranty and disclaimer Humana Press Inc. warrants the CD-ROM contained herein to be free of defects in materials and workmanship for a period of 30 days from the date of the book’s purchase. If within this 30-day period Humana Press receives written notification of defects in materials or workmanship, and such notification is determined by Humana Press to be valid, the defective disk will be replaced. In no event shall Humana Press or the contributors to this CD-ROM be liable for any damages whatsoever arising from the use or inability to use the files contained therein. The authors of this book have used their best efforts in preparing this material. Neither the authors nor the publisher make warranties of any kind, express or implied, with regard to these programs or the documentation contained within this CD-ROM, including, without limitation, warranties of merchantability or fitness for a particular purpose. No liability is accepted in any event, for any damages including incidental or consequential damages, lost profits, costs of lost data or program material, or otherwise in connection with or arising out of the furnishing, performance, or use of the files on this CD-ROM.

Index

Aberrant T-cell profile, 138–153, 245–246 Acute lymphoblastic leukemia/lymphoma (ALL/LL), 290 precursor B-ALL/LL, 35–36, 46, 67–68, 74–75, 81, 83–84, 86, 98–102, 221–222, 233, 290–291, 300 precursor T-ALL/LL, 63, 86, 103–105, 138, 221–222, 290 Acute myeloid leukemia (AML), 11, 224, 292–295 AML with maturation, 224, 228, 292, 294 AML with minimal differentiation/maturation, 76–77, 97, 292–293, 295 AML with monocytic differentiation (or, a monocytic component), 58, 67, 70, 77, 80, 81, 86–87, 89–92, 96, 223–224, 226, 294–295 AML-M2Eo, 224, 229, 292 AML-M3, 57, 60, 77–80, 82, 86, 88, 95, 97, 235, 269, 271–272, 292 AML-M3v, 77, 80–81, 86 AML-M4E, 224, 227, 292, 294 antigenic aberrancies in AML, 86, 96–97 Add-on testing, rationale, 34–36 bcl-2 protein, 34, 36, 185 cCD3 staining, 17, 34–36 cCD22 staining, 17, 34–35 cytoplasmic light chains, 34 MPO antibody, 34 TCR-Vβ repertoire, 35 TdT, 34, 35 Adult T-cell leukemia-lymphoma (ATLL), 151, 162, 321 Aggressive NK neoplasms, 150, 152, 164, 166, 169, 300, 321, 323 NK-like T-cell, 139, 151–152, 163–164, 321, 323, 326–329 true NK, 163–165, 323, 329 Analysis panels, FCM data display, 37 color dot plots, 37–41 FSC/fluorescence dot plots, 38–40, 50–54 Antibody panels, 22–31 disease oriented, 31–32 specimen type oriented, 32–34 BBS, 32–33 TF, 32–33 Anti-CD20 therapy, 3, 8, 27, 33, 67, 112, 131, 242, 245, 332 Antigenic patterns, 49–50, 175, 289 Artifacts (cytologic), 14–15, 52, 314–315 Auer rods, 290, 292 slow drying, 15, 52, 118, 295, 315 storage, 14, 302 Basophils, 230, 232, 265, 268, 276–278 Bcl-2 (FRFH vs. Follicular center cell lymphoma), 176, 185–187 Benign/reactive (“negative”) lymph node, 175–176 follicular hyperplasia (FRFH), 179–186 Biclonal/multiclonal acute leukemia, 285, 287, 300 Biphenotypic acute leukemia, 300–301 Blasts, 73–81 blast levels (hemodilution artifacts), 190–191 morphology, 289–290 Bone marrow, collection, 14–15 core biopsies, 14, 191 Bone marrow precursor B-cells, 73, 98–99, 192

B-prolymphocytic leukemia (B-PLL), 306–307, 321 Burkitt’s lymphoma, 13, 34, 48–49, 51, 116–117, 120, 300, 302 CD1 in Langerhan’s cell histiocytosis, 173 CD4/CD8 abnormalities, 138–146, 151–155, 157, 176 CD5, in: B-cell LPD/NHL, 33, 108–115, 123, 238, 242 benign, reactive B-cells, 176 CD10: bone marrow B-cell precursors, 9, 98–99 CD10/CD20 pairing, 27, 29, 37, 99, 179–186, 233 fluorochrome conjugation, 26–27 follicular center cell lymphoma, 112, 116, 302–306 granulocytes, 218, 231, 276 precursor B-ALL, 67, 74–75, 83, 98–99, 102, 222, 233–234, 290–291 reactive germinal center cells, 27 CD10/CD20 pattern, in: bone marrow B-cell precursors, 9, 98–99 FRFH (“hockey stick” pattern), 179, 184–185 CD11b/CD16: AML, 77, 292–293 mature granulocytes, 217–219, 221, 273 CD11c, wide dynamic range, 33 CD11c/CD20 pattern, in: CD11c/CD20 “trail” pattern (non-HCL B-cell LPD), 118, 122, 128 HCL, 122, 306 CD13/CD33 aberrant pattern, 293, 320, 325, 331 CD14 clones, 23, 86 upregulated (on granulocytes), 275–279 CD14/CD64 patterns, 86–90, 262 AML with monocytic differentiation, 86–90, 292, 294, 300 CMMoL, 86, 262 monocytes (normal), 89, 271 upregulated, 89, 271, 273–279 CD20: relationship of CD20 and CD19 intensities, 34, 40, 85, 108, 110, 112, 307 separating benign and malignant B-cells, 238 wide dynamic range, 23, 33, 71, 108 CD38, in: bone marrow B-cell precursors, 98, 101 plasma cells, 209, 212–213, 255–259, 318 CD45-negative, 20, 81, 83–84, 133, 171, 190, 214, 216–217, 235, 257, 261 CD56 expression: abnormal plasma cells, 33, 255–256, 318 AML, 11, 81–82, 86, 90, 92, 96, 292 monocytes, 80, 96 small cell carcinoma, 171 CD71, in: erythroid precursors, 196, 214, 216, 280 lymphoma grading, 106, 166 CD103, in: HCL, 33, 118, 122, 306, 315 T-cells, 176, 178, 323–325 CD117 expression, 172, 190, 220, 225, 255, 261–262, 265, 269 Cell size (FSC), 50–53 bimodal, 45, 51, 55, 133 variable, 51, 54 Cell viability, 14, 22

Chronic lymphocytic leukemia (CLL), 106–112 activated CLL, 110, 113–114, 306–307 with increased prolymphocytes (CLL/PL), 51, 55, 109, 111, 306 Chronic myelomonocytic leukemia (CMMoL), 87, 90, 92, 95, 175, 223–224, 260, 262–264, 273, 298, 300 Correlating FCM with other data, 6–7 Cytospin, 6, 16, 20, 34, 51, 82, 213 Differentiation and maturation (overview), 9–12 B-cell, 9, 11 erythroid, 10 myeloid lineage, 10 T-cells, 9–10, 11 DNA analysis: DNA index, 7, 42, 119 DNA histogram, 41–44, 121 FSC/DNA content dot plot, 41–42, 44, 121 dual parameter DNA-antigen analysis, 44 S-phase fraction, lymphoma grading, 7, 42–46, 48 TdT/DNA analysis, 35 DNA index: Prognostic significance, in ALL, 7, 42 Eosinophils, 187, 190–191, 216, 224, 227, 229, 251, 262, 265–267, 270, 292, 294 hypogranular, 265, 267 Erythroid hyperplasia, 214–215, 280, 282, 296 in AML, 224–225, 295 Fluorescence intensity determinations/reporting, 54–68 background fluorescence, autofluorescence, 57, 59–60 heterogeneous, 59, 63–64, 66–67 71–72, 75–78, 80, 98–103 bimodal, 64–70, 74 variable, 64–68 intensity of the expressed monoclonal light chain, 106–108 uniform/homogeneous, 64 Fluorescent beads, calibration, 19 Fluorochrome selection, 27, 30–31 FMC-7 antibody, 34 Follicular center cell lymphoma, 23, 27, 29–30, 43, 51, 53, 66, 72, 108–109, 112, 116–119, 149, 176, 179, 185–189, 302–306 FCC III lymphoma, S-phase fraction, 116 FSC, 50–53 bimodal, 45, 51, 55, 133 variable, 51, 54 Gamma-delta T-cell lymphoma-leukemia, 164 Gating strategies, 4–5, 16, 20, 22, 42, 71 G-CSF effect, 79, 217, 235, 260–261, 269, 271–273, 275, 277, 280, 282, 284 Granularity: hypergranularity, in AML-M3, 86, 234, 269, 272, 292 G-CSF, 217, 277 hypogranularity, 77, 216, 224, 265, 280 Hairy cell leukemia, 32–33, 117–118, 122, 306, 315 Hematogones, 26, 98–99, 192, 196, 237 High-grade B-cell lymphoma with plasma cell differentiation, 310, 318

344

INDEX

Histiocytes/macrophage proliferations, 172–173 HLA-DR, 3, 9, 11, 38–39, 60, 71, 77–78, 81, 86–88, 96, 138, 176, 178, 255, 260–261, 265, 269, 272–275, 280, 292 Hypereosinophilic syndrome, 265 Hypergammaglobulinemia, 128, 176, 201, 207–208, 257, 258, 260, 286, 316 monoclonal, 255, 315 polyclonal (in HIV), 207 Indolent NK proliferations, 248–249, 252–254, 256, 323 Infectious mononucleosis, 194, 206 Kappa and lambda antibodies: fluorochrome conjugation, 27–28 selection/titration, 23–26 Large B-cell lymphoma, 62, 140, 310–311, 313–314 CD5+, de novo, 310, 314 Large granular lymphocyte (LGL) proliferations, 204, 207–208, 233, 250, 323 List mode data, ungated, 4 doublet, 20 flow rate, 20 Lymphoplasmacytoid (LPC) disorders, 109, 112, 128– 129, 131, 242–243, 307, 312, 315–317 MALT, 112, 315, 318 Mantle cell lymphoma (MCL), 23–24, 49, 54, 112, 115, 123, 131, 136, 237, 242, 307, 309–310, 314–315 “blastic” variant, 290, 309–310 Marginal zone pattern (histologic), 306, 315, 318, 333 Mast cells, 265, 269 Megakaryocytic antigens, in AML, 295, 297 Megakaryocytic associated antigen, 295 Minimal residual disease (MRD), 1, 7, 21, 36, 187, 233–235, 242, 332 MRD in ALL, 233 MRD in AML, 36 Multiple myeloma, 175, 257, 260, 318 Mycosis fungoides, 56, 144, 151, 246, 321, 323, 329 Myelodysplastic syndromes (MDS), 38, 191, 220, 265, 273, 280, 292, 295, 297–298 high-grade, 175, 220–221, 223, 226, 230–231, 260, 265, 277, 285, 292, 295, 298 hypoplastic, 292, 297–298 low-grade, 189, 214, 260, 265, 270, 273, 277, 280, 281, 283

Myeloid maturation curves, abnormal, in: AML, 225, 230, 284 G-CSF, 271 impending (AML) relapse, 284 MDS/MPD, 226, 230, 273, 280, 297–298 Myeloproliferative disorders (MPD), 189, 220, 222– 223, 226, 230, 260, 265, 268–269, 273, 276, 285, 295, 297–298 Natural killer (NK) cells, 9, 12, 33, 198, 202, 204, 207– 211, 245, 252–253, 263 KIR antigen, 11–12, 204, 207 NK and NK-like T-cell malignancies, 150–152, 163– 165, 255, 260, 326–329 aggressive, 139, 150–152, 163–165, 171, 251, 323, 326–329 low-grade, 250, 252 Nondescript cytology (mononuclear cells), 290 Nondescript mature B-cell phenotype: large B-cell, 310–311, 313–314 small B-cell, 112, 314–315, 333 low level, in bone marrow, 318 Non-hematopoietic malignancies, 172 Paraimmunoblastic lymphoma, 307 Peripheral blood collection, 14 Peripheral T-cell lymphoma (PTCL), 21, 140, 142 ALK/CD20+ lymphoma, 54, 115, 140, 246 CD4+ T malignancies, 154–156, 246, 321–323 CD8+ T LPD/NHL, 323–324 Plasma cells: benign, 24, 209, 212–213 neoplastic, 112, 255, 257–258, 260, 318 Plasmacytoma, 318–320 Platelets, 10–11, 190, 217, 232 Post-thymic T-cell malignancies, identifying, 131 aberrant antigenic profile, 138, 141 CD4/CD8 abnormalities, 138–141 distinguishing from immature T-cell malignancies, 138 Propidium iodide, 16–17 Red cell lysis, 16, 213, 224, 280, 295 Reporting FCM data, 333 pitfalls, 2–4 Richter’s syndrome, 285, 307

Small lymphocytic lymphoma (SLL), 23–24, 49, 54, 65, 69, 106, 123, 131, 233, 238, 242, 306–307 Specimen handling: bone marrow, 14 mechanical dissociation, 14, 16 solid tissue, lymph node, 15–16, 21 SSC, 53–54 decreased, 53–54, 224, 230 increased, in AML-M3, 60 G-CSF, 217, 235, 269 SSC/CD45 patterns: AML with evidence of maturation, 77, 224 AML with minimal maturation, 77 AML with monocytic differentiation, 77, 81, 86–87, 262 AML-M3, 77–78, 82 AML-M3v, 77, 80 CD45-negative blasts, 81–83 HCL, 233 large cell LPD/NHL, 73, 86 MDS/MPD with increased blasts, 220, 223, 226–230 “negative” bone marrow, 187–190 “negative” peripheral blood, 187–190 Staining: DNA, 18 DNA dye DRAQ5, 17–19 intracellular antigens, 17 microtiter plate, 17, 22 surface antigens, 17–18 Surface light chain, evaluation: lack of surface light chain (in mature B-cells), 24 monoclonal light chain, intensity, 27, 64 polyclonal light chains, 24, 28–29, 180, 182–184, 212 T-cells, in normal blood, 192–193 TCR-Vβ repertoire, 193–194, 198, 200–204 abnormal distribution of Vβ antigens, 138, 147 oligoclonal expansions, 194, 255 TdT/CD19 pattern, 35, 86, 98 Thymocytes, 10, 99, 138, 141, 152, 221–222 T-prolymphocytic leukemia (T-PLL), 141, 143, 151–153, 321 Waldenstrom’s macroglobulinemia, 315

Sezary syndrome, 7, 138, 151, 233, 246, 250, 321 Small cell carcinoma, 171, 190

Zap-70, in CLL, 8, 114, 307, 309, 311

Plates

PLATES

a

b

c

d

e

f

Plate 1 Peripheral blood with chronic myelomonocytic leukemia (CMMoL). (a, b) The various cell clusters can be well appreciated on either the black-and-white (B&W) or color display. Pink, blasts (4%); green, monocytes; yellow, granulocytes; brown, lymphocytes. The HLA-DR+ population (d) is actually composed of the blast and monocytic populations (c). Lymphocytes, granulocytes and monocytes on the color display (e) appear as one continuous population on the B&W dot plot (f).

3

4

PLATES

a

b

c

d

Plate 2 Bone marrow with CMMoL. (a, b) Monocytes (red), granulocytes (blue) and lymphocytes (brown) can be well appreciated on either the color or B&W display. (c, d) The granulocytes, because of the bimodal reactivity for CD64, may be mistaken as two different populations on the B&W display.

a

b

Plate 3 Limitation due to color hierarchy. (a) Bone marrow with AML; in descending hierarchy, the colors are blue (blasts), green (granulocytes), red (monocytes) and brown (lymphocytes). (b) Very few of the red dots are seen through the green cluster.

PLATES

a

b

c

d

e

f

Plate 4 Precursor B-ALL (same patient as Figure 3.50). (a) Two subpopulations of blasts differing in cell size. (b–f) Blasts with lower FSC also display lower SSC. Their expression of CD45, CD10, CD19, CD34 and HLA-DR is less intense than (although overlapping with) that on the larger blasts.

5

6

PLATES

a

b

c

d

e

f

Plate 5 Parotid with low-grade B-cell NHL. (a, b) Mixture of T-cells (blue) and B-cells. (c) The larger (red) and smaller (green) B-cells both lack CD10. (d) The reagent combination kappa (1)–FITC and lambda (1)–PE reveals no overt abnormality. The kappa:lambda ratio is 2:1. (e, f) Correct results obtained with kappa (2) and lambda (2); the smaller B-cells are benign and the larger B-cells are malignant, monoclonal for kappa.

PLATES

a

b

c

d

e

f

Plate 6 Increased B-cell precursors (BCPs) in an 11-week-old infant with congenital thrombocytopenia. (a) Bone marrow with granulocytes (dark gray), lymphocytes (blue), erythroid cells (orange), few monocytes (green) and 25% BCPs (red). (b–f) Gated on mononuclear cells: BCPs coexpress bright CD10 and heterogeneous CD20. A small subset of BCPs is CD34+. Myeloblasts are less than 2%.

7

8

PLATES

a

b

c

d

e

f

Plate 7 Low-grade MDS. (a–f) Bone marrow with 2% myeloblasts (brown) coexpressing CD13, CD33 (not shown), CD34 and CD117. Erythroid precursors (orange) show normal coexpression of Gly-A and CD71. Myeloid precursors (pink) display several abnormalities including: (1) decreased SSC; (2) CD117 expression in a subset; and (3) altered CD11b/CD13 and CD16/CD13 maturation curves.

PLATES

a

b

c

d

e

f

Plate 8 AML with t(8;21)(q22;q22). (a–f) The leukemic blasts (red) in the bone marrow coexpress CD34, CD117, CD13, CD33, CD19 and CD56. CD15 is heterogeneous. Myeloid precursors (green) are hypergranular, with aberrant CD56 expression. A small subset displays abnormal retention of CD117.

9

10

PLATES

a

b

c

d

e

f

Plate 9 Peripheral blood with TCR-γδ T-cell lymphoma–leukemia (same patient as that in Figure 5.23). (a) A prominent cell cluster (pink) in the blast region with heterogeneous and downregulated CD45. (b–f) The leukemic cells are of medium cell size, coexpressing CD3, CD5, CD7 and TCR-γδ heterodimer. CD3 and CD5 are downregulated. CD2, CD4 and CD8 are absent. Residual T-cells (blue) are TCR-αβ+.

PLATES

11

Plate 10 Circulating blasts with scant cytoplasm in precursor B-ALL. The cytology (small cell size and condensed chromatin) in this case mimics that of a mature lymphoid malignancy.

Plate 13 AML with monocytic differentiation, showing a spectrum of blasts (arrow), promonocytes (thin arrow) and monocytes in the blood.

Plate 11 Myeloperoxidase cytochemistry (MPO cyto) in AML-M0. Blasts and a small number of hypogranular neutrophils are negative for MPO cyto. Staining is present in most neutrophils.

Plate 14 Activated CLL with increased prolymphocytes (arrow) and plasmacytoid lymphocytes (thin arrow), i.e., CLL/ PL by current morphologic criteria. The nucleolus is smaller and the chromatin is coarser in plasmacytoid lymphocytes than in prolymphocytes.

Plate 12 Agranular AML-M3v. The cytology (lack of granules, scant cytoplasm, coarse chromatin) in this case simulates that of circulating large lymphoma cells.

Plate 15 Florid reactive follicular hyperplasia.

12

PLATES

Plate 16 Hematogones (arrow) after bone marrow transplant for ALL.

Plate 19 Intense granulation in myeloid precursors secondary to G-CSF therapy.

Plate 17 EBV viral lymphadenitis. Sheets of transformed cells, mimicking the picture of large cell lymphoma.

Plate 20 T-ALL. Circulating blast with conspicuous azurophilic granules.

Plate 18 AML-M4E. Basophilic granules present in abnormal eosinophilic precursors (arrow).

Plate 21 Precursor B-cell ALL. The blasts in this case are cytologically indistinguishable from CLL cells.

PLATES

13

Plate 22 Nodal involvement by AML, morphologically simulating large cell lymphoma.

Plate 25 Blast with Auer rod (arrow) in AML-M1.

Plate 23 Lymphoblastic lymphoma (a) and “blastic” MCL (b), showing similarities in cell size, nuclear chromatin and mitotic activity. Increased apoptotic figures in (b).

Plate 26 AML-M2Eo. precursors.

Plate 24 Agranular M3v. Blasts with vacuolated cytoplasm and a few azurophilic granules, cytologically similar to blasts in AML with monocytic differentiation.

Plate 27 Blasts with vacuolated cytoplasm in AML with monocytic differentiation (see Figure 3.43), classified as AML-M5a.

Normal-appearing

eosinophilic

14

PLATES

Plate 28 AML-M6. Blasts (arrow) and abundant erythroid precursors. Hypogranular granulocytes and basophilic stippling present.

Plate 31 Follicular lymphoma with a so-called floral pattern at low-power magnification.

Plate 29 Hypoplastic MDS, morphologically similar to aplastic anemia. The combination of 7% blasts by FCM analysis and the cytogenetic abnormality of 7q− support the diagnosis of MDS.

Plate 32 Storage induced nuclear irregularities in CLL, simulating the blood picture of FCC or mantle cell lymphoma in leukemic phase.

Plate 30 Paratrabecular infiltration by PTCL.

Plate 33 Fresh blood smear from the same specimen as Plate 32. A few activated lymphoid cells with a small distinct nucleolus are present.

PLATES

15

Plate 34 (a) Hairy cell with abundant pale cytoplasm and bland-appearing nucleus. (b) In the thicker areas of the smear (which dry more slowly), the cell shrinks and appears smaller with more condensed chromatin.

Plate 37 Imprint from the same specimen as Plate 36. A mixture of small lymphocytes, plasmacytoid lymphocytes and prolymphocytes. The same picture in the blood is known as CLL/PL.

Plate 35 An occasional clefted lymphoid cell in a case of CLL/PL (under chemotherapy) resistant to multiple therapeutic regimens.

Plate 38 B-PLL (see Figure 3.78). To the untrained eye, the neoplastic cells may be confused with leukemic blasts.

Plate 36 SLL with increased and confluent pseudofollicles.

Plate 39 Nodal involvement by B-PLL from the same patient as Plate 38. H&E section and imprint showing sheets of transformed cells consistent with prolymphocytes.

16

PLATES

Plate 40 Mantle cell lymphoma. Nodular pattern with small residual germinal centers present.

Plate 41 MCL in leukemic phase (see Figure 3.64). The neoplastic cells in this case are bland appearing with round nuclei, similar to CLL cells.

Plate 42 CD5+ large cell lymphoma. The tumor cells display a plasmablastic appearance.

Plate 43 LPC lymphoma–leukemia. The cytology is better appreciated on the blood film (a) than on the bone marrow smear (b), where slow drying has resulted in hairs, blebs and a smaller cell size.

Plate 44 LPC lymphoma–leukemia. Optimal section (a); suboptimal section (b) with clearing artifacts often misinterpreted as clear pale cytoplasm.

Plate 45 MBCL. Interfollicular infiltration by large irregular clusters of pale cells.

PLATES

17

Plate 46 Imprint from the same specimen as Plate 45. Neoplastic cells showing a hairy cell-like cytology (abundant pale cytoplasm, bland reticulated chromatin). Neutrophils are commonly associated with monocytoid B-cells.

Plate 49 ATLL. Leukocytosis composed of lymphoid cells with marked nuclear irregularities.

Plate 47 Marginal zone pattern in SLL. The pale rings around residual reactive follicles are composed of loosely packed prolymphocytes and plasmacytoid lymphocytes. Occasional typical pseudofollicles (arrow) are also present.

Plate 50 T-PLL. The tumor cells display a small distinct nucleolus and are smaller than those in B-PLL.

Plate 48 Plasma cell leukemia. Predominantly immatureappearing plasma cells.

Plate 51 Advanced stage of Sézary syndrome, with abundant circulating tumor cells. The blood picture closely mimics ATLL. The patient has had a splenectomy.

18

PLATES

Plate 52 Earlier skin biopsy on the same patient as Plate 51. The epidermis is infiltrated by collections of neoplastic lymphoid cells (Pautrier abscess).

Plate 55 H&E section from the same specimen as Plate 54. Neoplastic cells with ample pale cytoplasm.

Plate 53 Indolent LGL leukemia (NK-like T-cell phenotype) with bland-appearing neoplastic large granular lymphocytes.

Plate 56 Bone marrow involved by γδ-T-cell lymphoma–leukemia. The cytology mimics that of AML with monocytic differentiation.

Plate 54 Aggressive NK lymphoma. Neoplastic cells with abundant azurophilic granules.

Plate 57 CD30+ lymphoma. The tumor population is relatively monotonous. ALK-1 is expressed.