Flow cytometric analysis of lymphomas and lymphoproliferative disorders

A cell-level discriminative neural network model for diagnosis of blood cancers

MOTIVATION

Precise identification of cancer cells in patient samples is essential for accurate diagnosis and clinical monitoring but has been a significant challenge in machine learning approaches for cancer precision medicine. In most scenarios, training data are only available with disease annotation at the subject or sample level. Traditional approaches separate the classification process into multiple steps that are optimized independently. Recent methods either focus on predicting sample-level diagnosis without identifying individual pathologic cells or are less effective for identifying heterogeneous cancer cell phenotypes.

RESULTS

We developed a generalized end-to-end differentiable model, the Cell Scoring Neural Network (CSNN), which takes sample-level training data and predicts the diagnosis of the testing samples and the identity of the diagnostic cells in the sample, simultaneously. The cell-level density differences between samples are linked to the sample diagnosis, which allows the probabilities of individual cells being diagnostic to be calculated using backpropagation. We applied CSNN to two independent clinical flow cytometry datasets for leukemia diagnosis. In both qualitative and quantitative assessments, CSNN outperformed preexisting neural network modeling approaches for both cancer diagnosis and cell-level classification. Post hoc decision trees and 2D dot plots were generated for interpretation of the identified cancer cells, showing that the identified cell phenotypes match the cancer endotypes observed clinically in patient cohorts. Independent data clustering analysis confirmed the identified cancer cell populations.

AVAILABILITY AND IMPLEMENTATION

The source code of CSNN and datasets used in the experiments are publicly available on GitHub (http://github.com/erobl/csnn). Raw FCS files can be downloaded from FlowRepository (ID: FR-FCM-Z6YK).

A cell-level discriminative neural network model for diagnosis of blood cancers

Motivation

Precise identification of cancer cells in patient samples is essential for accurate diagnosis and clinical monitoring but has been a significant challenge in machine learning approaches for cancer precision medicine. In most scenarios, training data are only available with disease annotation at the subject or sample level. Traditional approaches separate the classification process into multiple steps that are optimized independently. Recent methods either focus on predicting sample-level diagnosis without identifying individual pathologic cells or are less effective for identifying heterogeneous cancer cell phenotypes.

Results

We developed a generalized end-to-end differentiable model, the Cell Scoring Neural Network (CSNN), which takes the available sample-level training data and predicts both the diagnosis of the testing samples and the identity of the diagnostic cells in the sample, simultaneously. The cell-level density differences between samples are linked to the sample diagnosis, which allows the probabilities of individual cells being diagnostic to be calculated using backpropagation. We applied CSNN to two independent clinical flow cytometry datasets for leukemia diagnosis. In both qualitative and quantitative assessments, CSNN outperformed preexisting neural network modeling approaches for both cancer diagnosis and cell-level classification. Post hoc decision trees and 2D dot plots were generated for interpretation of the identified cancer cells, showing that the identified cell phenotypes match the cancer endotypes observed clinically in patient cohorts. Independent data clustering analysis confirmed the identified cancer cell populations.

Availability

The source code of CSNN and datasets used in the experiments are publicly available on GitHub and FlowRepository.

Contact

Edgar E. Robles: roblesee@uci.edu and Yu Qian: mqian@jcvi.org.

Supplementary information

Supplementary data are available on GitHub and at Bioinformatics online.

Core Needle Biopsy in Lymphoma Diagnosis: The Diagnostic Performance and the Role of the Multidisciplinary Approach in the Optimization of Results

To compare the diagnostic accuracy of core needle biopsies (CNBs) and surgical excisional biopsies (SEBs), samples of lymphoid proliferation from a single institution from 2013 to 2017 (N=476) were divided into groups of CNB (N=218) and SEB (N=258). The diagnostic accuracy of these samples was evaluated as a percentage of conclusive diagnosis, according to the World Health Organization Classification of Tumours of Haematopoietic and Lymphoid Tissues . The contribution of clinical data, the assessment of sample adequacy by a pathologist during the procedure, the number and size of fragments, the needle gauge, the ancillary tests, and the type of lymphoid proliferation were also examined. The diagnostic accuracy of SEB was 97.3% and CNB 91.3% ( P =0.010). Additional factors considered essential for establishing the final diagnosis in some cases were: clinical information (20.6% CNB, 7.4% SEB; P <0.001); immunohistochemistry (96.3% CNB, 91.5% SEB; P =0.024); flow cytometry (12% CNB, 6.8% SEB; P =0.165); and other complementary tests (8.2% CNB, 17.3% SEB; P =0.058). Factors that did not influence performance were the evaluation of sample adequacy during the procedure, the number and size of fragments, and the needle gauge. Increased percentage of nondiagnostic CNB was observed in T-cell lymphomas (30%), followed by classic Hodgkin lymphoma (10.6%). The main limitation of CNB was the evaluation of morphologically heterogenous diseases. CNB is useful and safe in lymphoma diagnosis provided it is carried out by a team of experienced professionals. Having an interventional radiology team engaged with pathology is an essential component to achieve adequate rates of specific diagnoses in CNB specimens.

Value of cerebrospinal fluid white cell count and protein level in predicting leptomeningeal involvement by systemic aggressive B-cell lymphoma

INTRODUCTION

Diagnostic cerebrospinal fluid (CSF) analysis for patients with newly diagnosed aggressive B-cell lymphoma at risk of secondary central nervous system involvement typically includes multiparametric flow cytometry (MFC), cytology (CC), white cell count (WCC) and total protein. The strength of relationships between MFC results and the remaining variables has been disputed in small studies. We explored these relationships in a large homogeneous cohort of patient samples, aiming to establish the relationship between WCC and protein level and MFC results.

METHODS

Adult patients with aggressive B-cell lymphoma at risk of CNS involvement who underwent staging CSF analysis by MFC were identified retrospectively from institutional electronic records between October 2011 and December 2020.

RESULTS

Three hundred and seventy eight samples, including 45 (11.9%) MFC+ samples, were analysed. The relative sensitivity of CC for MFC positivity was 0.38, with PPV of 0.68. Significantly higher median WCC (p < .001) and protein levels (p = .011) were seen in MFC+ vs. MFC- samples. MFC + CC+ (vs. MFC + CC- samples) demonstrated higher median neoplastic events and neoplastic cell concentration. WCC ≥36 × 10⁶ /L and protein ≥1.12 g/L cut-off values demonstrated the highest PPVs for MFC positivity (0.67 and 0.88, respectively).

CONCLUSIONS

Statistically significant associations exist between elevated WCC and protein and MFC positivity, and selected WCC and protein cut-off values have PPVs comparable to that of cytological assessment. Whilst routine WCC and protein analysis may be unnecessary, WCC/protein values above these levels could be regarded as reasonable evidence of CSF involvement in the appropriate setting.

Value of multi-parameter flow cytometry immunophenotyping in T/NK-cell neoplasms in cytology specimens: A retrospective study in Chinese patients

BACKGROUND

Innate limitations of morphological diagnosis of T/NK-cell neoplasms mean that they can be misdiagnosed or missed, especially when mixed with a variety of benign and reactive conditions. The aim of this study was to investigate the application value of multiparameter flow cytometry immunophenotyping (MFCI) in screening and diagnosing T/NK-cell neoplasms with cytology specimens.

MATERIAL AND METHODS

The clinical and pathological characteristics of 1028 newly diagnosed cases from Fudan University Shanghai Cancer Center who provided a cytology specimen between June 2010 and January 2016 with correlated histology diagnosis and clinical confirmation were retrospectively reviewed. MFCI was used for screening, diagnosis and typing. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) in diagnosis of T/NK-cell neoplasms were calculated.

RESULTS

There were 606 males and 422 females in 1028cases, with a mean age of 47.5 years (range 9-86 years). Specimens used for cytologic diagnosis included 996 FNAs, 2 US-FNAs, 13 EUS-FNAs and 17 effusions. Screening for types of lymphoma of MFCI, 139 (13.52 %) cases were T/NK cell lymphoma, 3 (0.29 %) cases were B cell lymphoma T-NHL and B-NHL coexist. A total of 146 suspected T/NK-cell neoplasms were screened out (sensitivity = 94.64 %, specificity = 95.63 % PPV = 72.60 %, NPV = 99.32 %) by MFCI, with 112 (76.71 %) histologically confirmed cases and 6 (4.11 %) false-negative cases identified (3 cases diagnosed as B-cell neoplasms and 1 case as T-cell neoplasm with B-cell neoplasm, which also were confirmed by gene rearrangement. 2 cases were suspicious T-cell-immunophenotypic abnormalities). When used at the diagnostic level, a total of 88 T/NK-cell neoplasms were identified (sensitivity = 68.75 %, specificity = 98.80 %, PPV = 87.50 %, NPV = 96.28 %) with 11 false-positive cases recognized, 9 of which showed typical immunophenotypic T-cell neoplasms features, and 2 exhibited aberrant T immunophenotype.

CONCLUSIONS

MFCI has high sensitivity and specificity in the screening and diagnosis of T/NK-cell neoplasms and may be useful as an alternative diagnosis method in cytology specimens.

Role of Epstein-Barr virus status and immunophenotypic studies in the evaluation of exfoliative cytology specimens from patients with post-transplant lymphoproliferative disorders

BACKGROUND

Post-transplant lymphoproliferative disorders (PTLDs) are well characterized in tissue sections, but their evaluation in exfoliative cytology specimens is limited. This study reports a 25-year experience with PTLDs in exfoliative cytology specimens.

METHODS

All solid organ or allogeneic stem cell transplant recipients with PTLDs and exfoliative cytology specimens from 1987 to 2011 were identified. The cytomorphology, Epstein-Barr virus (EBV) status, flow cytometry, immunohistochemistry, and molecular studies were reviewed from all exfoliative cytology specimens previously diagnosed as atypical lymphoid proliferations or PTLDs.

RESULTS

A total of 55 patients (age range, 1-72 years) with PTLDs had 434 exfoliative cytology specimens. Thirty-six of the 55 patients (65%) had 54 specimens with abnormal lymphoid proliferations (12% of the specimens), and 26 of these patients had 37 specimens available for review (15 cerebrospinal fluid specimens, 12 peritoneal fluid specimens, 9 pleural fluid specimens, and 1 bronchoalveolar lavage fluid specimen). Thirty percent of the reviewed cytology specimens were diagnostic of PTLDs, including 8 cases of monomorphic post-transplant lymphoproliferative disorder (M-PTLD) with abnormal B/T-cell populations identified with flow cytometry/immunohistochemistry and 3 EBV-positive specimens with a differential diagnosis of polymorphic PTLD versus M-PTLD. All cases diagnostic of a PTLD had 1 to 3 ancillary studies performed. Forty percent of the cytology specimens (15 of 37) were suspicious for a PTLD, but ancillary studies were performed for only a third of them, and they did not support a definitive diagnosis of a PTLD. Thirty percent of the cytology specimens (11 of 37) appeared reactive, but they lacked sufficient ancillary studies to exclude a PTLD.

CONCLUSIONS

Atypical lymphoid proliferations are common in exfoliative cytology specimens from patients with PTLDs, and they require ancillary studies at least including immunophenotyping and EBV evaluations for a definitive diagnosis. Cancer Cytopathol 2016;124:425-35. © 2016 American Cancer Society.

In vitro and in vivo studies of the immunomodulatory effect of Echinacea purpurea on dendritic cells

BACKGROUND

Extracts of Echinacea have been used traditionally for the treatment of diverse types of infections and wounds. They have become very familiar immunostimulant herbal medicine. However, the specific immunomodulatory effect of Echinacea remains to be elucidated.

AIM

In our study, the effect of Echinacea purpurea extract on the generation of immature DCs from monocytes was described, as well as its effect on DC differentiation. In addition, an in vivo experiment was conducted to investigate whether treatment of mice with extracts derived from E. purpurea has immunomodulatory effect on murine splenic DCs.

METHODS

Immature DCs were generated by incubating peripheral blood monocytes with cytokine cocktail (GM-CSF + IL-4) and matured by tumor necrosis factor-α (TNF-α). The cells were randomized to 5 groups to investigate E. purpurea effect in different stages. Phenotypic analysis of cell marker CD83-expressed on DCs was performed by flow cytometry. Mice were randomly divided into 3 groups; control, E. purpurea treated and E. purpurea-TNF-α treated group. The murine splenic DCs were isolated and phenotyped for CD83 and CD11c by flow cytometry.

RESULTS

Treatment of monocytes with E. purpurea prior to addition of the maturation factor TNF-α resulted in a significant decrease in the yield of DC expressing CD83. On the other hand, immature DCs generated in the culture in the presence of GM-CSF and IL-4, when treated simultaneously with E. purpurea and TNF-α, exhibited an insignificant change in the yield of CD83-expressing DCs compared with untreated control. The in vivo experiments showed that splenic DCs obtained from mice treated with E. purpurea with or without TNF-α did not exhibit significant changes in CD83 or CD11c compared with those obtained from control mice.

CONCLUSION

Our findings suggest that the immunomodulatory mechanisms of E. purpurea impact generation fate of DCs rather than differentiation stages. The results obtained in the in vivo study utilizing murine splenic DCs supported those observed in vitro.

Determination of the minimum inhibitory concentration of Cryptococcus neoformans and Cryptococcus gattii against fluconazole by flow cytometry

Recent studies have used flow cytometry (FCM) as an important alternative method to determine the antifungal susceptibility of yeasts compared to the broth microdilution Clinical and Laboratory Standards Institute (CLSI) reference procedure. We present a comparative study of the broth microdilution method and flow cytometry to assess the in vitro antifungal susceptibility of Cryptococcus neoformans (n = 16) and C. gattii (n = 24) to fluconazole. The minimum inhibitory concentration (MIC) assays by flow cytometry were defined as the lowest drug concentration that showed ∼50% of the count of acridine orange negative cells compared to that of the growth control. Categorical classification showed all C. neoformans isolates were susceptible to fluconazole. Three isolates of C. gattii were susceptible dose-dependent and the remaining 21 isolates were classified as susceptible. MICs comparison of both methodologies demonstrated 100% categorical agreement of the results obtained for C. neoformans and C. gattii. The MICs obtained with the CLSI-approved method and flow cytometry were compared by the Spearman correlation test and a significant Pv = 0.001. The flow cytometric method has the advantage of analyzing a large and constant number of cells in less time, i.e., 9 h incubation for fluconazole using acridine orange versus 72 h for broth microdilution method. In conclusion, the two methods were comparable and flow cytometry method can expedite and improve the results of in vitro susceptibility tests of C. neoformans and C. gattii against fluconazole and also allows comparative studies in vitro/in vivo more rapidly, which along with clinical data, could assist in selecting the most appropriate treatment choice.

Hepatosplenic gamma-delta T-cell lymphoma in a female patient after delivery

Hepatosplenic γδ T-cell lymphoma (HSTCL) is a very rare peripheral T-cell lymphoma characterized by extranodal infiltration of mature malignant post-thymic T-lymphocytes into sinusoids of the liver and spleen without lymphadenopathy and significant cytopenias. The aetiology of the disease is unknown. We describe the case of a female patient in whom HSTCL developed after delivery and who was previously without disease. Flow cytometry and liver puncture are essential for diagnosing HSTCL, especially in patients with unexplained pancytopenia and hepatosplenomegaly. Since phenotypic results can easily be misinterpreted as non-malignant, the examiner should have enough experience to recognize clonal changes of T-lymphocytes. Namely, in contrast to B-lymphocytes, T-lymphocytes do not have an efficient indicator of clonality and are recognized by flow cytometry based only on aberrant expression of commonly present antigens of T-cell and NK-cell subsets. At present, there is no known cure for HSTCL with a maximum survival up to 2 years.

Immunocytochemical diagnosis as inflammation by vitrectomy cell blocks in patients with vitreous opacity

PURPOSE

To describe the clinical and cytopathologic characteristics in patients with vitreous opacity of unknown cause or preceding inflammation, diagnosed cytopathologically as inflammation.

DESIGN

Retrospective case series.

PARTICIPANTS

Forty-three consecutive patients (61 eyes) who underwent vitrectomy for vitreous opacity of unknown cause or preceding inflammation and were diagnosed cytopathologically with inflammation at one institution in 6 years from 2005 to 2010. During the same period, 11 consecutive patients with vitreous opacity of unknown cause were diagnosed cytopathologically with lymphoma (large B-cell lymphoma) and were excluded from the study.

METHODS

Cell blocks were made by centrifugation of vitrectomy fluid and embedded in paraffin for immunocytochemistry.

MAIN OUTCOME MEASURES

Cytopathologic and immunocytochemical diagnosis using vitrectomy cell blocks.

RESULTS

Histiocytes (macrophages), small lymphocytes, neutrophils, and eosinophils were predominant cells, with no atypical large cells on hematoxylin-eosin staining. Immunocytochemically, most predominant cells were CD68-positive histiocytes (macrophages), followed by CD3-positive T cells, but CD20- or CD79a-positive B cells were rarely present. Epithelioid cells, positive for CD68, were found in 4 patients with or without an established diagnosis of sarcoidosis, and giant multinucleated cells were found in 2 patients with suspected preceding self-limiting Vogt-Koyanagi-Harada disease, based on the presence of depigmented red fundi. Inflammation was diagnosed in 2 patients with vitreous opacity who had a preceding onset of brain lymphoma or systemic lymphoma.

CONCLUSIONS

The presence of macrophages, combined with small T lymphocytes, was a major sign in intravitreal inflammation, manifesting as vitreous opacity. A simple technique of cytopathology and immunocytochemistry, using vitrectomy cell blocks, can be performed in most pathology laboratories.

Immunophenotypic pattern of myeloid populations by flow cytometry analysis

We present our experience with immunophenotypic characteristics of benign and malignant myeloid populations, with emphasis on differential diagnosis especially between eosinophils, dysplastic granulocytes, neoplastic promyelocytes, and monocytes. Eosinophils are characterized by bright CD45, high side scatter (SSC), very low forward scatter (FSC), positive CD11b, CD11c, CD13, CD15, and CD33. They are negative for CD10, CD14, CD16, CD56, CD64, and HLA-DR. Mature monocytes are positive for CD11b, CD11c, CD13, CD14, CD33, and CD64, and may express CD2 and CD4. Blasts in acute myeloid leukemias (AML) with minimal differentiation have low SSC and moderate CD45 expression and are positive for CD34, CD117, CD13, HLA-DR, and CD33 and may be positive for TdT, CD4, and CD11c. In acute promyelocytic leukemia (APL), four FC patterns can be recognized. The majority of cases represented classical (hypergranular) APL and were characterized by high SSC, positive CD117, usually negative CD34, heterogeneous CD13, and bright CD33 (pattern 1). The second most common type, corresponding to hypogranular (microgranular) variant of APL, differed from classical APL by low SSC and frequent coexpression of CD2 and CD34 (pattern 2). Rare cases of APL (pattern3) showed mixture of neoplastic cells (SSC(low)/CD2(+)/CD13(+)/CD33(+)/CD34(+)/CD117(+)) and prominent population of benign granulocytes/maturing myeloid precursors (SSC(high)/CD10(+/-)/CD16(+/()/CD117(()). One case showed two APL populations, one with hypogranular and one with hypergranular characteristics (pattern 4). Detailed phenotypic characteristics of neoplastic monocytes and dysplastic granulocytes with their differential diagnosis are also presented.

Simultaneous cytomorphologic and multiparametric flow cytometric analysis on lymph node samples is faster than and as valid as histopathologic study to diagnose most non-Hodgkin lymphomas

We evaluated the validity and accuracy of cytomorphology and multiparametric flow cytometry (C-FCM) in diagnosing oncohematologic disease in 223 consecutive lymph node biopsy specimens from patients with lymphadenopathy, from 2004 to 2007. C-FCM and histopathologic studies were interpreted independently by hematologists and pathologists, respectively. C-FCM detected neoplastic disorders in 133 samples (59.6%): 92 non-Hodgkin lymphomas (NHLs; 41.3%), 21 Hodgkin lymphomas (HLs; 9.4%), 19 malignant nonhematologic neoplasms (8.5%), and 1 multiple myeloma (0.4%). Sensitivity and specificity were 87.25% and 95.95%, respectively. Positive predictive value and negative predictive value (NPV) were 97.74% and 78.89%, respectively. Sensitivity and NPV were 94.79% and 96.81% upon excluding HL and malignant nonhematologic neoplasms from the analysis. Of the 92 NHLs, 89 (97%) were categorized according to the 2001 World Health Organization classification of hematolymphoid neoplasms with a concordance of 87%. The C-FCM study was significantly faster than the histopathologic study. C-FCM has high sensitivity and specificity, allowing for a valid and reliable diagnosis, especially in NHLs and enabling their subclassification. C-FCM is faster than the histopathologic examination, allowing for therapeutic decisions to be made quickly. However, in the samples in which C-FCM cannot establish a diagnosis, histopathologic results are needed.

Bone marrow aspirate and biopsy: a pathologist's perspective. II. interpretation of the bone marrow aspirate and biopsy

Bone marrow examination has become increasingly important for the diagnosis and treatment of hematologic and other illnesses. Morphologic evaluation of the bone marrow aspirate and biopsy has recently been supplemented by increasingly sophisticated ancillary assays, including immunocytochemistry, cytogenetic analysis, flow cytometry, and molecular assays. With our rapidly expanding knowledge of the clinical and biologic diversity of leukemia and other hematologic neoplasms, and an increasing variety of therapeutic options, the bone marrow examination has became more critical for therapeutic monitoring and planning optimal therapy. Sensitive molecular techniques, in vitro drug sensitivity testing, and a number of other special assays are available to provide valuable data to assist these endeavors. Fortunately, improvements in bone marrow aspirate and needle technology has made the procurement of adequate specimens more reliable and efficient, while the use of conscious sedation has improved patient comfort. The procurement of bone marrow specimens was reviewed in the first part of this series. This paper specifically addresses the diagnostic interpretation of bone marrow specimens and the use of ancillary techniques.

Generation of flow cytometry data files with a potentially infinite number of dimensions

Immunophenotypic characterization of B-cell chronic lymphoproliferative disorders (B-CLPD) is associated with the use of increasingly larger panels of multiple combinations of 3 to > or =6 monoclonal antibodies (Mab), data analysis being separately performed for each of the different stained sample aliquots. Here, we describe and validate an automated method for calculation of flow cytometric data from several multicolor stainings of the same cell sample--i.e., the merging of data from different aliquots stained with partially overlapping combinations of Mab reagents (focusing on > or =1 cell populations)--into one data file as if it concerned a single "super" multicolor staining. Evaluation of the performance of the method described was done in a group of 60 B-CLPD studied at diagnosis with 18 different reagents in a panel containing six different 3- and 4-color stainings, which systematically contained CD19 for the identification of B-cells. Our results show a high degree of correlation and agreement between originally measured and calculated data about cell surface stainings, providing a basis for the use of this approach for the generation of flow cytometric data files containing information about a virtually infinite number of stainings for each individual cellular event measured in a sample, using a limited number of fluorochrome stainings.

Predictive value of blood and bone marrow flow cytometry in B-cell lymphoma classification: comparative analysis of flow cytometry and tissue biopsy in 252 patients

OBJECTIVE

To study the effectiveness of peripheral blood (PB) and bone marrow flow cytometric immunophenotyping (FCIP) in predicting the histologic B-cell lymphoma type.

PATIENTS AND METHODS

We studied the FCIP results and tissue histopathology from 252 patients with B-cell lymphoma seen at Mayo Clinic's site in Rochester, MN, between January 1, 1997, and January 1, 2004, who had positive results on PB, bone marrow, or body fluid FCIP and a corresponding diagnostic tissue biopsy specimen.

RESULTS

Most of the B-cell lymphomas studied were low grade, with chronic lymphocytic leukemia (CLL)/small lymphocytic lymphoma being most common. Flow cytometric immunophenotyping histogram analysis was more informative than tabulated percentage antigen positivity; surface immunoglobulin and CD20 staining intensity, CD5 and CD23 positivity, CD10 positivity, and the coexpression of CD11c/CD22 and CD103 were the most pertinent markers. Using these FCIP parameters and strict immunophenotypic definitions for CLL, mantle cell lymphoma (MCL), and hairy cell leukemia, we obtained greater than 95% specificity for each diagnosis. However, we encountered the following exceptions to standard paradigms of B-cell lymphoma-associated FCIP: (1) CD5 expression by disorders distinct from CLL and MCL, (2) lack of uniform CD5 positivity in some CLL and MCL cases, (3) absence of CD10 in approximately 50% of follicular lymphomas, and (4) expression of CD103 by occasional marginal zone lymphomas.

CONCLUSION

Stringent interpretation of PB and bone marrow FCIP results enables identification of certain B-cell lymphoma types. However, the observed exceptions to accepted immunophenotypic paradigms highlight the occasional phenotypic overlap among diseases and emphasize that a systematic approach to FCIP interpretations is key to providing clinically useful diagnostic information.

A single unit lymphoma experience: outcome in a Cape Town academic centre

To document outcome in Hodgkin and other lymphomas from a privately based academic centre the clinical records from 253 consecutive referrals were analysed. Diagnosis was according to World Health Organization criteria, prognosis assigned by the international index and therapy risk-stratified with results subject to appropriate statistical methodology. None of these patients underwent transplantation. For the cohort the median age was 55 years (range 11-94) and 63% were male. Constitutional symptoms were present in 22%; a quarter had previous chemotherapy and a third some form of irradiation prior to referral. Fifty-seven percent were stage I or II and 21% had nodal disease above and below the diaphragm whilst in the remainder cells were present in the circulation and this included the subset of chronic lymphocytic leukaemia -- small lymphocytic lymphoma. Positron emission scanning was not available for these studies. Median survival for the cohort is 3.2 years and reduced to 1.3 years by the presence of unexplained fever, sweating or inappropriate weight loss. Further adverse factors included any prior treatment, intermediate or high-grade histopathology, risk factors defined by the International Prognostic Index as well as late Rai stages. Analysed by disease category Hodgkin lymphoma (n=17) when managed according to the German Study Group protocols and hairy cell leukaemia (n=10) treated with two chlorodeoxyadenosine -- both had a stable plateau in excess of 90%. The corresponding figures for follicular variants (n=31) was 72% in the low risk and 58% in the remainder when treated with cyclophosphamide, vincristine and prednisone. Curves for the aggressive or diffuse large B-cell lymphoma (n=44) fell initially to 48%, but relapse continued in stages III and IV to the current level of 18% when receiving cyclophosphamide, hydroxydaunorubicin, vincristine and prednisone on the 21-day schedule. Chronic lymphocytic leukaemia -- small lymphocytic lymphoma (n=58) were initially given pulsed chlorambucil and sustained response was over 90% with low bulk, but declined to reach 30% as prognostic score rose. The miscellaneous categories (n<5 each) managed variably, but using the same criteria, were pooled and are presently at 62% and 30% for high and low grades. It is concluded that precise diagnosis, accurate staging and therapy on standardised risk-stratified programmes, delivered uniformly by a single multidisciplinary group, creates the all-important centre effect; matching figures are unlikely to apply outside these disciplined circumstances. The expectation from patients and referring physicians alike is that, since lymphomas are potentially curable, such an approach to comprehensive management will be regarded as standard even in an under resourced or Third World country. It follows that late referral and prior therapy will adversely affect performance status and compromise life span: These alternative approaches are inappropriate and strongly discouraged.

9-color and 10-color flow cytometry in the clinical laboratory

CONTEXT

The development of commercial flow cytometers capable of detecting more than 10 simultaneous fluorescent signals presents opportunities for improved diagnosis and monitoring of patients with leukemia and lymphoma.

OBJECTIVE

To describe instrument and reagent characteristics necessary for successful 9-color and 10-color flow cytometry in a clinical setting.

DESIGN

Systematic review of issues related to instrument settings, reagent performance, and general principles of panel construction.

RESULTS

Nine-color and 10-color flow cytometry offers the possibility for increased accuracy in population identification, the ability to obtain detailed information from paucicellular specimens, improved laboratory efficiency, and the means to consistently detect abnormal populations at low levels. Careful attention to details of instrument and reagent performance allows for the development of panels suitable for screening of samples for leukemia and lymphoma in a clinical setting.

CONCLUSIONS

The characteristics of this technique are particularly well suited to the analysis of leukemia and lymphoma and have the potential to revolutionize and standardize this type of analysis in the clinical laboratory.

Ultrasound-assisted transthoracic biopsy: Cells or sections?

Physicians increasingly use transthoracic ultrasound (US) as an aid for diagnostic procedures. At the bedside, US helps to visualize neoplasms in the chest wall, pleura, peripheral lung, and anterior mediastinum involving or abutting the pleura. Histology specimens from cutting-needle biopsies have been shown to be safe and effective. This prospective study determined the yield and safety of US-guided fine-needle aspiration biopsy (FNAB) as a first-line bedside investigation. We recruited 97 consecutive patients, and of these, 85 underwent both cutting-needle biopsy and FNAB. These were adequate for diagnosis in 81.2% and 80% of cases, respectively, with a combined yield of 90%. Measured with a novel semiquantitative score, FNAB allowed a diagnosis with fewer special investigations than cutting biopsy. US-guided FNAB by pulmonologists performed best in lung carcinoma and can be recommended as a first-line investigation in patients with a high clinical suspicion of this diagnosis.

High incidence of occult leptomeningeal disease detected by flow cytometry in newly diagnosed aggressive B-cell lymphomas at risk for central nervous system involvement: the role of flow cytometry versus cytology

We assessed the cerebrospinal fluid (CSF) by flow cytometry and cytology in 51 newly diagnosed and 9 treated aggressive B-cell lymphomas at risk for central nervous system (CNS) involvement to examine the utility of flow cytometry, incidence of CSF disease, and clinical surrogates of CNS spread. Multicolor flow cytometry using multiple antibody panels for light chains and B- and T-cell antigens identified neoplastic clones that constituted as little as 0.2% of total CSF lymphocytes. Among 51 newly diagnosed patients, 11 (22%) had occult CSF involvement. All 11 were detected by flow cytometry but only 1 by cytology (P = .002). Among 9 treated patients, CSF involvement was detected by flow cytometry alone in 2 and also by cytology in 1 case. CSF chemistry and cell counts were similar in patients with and without CSF lymphoma. Only the number of extranodal sites was associated with occult CSF lymphoma in newly diagnosed patients by univariate (P = .006) or logistic regression analysis (P = .012). We hypothesize that the biologic phenotype associated with colonization of extranodal sites leads to CNS spread, possibly related to the microenvironment. Patients at risk for CNS spread should undergo staging CSF evaluation by flow cytometry.

Flow cytometry in the bone marrow staging of mature B-cell neoplasms

BACKGROUND

Even though flow cytometric (FC) analysis of bone marrow aspirates is often performed in hematolymphoid disorders at diagnosis and during disease monitoring, its role has not been defined during the staging of B-non-Hodgkin's lymphoma (B-NHL) and B-cell lymphoproliferative diseases. The goal of this study was to provide an objective evaluation of how FC might help in the detection of bone marrow involvement by the different types of B-cell malignant neoplasms.

METHODS

Fifty-four staging and 156 restaging bone marrow biopsies and bone marrow aspirates, obtained from 185 consecutive patients, were analyzed retrospectively. The results of the morphologic examination and FC were reviewed independently, and their ability to detect bone marrow involvement was compared.

RESULTS

FC and morphology agreed in 176 cases (83.8%), i.e., both showed 77 positive cases and 99 negative ones. Discrepant results were obtained in 30 cases (14.2%) in which morphologic examination showed 25 (11.9%) positive cases, whereas FC showed no evidence of disease. FC detected involvement in five cases (2.4%) in the presence of a histologically negative bone marrow biopsy. All morphologically undetermined bone marrow cases (four) were negative by FC.

CONCLUSIONS

Neither morphologic examination nor FC alone is adequate for the detection of all cases of B-lymphoid neoplasm bone marrow involvement. FC failed to detect bone marrow involvement in those B-NHL cases having focal paratrabecular infiltration, but proved to be more sensitive than histology in detecting small clonal B-cells in B-NHL, which demonstrated fewer than 5% neoplastic infiltrates. The clinical relevance of minimal disease detected by FC alone needs further evaluation because staging of lymphomas currently is based only on morphologic data.

Optimization of three- and four-color multiparameter DNA analysis in lymphoma specimens

BACKGROUND

Simultaneous analysis of DNA and immunophenotype of lymphoma cells by flow cytometry allows the calculation of the proliferative activity and aneuploidy in even a small lymphoma population. Unfavorable DNA binding characteristics or spectral features of DNA dyes impair the accuracy of multiparameter DNA analysis and limit their clinical application. We describe here a reliable and reproducible application of both three- and four-color multiparameter DNA analysis.

METHODS

After immunostaining of fresh samples of peripheral blood, bone marrow and single cell suspensions of lymph nodes from healthy and lymphoma patients, a methanol fixation for TO-PRO-3 and DRAQ5 staining was tested.

RESULTS

The red-excitable TO-PRO-3 on a FACSCalibur is limited to two-color antigen staining including fluorescein-isothiocyanate and phycoerythrin-labeled monoclonal antibodies due to its broad excitation spectrum. Although DRAQ5 is only applicable to flow cytometers equipped with a single argon laser emitting 488-nm light, its emission spectrum can be easily separated from the FITC, PE, and PE/Texas-Red emissions. DRAQ5 showed almost identical stoichiometric DNA binding characteristics as propidium iodide. Coefficient of variation produced by DRAQ5 staining is in the range of 3.5 and is adequate for detecting aneuploid amd near-diploid cells.

CONCLUSIONS

These advantageous features of DRAQ5 make it a reliable candidate for multiparameter clinical studies.

Specificity of polymerase chain reaction-based clonality analysis of immunoglobulin heavy chain gene rearrangement for the detection of bone marrow infiltrate in B-cell lymphoma-associated haemophagocytic syndrome

As a wide range of disorders underlie haemophagocytic syndrome, a rapid distinction between benign polyclonal and malignant monoclonal lymphoid proliferations is critical. We investigated whether polymerase chain reaction (PCR) amplification of immunoglobulin heavy chain gene rearrangement could efficiently detect clonal B-cell populations in non-diagnostic marrow for B-cell lymphoma-associated haemophagocytic syndrome (B-LAHS). On amplifying two DNA samples per biopsy, no reproducible monoclonal PCR result was found in reactive haemophagocytic marrows. In contrast, four out of nine assessable B-LAHS patients with histomorphologically and immunohistochemically lymphoma-free bone marrow showed a reproducible monoclonal immunoglobulin heavy chain gene rearrangement. At the molecular level, two B-LAHS patients had lymphoma-free marrow as demonstrated by patient-specific PCR, suggesting that haemophagocytic marrow is not always associated with lymphoma involvement. PCR-based demonstration of clonal B-cell populations in marrow would add an extra dimension to B-LAHS diagnosis.

Large B-cell lymphomas: fine-needle aspiration plays an important role in initial diagnosis of cases which are falsely negative by flow cytometry

Large B-cell lymphomas (LBCLs) have significant false-negative results when immunophenotyped by flow cytometry (FC). To clarify the role fine-needle aspiration (FNA) in reducing this false-negative rate, 28 cases ultimately diagnosed as LBCL that had FNA as part of the workup and a negative FC were identified. We examined their clinical and cytologic features, in comparison with cases of LBCL with FNAs that were positive by FC. In 24/28 FC-negative cases (86%) a cytologic diagnosis of suspicious or positive for malignancy was rendered. We conclude that cytologic analysis is more sensitive than FC in the diagnosis of malignancy in FNA of LBCL, particularly in aspirates with low cellularity and/or low viability. Examination of cytospin preparations of the actual material analyzed by FC may provide an indication that an FC result is falsely negative. It is important to recognize the potential of false-negativity by FC of LBCLs when interpreting FNAs with features suggesting lymphoma.

CD2 is expressed by a subpopulation of normal B cells and is frequently present in mature B-cell neoplasms

BACKGROUND

CD2 is expressed by T and natural killer (NK) cells and has been reported in T/NK cell lineage neoplasms as well as in immature B-lymphoblastic and myeloid leukemias. Although CD2+ B-cells have been identified in normal fetal and postnatal thymus, they have not been reported in adults.

METHODS

We retrospectively reviewed flow cytometric immunophenotypic data on consecutive low-grade B-cell leukemias and lymphomas to investigate the frequency of CD2 expression. We also reviewed samples from normal healthy donors to determine whether there is a normal CD2+ B-cell population.

RESULTS

CD2 expression (partial or complete) was observed in 13 of 83 (16%) chronic lymphocytic leukemias (CLL), 16 of 29 (55%) follicle center lymphomas (FCL), 3 of 12 (25%) hairy cell leukemias (HCL), 0 of 6 mantle cell lymphomas (MCL), 8 of 28 (29%) large cell lymphomas (LCL), and in 0 of 5 marginal zone/mucosa-associated lymphoid tissue lymphomas (MZL/MALT). We determined that 5.74 +/- 2.46% (mean +/- SD) of normal peripheral blood B cells and 6.48 +/- 1.62 % (mean +/- SD) of normal bone marrow B cells coexpress CD2.

CONCLUSIONS

CD2 expression in B-cell neoplasia is a more prevalent phenomenon than previously appreciated. Normal CD2+ B-cell populations are observed in adults and may represent the nonmalignant counterpart of CD2+ B-cell neoplasms.

Flow cytometric analysis of cerebrospinal fluid samples and its usefulness in routine clinical practice

Low volume and few cells have hampered the use of flow cytometry for studying cerebrospinal fluid (CSF) in routine clinical practice, although information about the cellular phenotypes present in this type of sample is of great value in many diseases. We developed a novel flow cytometric strategy capable of identifying total CSF T lymphocytes and the CD4+ subset, even in CSF samples with as few as 1 leukocyte per 3 microL of sample. We also showed that identification of CD8+ T cells could be achieved in most samples, while B lymphocytes are detectable only in samples with more than 5 cells per microliter. These findings demonstrate the reliability of this method to improve the diagnostic accuracy of classic cytologic studies in many neurologic disorders.

Differentiating lymphoblastic lymphoma and Ewing's sarcoma: lymphocyte markers and gene rearrangement

We encountered a child with an intraosseous small round cell tumor that was negative for LCA, CD20 (L26), and CD3 and positive for vimentin, CD99 (MIC-2), and periodic acid-Schiff. The tumor exhibited rosette-like formations. This case was initially interpreted as Ewing's sarcoma (ES); however, additional studies revealed positivity for CD79a, CD43, and TdT expression, and an immunoglobulin heavy chain gene rearrangement (IgH-R) by polymerase chain reaction (PCR) established this to be a precursor B-lymphoblastic lymphoma. Because the differential diagnosis of ES and lymphoblastic lymphoma can be difficult and the differential diagnostic value of leukocyte antigens and immunoglobulin heavy chain gene rearrangement studies have not been fully evaluated, we conducted a more extensive investigation on 33 (21 soft tissue and 12 intraosseous) ES cases. Cases were retrieved from the files of the Department of Pathology at Georgetown University and from the Soft Tissue Registry of the Armed Forces Institute of Pathology. The cases were studied by light microscopy, immunohistochemistry, and PCR for IgH-R and T cell receptor gamma chain gene rearrangement (Tgamma-R). There were 17 females and 16 males; the mean age was 29.3 years. Locations included the extremities (n = 17) and trunk (n = 16). All cases fit the ES spectrum by light microscopy and immunohistochemistry, as previously determined, and were negative for lymphoid markers (LCA, CD3, CD20, CD43, CD79a, and TdT), CD10 and CD34. CD99 was positive in 31/33 and bcl-2 was weakly positive in 13/33 cases. All 21 cases studied for gene rearrangements by PCR were negative for IgH-R and Tgamma-R. Distinction of intraosseous lymphoblastic lymphoma from ES may be difficult because lymphomas may occasionally exhibit unexpected morphologic and immunophenotypic properties including LCA, CD3 and CD20 negativity and cytokeratin positivity. Additional analysis using CD79a, CD43, TdT, and PCR should be performed to avoid misdiagnosis. True ES is negative for lymphoid markers including CD79a, CD43, and TdT, as well as for IgH-R and Tgamma-R.