Hematopathology / CYTOLOGY AND FLOW CYTOMETRY IN LYMPH NODE FINE-NEEDLE ASPIRATION

We studied flow cytometry in 156 fine-needle aspirations (FNAs) of lymph nodes performed between June 1993 and September 1998. Information from flow cytometry was combined with cytomorphologic evaluation, and the diagnosis determined by using combined modalities was compared with tissue biopsy results or clinical follow-up. In 74 cases, a combined cytopathologic-flow cytometric diagnosis of lymphoma was made; histologic material was available for 52 patients; in no case was a benign process found. The lymphoma grade assigned agreed with histopathologic findings in 45 of 48 cases with a specific cytologic diagnosis. Treatment was initiated on the basis of the FNA alone for 17 of 52 patients with a history of lymphoma and in 22 additional patients with no followup biopsy. Among 71 cases in which the diagnosis using both modalities was benign, the only false-negative was 1 case of Hodgkin disease. Of the 156 cases, 11 were considered atypical or suggestive of lymphoma; biopsies from 8 of 10 patients revealed lymphoma. A combination of flow cytometry and cytomorphology of cells obtained by FNA of lymph nodes can distinguish between benign and malignant lymphoid infiltrates and support a diagnosis of lymphoma that permits definitive therapy in most cases. The fine-needle aspiration (FNA) technique has been used in the diagnosis of lymph node disease since the early 20th century. The method was used originally for the identification of infections and in the diagnosis of metastatic cancers.1-3 Application to lymphoma diagnosis was viewed historically as problematic because of the central importance given to the evaluation of lymph node architecture. Some cytologic criteria, such as cellular monomorphism, have been considered useful for helping to distinguish malignant lymphoid infiltrates from benign infiltrates,4,5 but widespread use of FNA as a primary diagnostic modality in lymphoma has been limited. Ancillary studies have been applied widely to the classification of lymphomas in cytopathology and in histopathology. Many early cytologic studies used standard immunocytochemical methods6-10 because the perceived need for high cellularity in multiparameter flow cytometry limited its application in cytopathology.11,12 However, modern flow cytometry seems ideally suited to the study of FNA specimens because this technique gives rise to single-cell suspensions of well-preserved cells, and with care, adequate numbers of cells can be harvested and studied by multiparameter analysis to accurately diagnose lymphoma.13,14 Several studies have documented the value of flow cytometry in lymphoma diagnosis,9,14 although many have been descriptive or focused specifically on comparison of immunocytochemistry and flow cytometry.10,13,15 We describe our experience with the use of flow cytometric testing and routine cytomorphologic examination during a 5-year period. We deliberately did not design the study to compare the accuracy of one modality with another, because we believe that they are complementary techniques that provide information that is best used Hematopathology / ORIGINAL ARTICLE Am J Clin Pathol 2000;114:18-28 19 © American Society of Clinical Pathologists together. We thus tried to formulate the most clinically useful diagnosis we could by combining information from both techniques and studied how this correlated with available histopathologic diagnoses or with clinical follow-up. Materials and Methods Flow cytometry was performed on 156 FNA specimens that were clinically or morphologically suggestive of lymphoma between June 1993 and September 1998. Cases were largely unselected, although those in which cellularity or cell preservation was insufficient for flow cytometric analysis and cases of nonhematolymphoid processes were excluded. Specimen collection was performed in the clinic or at the patient’s bedside. After the patient was examined, the procedure and risks were discussed, informed consent was obtained, and the area was cleaned and draped. Aspirations were performed using a 23or 25-gauge needle, a 10-mL syringe, and a syringe holder (Cameco, London, England). Paired slides were prepared, with 1 slide air dried and stained with a modified Wright-Giemsa stain (Diff-Quik stain set, Dade Behring, Newark, DE), and the other alcohol-fixed for Papanicolaou staining. Material was sent for flow cytometric analysis based on clinical suspicion or from on-site evaluation of smears stained with a modified Wright-Giemsa. An average of 3 passes was made to provide a sufficient number of cells. Material from these passes was placed in Hank balanced salt solution and transported immediately to the flow cytometry laboratory. Of the total 450 lymph node aspirations performed during this time, 156 were sent for flow cytometric evaluation. Flow cytometric evaluation was performed on aliquots of aspirate material by 3-color immunofluorescence using the combinations kappa-fluorescein isothiocyanate (FITC)/lambda-phycoerythrin (PE)/CD19-PE-cyanine 5 (cychrome), CD5-FITC/CD22-PE/CD3-perdinin chlorophyll protein (PerCP), and CD71-FITC/CD33-PE/CD45PerCP as a screening panel. When sufficient cells were available and in cases in which an abnormal population was identified, additional antibodies, including CD10, CD20, CD23, FMC7, CD11c, CD7, CD4, CD8, CD2, and, occasionally, others also were used for improved subclassification. All antibodies were from Becton Dickinson, San Jose, CA, except CD19-cychrome (Pharmingen, La Jolla, CA) and FMC7 (Immunotech, Westbrook, ME). Flow cytometric analysis was performed on a flow cytometer (FACScan, or, in later cases, FACSCalibur flow cytometer, Becton Dickinson). Data analysis was performed using PAINT-A-GATE PRO software (Becton Dickinson). Populations were identified by “painting” (arbitrarily assigning color, for analysis) the constituent populations, and the phenotype of normal and abnormal cell components was inferred from the distribution of the painted populations in multiparameter space. To determine the presence of B-cell clonality, all B cells in the sample were assigned 1 color, and the relative distribution of kappa and lambda was visualized in displays of kappa vs lambda and in displays of forward-angle light scatter (a measure of cell size) vs light chains. In this manner, visual determination of clonal populations was readily made even in the presence of normal polyclonal B cells in the background. Flow cytometric data were reviewed in conjunction with the morphologic features to determine a diagnosis, and both pieces of information were incorporated into the final pathologic report. However, no specific mandate was given to the pathologist signing out the case about how the information should be used. In the first portion of the study, cytopathologic reports were reviewed retrospectively for all cases and initially categorized as positive for lymphoma, negative for lymphoma, or atypical or suggestive of lymphoma. The latter category included cases in which the report contained the terms “atypical” or “suspicious for lymphoma,” without giving a definitive diagnosis. Although no single explanation could be identified for this diagnosis, most of these cases had insufficient cell number, poor cellular preservation, or indefinite flow cytometric findings, and patients given this diagnosis generally underwent further more definitive examination. As the retrospective review proceeded, however, it became clear that among cases diagnosed as lymphoma, some of the reports gave more precise diagnoses than did others. For example, some specified the grade of lymphoma and some the histologic subtype, while others simply diagnosed lymphoma without further elaboration. Because it was unclear how much of this variability could be due to lack of standardized reporting, all cases diagnosed as lymphoma without mention of specific grade were re-reviewed blindly in an attempt to assign a definitive grade. We chose to focus on grade because we believed that attempting to reproduce histopathologically defined classification systems, such as the Revised European-American classification of lymphoid neoplasms (REAL),16 was likely to be difficult or impossible, while lymphoma grade provides clinically useful information. Once cases were classified, additional pathology records were reviewed to identify evidence of definitive tissue biopsy for correlation. In cases in which there was no additional pathology material, medical records were reviewed when available to determine clinical outcome. Pathologic or clinical follow-up was available for 132 (84.6%) of 156 patients. Nicol et al / CYTOLOGY AND FLOW CYTOMETRY IN LYMPH NODE FINE-NEEDLE ASPIRATION 20 Am J Clin Pathol 2000;114:18-28 © American Society of Clinical Pathologists Results Correlation Between Cytopathologic and Flow Cytometric Diagnoses and Histopathologic Diagnosis ❚Table 1❚ shows the correlation between the combined cytopathologic and flow cytometric diagnosis and histopathologic diagnosis. A total of 74 specimens were considered positive for lymphoma; 52 of these cases also had histologic diagnoses of lymphoma. One patient had a follow-up biopsy that was of poor quality and nondiagnostic, and treatment was instituted without additional procedures. No case diagnosed as lymphoma had a benign disease or nonhematolymphoid malignant neoplasm revealed by tissue biopsy. Seventy-one specimens were considered negative for lymphoma; for 12 of these, results of follow-up biopsies were negative, while 58 patients did not undergo biopsy, although no other clinical or pathologic evidence of lymphoma was ever obtained. For 1 patient who underwent biopsy, Hodgkin disease was found (see “Cases Diagnosed as Benign”). Of the 11 specimens considered atypical or suggestive of lymphoma, 10 patients had undergone biopsy and 8 were found to have lymphoma. Cases Diagnosed as Lymphoma As noted, histologic material was available for comparison in 52 cases in which a definitive diagnosis of lymphoma was made by the combination of cytopathologic examination and flow cytometry. Treatment decisions were made for 21 patients diagnosed as having lymphoma without tissue biopsy ever having been obtained. In 17 cases, a previous diagnosis of lymphoma had been given, and the combined cytopathologic and flow cytometric diagnosis was considered evidence of recurrence; treatment was initiated without additional biopsy. In 31 of the remaining 35 cases, the combined cytopathologic and flow cytometric diagnosis also provided grade or precise classification of lymphoma, and in 28 of these 31 cases, this information proved correct by the subsequent biopsy ❚Table 2❚. ❚Image 1❚ and ❚Figure 1❚ show an example of the cytologic and flow cytometric findings, respectively, in a case of a small lymphocytic lymphoma that was confirmed by tissue biopsy. Cytologically, the cells are small and uniform, and the flow cytometric findings show a dimly kappa-positive clonal B-cell population with the characteristic CD5+, CD23+ B-cell phenotype expected for chronic lymphocytic leukemia/small lymphocytic lymphoma. By contrast, ❚Image 2❚ and ❚Figure 2❚ show the cytologic and flow cytometric findings, respectively, of a large B-cell lymphoma, again confirmed by tissue biopsy. In this case, the cells are large cytologically and by forward-angle light scatter, and these large cells show clonal kappa light chain expression. This population showed moderate density expression of CD71, the transferrin receptor, a finding that supports a conclusion of an aggressive malignant neoplasm (data not shown). In 3 cases, subsequent biopsy revealed that the grade of lymphoma on the tissue biopsy specimen was substantially different from that suggested by the combined cytopathologic and flow cytometric diagnosis. In 2 cases, the combined cytopathologic and flow cytometric diagnosis was low-grade lymphoma, while subsequent biopsies revealed large cell lymphomas arising from underlying low-grade lymphomas. One of these is illustrated in ❚Image 3❚ and ❚Figure 3❚. The majority of cells are small, and flow cytometry studies showed a typical kappa-positive CD10+ B-cell phenotype characteristic of follicle center cell lymphoma, with only a few large cells by light ❚Table 1❚ Correlation of Combined Cytopathologic and Flow Cytometric Diagnoses and Histologic Findings Cytopathologic/Flow Cytometric Diagnosis Biopsy or Other Follow-Up No. of Cases Positive for lymphoma (n = 74) Grade agreed with subsequent biopsy findings 28 No follow-up biopsy result available 21 Grade agreed with previous diagnosis 17 No grade given in cytology; follow-up biopsy findings diagnostic 4 Grade different in follow-up biopsy findings 3 Follow-up biopsy nondiagnostic 1 Negative for lymphoma (n = 71) No follow-up biopsy 58* Follow-up biopsy result negative 12 Follow-up biopsy result positive (Hodgkin disease) 1 Atypical or suggestive of lymphoma (n = 11) Follow-up biopsy result positive 8 Follow-up biopsy result negative 2 No follow-up 1 * Review of available clinical records for 48 patients revealed no evidence of development of lymphoma. Hematopathology / ORIGINAL ARTICLE Am J Clin Pathol 2000;114:18-28 21 © American Society of Clinical Pathologists scatter; by contrast, focal sheets of large cells were seen on the histologic section. In these cases, it is likely that tissue sampling limited the accuracy of the combined cytopathologic and flow cytometric diagnosis. In the remaining case, a diagnosis of a high-grade lymphoma was given, but subsequent tissue biopsy revealed a follicular mixed lymphoma with insufficient large cells to warrant classification as a large cell lymphoma. In 4 cases, the combination of cytomorphologic examination and flow cytometry was sufficient to arrive at a diagnosis of lymphoma, but no further subclassification or grading was attempted. Histologic findings in these cases included 1 case of follicular lymphoma, mixed cell type, 2 cases of large cell lymphoma (1 partially involving a lymph node), and 1 case of a marginal zone lymphoma in which an unusually high number of large cells and focally high (Ki67) proliferative index was noted. Cases Diagnosed as Benign In 58 of 71 cases diagnosed as benign, no corresponding tissue biopsy was obtained. Review of available clinical information in these cases showed no meaningful clinical suspicion of lymphoma and no evidence of development of lymphoma in any patient. For 12 patients, results of follow-up biopsies were negative. Diagnoses in these cases included the following: reactive follicular hyperplasia, 5 cases; atypical hyperplasia, 3 cases; granulomatous inflammation, 2 cases; acute necrotizing lymphadenitis, 1 case; and incidental benign intraparotid lymph node, 1. One patient given a definitive diagnosis of ❚Table 2❚ Comparison of Combined Cytopathologic and Flow Cytometric Diagnoses of Lymphoma With Subsequent Histopathologic Material Cytopathologic/Flow Cytometric Diagnosis Histologic Diagnosis No. of Cases Aggressive B-cell lymphoma Large B-cell lymphoma 16 Low-grade B-cell lymphoma (n =7) Small lymphocytic lymphoma Small lymphocytic lymphoma 2 Follicle center cell lymphoma Follicular small cleaved lymphoma 3 Low-grade B cell, not otherwise Follicular small cleaved lymphoma 1 specified (n = 2) Marginal zone lymphoma 1 Peripheral T-cell lymphoma Peripheral T-cell lymphoma 2 Anaplastic large cell lymphoma Anaplastic large cell lymphoma 1 Low-grade B-cell lymphoma (n = 2) Follicular small cleaved lymphoma 1 transforming to large cell lymphoma Marginal zone lymphoma transforming 1 to large cell lymphoma Aggressive B-cell lymphoma Malignant lymphoma, follicular mixed 1 small cleaved and large cell