The Value of Transformed Lymphocyte Count in Subclassification of Non-Hodgkin's Lymphoma by Fine-Needle Aspiration

Transcription

1 HEMATOPATHOLOGY Original Article The Value of Transformed Lymphocyte Count in Subclassification of Non-Hodgkin's Lymphoma by Fine-Needle Aspiration NANCY A. YOUNG, MD, 1 TAHSEEN I. AL-SALEEM, MD, 1 ZINA AL-SALEEM, 1 HORMOZ EHYA, MD, 1 AND MITCHELL R. SMITH, MD, PhD 2 No established criteria exist for predicting lymphoma grade or transformation in cytologic material. We counted transformed lymphocytes in fine-needle aspiration (FNA) biopsy specimens to determine whether the percentage of these cells in the smear could predict the histologic grade, the biologic behavior, or both. The percentage of transformed lymphocytes out of total lymphoid cells was determined on Papanicolaou-stained smears. Afterward, a cytodiagnosis was based on clinical information available at the time of the FNA, cytomorphologic data, and flow cytometry data. Results were correlated with results of examination of the surgical biopsy specimen, clinical behavior of the lymphoma, or both. The percentage of transformed lymphocytes was 10% or less in all low-grade or indolent lymphomas. Aspirates with transformed lymphocyte counts of 20% or greater were aggressive lymphomas. We also report our experience in the diagnosis of non-hodgkin's lymphoma by FNA using cytomorphologic examination and immunophenotyping by flow cytometry at a cancer referral hospital. This is a preliminary study, and larger series may help establish the ranges of transformed lymphocyte counts that correlate with the lymphoma subtype. (Key words: Lymphoma; Classification; Diagnosis; Pathology; Needle biopsy; Fine-needle aspiration; Lymphocyte transformation; Flow cytometry; Cytology) Am J Clin Pathol 1997,108: The application of fine-needle aspiration (FNA) cytology in the diagnosis and management of malignant lymphoma is still controversial. 1-4 Major limitations of this technique include overlap of cytologic patterns between some benign and malignant processes and the inability to determine architectural patterns. 5,6 The revised European-American classification system of lymphoid neoplasms (REAL) places a greater emphasis on cytomorphologic examination in combination with immunophenotyping and molecular studies rather than on architectural pattern. 7 This new classification system may expand the role of cytology in lymphoma diagnosis, particularly for From the Departments of 1 Pathology and 2 Medical Oncology, Fox Chase Cancer Center, Philadelphia, Pennsylvania. This research was supported in part by a grant from Bristol- Myers Squibb, Oncology, Princeton, NJ, and was presented at the 44th Annual Scientific Meeting of the American Society of Cytopathology, November 6-7,1996. Manuscript received December 4, 1996; revision accepted January 23,1997. Address reprint requests to Dr Young: Fox Chase Cancer Center, 7701 Burholme Ave, Philadelphia, PA those practicing cytology with expertise in hematopathology. Although relatively limited, aspirated material can be adequate for immunophenotypic and molecular studies. Another challenge facing those interpreting FNA specimens from patients with lymphoma is determining when transformation to a higher grade has occurred in a patient with a history of low-grade lymphoma. No guidelines exist in cytology for diagnosing transformation, particularly when there is a mixture of small and large lymphoma cells. The Rapapport, 8 Lukes and Collins, 9 Kiel, 10 and Working Formulation 11 classification systems all graded follicular lymphomas according to the estimated percentages of large cells in the surgical biopsy specimen. The Mann-Berard 12 system grades follicular lymphomas based on precise counting of large noncleaved cells observed within a field of the neoplastic nodule. We report our experience with a retrospective study in which we attempted to classify non- Hodgkin's lymphoma (NHL) by cytomorphologic examination and immunophenotypic markers at a cancer referral center. We also applied Mann-Berard's concept of counting large cells to our FNA biopsy specimens of 143

2 144 HEMATOPATHOLOGY 1 Article lymphoproliferative lesions to determine whether the percentage of transformed lymphocytes (TL) correlated with the results of examination of the surgical biopsy specimen, the biologic behavior of the lymphoma, or both. MATERIALS AND METHODS We reviewed our records for the period April 8,1993, to August 23,1995. During this period, 1,270 FNA biopsies were performed. Of these, 59 (4.6%) specimens resulted in a diagnosis of lymphoma or a lymphoproliferative or reactive lymphoid process. One case was excluded because of specimen inadequacy, leaving 58 FNA specimens from 54 patients for the study. Corresponding surgical biopsy slides and reports were also obtained. Cytotechnologists assisted the clinicians on site during the performance of the FNAs to prepare cytologic smears, collect material for flow cytometry, and determine specimen adequacy. To optimize our yield for flow cytometry in cases of suspected lymphoma, the needle, syringe, and cell culture medium for flow cytometry were heparinized to prevent bloody specimens from clotting and to facilitate cell retrieval for flow cytometry. Two to four needle passes provided adequate material in most cases. For each needle pass when possible, an alcohol-fixed Papanicolaou-stained smear and an air-dried Diff-Quik-stained (Baxter Healthcare, Miami, Fla) smear were prepared, and a drop or two of material was submitted in 10 ml of RPMI-1640 medium (Life Technologies, Grand Island, NY) for flow cytometry. The needle was flushed in 10 ml of Cyto Lyt (Cytyc, Marlborough, Mass) for the cell block preparation. After centrifugation, the sediment was fixed in formalin, embedded in paraffin, and 4-um sections were hematoxylin-eosin stained. Of the 54 patients, material was submitted for flow cytometry for 39. Cells were separated for flow cytometry using a gradient (Ficoll-Paque, Pharmacia Biotech, Uppsala, Sweden). Viable cells were counted, and the number was adjusted to 0.5 to 1 million cells per milliliter in aliquots of 100 ul with a 1% Hanks' balanced salt solution (Gibco BRL Life Technologies, Gaithersburg, Md). Directly labeled antibodies were added, incubated for 20 minutes in a covered ice bath, washed in 1% azide, and resuspended, and the fluorescence was measured with a flow cytometer (FACScan, Becton-Dickinson, San Jose, Calif). The basic panel of antibodies used was CD45/CD14, KA, CD20/CD5, and CD19/CD10 with isotype and negative controls. This basic panel was reduced or extended according to the number of cells available, the predicted subtype from preliminary examination of the slides, the patient's previously diagnosed lymphoma, or the clinical situation. After examination of the forward and side scatter, cells within the lymphocyte gate were examined, excluding the granulocytes and monocytes. However, if the cells were morphologically mixed, small and large, "backgating" on the two sizes was used, and each gate was examined separately. When available, flow cytometry reports on the aspirate, surgical biopsy specimen, blood sample, or bone marrow specimen were retrieved on each case. Relevant clinical and follow-up information was obtained from the medical records. A flow chart containing the laboratory and clinical information was maintained on each patient by one member of the team (Z.A-S.) so that the pathologists reviewing the cytology (N.A.Y., T.A-S.) for this study examined all the cases together on a multiheaded microscope and were unaware of the original cytologic or surgical biopsy diagnoses. On each case, the most cellular Papanicolaou-stained smear was selected, and 100 lymphoid cells were randomly counted under x400 magnification. The percentage of transformed lymphoid cells (transformed lymphocyte count or TLC) was determined except when the Papanicolaou-stained material was not available. A transformed lymphocyte (TL) is defined as a large lymphoid cell that is two to three times the size of a normal lymphocyte, with a vesicular noncleaved nucleus. Nucleoli could be large and prominent or smaller and less distinct (Fig 1). Smaller lymphoid cells were also considered as transformed if they had prominent nucleoli (Fig 2). These should be differentiated from small lymphocytes with chromatin clumping seen in small lymphocytic lymphoma (Fig 3). We paid special attention to differentiate macrophages from the large lymphoid cells. Macrophages have large round to oval or somewhat coffee bean-shaped nuclei with smooth regular nuclear membranes, fine chromatin, small nucleoli, and wispy cytoplasm that may contain phagocytized debris (Fig 4). After a TLC was performed, the record keeper provided the reviewers with the specimen source and the clinical information that was known at the time of the aspiration. The cell block and Diff-Quik-stained smears were also reviewed. An initial cytologic diagnosis without the flow cytometry information was determined. If more than 25% of the lymphocytes were transformed, a provisional diagnosis favoring large cell lymphoma was made. Lymphoblastic, immunoblastic, or Burkitt-like features were also recorded. Conversely, if there were fewer transformed cells admixed with small lymphocytes or predominantly small lymphoid cells with uniform clumped chromatin, a preliminary diagnosis of suggestive of lymphoma was given. If sheets of transformed cells were present in the cell AJCP At igust!997

3 YOUNG ET AL 145 Transformed Lymphocytes in Fine-Needle Aspiration Biopsy Specimens block, the lymphoma was also considered transformed even when the TLC was less than 25% in the smears. Subsequently, cytomorphologic findings were reviewed, with immunophenotyping studies performed on the FNA specimen in an attempt to arrive at a more definitive diagnosis and subtype. After a final cytologic-immunophenotypic diagnosis was determined, the pathologists reviewed the corresponding surgical biopsy slides and results of additional immunophenotyping studies. Classifications according to the Working Formulation and the REAL were recorded. However, because most clinical protocols for lymphoma therapy apply the Working Formulation, this system was used for the clinicalpathologic correlation and for predicting biologic behavior. Clinical follow-up data were obtained from the oncologist (M.R.S.) to determine how well the cytologic diagnoses compared with the surgical biopsy result or the clinical behavior of the lymphoma in each case. Clinical behavior compatible with indolent and aggressive lymphomas was defined (Table 1). Results were tabulated and evaluated with particular focus on the correlation between the TLC and lymphoma grade or behavior. Statistical analysis was done by comparing the means of TLC using the Student's t test. RESULTS The slides of 58 FNAs for 54 patients were available for review. Of those, 9 FNAs did not have Papanicolaou smears adequate for evaluation. However, all had Diff- Quik smears or cell blocks considered adequate for diagnostic evaluation. Of the 54 patients, the FNA specimens from 39 were submitted for flow cytometry. Nine (23%) of those 39 were inadequate for immunophenotypic studies because they did not have enough viable cells. Thirteen had at least a basic panel of antibodies, 11 had a more extensive panel as considered necessary, and 5 had only enough cells to perform K and X light-chain analysis. Our diagnoses based on cytomorphologic findings with flow cytometry results when available were as follows: high-grade lymphoma, 3; intermediate-grade lymphoma, 19; low-grade lymphoma, 9; suggestive of but not diagnostic for lymphoma, 18; diagnostic for lymphoma but grade undetermined (all between low and intermediate grade), 5; and reactive, 4. Of the 54 patients, 23 (42%) had a previous diagnosis of lymphoma or leukemia in which the FNA was performed to diagnose recurrence of disease or transformation to a higher grade. Of the 18 diagnoses that were suggestive of but not diagnostic for lymphoma, 13 (72%) were in patients without a previous diagnosis of lymphoma. 1. V* 9 V FIG 1. Transformed lymphocytes. Large noncleaved lymphocytes with prominent nucleoli surrounded by chromatin clearing (arrows) (Papanicolaou stain, x100). FIG 2. Transformed lymphocytes. Large noncleaved cell with open chromatin. Nucleolus is not prominent in the cell in the center (short arrow); a small transformed lymphocyte containing a prominent nucleolus is surrounded by a halo (long arrow) (Papanicolaou stain, xloo). Vol. 108 No. 2

4 146 HEMATOPATHOLOGY Article Fie 3. Mantle zone lymphoma. Small lymphoid cells with irregular nuclei and coarse chromatin clumping (Papanicolaou stain, xloo). FIG 4. Macrophages. Large round and coffee bean-shaped nuclei with fine chromatin, small nucleoli, and wispy cytoplasm (Papanicolaou stain, xloo). Of the four patients given a diagnosis of reactive lymphoid process, one had no follow-up, examination of a biopsy specimen confirmed the diagnosis in one, and in two patients, the reactivity of the lymph nodes resolved. One of these had positive Epstein-Barr virus titers. The flow cytometry, biopsy, and clinical follow-up data for the cases initially assessed as diagnostic of lymphoma are outlined in Table 2. Large cell, lymphoblastic, and Burkitt's lymphomas could be diagnosed correctly after the initial assessment by cytomorphologic examination and flow cytometry. These diagnoses were compatible with the clinical evaluation and the results of examination of the biopsy specimens when available. Of 44 FNA specimens from patients with a final diagnosis of lymphoma, 26 were examined by flow cytometry, and 18 were not. A definitive diagnosis of lymphoma was made on 24 (92%) of the 26 FNA specimens in which flow cytometry was performed, and only 2 were inconclusive. Both cases were judged inconclusive because they were iirvrnunoglobulin negative. Of 18 FNA specimens without concurrent immunophenotyping, in 12 (67%) a diagnosis of lymphoma could be made, and 6 were inconclusive. Of the 12 FNA specimens that were diagnostic of lymphoma even without immunophenotyping, 9 were large cell lymphomas. One was low-grade follicular lymphoma consisting of predominantly small, irregular, cleaved lymphoid cells. Subsequent immunophenotyping and cytogenetics performed on a biopsy specimen proved it to be a follicular lymphoma. In two aspirates, the diagnosis of lymphoma could be determined based on morphology alone, but the subtype could not be determined. The final diagnoses and follow-up on the 18 FNAs considered suggestive of lymphoma on initial assessment are listed in Table 3. Only 2 of the 18 aspirates that were suggestive of lymphoma had corresponding flow cytometry data for the FNA specimen. One had a slight X predominance, but material was scant, and the other proved to be a lymphoma with absent expression of immunoglobulins. Four of 18 lymph nodes that were suggestive of lymphoma were reactive. Interestingly, two of these that contain many atypical cells were progressive transformation of germinal centers, which is a reactive condition that may be associated with a concomitant or subsequent neoplasia. The final diagnoses according to the Working Formulation and the REAL classification obtained for each lymphoma based on cytomorphologic examination, subsequent examination of a biopsy specimen, or clinical follow-up and immunophenotyping are listed in Table 4. Transformed Lymphocyte Count For nine aspirates, the Papanicolaou-stained smear was unavailable or inadequate for a cell count. In the AJCP August 1997

5 YOUNG ET AL 147 Transformed Lymphocytes in Fine-Needle Aspiration Biopsy Specimens remaining 49 smears, the TLC was determined. Smears were more suitable than the cell block for counting TLs because of more even distribution of the cell population and better nuclear morphology. Results were tabulated for the 40 patients with adequate follow-up information (Fig 5). Included are the cases that were initially suggestive of lymphoma and the cases of lymphoma of undetermined grade with biopsy results or adequate clinical follow-up data for appropriate categorization. The TLs were clustered in sheets in the cell block in four cases of large cell lymphoma. In two cases, the distribution of TLs in the smears was uneven, requiring two separate cell counts. Both were proved by examination of the biopsy specimen to be reactive follicular hyperplasia with focal progressive transformation of germinal centers. Of eight reactive lymph nodes, four had 4% or fewer transformed cells. One case lacked a Papanicolaou-stained smear, and, thus, a TLC was not performed. In the two cases of biopsy-proved progressive transformation of germinal centers, the cellular distribution was not uniform, requiring two separate cell counts yielding counts of 10% and 44% in one case and 8% and 36% in another. One case diagnosed as reactive had a TLC of 12%. The nine low-grade lymphomas, including one case of mantle cell lymphoma, had 10% or fewer TLs (mean, 4.4±3.1). Of the 26 large cell lymphomas (biopsy-proved or clinically aggressive), 22 had Papanicolaou-stained smears for which a TLC could be performed. This number includes one case of immunoblastic lymphoma, considered high grade according to the Working Formulation, that had a TABLE 1. CRITERIA FOR INDOLENT AND AGGRESSIVE CLINICAL BEHAVIOR Indolent lymphoma Stable or slowly progressive disease including one or more of the following: Stable with expectant observation Stable or improving with nonanthracycline chemotherapy Stable after localized radiation Stable after single course of monoclonal antibody Aggressive lymphoma Rapid progression before or during treatment including one or more of the following: Primarily refractory to anthracycline-containing chemotherapy Refractory to salvage regimens for aggressive lymphoma Responds to anthracycline-containing chemotherapy and then relapses Responds to anthracycline-containing chemotherapy; in remission TLC of 23%. In 18 of 21 lymphomas, the TLC ranged between 20% and 94%. The three exceptions were as follows: (1) a TLC of 18% in a mediastinal large B-cell lymphoma with sclerosis; (2) a grade 3 transformed follicular lymphoma (follicular large cell lymphoma) in which the smear showed a transformed cell count of 11%, but the cell block contained sheets of transformed cells accounting for about 37% of the cells; and (3) a cervical lymph node aspirate with a TLC of 9% that was initially called suggestive of but not diagnostic of lymphoma by cytology. Clinically, this patient had an aggressive lymphoma with bulky mediastinal adenopathy. Examination of a subsequent jugular lymph node biopsy specimen showed mediastinal large B-cell lymphoma with sclerosis. A TLC of 44% was obtained in a high-grade Burkitt-like B-cell lymphoma, but a TLC of only 5% was seen in a case of T-lymphoblastic lymphoma. The diagnosis of T-lymphoblastic lymphoma was documented by flow cytometry and examination of the bone marrow biopsy specimen. The mean TLC for aggressive (intermediate- and high-grade) lymphomas was 47.6%±22.3 %. The difference between TABLE 2. FLOW CYTOMETRY, BIOPSY, AND CLINICAL FOLLOW-UP ON FNA DIAGNOSTIC OF LYMPHOMA ON INITIAL ASSESSMENT AS GRADED ACCORDING TO THE INTERNATIONAL WORKING FORMULATION Grade Low 9 Initial Assessment Intermediate 19 + High 3 Undetermined 5 Total 36 Monoclonality by Flow Cytometry 8 9 3* 3 23 Biopsy 4* Clinical Assessment All stable compatible with low grade All aggressive course All very aggressive course 1, aggressive course; 2, stable under observation FNA = fine-needle aspiration. "Includes the case in which flow cytometry was not performed. + One patient underwent two FNA biopsies because the first aspirate was not satisfactory. *One T-cell lymphoma confirmed by flow cytometry; one non-burkitt's lymphoma. Two separate FNAs from one patient revealed low-grade assessment on a neck mass (low grade on biopsy) and a higher-grade component on FNA of the tonsil (biopsy of tonsil not performed). Vol. 108 No. 2

6 148 HEMATOPATHOLOGY Original Article TABLE 3. FINAL DIAGNOSIS ON 18 FINE-NEEDLE ASPIRATES CONSIDERED SUGGESTIVE OF LYMPHOMA ON INITIAL ASSESSMENT Diagnosis Number Remarks Reactive lymphoid process 4 Examination of biopsy specimen revealed reactive hyperplasia, 3; lymph node regressed, 1 Intermediate grade 5 Represents 5 aspirates from 3 patients (three aspirates from different areas were from 1 patient with transformed mucosal-associated lymphoid tissue[malt] lymphoma) Lymphoma, grade undetermined 2 1 interpretation of a biopsy specimen was complicated by tuberculosis and previous radiation for lymphoma in the same lymph node; 1 retroperitoneal aspirate indicated possible tumor, but examination of the lung biopsy specimen was diagnostic of lymphoma High grade 1 Immunoblastic with plasmacytic differentiation proved by examination of biopsy specimen Thymoma 1 Proved by examination of biopsy specimen Hodgkin's disease 2 Both specimens were considered polyclonal by flow cytometry; Hodgkin's disease was proved by examination of biopsy specimens No follow-up 3 Total 18 the TLCs of aggressive lymphomas and those of lowgrade lymphomas or reactive lymphoid processes was significant (P<.0001). The difference between the TLCs of reactive lymphoid processes and low-grade lymphomas was not significant. (P=.9) DISCUSSION Fine-needle aspiration (FNA) has been accepted as a screening test in patients with unexplained and persistent lymphadenopathy and for staging and monitoring recurrence or response to treatment in patients with a history of lymphoma However, the indications for using FNA cytology as a primary diagnostic test are controversial and have not been clearly delineated. Limitations of the technique include inability to determine architectural patterns, overlap of cytologic patterns between some benign and malignant processes, and the potential for sampling error. 5-6 Furthermore, accuracy depends on the experience of the diagnostician Experienced pathologists report that certain types of lymphomas can be reliably diagnosed by FNA when cytomorphologic examination is combined with ancillary studies, such as immunohistochemistry, flow cytometry, or molecular studies, when indicated. 16 " 20 However, when the morphology is that of predominantly cleaved cells with light chain restriction, as that from a follicular center cell lymphoma, it is recommended that a tissue biopsy specimen be obtained to identify the pattern so the lymphoma grade can be established according to the Working Formulation of NHL. 15 The REAL 7 contains newly recognized entities, such as MALT (mucosalassociated lymphoid tissue) and mantle cell lymphoma. It depends heavily on immunophenotyping and molecular studies rather than morphology alone and may expand the role of cytology in the diagnosis and subtyping of lymphoma. How cytometry can be used to establish monoclonality in B-cell lymphomas when the result of the cytomorphologic examination is ambiguous for a conclusive diagnosis of malignancy. The technique is also valuable for determining immunophenotypic markers for subclassifying lymphomas. 21 Working as a team, the cytopathologist and hematopathologist can make the best decisions about the choice of antibodies and selection of gates to obtain the most information possible from limited material, as usually obtained from FNAs. One major criticism of lymphoma cytology is that it cannot provide information about nodal architecture. For instance, most follicular lymphomas are considered low grade and have a different prognosis and treatment from their higher-grade diffuse counterparts. This distinction is important, not only in the primary diagnosis of NHL, but also in patients with previous low-grade lymphoma in whom new lesions develop. Transformation to a higher-grade lymphoma that would require more aggressive therapy must be determined. We believe that the architectural differences and, thus, the differences in behavior between follicular and diffuse lymphomas frequently correlate with the cellular composition of these two entities. The nodularity of follicular lymphomas is due to the tendency of the neoplastic cells in the more "differentiated" follicular center cell lymphomas to form follicular structures similar to their nonneoplastic counterparts. The small cleaved AJCP August 1997

7 YOUNG ET AL 149 Transformed Lymplwcytes in Fine-Needle Aspiration Biopsy Specimens TABLE 4. RECLASSIFICATION (BY REAL) OF LYMPHOMA IN 38 PATIENTS WITH ADEQUATE FOLLOW-UP Working Formulation REAL Classification High-grade lymphoma (n = 4) Immunoblastic large cell (n = 2) Burkitt-like small noncleaved (n = 1) T-cell lymphoma/leukemia (n = 1) Intermediate-grade lymphoma (n = 25) Mixed small and large cell, diffuse (n = 4) Large cell "de novo" (n = 12) Large cell (transformed from low grade) (n = 9) Low-grade lymphoma (n = 9) Follicular (n = 6) Small lymphocytic(n = 3) Large cell lymphoma (n = 2) Burkitt(n = l) Lymphoblastic lymphoma (n = 1) Transformed MALT (n = 4) Large cell lymphoma (n = 21) Follicular center cell lymphoma (n = 4), grade 1-2 Marginal zone lymphoma (n = 1) Mantle cell lymphoma (n = 1) Extranodal small B lymphocytic/chronic lymphocytic leukemia(n = 1) MALT (n = 2) REAL = revised European-American classification system of lymphoid neoplasms; MALT = mucosal-associated lymphoid tissue. cells proliferate slowly but are more likely to disseminate than the large cells. 22 This accounts for the indolent but slowly progressive course of the low-grade follicular lymphomas. Conversely, lymphomas containing a high percentage of noncleaved cells, both large and small, are considered more aggressive neoplasms. These cells are the proliferative component of the tumor and are referred to as centroblasts (Kiel classification) 10 or transformed cells (Lukes and Collins classification). 9 In the Working Formulation, follicular lymphomas are graded as low or intermediate according to the cellular composition of the nodules. 23 Conversely, diffuse mixed small and large cell lymphomas are considered intermediate grade, but nodular mixed small and large cell lymphomas are considered low grade. This created the concept that in FNA, a mixture of small and large cells can be follicular low grade or diffuse intermediate grade. According to the REAL, follicular lymphomas are graded according to their cytologic components, while diffuse mixed small and large cell lymphoma is not recognized as a separate entity. Another relatively new concept, especially in the United States, is that the diffuse small cleaved cell lymphoma in the intermediate grade of the Working Formulation is now recognized as one or more of the new entities rather than a subtype of follicular center cell lymphoma. Most do represent CD10", CD5 + mantle cell lymphomas. 7 Therefore, although nodal architecture is important, it is not always essential for the diagnosis and subtyping of lymphomas when cytomorphologic examination is used in combination with immunophenotyping. The Rapapport, 8 Lukes and Collins, 9 Keil, 10 and the Working Formulation 11 classifications all graded follicular lymphomas according to the estimated percentages of large cells in the surgical biopsy specimen. Problems that exist in grading lymphomas purely by estimating the percentage of large cells involve subjectivity and determining the appropriate numerical cutoff. 24 The Mann- Berard subclassification of follicular lymphomas 12 attempted to more accurately and objectively subdivide follicular lymphomas based on precise counting of large noncleaved cells observed within a field of the neoplastic nodule. We modified the Mann-Berard method to apply it to FNA. They counted numbers of large cells per highpower field, whereas we used percentages of transformed cells in the total number of lymphoid cells in the Papanicolaou-stained smear. We also did not limit our examination to follicular lymphomas but included all other types of NHLs. Cytologic preparations do not show architecture, but they are ideal for identifying nuclear morphology. Although most hematopathologists are more familiar with the Wright-Giemsa stain, we found the Papanicolaou stain to be superior for demonstrating nuclear morphology in lymphoid neoplasms, and it complemented our use of the Diff-Quik-stained smears and hematoxylin-eosin-stained sections of the cell block. However, the cell block preparations are sometimes useful as minibiopsy specimens, giving a rough idea of the architectural pattern to augment the cytomorphologic and immunophenotypic findings. Immunophenotyping is critical in the cytologic diagnosis of lymphoma when a mixture of small and large lymphoid cells is seen, and a reactive condition would be considered. 21 In our study, the lack of adequate material for flow cytometry was the most common reason Vol. 108 No. 2

8 150 HEMATOPATHOLOGY Original Article M 80- at o * o s: a. E >. _i a o E w "Si c RI # L Intermediate and High Grade Low Grade Reactive FlG 5. Correlation of transformed cell count with lymphoma grade. The transformed cell count was plotted against lymphoma grade as determined from the surgical biopsy specimen or clinical behavior. for judging a specimen as suggestive of lymphoma or nondiagnostic. Only 67% of the FNA specimens without immunophenotyping were diagnostic of lymphoma, in contrast to 92% in those with cell marker studies. When there is a mixture of small and large lymphoid cells, the diagnosis of lymphoma can usually be confirmed by flow cytometry. In addition, immunophenotyping helps to subclassify lymphoma. For instance, if the small cells are CD5 + T-lymphoid cells but the large cells are monoclonal B cells, then the tumor is a large cell lymphoma. Although the large cells may be only a small portion of the total cell population, they represent tumor cells. However, if the small and large cells are both part of the monoclonal population of B cells and CD10 antigen is expressed, then the tumor is most likely a follicular lymphoma. If the tumor cells are CD10", the disease is marginal zone, mantle cell, or small lymphocytic lymphoma. The CD5 positivity of the B cells differentiates the latter two from marginal zone lymphoma, while small lymphocytic lymphoma is usually CD23 + with dim expression of CD20 and immunoglobulins. The results of our study indicate that it may be possible to determine lymphoma grade in a cytologic preparation by counting the percentage of TLs in the Papanicolaou-stained smear. After a diagnosis of lymphoma is established by cytomorphologic examination and immunophenotyping, with proper clinical conditions, a TLC of 20% or more represents a large cell or other aggressive lymphoma with a predictive value of 100%. The predictive value of a TLC of 10% or less representing a low-grade lymphoma is 80%. Most reactive processes also have a TLC of 10% or less but can be differentiated from low-grade lymphoma on the basis of cytomorphologic examination and flow cytometry. All low-grade lymphomas in our study had TLCs of 10% or less. However, two cases of aggressive lymphoma also had a TLC of less than 10%. One was a high-grade lymphoma (T-lymphoblastic lymphoma) AJCP August 1997

9 YOUNG ET AL 151 Transformed Lymphocytes in Fine-Needle Aspiration Biopsy Specimens with a TLC of 5%. The cells in this unusual case were smaller than large cell lymphoma and had ill-defined nucleoli accounting for the low transformed cell count in our original examination. The other case was a cervical lymph node FNA with a TLC of 9%. This was a clinically aggressive lymphoma with bulky mediastinal adenopathy. Examination of a jugular lymph node biopsy specimen revealed large cell lymphoma. A low TLC does not guarantee the presence of low-grade lymphoma, particularly in cases of small but highly malignant lymphoblasts, which often have indistinct nucleoli, in lymphoblastic lymphoma. Also mantle cell lymphoma generally does not contain TLs, although it has a more aggressive clinical behavior than do typical low-grade lymphomas. However, lymphoblastic and mantle cell lymphoma have a characteristic immunophenotype that should distinguish them from follicular center cell and MALT lymphomas in which TLC is an indicator of transformation. In our hands, FNA biopsy with immunophenotyping by flow cytometry is an extremely valuable way to confirm or exclude the suspected relapse of NHL. Biopsy confirmation was required when cytomorphologic examination was not diagnostic, and the material for flow cytometry was inadequate or the examination results were inconclusive. The TLC seems to be a promising method to determine the presence of transformation, especially in patients experiencing a relapse. Fine-needle aspiration can furnish adequate material for immunophenotyping in patients whose original biopsy specimens could not be satisfactorily classified into one of the recently recognized entities because of technical or other limitations. Furthermore, cells immunophenotyped by flow cytometry can be retrieved if needed for further molecular studies. Larger series including patients in a primary care setting with longer follow-up periods are desirable to confirm our results and to further delineate the role of FNA in the diagnosis and monitoring of lymphoma. REFERENCES 1. Carter T, Feldman P, Innes D, Frierson H, Frigy A. The role of FNA cytology in the diagnosis of lymphoma. Acta Cytol. 1988;32: Pitts W, Weiss L. Fine needle aspiration biopsy of lymph nodes. In: Rosen PP, Fechner RE, eds. Pathol Annu. 1988;23: Hajdu SI, Melamed M. Limitations of aspiration cytology in the diagnosis of primary neoplasm. Acta Cytol. 1984;28: Chernoff WG, Lampe HB, Cramer H, Banerjee D. The potential clinical impact of the fine needle aspiration/flow cytometric diagnosis of malignant lymphoma. / Otolaryngol. 1992;21: Gupta AK, Nayar M, Chandra M. Reliability and limitations of fine needle aspiration cytology of lymphadenopathies: an analysis of 1,261 cases. Acta Cytol. 1991;35: Pontifex AH, Klimo P. Application of aspiration biopsy cytology to lymphomas. Cancer. 1984;53: Harris NL, Jaffe ES, Stein H, et al. A revised European-American classification of lymphoid neoplasms: a proposal from the International Lymphoma Study Group. Blood. 1994;84: Rappaport H. Tumors of the hematopoietic system. In: Atlas of Tumor Pathology, Section III, Fascicle 8. Washington, DC: Armed Forces Institute of Pathology; 1966: Lukes RJ, Collins RD. Immunologic characterization of human malignant lymphomas. Cancer. 1974; 34: Lennert K, Mohri N. Histopathology and diagnosis of non- Hodgkin's lymphomas. In: Lennert K, Stein H, Mohri N, Kaiscerling E, Muller-Hermelink HK, eds. Malignant Lymphomas Other Than Hodgkin's Disease. Berlin, Germany: Springer-Verlag; 1978: The Non-Hodgkin's Lymphoma Pathologic Classification Project. National Cancer Institute-sponsored study of classification of non-hodgkin's lymphoma: summary and description of a working formulation for clinical usage. Cancer. 1982;49: Mann RB, Berard CW. Criteria for the cytologic subclassification of follicular lymphomas: a proposed alternative method. Hematol Oncol. 1983;1: Kardos T, Maygarden S, Blumberg A, Wakely P, Frable W. FNA biopsy in the management of children and young adults with peripheral lymphadenopathy. Cancer. 1989;63: Pontiflex AH, Klimo P. Application of aspiration biopsy cytology to lymphomas. Cancer. 1984;53: Russel J, Orell S, Skinner J, Seshadri R. Fine needle aspiration cytology in the management of lymphoma. Aust NZj Med. 1983;13: Cartagena N Jr, Katz RL, Girsch-Ginsberg C, Childs CC, Ordonez NG, Cabanillas F. Accuracy of diagnosis of malignant lymphoma by combining fine-needle aspiration cytomorphology with immunocytochemistry and in selected cases, Southern blotting of aspirated cells: a tissue-controlled study of 86 patients. Diagn Cytopathol. 1992;8: Tani E, Christenson B, Powit A, Skoog L. Immunocytochemical analysis and cytomorphologic diagnosis on FNA of lymphoproliferative disease. Acta Cytol. 1988;32: Joensuu H, Pekka Y, Eerola E. Diagnostic value of DNA flow cytometry combined with FNA biopsy in lymphomas. / Pathol. 1988;154: Lubinski J, Chosia M, Huebner K. Molecular genetic analysis in the diagnosis of lymphoma in fine needle aspiration biopsies, part 2: lymphomas vs non-lymphoid malignant tumors. Anal Quant Cytol Histol. 1988;10: Katz RL, Hirsch-Ginsberg C, Childs C, et al. The role of gene rearrangements for antigen receptors in the diagnosis of lymphoma obtained by fine-needle aspiration: a study of 63 cases with concomitant immunophenotyping. Am j Clin Pathol. 1991;96: Al-Saleem T, Schilder RJ, Millenson M, Smith MR. Reclassification of lymphoid neoplasms into newly recognized entities by flow cytometric immunophenotyping. Blood. 1995;86:533a. Abstract. 22. Mann RB, Jaffe ES, Berard CW: Malignant lymphomas: a conceptual understanding of morphologic diversity. Am j Pathol. 1979;94: Jaffe ES, Raffeld M, Medeiros J. Histopathologic subtypes of indolent lymphomas: caricatures of the mature B-cell system. Semin Oncol. 1993;20: Metter GE, Nathwani BN, Burke JS, et al. Morphological subclassification of follicular lymphoma: variability of diagnoses among hematopathologists: a collaborative study between the repository center and pathology panel for lymphoma clinical studies. / Clin Oncol. 1985;3: Vol. 108 No. 2