Fine-needle cytology and flow cytometry immunophenotyping and subclassification of non-hodgkin lymphoma

A critical review of 307 cases with technical suggestions

Authors

  • Pio Zeppa M.D.,

    1. Dipartimento di Anatomia Patologica e Citopatologia, Facoltà di Medicina e Chirurgia, Università di Napoli “Federico II,” Napoli, Italia
    Search for more papers by this author
  • Gilda Marino M.D.,

    1. Dipartimento di Anatomia Patologica e Citopatologia, Facoltà di Medicina e Chirurgia, Università di Napoli “Federico II,” Napoli, Italia
    Search for more papers by this author
  • Giancarlo Troncone M.D.,

    1. Dipartimento di Anatomia Patologica e Citopatologia, Facoltà di Medicina e Chirurgia, Università di Napoli “Federico II,” Napoli, Italia
    Search for more papers by this author
  • Franco Fulciniti M.D.,

    1. Dipartimento di Anatomia Patologica e Citopatologia, Facoltà di Medicina e Chirurgia, Università di Napoli “Federico II,” Napoli, Italia
    Search for more papers by this author
  • Amalia De Renzo M.D.,

    1. Dipartimento di Ematologia, Facoltà di Medicina e Chirurgia, Università di Napoli “Federico II,” Napoli, Italia
    Search for more papers by this author
  • Marco Picardi M.D.,

    1. Dipartimento di Ematologia, Facoltà di Medicina e Chirurgia, Università di Napoli “Federico II,” Napoli, Italia
    Search for more papers by this author
  • Giulio Benincasa M.D.,

    1. Dipartimento di Anatomia Patologica e Citopatologia, Facoltà di Medicina e Chirurgia, Università di Napoli “Federico II,” Napoli, Italia
    Search for more papers by this author
  • Bruno Rotoli M.D.,

    1. Dipartimento di Ematologia, Facoltà di Medicina e Chirurgia, Università di Napoli “Federico II,” Napoli, Italia
    Search for more papers by this author
  • Antonio Vetrani M.D.,

    1. Dipartimento di Anatomia Patologica e Citopatologia, Facoltà di Medicina e Chirurgia, Università di Napoli “Federico II,” Napoli, Italia
    Search for more papers by this author
  • Lucio Palombini M.D.

    Corresponding author
    1. Dipartimento di Anatomia Patologica e Citopatologia, Facoltà di Medicina e Chirurgia, Università di Napoli “Federico II,” Napoli, Italia
    • Dipartimento di Anatomia Patologica, Facoltà di Medicina e Chirurgia, Università di Napoli “Federico II,” Via Pansini no. 5, 80131 Napoli, Italia
    Search for more papers by this author
    • Fax: (011) 39 0817463679


Abstract

BACKGROUND

Flow cytometry (FC) is a useful adjunct to fine-needle aspiration cytology (FNC) in evaluating lymphoproliferative disorders. The authors present a critical review of 307 lymph nodal and extra lymph nodal lymphoproliferative disorders that were diagnosed with FNC and FC.

METHODS

FC was performed over a 4-year period on 185 palpable and 122 impalpable lymph nodal and extra lymph nodal lymphoproliferative processes under ultrasound or computed tomography guidance. FC was performed using the following fluoresceinated antibodies: CD3, CD4/CD8, CD2/CD7/CD3, CD5/CD10/CD19, CD19/κ/λ, FMC7/CD23/CD19, CD38/CD56/CD19, and bcl-2. The series included 15 inadequate, 10 suspicious, and 135 benign reactive hyperplasias (BRHs); 70 primary non-Hodgkin lymphomas (NHLs), and 77 recurrent NHLs (rNHLs). FC/FNC diagnoses of suspicious, NHL, and rNHL were controlled either histologically or clinically or by the interphase fluorescence in situ hybridization demonstration of t(11;14)(q13;q32) in two cases of mantle cell lymphoma. BRHs were controlled by follow-up. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of the FC/FNC diagnoses of NHL, rNHL, and BRH were calculated as well as the identification of specific subtypes among the small- and medium-sized cells.

RESULTS

Statistical analysis showed 93% sensitivity, 100% specificity, 100% PPV, and 91% NPV in NHL, rNHL, and BRH discrimination. The subclassification of small cell and medium-sized NHLs showed 63% sensitivity, 88% specificity, 95% PPV, and 37% NPV.

CONCLUSIONS

FC applied to FNC enhanced the precision of cytologic diagnosis in lymph nodal and extra lymph nodal lymphoproliferative disorders and allowed further subclassification in more than half of the cases, thus avoiding invasive surgical biopsies in many patients. Cancer (Cancer Cytopathol) 2004;102:55–65. © 2003 American Cancer Society.

Immunophenotyping is a fundamental step in the diagnosis of lymph nodal and extra lymph nodal lymphoproliferative disorders. It becomes virtually compulsory when the diagnostic procedure is performed on cytologic material. In recent years, flow cytometry (FC) has proven useful in the evaluation of mainly lymph node lymphoproliferative disorders on samples obtained by surgical specimens or fine-needle cytology (FNC);1–12 therefore, FC applied to FNC has become diffused widely, often replacing classic immunocytochemistry on cytospins.1, 2, 13–15 The possibility of applying a complete panel of antibodies and diagnostic algorithms is the most appreciated advantage of the technique; 4, 6, 8 consequently, many reports stress the high sensitivity in detecting non-Hodgkin lymphoma (NHL) through the combined application of FNC and FC.1–15 The most enthusiastic reports hypothesize that, in the near future, histology will not be necessary strictly for the diagnosis of NHL. In fact, in many institutions, FNC coupled with ancillary techniques like FC routinely has replaced histology in the diagnosis of recurrent NHL (rNHL).1, 2, 4, 8, 10 However, the possibility of a definite cytologic diagnosis even for primary NHL is related to the possibility of a correct subclassification of the single entities, according to the most recent Revised European–American Classification of Lymphoid Neoplasms (the REAL classification system).16, 17 The possibility of identifying a follicular NHL through CD10/CD19 coexpression without microscopic visualization of follicular structures is a stimulating idea; and, in recent years, there have been some reports using FC or immunocytochemistry to exploit the possibility of subclassifying NHL.4, 8, 10, 13, 18, 25–27 The objective of this article was to report our experience concerning the application of FC and FNC to the diagnosis of lymph nodal and extra lymph nodal lymphoproliferative disorders emphasizing technical aspects and the possibility of subclassifying NHL.

MATERIALS AND METHODS

Patients and FNC

FC was used to analyze 307 FNC specimens of lymphoproliferative processes that were obtained between January, 1999 and December, 2002. FNC was performed on 148 palpable lymph node lesions and on 37 palpable extra lymph node lesions in the outpatient clinic of our department. FNC of impalpable and/or deeply located lymph nodes (n = 98 FNCs) and organs (n = 24 FNCs) was performed under ultrasound or computed tomography guidance in the Radiology and Hematology Departments. Extra lymphatic organs included the thyroid (n = 13 cases), parotid (n = 15 cases), breast (n = 4 cases), soft tissue (n = 5 cases), liver (n = 2 cases), small bowel (n = 1 case), and spleen (n = 20 cases); of these organs, to date, only splenic FC/FNCs have been described extensively19; no bone marrow samples were examined primarily. The series accounted for patients with (n = 77 patients) or without (n = 230 patients) a history of lymphoma or other hematologic disorders. At the time of FNC, the diagnostic procedure and its related risks were discussed first with the patients, and their informed consent was obtained; then, the FNC procedure was performed as described previously.18 In all patients, the first pass was used to prepare two traditional smears, the first of which was stained immediately with the Diff-Quik method and evaluated to select smears that were suitable for immunophenotyping. The remaining material left in the hub of the needle was flushed carefully with phosphate-buffered saline solution (PBS) or RPMI and implemented by a second or third pass in smears with scant cellularity. An immediate microscopic evaluation of Diff-Quik-stained smears revealed whether the sample was adequate and whether FC immunophenotyping was required. In these smears, one or more further passes were performed to provide sufficient cells in the suspension. Clinical data and microscopic features also were used to lay out the panel of antibodies to apply in each case. Insufficient smears showed scant cellularity and/or abundant necrosis and blood; in these cases, another pass was performed and was processed in the same manner. Cases concerning unequivocal clinical and microscopic reactive processes, Hodgkin disease, and metastases were not processed; and, in these cases, additional FNC samples were used for immunocytochemical stains on cytocentrifuged specimens. In selected cases, with the authorization of the patient, another pass was used for storing cells that were suitable for further ancillary techniques. In three patients, in particular, immunocytochemistry was performed on cytospins for cyclin-D1 (Dakopatts, Glosstrup, Denmark) using a three-step, peroxidase-based method but, similar to another experience,18 without obtaining significant results; moreover, in the last two patients, who had mantle cell lymphoma (MCL), stored cells were used for the detection of t(11;14)(q13;q32) by interphase fluorescence in situ hybridization (FISH). Dual-color FISH analysis was performed using an IGH-CCND1 probe set (Vysis Inc., Downer's Grove, IL), as described elsewhere.20, 21

FC

FNC specimens were processed within 2 hours, and the cell suspensions were washed twice by centrifugation for 5 minutes at 2500 revolutions per minute, removing the supernatant fluid and adding 400 μL of PBS. The final suspension was divided into four or more tubes when sufficient cells were available. Starting from the first 60 cases, 1 or 2 tubes of enriched cell suspension were stored until the end of the procedure so that additional material would be available if unsatisfactory results were achieved or if additional tests were needed. Samples were then incubated for 15 minutes in the dark with 10 μL of the following basic combinations of fluorescein isothiocyanate (FITC), phycoerythrin (PE), and perdin chlorophyll protein (PerCP) antibodies: CD3, CD4/CD8, CD2/CD7/CD3, CD5/CD10/CD19, CD19/κ/λ, FMC7/CD23/CD19, CD38/CD56/CD19, and bcl-2. All antibodies were purchased from Becton Dickinson (San Jose, CA) except for bcl-2, which was purchased from Pharmingen; all antibodies used and the corresponding conjugated fluorochromes are listed in Table 1. After incubation, red blood cells were lysed with ammonium chloride lysing solution (diluted to 10%) for 15 minutes and then washed. When small fragments still were present, the suspension was filtered through 50-μm filters; finally, an equal part of 1% paraformaldehyde was added to each tube for cell fixation. In cases in which intracytoplasmic antigens were to be detected, such as bcl-2 or light chains when the routine technique failed, cells were suspended in permeabilizing solution and incubated for 30 minutes in the dark; FC was then performed using a three-color analysis technique on a Becton Dickinson FACS scan, as described previously.19 For data evaluation, an antibody was considered expressed when a minimum of 20% of the gated cells were positive; for the light-chain evaluation, κ:λ ratios > than 4:1 or > 1:2 were considered definite evidence of monoclonality.3, 4 In cases of equivocal results or technical difficulties, residual material in the tubes was suitable for further analysis within 24 hours; and aliquoted, stored tubes also were available for further phenotyping. FC data were reviewed along with the cytologic features, and both were incorporated into the final report. A definitive diagnosis of NHL or rNHL was reached by evaluating the combination of cytologic features, light-chain restriction, or abnormal expression of specific T antigens, such as CD2/CD7/CD3, CD4/CD8, and CD56. The cytologic features were then matched to the different expression and coexpression of CD5, CD10, CD23, FMC7, CD38, and CD56 in different combined phenotypes to classify the specific subtype, when possible. According to other reports,1, 2, 4, 6, 8, 10–12, 24, 25 the following phenotypes, coupled with proper microscopic features, were considered specific for the corresponding reported NHL: CD5/CD19+, CD19/23+ phenotypes were considered specific for small lymphocytic lymphoma/chronic lymphatic leukemia (SLL/CLL) (Fig. 1); CD5/CD19+, CD23−, FMC7/CD19+ phenotypes were considered specific for MCL (Fig. 2); and CD5/19−, CD23/19−, CD19/CD10+ phenotypes were considered specific for follicle center cell lymphoma (FCCL) (Fig. 3). Moreover, CD19/CD38+ phenotypes were considered specific for lymphoplasmacytoid lymphoma (LpcL); and CD10/CD19+ phenotypes, even without light-chain expression, in the proper clinical-cytologic setting, were considered specific for small noncleaved cell lymphoma (SNCL) (Fig. 4). The phenotype CD19+, CD19/CD5−, CD19/CD10−, CD19/CD23− CD38/CD56− was observed the most frequently in small cell and medium-sized NHL but was considered specific only for marginal zone/mucosa-associated lymphoid tissue (MALT) B-cell lymphomas (MZL/MALT) with specific cytologic features when it was observed in extra lymphatic organs, such as the parotid, small bowel (Fig. 5), and thyroid; in all the other cases of small cell and medium-sized NHL with the CD19+, CD19/CD5−, CD19/CD10−, CD19/CD23− phenotype, a diagnosis of NHL not otherwise specified (NHL NOS) was performed. The phenotype CD5+, CD19−, CD4+/CD8−, CD2/CD3−, and CD7+, or CD2/CD3+ and CD7−, in a proper cytologic setting, was considered specific for peripheral T-cell lymphoma (PTCL) (Fig. 6).19, 21 The phenotype CD3+, CD56+, CD19−, CD10−/CD23− was specific for natural killer cell lymphoma.22, 25 In large cell and anaplastic NHL, the phenotype CD19+, CD19/CD5−, CD19/CD10−, CD19/CD23−, CD38/CD56− was observed the most frequently. In these cases, the diagnoses were reached on the basis of cytologic features, CD19 positivity, and light-chain restriction. The diagnostic FC algorithm for small cell and medium-sized NHL is summarized in Table 2. FC/FNC diagnoses were controlled by histology, clinical control, or follow-up. Clinical control consisted of lactate dehydrogenase and β2-microglobulin assay evaluation, ultrasound echo-color Doppler, Gallium 67 scan for supradiaphragmatic localizations, and positron emission tomography; the latter was preferred in patients with subdiaphragmatic localizations. Therefore, FC/FNC diagnoses were controlled as follows: 44 patients with NHL and 10 patients with suspicious results were controlled histologically; 24 patients with NHL were controlled clinically alone or were confirmed by the FISH demonstration of t(11;14)(q13;q32) in 2 patients with MCL. The 24 clinically controlled cases concerned patients with Stage IV disease and/or older patients, patients with bleeding diathesis, or patients with deeply located lymph nodes who were treated without histologic examination. Patients with rNHL were controlled clinically, and the specific FC/FNC subtypes were compared with the former histologic diagnoses. One hundred fifteen of 135 patients with BRH were controlled by follow-up; 20 patients with BRH and 1 patient with inadequate results were lost to follow-up control. Data concerning the whole series are summarized in Table 3.

Table 1. Primary Antibodies Used for Flow Cytometry
AntibodySourceDilutionIncubation
  • PE: phycoerythrin; RT: room temperature; FITC: fluorescein isothiocyanate; PERCP: perdin chlorophyll protein.

  • a

    One hundred milliliters of cell suspension.

  • b

    Associated with Becton Dickinson.

CD3 (PE)Becton Dickinson, (San Jose, CA)1 μL:100 mLa15 minutes in the dark, RT
CD4 (FITC)/CD8 (PE)Becton Dickinson1 μL:100 mLa15 minutes in the dark, RT
CD2 (FITC)/CD7 (PE)/CD3 (PERCP)Becton Dickinson1 μL:100 mLa15 minutes in the dark, RT
CD5(FITC)/CD10(PE)/CD19(PERCP)Becton Dickinson1 μL:100 mLa15 minutes in the dark, RT
FMC7(FITC)/CD23(PE)/CD19(PERCP)Becton Dickinson1 μL:100 mLa15 minutes in the dark, RT
CD38(FITC)/CD56(PE)/CD19(PERCP)Becton Dickinson1 μL:100 mLa15 minutes in the dark, RT
CD19 (FITC)/K (PE)/λ (PERCP)Becton Dickinson1 μL:100 mLa3 minutes in the dark, 4° C
Bcl-2 (FITC)Pharmigenb1 μL:100 mLa3 minutes in the dark, 4° C
Figure 1.

Fine-needle aspiration cytology (FNC) and flow cytometric (FC) analysis of small lymphocytic lymphoma/chronic lymphatic leukemia. (A) FNC smear shows an atypical, monomorphous population of small, immature lymphocytes with granular chromatin and inconspicuous nucleoli (Diff-Quik stain; original magnification, × 430). (B) FC coexpression of CD19/CD5 is seen in the upper right quadrant. (C) FC coexpression of CD19/CD23 is seen in the upper right quadrant. (D) FC clonal population of B-cells showing κ light-chain restriction in lower right quadrant. PerCP-Cy5.5: perdin chlorophyll protein-indocarbocyanine; FITC: fluorescein isothiocyanate; PE: phycoerythrin.

Figure 2.

Fine-needle aspiration cytology (FNC) and flow cytometric (FC) analysis of mantle cell lymphoma. (A) FNC smear shows medium-sized lymphoid cells with irregularly shaped and cleaved nuclei (Diff-Quik stain; original magnification, × 430). (B) FC coexpression of CD19/CD5 is seen in the upper right quadrant. (C) FC coexpression of CD19/FMC7 is seen in the upper right quadrant. (D) FC λ light-chain restriction is seen in the upper left quadrant. (E) Fluorescence in situ hybridization evidence of t(11;14)(q13;q32). Note the 2 single and 1 double-fusion signals (yellow) in which the green signals (11) overlap the orange signals (14). PerCP-Cy5.5: perdin chlorophyll protein-indocarbocyanine; FITC: fluorescein isothiocyanate; PE: phycoerythrin.

Figure 3.

Fine-needle aspiration cytology (FNC) and flow cytometric (FC) analysis of follicle center cell lymphoma (FCCL) non-Hodgkin lymphoma. (A) FNC smear of a small-cell FCCL shows a population of small lymphocytes with dense chromatin and inconspicuous nucleoli (Diff-Quick stain; original magnification, × 430). (B) FCCLs show CD19/CD10 coexpression of the cell populations. (C) FC clonal population of B-lymphocytes showing κ light-chain restriction in the lower right quadrant. (D) Histogram of bcl-2 expression: The threshold of positivity is between the second and the fourth decades (M1); here, > 90% of lymphocytes show positivity in the second decade (101–102) corresponding to the bcl-2 fluorescein isothiocyanate (FITC) channel of fluorescence. PerCP-Cy5.5: perdin chlorophyll protein-indocarbocyanine; PE: phycoerythrin.

Figure 4.

Fine-needle aspiration cytology (FNC) and flow cytometric (FC) analysis of small noncleaved cell lymphoma. (A) This FNC smear shows a monomorphic population of small, undifferentiated lymphoid cells with coarse chromatin and frequent mitoses (Diff-Quik stain; original magnification, × 430). (B) FC analysis shows a large B-cell population with CD19/CD10 coexpression. Light chains were not expressed. PerCP-Cy5.5: perdin chlorophyll protein-indocarbocyanine; PE: phycoerythrin.

Figure 5.

Fine-needle aspiration cytology (FNC) and flow cytometric (FC) analysis of marginal zone/mucosa-associated lymphoid tissue B-cell lymphomas. (A) Ultrasound imaging of extremely thick intestinal wall; note the tip of needle. (B) FNC smear shows a monomorphic population of small lymphocytes with compact or coarse chromatin; some interspersed large cells also are present (Diff-Quik stain; original magnification, × 430). (C0 FC shows CD19 positive (CD19+) lymphocytes in the upper left quadrant and CD5+, reactive T-lymphocytes in the lower right quadrant. (D) FC clonal population of B-lymphocytes showing λ light-chain restriction in the upper left quadrant. PerCP-Cy5.5: perdin chlorophyll protein-indocarbocyanine; FITC: fluorescein isothiocyanate; PE: phycoerythrin.

Figure 6.

Fine-needle aspiration cytology (FNC) and flow cytometric (FC) analysis of peripheral T-cell lymphoma. (A) FNC smear shows a relatively polymorphous population of medium-sized lymphoid cells with coarse chromatin and irregular shape (Diff-Quik stain; original magnification, × 430). (B) FC shows a large T-cell population that is positive for CD5 (CD5+); CD19+ cells are absent. (C) FC shows a T-cell population that is CD2+ and CD7 negative. PerCP-Cy5.5: perdin chlorophyll protein-indocarbocyanine; FITC: fluorescein isothiocyanate; PE: phycoerythrin.

Table 2. Flow Cytometric Correlation between Phenotypes and Small Cell and Medium-Sized Non-Hodgkin Lymphoma Subtypes
PhenotypeLpcLSLL/CLLMCLFCCLMZL/MALTNKLPTCLSNCL
  1. LpcL: lymphoplasmacytoid lymphoma; SLL/CLL: small lymphocytic lymphoma/chronic lymphatic leukemia; MCL: mantle cell lymphoma; FCCL: follicle center cell lymphoma; MZL/MALT: marginal zone lymphoma/mucosa-associated lymphoid tissue B-cell lymphoma; NKL: natural killer non-Hodgkin lymphoma; PTCL: peripheral T-cell lymphoma; SNCL: small noncleaved cell lymphoma; −: negative; +: positive.

κ:λ > 4:1 or < 1:2++++++/−
CD19/CD5+++
CD19/CD10+++/−+/−
CD19/CD23++/−+
CD19/FMC7++/−++/−
CD19/CD38++
bcl-2++/−+/−+/−++/−
CD4+/CD8− or CD4−/CD8++
CD5+/CD19−+++/−
CD2/CD3+CD7− or CD2/CD7+CD3−+
CD3+ CD56+/CD19−+
Table 3. Correlation between Fine-Needle Aspiration Cytology/Flow Cytometry Diagnoses and Histology, Clinical Control, and Follow-Up
FNC/FCNo. of patients (%)Histology, clinical control, and follow-up
NHLrNHLBRHHDMetastasisLost
  1. FNC/FC: fine-needle aspiration cytology/flow cytometry; NHL: non-Hodgkin lymphoma; rNHL: recurrent non-Hodgkin lymphoma; BRH: benign reactive hyperplasia; HD: Hodgkin disease (Hodgkin lymphoma).

Inadequate5 (5)165021
rNHL77 (25)0770000
NHL70 (23)7000000
Suspicious10 (3)243100
BRH135 (44)001150020
Total307 (100)73871231221

Statistical analysis was performed with two goals: The first goal was to assess the sensitivity, specificity, PPV, and NPV of combined FNC/FC in the diagnosis of NHL and rNHL considering both results truly positive and considering BRH results truly negative; because 10 patients with suspicious results were referred to histology for a definite diagnosis, we considered the NHL results in this group (6 of 10 patients) false-negative and considered the BRH results (4 of 10 patients) false-positive. The second goal was to assess sensitivity, specificity, PPV, and NPV of FNC and FC in the subclassification of NHL according to the REAL classification. In this analysis, we considered only small cell and medium-sized NHLs, because large cell and anaplastic NHLs are a quite heterogeneous group of NHLs that lack specific phenotypes. From this perspective, for cases in which a definite subclassification was achieved by combining cytologic features and specific phenotypes, 71 of 147 results were considered truly positive, whereas 55 of 147 diagnoses of NHL or rNHL without further subclassification were considered false-negative. Twenty-one of 147 large cell and anaplastic NHLs were considered truly negative.

RESULTS

The anatomic distribution of palpable and impalpable lymph nodes and other organs with lymphoproliferative disorders is summarized in Table 4. Combined FC and FNC provided a definitive diagnosis of BRH in 135 cases (44%), primary NHL in 70 cases (23%), and rNHL in 77 cases (25%). First, NHL and rNHL were divided cytologically into small cell and medium-sized NHLs (115 cases) and large cell or anaplastic NHLs (21 cases) according to currently accepted cytologic criteria.6, 12, 23 The clinical control confirmed the diagnoses of rNHL; histology (44 cases), clinical control (24 cases), and t(11;14)(q13;q32) (2 cases) confirmed the 70 primary NHL diagnoses. The t(11;14)(q13;q32) was detected by the contact or juxtaposing of IGH locus probe (spectrum green) and CCND1 locus probe (spectrum orange) resulting in 1 or 2 yellow fusion signals in 60% of the cells (Fig. 2E). Follow-up confirmed the FN/FNC diagnoses of BRH (115 of 135 cases); 20 BRHs and 1 inadequate case were lost to follow-up. Fifteen inadequate cases (5%) were represented by scantly cellular FNC and/or extensive hemorrhagic or necrotic material; they concerned organs or deeply located lymph nodes (12 cases) and palpable lymph nodes (3 cases). In these cases, immediate repetition of FNC did not provide sufficient cells, and repetition or excision was required for diagnosis. In 10 suspicious cases (3%), the microscopic features were suggestive only of NHL, and FC had yielded unsatisfactory results and/or had failed to demonstrate a light-chain restriction; in these cases, the subsequent histologic control revealed 6 small cell NHLs and 4 BRHs; no false-positive results were detected. Of 115 cases that were diagnosed cytologically as small cell and medium-sized NHL, the evaluation of different expression and coexpression of CD3, CD2/CD7/CD3, CD4/CD8, CD5, CD10, FMC7, CD23, CD38, and CD56, applied to the cytologic features, suggested a specific subtype in 70 cases. All other cases with negative or equivocal phenotypes were diagnosed only as NHL NOS. Among the specific phenotypes, CD19+, CD10−, CD23−, CD38−, was observed most frequently and was considered specific for MZL/MALT only in patients with primary NHL of the liver (1 patient), parotid (2 patient), and thyroid (1 patient). Among the remaining 55 patients who had this phenotype with a diagnosis of NHL NOS, histologic controls revealed LpcL (2 patients), SLL/CLL (7 patients), MCL (3 patients), FCCL (21 patients), MZL/MALT (11 patients), PTCL (1 patient), SNCL (1 patient), and large cell lymphoma (LCL) (3 patients) (Tables 5, 6). rNHL and 24 NHLs were treated without subsequent histologic controls. Patients who had NHL without a specific phenotype, extra lymph node NHL in which surgery was part of the therapeutic procedure, suspicious results, and some of the patients who had inadequate samples underwent surgical excision and subsequent histologic control. Clinical, cytologic, and cytofluorimetric data of the series are summarized in Tables 3, 5, and 6. Statistical analysis showed 93% sensitivity, 100% specificity, 100% PPV, and 91% NPV in NHL, and BRH discrimination. In the subclassification of small and medium-sized NHL and rNHL, the following values were obtained: 63% sensitivity, 88% specificity, 95% PPV, and 37% NPV.

Table 4. Anatomic Distribution of 307 Palpable and Impalpable, Ultrasound-Guided/Computed Tomography-Guided Lymphoproliferative Processes Diagnosed by Fine-Needle Aspiration Cytology/Flow Cytometry
PalpableNo. of patientsUS/CT guidedNo. of patients
  1. US: ultrasound; CT: computed tomography.

Lymph node148Lymph node98
Thyroid13Spleen20
Parotid15Liver2
Breast4Small bowel1
Soft tissue5Soft tissue1
Total185Total122
Table 5. Correlation between Fine-Needle Aspiration Cytology/Flow Cytometry and Histologic Subtypes of 44 Non-Hodgkin Lymphomas and 10 Suspicious Lymphoproliferative Processes
FNC/FCNo. of patientsSubsequent histology
BRHLpcLSLL/CLLMCLFCCLMZL/MALTPTCLSNCLNKLLCLALHD
  1. FNC/FC: fine-needle aspiration cytology/flow cytometry; BRH: benign reactive hyperplasia; LpcL: lymphoplasmacytoid lymphoma; SLL/LLC: small lymphocytic lymphoma/chronic lymphatic leukemia; MCL: mantle cell lymphoma; FCCL: follicle center cell lymphoma; MZL/MALT: marginal zone lymphoma/mucosa-associated lymphoid tissue B-cell lymphoma; PTCL: peripheral T-cell lymphoma; SNCL: small noncleaved cell lymphoma; NKL: natural killer lymphoma; LCL: large cell lymphoma; AL: anaplastic lymphoma; HD: Hodgkin disease (Hodgkin lymphoma); NHL, NOS: non-Hodgkin lymphoma, not otherwise specified.

Suspicious103   321    1
LpcL2 2          
SLL/LLC6  6         
MCL3   3        
FCCL6    6       
MZL/MALT2     2      
PTCL1      1     
SNCL2       2    
NKL1        1   
NHL, NOS13 1113411 1  
LCL5         5  
AL3          3 
Total543374128331631
Table 6. Correlation between Fine-Needle Aspiration Cytology/Flow Cytometry and Histological Subtypes of 77 Patients with Recurrent Non-Hodgkin Lymphoma and 24 Patients with Non-Hodgkin Lymphoma
FNC/FCNo of patientsHistologic or FISH resultsClinical control
LpcLSLL/CLLMCLaFCCLMZL/MALTPTCLSNCLNKLLCLAL
  • FISH: fluorescence in situ hybridization; FNC/FC: fine-needle aspiration cytology/flow cytometry; LpcL: lymphoplasmacytoid lymphoma; SLL/LLC: small lymphocytic lymphoma/chronic lymphatic leukemia; MCL: mantle cell lymphoma; FCCL: follicle center cell lymphoma; MZL/MALT: marginal zone lymphoma/mucosa-associated lymphoid tissue B-cell lymphoma; PTCL: peripheral T-cell lymphoma; SNCL: small noncleaved cell lymphoma; NKL: natural killer lymphoma; LCL: large cell lymphoma, AL: anaplastic lymphoma; NHL, NOS: non-Hodgkin lymphoma, not otherwise specified.

  • a

    Two cases of mantle cell lymphoma were confirmed by fluorescence in situ hybridization, which demonstrated t(11;14)(q13;q32) on cytologic samples.

LpcL44          
SLL/LLC14 12        2
MCL7  7        
FCCL8   14      4
MZL/MALT2    2     2
PTCL2     1    1
SNCL1      1    
NKL          
NHL, NOS42162187   2 6
LCL11        1 10
AL2         11
Total1035189329113124

DISCUSSION

FNC has been viewed with skepticism when applied to lymphoproliferative disorders, because the lack of information regarding histologic architecture always has been considered a limitation in the diagnostic interpretation of smears. Starting in the 1980s, in the light of the increasing genetic and molecular knowledge of NHLs, the corresponding morphologic classifications progressively have focused on phenotypes instead of histologic features; consequently, the so-called ancillary techniques (immunocytochemistry, cytogenetics, and molecular biology) have gained importance in the diagnosis of lymphoproliferative disorders.

In fact the latest REAL classification of lymphomas16, 17 enhances immunophenotyping and cytologic features of lymphoid populations rather than the conventional nodular or diffuse growth patterns that characterized previous NHL classifications. From this perspective, FNC has gained credibility, mainly when contextual immunophenotyping of the cell population is performed. The main advantages of FC immunophenotyping are the timely application of a complete panel of antibodies and contextual triage of fluorosceinated antibodies with detection of specific antibody expression and coexpression patterns that allow the application of diagnostic algorithms.1, 4–10 Although these advantages cannot be overemphasized, many enthusiastic reports were based on different kinds of samples: FNC,1, 2, 5, 7, 8, 10, 22, 24 thick-needle or surgical biopsies with manual disaggregation of tissue fragments,4, 9, 23, 28, 29 or the evaluation of FC and immunocytochemical data together.6, 24, 30 Moreover, in some experiences reported in the literature, cytopathologists performed both FNC and FC12, 16, 22; in other reports, FC was performed in different laboratories by immunohematologists.1, 2, 10, 13, 28 These differences have generated heterogeneity of sample types and results. Furthermore, recent articles have reported false-negative diagnoses by FC in large B-cell NHLs31, 32 and cases that lack expression of surface light chains in NHL.33 These reports suggest that technical aspects of FC other than interpretation of the results and microscopic data still are fundamental steps in FNC-FC diagnosis of lymphoproliferative disorders.

In this study, we tested FNC samples that were associated uniquely with FC to evaluate the real impact of both techniques on the diagnosis of lymphoproliferative processes; moreover, because we performed both FNC and FC, we were aware of technical and/or microscopic problems immediately. From a technical point of view, in our experience, paraformaldehyde fixation allows sample storage for a couple of days to postpone or repeat evaluation in equivocal cases. We also found it very useful to preserve some of the cells in one or two tubes for further immunophenotyping, mainly to detect intracytoplasmic light chains in cases in which routine procedure had not detected light-chain expression or for adding other antibodies to the basic panel.

Despite this foresightedness, we still had 15 inadequate cases (5%) and 10 suspicious cases (3%) in which the patients had to be referred to surgical excision for histologic examination. Inadequate and suspicious case rates in the literature range from 4% to 30%4, 7, 8, 10, 22; scant cellularity, necrosis, and excessive blood contamination are the most frequent causes of inadequate cases; as reported earlier, location in splanchnic organs or deeply located lymph nodes also contributed to this group because of the difficulty in performing two or more passes, as in FNCs of the spleen.21 Moreover, most inadequate cases were obtained in the first year of the examiners' experience and decreased progressively with time. Improvement was due mainly to more careful examination of cell suspensions and Diff-Quik-stained smears, rejecting samples that would be inadequate or useless for FC evaluation. Suspicious cases were due mainly to the lack of demonstration of light-chain restriction in samples that were suggestive cytologically for NHL. The lack of expression of surface immunoglobulin light chains in B-cell NHL is an unusual phenomenon.30 In the current study, it occurred 12 times; intracytoplasmatic immunoglobulins were detected in 5 of these cases using the aliquoted tubes; however, in the other 7 cases, even this method failed to detect immunoglobulins. Five of those seven cases were diagnosed as suspicious, and two were diagnosed as SNCL because of their cytologic features and pathologic coexpression of CD19/CD10.

In equivocal or incomplete cases, FC also may play a role in a comprehensive evaluation of the phenotype and cytologic features. In fact, it has been reported that CD19/CD5 coexpression is not a feature of benign lymphoid proliferations, suggesting involvement by SLL/CLL or MCL,4 and that CD20 positivity in excess of 85% and CD19/10 coexpression > 18%, independently, are diagnostic of B-cell lymphoma.22 In our experience, CD19/CD5 coexpression also has been observed in NHL only (SLL/CLL or MCL). CD19/CD10 coexpression was observed in FCCLs and SNCLs as well as in some cases of BRH (although it was < 20%). Therefore, we considered CD19/CD10 coexpression diagnostic for FCCL only in cases with light-chain restriction, with the exception of 2 SNCLs in which cytologic features and clinical settings clearly indicated a lymphoproliferative process (Fig. 4). Combined microscopic features and FC data, therefore, also may allow the diagnosis of incomplete cases, such as large cell B-NHL with equivocal FC or MCL or SLL/CLL without light-chain restriction. Therefore, we believe that the correlation among clinical data, FNC, microscopic features, and FC is most important for resolving some of these cases. In fact, in our department, cytopathologists perform FNC, evaluate the smears, decide on the antibody panel, and participate in FC acquisition and data evaluation. Inadequate and suspicious cases probably are the price paid, in the beginning, for the limited experience in FC; however, we believe that that price is returned through a more comprehensive evaluation of the clinical, cytologic, and phenotypic data.

With regard to the efficiency of the method, we obtained very high sensitivity (93%), specificity (100%), PPV (100%), and NPV (91%) by dividing diagnostic categories into suspicious, BRH, NHL, and rNHL. The results were quite similar to those from other large series4, 8, 10; all of those groups obtained excellent results, probably because of the synergic effect of the two techniques. In fact, as observed by Liu et al.,2 when FNC is limited (e.g., in differentiating between low-grade NHL and BRH), FC is helpful; and, when FC is limited (e.g., in case of high-grade, large cell NHL), FNC often is useful.

With regard to the diagnosis of specific subtypes of small cell and medium-sized NHLs through cytologic features and specific phenotypes, we were able to subclassify 70 of 115 cases (61%); in fact, as reported above, the remaining 55 of 115 small cell and medium-sized NHLs were diagnosed as NHL NOS on the basis of microscopic features and FC light-chain restriction or abnormal expression of T-cell antigens. Histologic controls, when available, revealed 2 LpcLs, 7 SLL/CLLs, 3 MCLs, 21 FCCLs, 10 MZL/MALTs, 1 PTCL, 1 SNCL, and 3 LCLs (Tables 5, 6). The histologic distribution of these cytologic NHLs NOS was representative of the incidence of the corresponding phenotypes in the series, except for MZL/MALTs, which were identified less frequently because of their specific negative phenotype. These results are quite similar to others reported in literature8, 10 and lower compared with the results of Liu et al.2 and Kaleem et al.4 Technical problems and partial lymph node involvement probably were the causes of these limitations, but the lack or equivocal coexpression of specific antibodies also limited the subclassification of NHL. In fact, considering the specific coexpression of antibodies in the whole series, we found CD19/CD5 coexpression ≥ 20%, the only truly pathologic level observed in SLL/CLLs or MCLs only, whereas CD19/CD5 coexpression was absent in 2 subsequently proven MCLs. Conversely, no BRHs showed CD19/CD5 coexpression. The coexpression of CD10 and CD23 with CD19 was lacking in four histologically proven FCCLs and in three SLL/CLLs, respectively. Moreover, CD19/CD10 coexpression (< 20%) was observed in 12 BRHs (Fig. 7), probably corresponding to florid follicular hyperplasia; and CD19/CD23 and CD19/FMC7 coexpression was observed in 18 BRHs and 20 BRHs, respectively. In PTCLs, as reported above, the cytologic features, the reduction or absence of B-cell antigens, and the lack of coexpression of T-cell markers were the hallmarks of this heterogeneous group of NHLs. On cytologic smears that clearly demonstrated malignancy, the over-expression of T-cell markers was considered specific for PTCL; in less atypical and more cytologically polymorphous cases, the triage of CD2/CD7/CD3 or CD4/CD8 showed a dissociation of the phenotype with expression of only one or two antibodies, lacking the expression of the other phenotype that was included in the panel. These data suggest that subclassification should be performed only when a definite diagnosis of NHL has been reached on the basis of cytologic pattern, light-chain restrictions, or a specific T-cell phenotype.

Figure 7.

Fine-needle aspiration cytology (FNC) and flow cytometric (FC) analysis of benign reactive hyperplasias. (A) FNC smear shows a relatively polymorphous cell population of small, medium, and large lymphoid cells (Diff-Quik stain; original magnification × 430). (B) FC analysis shows two cell populations: CD5 positive (CD5+) and CD19+. (C) CD19/CD10 coexpression probably corresponding to florid follicular hyperplasia. (D) FC polyclonal population of B-lymphocytes showing balanced κ and λ light-chain expression. (E) CD2/CD7 coexpression of reactive T-lymphocytes. PerCP-Cy5.5: perdin chlorophyll protein-indocarbocyanine; FITC: fluorescein isothiocyanate; PE: phycoerythrin.

In conclusion, as foreseen by Zaijcek,34 FNC offers vital cells to FC suitable for an easy and timely phenotyping, allowing a quick and sensitive diagnosis of lymphoproliferative processes thanks to a highly sensitive light-chain assessment and T-cell marker evaluation. This technique is so accurate in diagnosing and subclassifying NHL that, in the near future, histologic evaluation probably will no longer be necessary for the diagnosis and follow-up of NHL, thus avoiding useless and expensive surgical biopsies.

Acknowledgements

The authors thank Dr. Antonino Iaccarino, Dr. Carmela Frangella, and Dr. Maria Russo for technical assistance and Dr. Diana Roberts for reviewing the article.

Ancillary