Tissue flow cytometry immunophenotyping in the diagnosis and classification of non-Hodgkin's lymphomas: A retrospective evaluation of 1,792 cases

Authors


  • How to cite this article: Demurtas A, Stacchini A, Aliberti S, Chiusa L, Chiarle R, Novero D. Tissue flow cytometry immunophenotyping in the diagnosis and classification of non-Hodgkin's lymphomas: A retrospective evaluation of 1,792 cases. Cytometry Part B 2013; 84B: 82–95.

Abstract

A retrospective analysis of 1,792 solid tissues suggestive of lymphoma, submitted over a 12-year period, was carried out and flow cytometry (FC) results were compared with histologic findings. The final histologic diagnosis of cases documented in this report is as follows: 1,270 non-Hodgkin's lymphomas (NHL); 17 composite lymphomas; four NHL plus carcinomas; five post-transplant lymphoproliferative disorders; 105 Hodgkin's lymphomas (HL); eight acute leukemias; 42 tissue cancers; and 341 non-neoplastic diseases. A strong correlation between morphology and FC data was observed among hematological malignancies (1,268/1,304, 97.2%) with the exception of HL. Among B-NHL, FC detection of clonally restricted B-cell allowed the identification of lymphomas that were not histologically clear and the differential diagnosis between follicular lymphoma and reactive hyperplasia. A high correlation level (r = 0.83; P < 0.0001) was obtained in comparing proliferation results obtained by FC and immunohistochemistry. Among T-NHL, FC detection of an aberrant phenotype direct histologic diagnosis in cases having less than 20% of neoplastic cells. In nine cases, FC suggested the need to evaluate a neoplastic population, not morphologically evident. Results show that FC routinely performed on tissue samples suspected of lymphomas is a fundamental adjunct to morphology in the diagnosis of NHL and may enhance the performance of the histologic evaluation so as to achieve the final diagnosis. To the best of our knowledge, this is the first report in the literature of a wide series of tissues also studied by FC. © 2013 International Clinical Cytometry Society

INTRODUCTION

The WHO classification of tumors of hematopoietic and lymphoid tissues recommends a multiparametric approach which uses morphological, immunophenotypic, genetic, and clinical features to define diseases (1). The phenotypic characterization of tissues performed by immunohistochemistry (IHC) on histological sections of frozen or fixed tissues is considered a fundamental tool for the achievement of the final diagnosis (2). The main advantages of IHC are the possibility to correlate antigen expression with cell morphology and tissue architecture and the ability to detect a relatively low number of neoplastic cells, such as in Hodgkin's lymphoma (HL) or T-cell-rich large B-cell lymphoma (TCRBCL). Some antibodies may be well evaluated in paraffin tissue (e.g., Bcl-2, Bcl-6, Cyclin-D1, or ALK-1) and they can be very useful in the differential diagnosis of specific lymphoma types. The scarce resolution of light-chain restriction as a marker of B-cell clonality obtained by using IHC techniques, or the identification of an aberrant T-cell phenotype, particularly when there is an abundance of reactive cells, however, makes it difficult for the morphological interpretation to reach the final non-Hodgkin's lymphoma (NHL) diagnosis in certain cases (3).

Flow cytometry (FC) immunophenotyping has become a standard practice in the evaluation and monitoring of patients with hematopoietic neoplasia and it has been proven to be superior to IHC, where the analysis depends on the evaluation of a single or double marker on each tissue section, with a dose of subjectivity in equivocal antigen expressions (4). FC studies are considered a very useful tool for the diagnosis of mature B-cell lymphoid neoplasms and may assist in the diagnosis and classification of mature T- and NK-cell lymphoid neoplasms (5). Although FC is able to provide objective and quantitative results, even on very small samples, within few hours, it is not routinely applied in the evaluation of solid tissues suspected for lymphoma. The aim of this retrospective study was to analyze the contribution of FC immunophenotyping in the detection of neoplastic cells in solid tissues suspicious for NHL.

MATERIALS AND METHODS

A total of 1,792 consecutive excisional biopsies obtained from patients clinically suggestive of lymphoma were collected at our institute, between July 1998 and December 2010. The specimens corresponded to lymph nodes (1,387 cases), spleens (215 cases), abdominal tissues (67 cases), head–neck tissues (52 cases), soft tissues (42 cases), intrathoracic tissues (23 cases), and skin (6 cases). Samples included 1,342 primary diagnosis and 450 recurrent lymphomas.

Histologic and Immunohistochemical Studies

For routine histology, fresh tissues were fixed in Lilly's fixative and morphology evaluated in processed H&E and Giemsa-stained tissue sections. The histopathologic diagnosis had been made by expert oncohematology pathologists (DN and RC), using the WHO classification; 297 cases collected before 2001 were revised according to WHO. Immunostaining was performed as previously described (6). The proliferative index was assessed by immunostaining, counting the percentages of the cell expressing the Ki-67 antigen (clone MIB1; Dako, Glostrup, Denmark), on total cells in lymphoma cases with diffuse growth patterns, on nodules or follicules in cases having a follicular pattern.

Multiparameter FC

FC analysis was performed on cases selected on the basis of clinical NHL suspicion, of the pathologist's macroscopic evaluation, and of cytologic analysis on touch imprints. All the cases having one or more of the previous NHL-suggestive features underwent FC evaluation. Generally, a fresh tissue piece (0.5–2 cm3) was allocated on saline solution and quickly mechanically disaggregated with the BD Medimachine System (BD Biosciences, San Josè, CA). Mononuclear cell suspension was obtained by the Ficoll-Hypaque density gradient method and resuspended in RPMI 1640 medium with 10% fetal calf serum. Enzymatic techniques were performed only in the case of skin tissue samples as described in Ref. (7). A cytocentrifuge slide was performed and stained with May-Grünwald-Giemsa for morphologic review in all cases. Viability assessment was determined manually using Trypan Blue until 2009 or by FC using the fluorescent dye 7AAD. Data acquisition was performed on a FACSCalibur flow cytometer (BD Biosciences) and analyzed by CellQuest Pro software until 2008, or afterward on FACSCanto II (BD Biosciences) using FACSDiva software. Sequential panels were used (Table 1): first, we performed a screening panel in order to detect the presence of lymphoma; if the lymphoma was detected, an additional panel followed to narrow the differential diagnosis according to B/T-cell lineage (8). If epithelial tumors were of concern, a tube containing monoclonal antibody anti-cytokeratin was added. Myeloid or lymphoid additional markers were used if blast cells were identified by low SSC and weak to intermediate expression of CD45 distribution. For detection of myeloma, the plasma cells were identified by their reactivity pattern of CD138 and CD38. Detection of cytoplasmic/nuclear antigens (e.g., cytoplasmic immunoglobulin, cytoplasmic CD3, Bcl-2, and TdT) was performed by using a fixing-permeabilizing agent (Fix & Perm; Invitrogen, Camarillo, CA), according to the manufacturer's instructions. Detection of the nuclear antigen Ki-67 was performed by a modified Fix & Perm procedure, adding 400 μl cold methanol after reagent A and performing a 10-min incubation at 4°C before reagent B. A minimum of 20,000 events were collected for each sample and dead cells were gated out of the analysis. Multiple gating strategies were utilized to define neoplastic populations, for example forward-angle light scatter (FSC) versus right-angle light scatter (SSC), CD45 versus SSC, CD19 or CD3 versus SSC or FSC, or any other useful parameter (Figs 1 and 2). If an abnormal combination of antigen expression was found, back-gating was used in the attempt to demonstrate clustering of the cells as regards light scatter characteristics. A small-cell population was defined as complete overlap of the neoplastic population with the normal T-cell population on a forward scatter histogram. A large-cell population was defined as a complete shift of the neoplastic population from the normal T-cell population without overlap. If the neoplastic population exhibited a significant shift but substantial overlap with the normal T-cell population, the cells were defined as intermediate. A pattern of “normality” was established on the basis of the presence of a mixture of T- and B-cells without evidence of monoclonality or aberrant immunophenotype (9). For B-cell data, the normal range of κ : λ ratio was previously determined in the laboratory to be >0.5:1 or <3:1. In select cases, B-cell clonality was determined by the quantification of κ and λ chains on CD10- or CD5-positive B cells. The absence of surface immunoglobulin (sIg) light-chain restriction in a mature (CD20+) B-cell proliferation has also been used as a surrogate marker to diagnose B-cell lymphoma (10). For T-cell analysis, an abnormal pattern of CD4 and CD8 expression (e.g., coexpression of both these antigens, lack of both of them, or altered CD4/CD8 ratio), lack of expected T-cell markers (e.g., CD2, CD5, CD7), aberrant expression (e.g., CD10), or staining intensity for any T-cell marker was considered suggestive of T-cell neoplasia (11, 12). For Bcl-2 expression, the normal ratio of mean fluorescence intensity of normal B-cells versus T-cells in reactive lymph nodes was previously determined in the laboratory to be 1.21 ± 0.18; an overexpression was considered when a ratio value is greater than or equal to the mean +2SD (standard deviation) was detected. The FC proliferation value was usually assessed by evaluating Ki-67 on all cells, on large cells when clearly distinguishable from the small ones. Whenever possible, the entire procedure was routinely completed in less than 3 h. All cases were reviewed and verified independently by two experienced flow cytometrists (AS and AD) involved in hematologic malignancy diagnosis, unaware of morphologic results. An FC report was performed for each sample using the recommendations of the 2006 Bethesda Consensus Conference (13) and incorporated into the histology report.

Figure 1.

Example of FC gating strategies applied to a case of TCRLBCL in a paratracheal lymph node. Gating on CD20+/CD45bright lymphocytes (red) shows a poorly represented neoplastic B-population displaying medium-large size and κmedium light-chain restriction. Polyclonal B-cells (blue) are defined as CD20+/CD45medium lymphocytes; T-lymphocytes with negative CD20 expression are depicted in green. APC-Cy7, allophycocyanin-cyanine 7; FITC, fluorescein isothiocyanate; FSC, forward scatter; PE, phycoerytrin; PerCP-Cy5.5, peridinin chlorophyll protein–cyanine 5.5; SSC, side scatter; TCRLBCL, T-cell/histiocyte-rich large B-cell lymphoma. [Color figure can be viewed in the online issue which is available at wileyonlinelibrary.com]

Figure 2.

Flow cytometric analysis of nodal MZL detected in an inguinal lymph-node. Neoplastic B-cells show a partial CD10 expression on both small-size (green) and medium-size (red) cells with κ light-chain restriction. APC, allophycocyanin; PE, phycoerytrin; FITC, fluorescein isothiocyanate; FSC, forward scatter; MZL, Marginal zone lymphoma; PC5, phycoerythrin-cyanin 5. [Color figure can be viewed in the online issue which is available at wileyonlinelibrary.com]

Table 1. Monoclonal Antibodies Used in the Flow Cytometry Study
  1. s, surface; c, cytoplasmic; BD, Becton Dickinson San Jose, CA; Coulter Immunology, Hialeah, FL; Caltag, Burlingame, CA; Biosource, Camarillo, CA; Dako; Glostrup, Denmark; IQP, Groningen, The Nederlands; Southern Biotechnologies, Birmingham, UK; Supertechs, Bethesda.

Screening panel (clone, source)sCD3 (SK7, BD), CD4 (SK3, BD), CD5 (BL1A, Coulter), CD8 (SK1, BD), CD10 (MEM 78, Caltag), CD56 (B159, BD), CD19 (SJ25C1, BD), CD20 (B9E9, BD), CD30 (BER-H2, Dako), CD45 (2D1, BD), κ (TB28–2, BD or polyclonal Dako), λ (I-155–2, BD or polyclonal Dako),
Additional panel in case of B-cell lymphomaCD11c (D12, BD), CD22 (B-ly8, IQP), CD23 (EBVC5–5, BD), CD25 (ACT-1, Dako), CD38 (HB7, BD), CD43 (DFT1, Coulter), CD79b (CB3–1, Southern B), CD103 (Bly-7, IQP), CD138 (BB4, IQP), CD200 (MRC OX-104, BD), IgA - IgG - IgD - IgM (polyclonal, Biosource), FMC7 (FMC7,BD), Ki-67 (MIB1, Dako), Bcl-2 (124, Dako), ZAP-70 (1E7.2, Caltag or SBZAP, Coulter)
Additional panel in case of T-cell lymphomaCD1a (VIT6B, Caltag), CD2 (S5.2, BD), CD7 (CD7–6B7, Caltag), CD16 (3G8, Caltag), CD45RA (ALB11, Coulter), CD45RO (UCHL1, Caltag), CD52 (CF1D12, Caltag), CD57 (TB01, Dako), ALK (ALK1, BD), TCR α-β (WT31, BD), TCR γ-δ (11F2, BD), TCR Vβ repertoire (IOtest kit, Coulter), TdT (Mixture, Supertech), cCD3 (UCHT1, Dako)

Statistical Methods

The statistical package SPSS 19.0 (SPSS, Inc., Chicago, IL, USA) was used to evaluate mean, median, and SD. The Mann–Whitney test was used to analyze Ki-67 expression in follicular lymphoma (FL) versus diffuse large B-cell lymphoma (DLBCL). The linear regression test was used to analyze Ki-67 expression in B-NHL subtypes. The Student t-test was applied to evaluate the T-cells distribution between HL and normal/reactive tissues. Data were considered statistically significant when P ≤ 0.05.

RESULTS

The 1,792 cases of biopsy specimens belonged to 1,706 patients (954 males and 752 females) with an age range of 1 to 93 years (mean age: 59 years).

Technical Considerations

A critical step in the FC analysis of tissue biopsies is the disaggregation of samples into a single-cell suspension of viable cells, where a primary concern involves the preservation of cell surface antigens. Enzymatic methods (tripsin, pepsin, and collagenase digestion) performed to increase the release of tumor cells from certain types of tissue, such as the skin, breast, and the gastrointestinal tract, may cause high cellular damage and low cell recovery (7, 14, 15). We used a mechanical plus density gradient separation technique able to provide adequate and representative cellular recovery in the majority of samples; only in 19 of the 1,792 (1%) cases, cell recovery was insufficient to perform both the screening and the additional antibody panel to complete the immunophenotyping characterization. In these cases, our screening approach was able to obtain useful information regarding the presence of a B- or T-cell lymphoma in 16 of the 18 cases (one was a reactive lymphoadenopathy). In 1,773 of the 1,792 (99%) cases, FC was performed by means of a double-step immunophenotyping procedure. The FC immunophenotyping was able to detect 1,290 diseases (72%) corresponding to 1,268 hematological malignancies (including composite lymphomas and lymphoma plus carcinoma) and 22 epithelial tumors. The distribution of the 1,792 cases among the different diagnostic categories documented in this report is summarized in Table 2 and illustrated as follows.

Table 2. Agreement Histology/Flow Cytometry of all 1,792 Samples Evaluated in the Study
Diagnosis (subclassification)Histologic diagnosisN° lymphomaFC identification agreement (%)
  • FC, flow cytometry; NHL, non Hodgkin lymphoma.

  • *

    *FC did not detect any immunophenotypic alterations among lymphocyte populations.

Precursor lymphoid neoplasms999 (100%)
 T-lymphoblastic leukemia/lymphoma 99 (100%)
Mature B-cell neoplasms1,1721,1721,141 (97.3%)
 Chronic lymphocytic leukemia / small lymphocytic lymphoma 144144 (100%)
 B-cell prolymphocytic leukemia11 (100%)
 Splenic marginal zone lymphoma8180 (98.8%)
 Hairy cell leukemia77 (100%)
 Splenic lymphoma/leukemia, unclassifiable22 (100%)
 Lymphoplasmacytic lymphoma99 (100%)
 Plasma cell neoplasms1010 (100%)
 Extranodal marginal zone lymphoma of mucosa-associated lymphoid tissue2624 (92.3%)
 Nodal marginal zone lymphoma4545 (100%)
 Follicular lymphoma430424 (98.6%)
 Mantle cell lymphoma7474 (100%)
 Diffuse large B-cell lymphoma (DLBCL)299291 (97.3%)
 T-cell/histiocyte-rich large B-cell lymphoma184 (22.2%)
 EBV positive DLBCL of the elderly11 (100%)
 Lymphomatoid granulomatosis11 (100%)
 Primary mediastinal large B-cell lymphoma1010 (100%)
 Burkitt lymphoma1212 (100%)
 B-cell lymphoma with features intermediate between DLBCL and Burkitt lymphoma22 (100%)
Mature T- and NK-cell neoplasms898974 (83.1%)
 T-cell prolymphocytic leukemia 33 (100%)
 Extranodal NK/T-cell lymphoma, nasal type11 (100%)
 Entheropathy-associated T-cell lymphoma54 (80%)
 Hepatosplenic T-cell lymphoma22 (100%)
 Mycosis fungoides / Sézary syndrome55 (100%)
 Peripheral T-cell lymphoma3325 (75.7%)
 Angioimmunoblastic T-cell lymphoma1715 (88.2%)
 Anaplastic large cell lymphoma, ALK positive64 (66.7%)
 Anaplastic large cell lymphoma, ALK negative1715 (88.2%)
Composite lymphoma (previously reported in Ref.6)173430 (88.2%)
 Diffuse large B-cell lymphoma + Nodal marginal zone B-cell lymphoma 22 (100%)
 Follicular lymphoma + Mantle cell lymphoma22 (100%)
 Follicular lymphoma + Nodal marginal zone B-cell lymphoma22 (100%)
 Follicular lymphoma + Splenic marginal zone B-cell lymphoma22 (100%)
 Follicular lymphoma + B-cell chronic lymphocytic leukemia/small lymphocytic lymphoma22 (100%)
 Diffuse large B-cell lymphoma + Peripheral T-cell lymphoma22 (100%)
 Diffuse large B-cell lymphoma + Angioimmunoblastic T-cell lymphoma44 (100%)
 Nodal marginal zone B-cell lymphoma + Peripheral T-cell lymphoma22 (100%)
 Nodal marginal zone B-cell lymphoma + Angioimmunoblastic T-cell lymphoma22 (100%)
 Nodal marginal zone B-cell lymphoma + Sézary syndrome22 (100%)
 Lymphoplasmacytic lymphoma + Peripheral T-cell lymphoma22 (100%)
 NK/T-cell, nasal type lymphoma + Precursor T-lymphoblastic lymphoma22 (100%)
 Nodal marginal zone B-cell lymphoma + Hodgkin lymphoma21 (50%)
 B-cell chronic lymphocytic leukemia/small lymphocytic lymphoma + Hodgkin lymphoma21 (50%)
 Follicular lymphoma + Hodgkin lymphoma21 (50%)
 Peripheral T-cell lymphoma + Hodgkin lymphoma21 (50%)
NHL plus Carcinoma443 (75%)
 Diffuse large B-cell lymphoma + Carcinoma 10 (0%)
 B-cell chronic lymphocytic leukemia/small lymphocytic lymphoma + Carcinoma22 (100%)
 Nodal marginal zone B-cell lymphoma + Carcinoma11 (100%)
Post-transplant lymphoproliferative disorders553 (60%)
Hodgkin lymphoma105 0 (0%)
Acute leukemia8 8 (100%)
Carcinoma42 22 (52.4%)
Non-neoplastic diseases341 *
 Reactive hyperplasia157  
 Necrotizing granulomatous lymphadenitis36  
 Piringer Kuchinka's lymphadenitis14  
 Kimura's disease1  
 Castleman's disease8  
 Extramedullary hematopoiesis4  
 Sarcoidosis15  
 Normal and other reactive entities109  
Total1,792  

Precursor Lymphoid Neoplasms

Nine cases were diagnosed as T-lymphoblastic leukemia/lymphoma (T-ALL) and all (100%) were identified by FC. Neoplasms of immature T-cell precursor showed variable expression of CD1a (six positive cases), CD2 (six positive cases), CD4 (six positive cases), CD5 (eight positive cases, in four dimly expressed), CD7 (eight positive cases), CD8 (seven positive cases), and surface CD3 (three positive cases). All were TdT positive with a variable fluorescence intensity from dim to bright and CD3 positive at cytoplasmic level. In four of the nine cases, the expression of CD10 and CD79a at cytoplasmic level was detected.

Mature B-Cell Neoplasms

A B-NHL was identified by FC in 1,141 of the 1,172 cases (97.3%). Among B-NHL subtypes, a high concordance was found between FC and histologic findings, with the exception of the TCRLBCL (only 22% identified by FC). In three cases, two splenic marginal zone lymphomas (SMZL) and one intraparotid gland FL, FC detected a clonally restricted B-cell population, not morphologically evident (Table 3, cases 1–3). Antigen expression in the 1,141 mature B-cell neoplasms detected by FC is reported in Table 4.

Table 3. Cases in which lymphoma was detected by FC but not by initial morphology
CaseAgeSexClinical manifestationsBiopsy siteFC (nBc%)Morphology cell sizeFinal diagnosis
  • AIL, Angioimmunoblastic T-cell lymphoma; c, cytoplasmatic; FC, flow cytometry; FL, follicular lymphoma; LN, lymph-node; MCL, mantle cell lymphoma; nBc, neoplastic B-cells; s, surface; SMZL, splenic marginal zone lymphoma; PTCL, peripheral T-cell lymphoma.

  • *

    *Neoplastic population firstly detected by FC.

  • §

    Diagnosis rendered after FC suggestion.

177MSplenomegalySpleenCD19+/ CD5−/CD10−/κ (4%)SmallSMZL
260FSplenomegalySpleenCD19+/ CD5−/CD10−/λ (1.5%)SmallSMZL
329MLaterocervical LymphadenopathyParotid glandCD19+/ CD10+/κ (10%)MediumFL
466MLymphadenopathySupraclavicular LNsCD3−/ cCD3+/ CD4+/CD10+ (8%)SmallPTCL
563FLaterocervical LymphadenopathyLaterocervical LNsCD3+/CD4+/CD10+ (13%)Small-mediumAIL
661FLaterocervical LymphadenopathyLaterocervical LNsCD3−/CD4+/CD8− (15%)SmallAIL
756Mgeneralized LymphadenopathyInguinal LN(41%) CD19+/ CD10+/κSmall-mediumFL
*CD19+/ CD5+/κbright (13%)Small§MCL
860FSplenomegalySpleen*CD19+ CD10+/λ (4%)Small§FL
CD19+/ CD5+/κ (42%)MediumSMZL
977FLymphadenopathyInguinal LNCD19+/ CD5−/CD10−/λ (9%)Small-mediumMZL
*CD3+/ CD4+/CD10+ (4%)Medium§AIL
Table 4. Antigen Expression in 1,141 Mature B-cell Neoplasms Detected by Flow Cytometry (Cases of Composite Lymphoma and Lymphoma Plus Carcinoma are not Included)
B-NHL (*)Frequency (%)
CD4CD5CD8CD10CD11cCD19CD20CD22CD23CD25
  • BL, Burkitt lymphoma; B-PLL, B-cell prolymphocytic leukemia; CLL/SLL, chronic lymphocytic leukemia/small lymphocytic lymphoma; DLBCL, diffuse large B-cell lymphoma; DLBCL/BL, B-cell lymphoma with features intermediate between DLBCL and Burkitt lymphoma; EBV+DLBCL, EBV positive DLBCL of the elderly; FL, follicular lymphoma; HCL, hairy cell leukemia; LPL, lymphoplasmacytic lymphoma; LYG, Lymphomatoid granulomatosis; MALT, extranodal marginal zone lymphoma of mucosa-associated lymphoid tissue; MCL, mantle cell lymphoma; NHL, non Hodgkin lymphoma; NMZL, nodal marginal zone lymphoma; PLNs, Plasma cell neoplasms; PMBCL, Primary mediastinal large B-cell lymphoma; SL, Splenic lymphoma/leukemia, unclassifiable; SMZL, splenic marginal zone lymphoma; TCRBCL, T-cell/histiocyte-rich large B-cell lymphoma.

  • *

    *Values in parentheses refer to the number of cases. For each column the first number is referred to positive cases, the second one to the number of cases evaluated for each antigen; when the number of cases was lower than 5, the percentages were not calculated; w: weak expression; m: medium expression; s: strong expression; fw: frequently week; nt: not tested; c, cytoplasmic.

CLL/SLL (144)0/144 (0%)144/144 (100%)2/144 (1.4%)1/144 (0.7%)67w/95 (70.5%)144/144 (100%)144w/144 (100%)112w/125 (89.6%)111/128 (86.7%)81w/102 (79.4%)
B-PLL (1)0/10/10/10/1nt1/11/1ntntnt
SMZL (80)0/80 (0%)12/80 (15%)0/80 (0%)1/80 (1.2%)46w/57 (80.7%)80/80 (100%)80/80 (100%)63/65 (96.9%)26/65 (40%)41/57 (71.9%)
HCL (7)0/7 (0%)0/7 (0%)0/7 (0%)1/7 (14.3%)7s/7 (100%)7s/7 (100%)7s/7 (100%)7s/7 (100%)2/7 (28.6%)6/7 (85.7%)
SL (2)0/20/20/20/2nt2/22/2ntntnt
LPL (9)0/9 (0%)0/9 (0%)0/9 (0%)0/9 (0%)1/49/9 (100%)9/9 (100%)5/5 (100%)0/5 (0%)nt
PLNs (10)0/10 (0%)0/10 (0%)0/10 (0%)0/10 (0%)nt0/10 (0%)0/10 (0%)ntntnt
MALT (24)0/24 (0%)0/24 (0%)0/24 (0%)1/24 (4.2%)8w/10 (80%)24/24 (100%)24/24 (100%)17/17 (100%)4/18 (22.2%)10/14 (71.4%)
NMZL (45)0/45 (0%)1/45 (2.2%)0/45 (0%)1/45 (2.2%)20w/30 (66.7%)45/45 (100%)45/45 (100%)39/39 (100%)13/38 (34.2%)23/35 (65.7%)
FL (424)0/424 (0%)4/424 (0.9%)1/424 (0.2%)370/424 (87.3%)53/283 (18.7%)423fw/424 (99.8%)422/424 (99.5%)344/350 (98.3%)245/355 (69%)85/316 (26.9%)
MCL (74)0/74 (0%)74/74 (100%)0/74 (0%)0/74 (0%)15w/52 (28.9%)74/74 (100%)74s/74 (100%)70/70 (100%)1/70 (1.4%)56/63 (88.9%)
DLBCL (291)1/291 (0.3%)28/291 (9.6%)1/291 (0.3%)103/291 (35.4%)128w/166 (77.1%)285/291 (97.9%)284/291 (97.6%)195/214 (91.1%)50/224 (22.3%)139/186 (74.7%)
TCRBCL (4)0/40/40/40/42/24/44/42/20/2nt
EBV+DLBCL (1)0/10/10/10/11/11/11/11/10/11/1
LYG (1)0/10/10/10/1nt1/11/1ntntnt
PMBCL (10)0/10 (0%)0/10 (0%)0/10 (0%)0/10 (0%)4/410/10 (100%)10/10 (100%)5/5 (100%)5/6 (83.3%)nt
BL (12)0/12 (0%)0/12 (0%)0/12 (0%)12/12 (100%)nt12/12 (100%)12/12 (100%)12/12 (100%)0/8 (0%)3/6 (50%)
DLBCL/BL (2)0/20/20/22/2nt2/22/22/20/2nt
Total (1141) 

Most of the 144 chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL) cases detected by FC showed a CD19, CD20, CD22, and CD79b weakly positive cell population with CD5, CD23, and monotypic “dim” expression of the surface or cytoplasmic light chains, the FMC7 antigen was negative or expressed at dim intensity. The 74 mantle cell lymphoma (MCL) cases (all confirmed by cyclin-D1 detection) showed monotypic “medium-bright” expression of surface immunoglobulins light chains with CD19 and CD20 antigens brightly coexpressed with CD5 and FMC7 (five cases were FMC7 negative); CD22 and CD79b were clearly expressed; in one case also the CD23 expression was detected. CD200, used in our protocols since 2009, was the most useful antigen to perform the differential diagnosis between SLL and MCL (17 SLL cases always resulted CD200 positive and 9 MCL cases always resulted CD200 negative among a total of 26 CD5 positive small-cell lymphomas collected in 2009 and 2010). In 80 SMZL, one B-cell prolymphocytic leukemia, two splenic lymphomas (SL), 9 lymphoplasmacytic lymphomas (LPL), 24 extranodal marginal zone lymphomas of mucosa-associated lymphoid tissue (MALT), 45 nodal marginal zone lymphomas (NMZL), and 54 CD10 negative (CD10) FL cases, differential diagnosis by FC relied heavily on exclusion of other low-grade lymphomas and on the type of tissue involved. The 54 CD10 FLs, which had the typical follicular histology, were CD10 negative also by IHC and considered belonging to germinal center origin, being Bcl-6+. The seven hairy cell leukemia (HCL) cases displayed the classical CD103 and bright CD11c expression; one was CD25 negative, one was CD10 positive. The 12 Burkitt lymphoma (BL) cases were recognized by FC being CD10 positive, Bcl-2 negative with the highest proliferation index (Fig. 3). The 10 plasma cell neoplasms detected by FC were all CD19 and CD20 negative, CD38 and CD138 positive with light-chain restriction at cytoplasmic level in 80% of cases. FC immunophenotyping was able to detect lymphoma but not able to make differential diagnosis between 291 DLBCL, 4 TCRLBCL, 1 EBV positive DLBCL of the elderly, 1 lymphomatoid granulomatosis (LYG), 10 primary mediastinal large B-cell lymphomas (PMBCL), and two B-cell lymphomas with features intermediate between DLBCL and BL. Among the cases detected by FC, CD10 expression was observed in 370 of the 424 FL (87.3%) and in 103 of the 291 DLBCL (35.4%), respectively, 29 of the 370 FL (7.8%) and 11 of the 103 DLBCL (10.7%) resulted sIg negative. The cytometric distinction between FL and DLBCL CD10+ mainly rely on cell size and FC proliferation value (Fig. 3). CD19 was frequently expressed at dim intensity in FL but, in our experience, this was not the rule. No significant differences were noted among FL CD10+ and FL CD10 about CD19 expression. The presence of antigens not normally found on B-cells (such as T-cell or NK-cell antigens) or unusual immunophenotype (such as CD5/CD10 coexpression) was observed in 222 of the 1,141 (19.4%) of B-cell lymphomas detected by FC. Among B-NHL, DLBCL displayed the highest number of atypical expression detected (Table 4).

Figure 3.

Ki-67 evaluation on B-NHL estimated by FC and IHC (in brackets are reported the number of cases respectively evaluated by the two methods for statistical purposes). BL, Burkitt's lymphoma; DLBCL, diffuse large B-cell lymphoma; FL G1-2, follicular lymphoma grade 1-2; FL G3a-b, follicular lymphoma grade 3a-b; MALT, mucosa associated lymphoid tissue lymphoma; MCL, mantle cell lymphoma; MZL, marginal zone lymphoma; PMBCL, primary mediastinal large B-cell lymphoma; SLL, small lymphocytic lymphoma.

Proliferation index assessed by Ki-67 antigen was examined by FC in most of the B-NHLs of our study and confrontation with IHC was evaluated in different NHL subtypes in which we had a sufficient number of cases to compare (Fig. 3). Distribution of the overall Ki-67 FC values was significantly correlated with those estimated in IHC (r = 0.83; P < 0.0001); P values were however not statistically significant in MALT, MCL blastoid variant, and PMBCL due to the low numerousness of the subgroups. FC data were lower than those obtained by IHC but displayed the same progressive increase from low- to high-grade lymphomas. Moreover, FC proliferation was analyzed in detail in FL and DLBCL (Table 5). Distribution of the cytometric Ki-67 values was significantly correlated with those estimated in IHC (r = 0.81; P < 0.0001). As expected, proliferation index, estimated with both techniques, was lowest in FL-G1 and highest in DLBCL. Among FLs and DLBCL, the difference in proliferation index between FC and IHC values was statistically significant in comparison between the various groups.

Table 5. Comparison of Proliferation Index (Assessed by Ki-67 Antigen) Evaluated by FC and IHC in FL and DLBCL
Lymphoma (Group)Ki-67 (FC)Ki-67 (IHC)
Cases numberMeanSDPCases numberMeanSDP
  1. DLBCL, diffuse large B-cell lymphoma; FC, flow cytometry; FL, follicular lymphoma; grade 1 (G1), grade 2 (G2), grade 3a (G3a), grade 3b (G3b); IHC, immunohistochemistry; SD, standard deviation.

FL-G1 (1)624.352.82(1) vs (2) < 0.00014811.815.49(1) vs (2) < 0.0001
FL-G2 (2)1489.397.0512423.447.71
FL-G3a (3)5021.0615.36(2) vs (3) < 0.00014539.6713.07(2) vs (3) < 0.0001
FL-G3b (4)1635.3821.891559.0018.73
DLBCL (5)22852.1125.28(3) vs (4) = 0.01415271.4516.63(3) vs (4) = 0.001
(4) vs (5) = 0.01(4) vs (5) = 0.015

Mature T- and NK-Cell Neoplasms

Seventy-four of the 89 mature T- and NK-cell LNH (83.1%) were detected by FC by means of an altered immunophenotype; 42/74 (56.7%) also had altered cell size (medium to large cells). The neoplastic T-cell population ranged from 1.5 to 92% of the total cell sample; 39.2% of cases (29/74) had a neoplastic T-cell amount lower than 20%. In three cases, one peripheral T-cell lymphoma (PTCL) and two angioimmunoblastic T-cell lymphoma (AIL), FC detected a neoplastic population not morphologically evident suggesting to perform additional investigations (Table 3, cases 4–6). In 10 cases, the analysis of the beta-chain variants of the T-cell receptor (TCR) was used to confirm the presence of malignant T-cell population. A summary as regards size and antigen expression in the total 74 mature T- and NK-cell neoplasms detected by FC is reported in Table 6. NHLs with increased numbers of medium–large cells included mainly extranodal NK/T-cell lymphoma, nasal type (NTENL), entheropathy-associated T-cell lymphoma (EATL), and anaplastic large-cell lymphoma (ALCL), anaplastic lymphoma kinase positive (ALK+) or negative (ALK). As previously described (11), CD10+ T-cells were observed in cases of PTCL and AIL, while γδ T-cell receptor was expressed in the two hepatosplenic T-cell lymphomas. FC demonstrated altered T-cell phenotype CD3+dimCD4+CD26 in all cases of mycosis fungoides/Sézary syndrome (MF/SS) (16). In ALCL, the CD30 antigen was positive in all cases of ALK+ and in most cases of ALK. The characteristic CD56 expression in NTENL was not observed by FC in one of the two cases studied and was undetected also by IHC. The CD103 was used as a specific and useful marker to recognize EATL.

Table 6. The Most Common Information on the Size and Antigen Expression in the Total 74 Mature T- and NK-cell Neoplasms Detected by Flow Cytometry (Cases of Composite Lymphoma are not Included)
T-NHL(*)Frequency (%)
Medium-large cellsCD2sCD3CD4CD5CD7CD8CD10CD30CD56TCR γ-δ
  • AIL, angioimmunoblastic T-cell lymphoma; ALCL, anaplastic large cell lymphoma; ALK, anaplastic lymphoma kinase; EATL, entheropathy-associated T-cell lymphoma; HSTCL, hepatosplenic T-cell lymphoma; MF/SS, mycosis fungoides/Sézary syndrome; NHL, non Hodgkin lymphoma; NTENL, extranodal NK/T-cell lymphoma, nasal type; PLL, cell prolymphocytic leukemia; PTCL, peripheral T-cell lymphoma; sCD3, surface CD3; TCR, T-cell receptor.

  • *

    Values in parentheses refer to the number of cases.

T-PLL (3)0 (0%)3 (100%)3 (100%)2 (66.7%)3 (100%)3 (100%)1 (33.3%)0 (0%)0 (0%)0 (0%)0 (0%)
NTENL (1)1 (100%)1 (100%)0 (0%)0 (0%)0 (0%)0 (0%)0 (0%)0 (0%)0 (0%)1 (100%)0 (0%)
EATL (4)3 (75%)1 (25%)1 (25%)0 (0%)0 (0%)4 (100%)4 (100%)0 (0%)0 (0%)2 (50%)0 (0%)
HSTCL (2)0 (0%)2 (100%)2 (100%)0 (0%)0 (0%)2 (100%)0 (0%)0 (0%)0 (0%)0 (0%)2 (100%)
MF/SS (5)2 (40%)4 (80%)5 (100%)5 (100%)3 (60%)1 (20%)0 (0%)0 (0%)0 (0%)0 (0%)0 (0%)
PTCL (25)9 (36%)17 (68%)13 (52%)16 (64%)20 (80%)10 (40%)5 (20%)5 (20%)2 (8%)1 (4%)2 (8%)
AIL (15)10 (66.7%)12 (80%)8 (53.3%)12 (80%)12 (80%)6 (40%)2 (13.3%)8 (53.3%)1 (6.7%)2 (13.3%)0 (0%)
ALCL, ALK+ (4)4 (100%)3 (75%)1 (25%)1 (25%)1 (25%)2 (50%)0 (0%)0 (0%)4 (100%)1 (25%)0 (0%)
ALCL, ALK- (15)13 (86.7%)4 (26.7%)2 (13.3%)8 (53.3%)4 (26.7%)0 (0%)0 (0%)0 (0%)13 (86.7%)0 (0%)0 (0%)
Total (74) 

Among T-cell lymphomas, eight of the 33 PTCL, two of the 17 AIL, one of the five EATL, two of the six ALCL ALK+, and two of the 17 ALCL ALK- were not detected by FC because no aberrant population was observed.

Composite Lymphoma

In 17 of the 1,792 cases, a composite lymphoma was identified. All the detailed results have been previously described (6). In three cases, in adjunct to the neoplasia histologically detectable, FC indicated the need to evaluate an additional malignant T or B population that was not morphologically evident (Table 3, cases 7–9). A strong correlation between morphologic findings and FC was observed in 13 cases (76.4%). In the four cases diagnosed as NHL plus HL, although FC did not detect Reed-Sternberg cells, it accurately identified the neoplastic B- or T-cell component.

NHL Plus Carcinoma

Among the four cases of carcinoma associated with B-NHL (two CLL/SLL, one NMZL, and one DLBCL), three (75%) were identified as lymphomas by FC.

Post-transplant lymphoproliferative disorders

In this study, three of the five post-transplant lymphoproliferative disorder (PTLD) cases (60%) were identified by FC. FC analysis revealed the presence of a medium to large B-cell population (CD19+CD20+CD5CD10) with monotypic or absent expression of surface light chains in one case of polymorphic PTLD (post-liver transplant) and two cases of monomorphic PTLD (one case post–heart-kidney transplant and one case post allogeneic bone marrow transplant for acute myeloid leukemia [AML]).

Hodgkin's Lymphoma

As expected, no immunophenotypic abnormalities were detected on B-cells by FC in the 105 HL cases diagnosed by histology. These cases, in comparison with the 109 cases of normal and reactive tissues, had a higher percentage of CD3 (64.7 ± 18.6 SD versus 52.2 ± 17.5 SD) and a prevalence of the CD4 subset (49.8 ± 16.8 SD versus 34.2 ± 16.2 SD); the difference was statistically significant, P = 0.00001. In eight of the 16 cases diagnosed as nodular lymphocyte-predominant Hodgkin lymphoma (NLPHL), FC detected an increase in the CD4+CD8dim+ T-cell population, ranging from 3 to 20% of T-cells, with no pan-T antigen loss or CD1a expression.

Acute Leukemia

All the eight cases of acute leukemia (AL) were easily identified by FC (100%): seven were cases of AML and one was acute biphenotypic leukemia (T-ALL plus AML); all showed CD45 dim expression and variable expression of CD13, CD33, CD34, CD117, and HLA-DR.

Carcinoma

Forty-two cases, initially suspected for lymphoma, were morphologically identified as metastatic carcinoma. In 22 of the 42 cases (52.4%), a variable amount (between 30% and 90%) of CD45 negative cells were observed by FC; seven cases also displayed the CD56 antigen (17). Cytokeratin detection performed by FC confirmed their non-hematopoietic nature.

Non-Neoplastic Diseases

Absence of an altered cytometric immunophenotypic profile was observed in 341 cases of non-neoplastic diseases. The expansion of CD10+ B-cell population was observed in 25 of the 157 cases of reactive follicular hyperplasia (RFH), ranging from 7 to 61% of all B-cell population (median 27.8); the politypic light-chain expression, and the reduced pattern of expression of Bcl-2 on these cells confirmed their benign nature.

DISCUSSION

Several studies have demonstrated the usefulness of FC in diagnosing lymphoma by means of fine needle aspiration (FNA) cytology as well as in staging and follow-up (18–24). However, only few reports are available regarding the role of FC in tissue diagnosis of lymphoma (25–31), and literature mainly reports tissue FC studies in selected lymphoma subtypes (12, 14, 32–48). In the present study, a high concordance between morphology and an abnormal immunophenotypic pattern was reported on hematological malignancies detected by FC (1,268/1,304, 97.2%, including composite pathologies such as composite lymphomas and lymphoma plus carcinoma) with the exception of HL. Among all the 1,196 B-cell lymphomas histologically diagnosed in our study (1,172 mature B-cell neoplasms, 20 B-NHL among composite lymphomas, and 4 B-NHL among cases of NHL plus carcinoma), 1,164 (97.3%) were identified by FC. The surface or cytoplasmic light-chain evaluation was the most useful parameter in identifying a B-cell lymphoma, whereas CD5 and CD10 are useful in identifying the different subtypes (1). However, their specificity is not absolute on neoplastic cells, as stands out from data also reported by us in Table 4; in fact they may be absent in FL (CD10FL) (37) or can be unexpectedly expressed elsewhere, such as CD5 on some SMZL cases (5, 49) and FL (36, 39, 50, 51), CD10 on SLL (50, 51), NMZL (43), MCL and MALT (37), and HCL (52), CD5/CD10 coexpression on DLBCL (35, 50, 53). These kinds of B-NHL still remain challenging for flow cytometrists, because the lack of a specific phenotype and their great heterogeneity, especially observed among MZL (nodal, splenic, extranodal), SL and LPL, determine their identification mainly on the κ/λ ratio. It has to be considered that certain NHL subtypes represent a challenge also for pathologists, as evidenced in a study of the International Lymphoma Study Group (54). In this retrospective clinical evaluation of 1,403 NHL cases, performed by five expert hemopathologists, the overall consensus diagnosis was 85%, with a low level of reproducibility in histologic diagnosis especially reported in MZL, LPL, and high-grade B-cell Burkitt-like lymphoma. The superior power in detecting clonal light-chain expression of FC with respect to IHC confirms its utility in the diagnosis of B-NHL especially in situation in which morphology is confounding. Among cases reported in this study, the differential diagnosis between FL and RFH mainly relied on FC data, as none of the RFH showed neither light-chain clonality nor Bcl-2 expression. The role of Bcl-2 detection by FC was examined in few studies on lymph-node biopsies and FNA (46, 47, 55). We have extensively analyzed the Bcl-2 FC expression pattern on RFH and B-NHLs. In our study, 25/157 reactive hyperplasia showed a CD10+ B-cell expansion. In these cases, cytometric evaluation of Bcl-2 (together with polyclonal light-chain detection and CD20 expression pattern) was useful in excluding a neoplastic process, as normal germinal center B-cells displayed a Bcl-2 expression level below that detected on normal T- or mantle B-cells (46). None of these cases considered as “reactive” developed a lymphoma.

Our report also documents the unusual expression of T- or NK-cell antigens on B-cell NHLs: CD4, CD8, and CD56 were the most common aberrant antigens observed. In our opinion, the overall frequency of unusual phenotype in B-NHLs reported in the literature (34, 44, 45, 56) might be underestimated, because cases without FC studies might have been ignored owing to the lower sensitivity in detecting aberrant expressions by IHC and also because of more limited panel of antibodies generally used on routine workup. The identification of aberrant expressions may be very useful in detecting and monitoring of minimal residual disease.

By using a modified fixing and permeabilizing procedure, we have been able to evaluate by FC the Ki-67 value in B-NHLs. As expected, proliferation increased with the level of malignancy. When FC proliferation results were compared with IHC data, we obtained a high correlation level. FC data were particularly interesting among groups for FL and DLBCL. In MCL, FC proliferation clearly differentiated the blastoid form from the classical one, although data were not statistically significant due to the low numerousness of the subgroup. Among FL and DLBCL, we obtained a high level of correlation with the IHC proliferation results and statistically significant data among groups. To the best of our knowledge, this is the first report in which proliferation values obtained by FC were systematically compared to those obtained by IHC. We believe that by adopting the approach proposed by Mourad et al. (57), which uses a CD19/FSC dot plot to identify centroblasts in FL, in combination with the cytometric Ki-67 evaluation, might be possible to grade FL with a certain degree of success. The possibility to correlate proliferation with cell size and immunophenotype might allow for a better sub-classification of B-NHL by FC. Our findings need to be validated by others and might represent a useful method to distinguish by FC low-grade from high-grade lymphomas especially in small samples and fine needle aspirates (58).

FC has been useful in identifying T-cell lymphoma. From the histological point of view, neoplastic T-cells may be scarce and difficult to recognize, leading to diagnostic errors in >50% of cases (12). The differential diagnosis includes atypical reactive processes, Hodgkin's lymphomas (HL) or B-cell lymphomas so that reaching a definitive diagnosis requires additional immunological, cytogenetic, and molecular data. In our study, FC easily detected aberrant or atypical antigen expression, was able to distinguish the surface from the cytoplasmic expression (like CD3 not appreciable by IHC), and to correlate the altered phenotype with abnormal physical parameters (forward and side light scatter) also in those situations in which neoplastic cells represented less than 20%. In all these cases, information given by FC were used by pathologists to direct diagnosis.

Among all NHLs histologically diagnosed, FC failed to identify 47of the 1,291 NHL cases (3.6%) respectively corresponding to 32 of the 1,196 (2.7%) B-NHL and 15 of the 98 (15.3%) T-NHL. It must be considered that among these cases, the majority (20/47) belonged to the first three years of our study, when the multiparameter analysis was still limited as we used a three-color immunophenotyping protocol on a Calibur instrument, so antibody panel selection could have been misdirected and gating strategies would be too restrictive or too inclusive to adequately detect tumor population. Adverse storage conditions, inadequate sampling, or manipulation due to the difficulty encountered in obtaining sufficient cells from certain tissue types could also be invoked to explain false-negative results (8). Moreover, many neoplastic nodes or masses can show dishomogeneous involvement and inaccurate sampling can lead to a biopsy which is devoid of malignant cells and not representative of total mass (59). Finally, we cannot exclude that density-gradient separation might have been responsible for selective malignant cell loss in certain types of lymphomas. Working with solid tissue needs to be aware of all the factors that may affect the quality of the result. Undoubtedly, multicolor and multiparameter analysis offered by more sophisticated instruments (such as Canto instrument) and the experience acquired through the years have improved our ability in detecting lymphomas. TCRLBCL concentrate the false-negative FC results. Large cells are particularly fragile and often scarcely represented in TCRLBCL characterized by the presence of numerous residual non-neoplastic lymphocytes among the neoplastic cells. Bertram et al. (33) also reported that apoptosis, necrosis, and sclerosis were histologically present in many large B-cell lymphomas, and FC performed on solid tissue samples was often difficult. In our experience, the discordance between morphologic features and immunophenotyping results also remains problematic for ALCL. The most effective means to detect ALCL by FC was an analysis strategy that examines total ungated cells on plots displaying CD30 and light scatter properties (60) (Fig. 4); the inclusion of CD30 in our routine lymphoma panel, since 2005 has consistently reduced false-negative results. Moreover, we experienced that in suspected cases of large B- or T-cell lymphoma, the acquisition of larger-than-usual numbers of cells can be very helpful in detecting a low-represented neoplastic population.

Figure 4.

Example of FC analysis of an ALCL ALK+ in an inguinal lymph-node. A high number of events (>100,000) were collected to improve resolution of anaplastic cells. Medium-large CD30+ cells, extremely dispersed in the FSC/SSC plot, show cytoplasmic ALK-1 expression. ALCL, Anaplastic large cell lymphoma; ALK, anaplastic lymphoma kinase; FITC, fluorescein isothiocyanate; FSC, forward scatter; PE, phycoerytrin; SSC, side scatter. [Color figure can be viewed in the online issue which is available at wileyonlinelibrary.com]

We did not attempt to detect Hodgkin's lymphoma by FC, inasmuch as there is scarce representation of Reed-Sternberg cells in the cell suspension obtained from the tumor. A nine-color, single-tube FC assay to diagnose classical HL in tissues has recently been proven useful (42); nevertheless, this method cannot be routinely used as it requires a high dose of experience in detecting Reed-Sternberg cells by FC. Our data together with other previously published reports (61, 62) demonstrated that the prevalence of CD4+T-cell subset, in absence of others immunophenotypic alterations on lymphocyte populations, is highly suggestive of HL. We experienced moreover that results of FC analysis may help the pathologist in the differential diagnosis of certain HL subtypes, particularly in those cases in which a diffuse growth pattern makes the recognition of an NLPHL from a TCRBCL difficult. The high amount of CD4+CD8dim+ T-cells detected by FC and previously described in NLPHL (63, 64) oriented the histologic diagnosis toward NLPHL, in three HL cases recently observed at our institution (personal observation).

Finally, there are situations where FC plays a diagnostic role and proves to be more informative and useful than IHC (6, 44, 45). Among cases reported in this study, nine cases were identified, in which FC detected the presence of a small neoplastic population not morphologically evident. IHC or molecular investigations performed after FC suggestion confirmed the presence of neoplastic cells respectively in two SMZL, one FL (inside the parotid gland), two AIL, one PTCL, and three cases of composite lymphomas (described in a previous paper, 6). It can be observed that two cases were SMZL with a neoplastic population lower than 5%; as a matter of fact the differential diagnosis between “early” SMZL (lacking evident red pulp infiltration) and RFH with enlarged marginal zone, cannot be made without monotypic B-cell demonstration.

In conclusion, FC data interpretation of hematologic neoplasia is a high-complexity assay that needs analyst training and education in several disciplines (e.g., histology, cytology, IHC, FISH, molecular biology) (66). It is important that FC data interpreters have thorough knowledge of the phenotypes of the different normal cell populations, can recognize deviations from normal, and are able to discuss the potential clinical significance of the FC findings (5). The role that FC may play in tissues is not only of diagnostic support and classification, detection of antigenic expression that may be important for detection of aberrancies, prognosis (CD38, ZAP-70) or therapy (CD20, CD52), but may be crucial in situations that are non-histologically clear, like RFH, certain types of B-NHL or T-NHL and composite lymphomas. FC performed in conjunction with conventional fine needle aspiration cytology may be fundamental in detecting lymphomas, particularly in situations in which surgical biopsy could not be performed because of risk or of difficult access (21, 22); valuable results we obtained evaluating proliferation by FC offers intriguing perspectives in this field. The availability of more sophisticated cytometers, wider selection of antibodies in routine diagnostic laboratories and recently proposed mathematical algorithms, which allow direct assignment to specific diagnostic categories (67), will lead to the resolution of those more complex disease entities whose diagnosis and classification remains problematic.

Acknowledgements

The authors thank O. Gaiola, A. Di Tomasso, S. Monticone, and D. Corino for their invaluable and excellent technical assistance. The authors are grateful to M.G. Teriaca for proofreading the manuscript.

Ancillary