B-cell lymphomas with concurrent IGH-BCL2 and c-MYC rearrangements (so-called “double-hit lymphomas” [DHL]) are a relatively rare, recently described category in the 2008 World Health Organization classification of hematopoietic neoplasms. Response to chemotherapy and survival are poor.
The authors reviewed files of cytogenetically documented DHL to identify cytologic features that would allow its possible recognition.
Twelve fine-needle aspirates (FNAs), 2 pleural fluids, and 1 touch imprint of cytogenetically proven DHL were uncovered. Primary DHL was correctly recognized in 3 of 12 FNA cases using Ki-67 staining coupled with a positive bcl-2 result as the basis for performing fluorescence in situ hybridization (FISH) analysis of c-MYC and IGH-BCL2 rearrangements. Remaining FNAs and non-FNA cases were diagnosed as non-Hodgkin lymphoma, B-cell lymphoma, or atypical lymphocytosis. Ten cases had cell block material available. All cases had high cellularity with a dissociated smear pattern and background lymphoglandular bodies. Cell size ranged from intermediate to large. Nuclei were predominantly rounded or slightly irregular in contour; 4 FNAs had markedly cleaved nuclei. Some nuclei harbored discrete but small nucleoli, whereas in others coarse chromatin and indistinct or multiple small nucleoli existed. A variable number of mitotic figures, tingible body macrophages, and background apoptotic cells were also present.
Chromosomal translocations occur frequently in B-cell lymphomas and often involve the deregulation of an oncogene. The more information gleaned from tumors cytogenetically, the less specific certain formerly “classic” translocations have become. For example, the trademark t(14;18)(q32;q21) of follicular lymphoma (FL) and the 8q24 c-MYC translocation of Burkitt lymphoma (BL) are found in up to 30% and 15%, respectively, of cases of de novo diffuse large B-cell lymphoma (DLBL). 1-5
Recently, investigators identified non-Hodgkin lymphomas (NHLs) with multiple reciprocal translocations, including those involving the c-MYC, B-cell lymphoma 2 (BCL2), BCL6, and CCND1 genes. 6 The most common of these combinations (62%) appears to be that of c-MYC and BCL2, producing the appellation “double-hit lymphoma” (DHL) for these neoplasms. 6 DHL is a diagnostic category of unclassifiable lymphomas introduced in the 2008 World Health Organization (WHO) classification. 7 These patients appear to have more advanced disease at the time of presentation with a higher incidence of bone marrow and central nervous system (CNS) involvement, higher lactate dehydrogenase (LDH) levels, and a markedly shorter median overall survival. 8
Histologically, DHL is a heterogeneous group, having features of both BL and DLBL. Cells are intermediate in size with often irregular nuclear contours and intermixed larger cells. Chromatin may be finely or coarsely granular, or discrete nucleoli may be identified. Similar to BL, DHL has a high Ki-67 proliferation index (PI) ranging from 50% to 100% and a variable “starry sky” appearance with apoptotic cells and numerous mitotic figures. 9 Immunophenotypically, DHL expresses B-cell markers, usually with a germinal center phenotype (bcl-6+, CD10+). 9 Immunohistochemical expression of bcl-2 in a case that otherwise resembles BL or Burkitt-like lymphoma (BLL) morphologically may also suggest both the c-MYC and IGH/BCL-2 translocations. The c-MYC translocation is believed to be a second “hit” in these cases because of the finding that variant c-MYC translocations (involving 2p and 22q) are more common in DLBL than BL because the IGH locus is already involved with the primary t(14;18). 3
Accurate and timely diagnosis is clinically important in classifying aggressive lymphomas such as DHL. Once histologic and/or immunophenotypic features of potential DHL are recognized, fluorescence in situ hybridization (FISH) probes for IGH/BCL-2 and c-MYC are commercially available to prove the “double-hit” nature of this lymphoma using paraffin-embedded cell blocks, unstained cytologic slides, or tissue biopsy. Because cytopathology is often a first step in the investigation of patients with clinically presumptive malignant lymphoma, in the current study we reviewed our experience with 15 histologically/cytogenetically proven DHL cases to determine whether any morphologic or immunophenotypic clues exist with regard to its recognition using cytologic specimens.
MATERIALS AND METHODS
We searched our cytology files for cases diagnosed as DHL, DLBL with a high proliferation rate, and atypical BL, and our surgical pathology files for any cases with these diagnoses that had corresponding cytopathology findings. Slides from cases seen in consultation were obtained from the original submitting institution.
Percutaneous fine-needle aspiration (FNA) biopsy was performed without local anesthesia using standard technique with 21-gauge or 22-gauge needles. Three or 4 passes were made into the lesion, and each pass was rinsed into an RPMI-balanced salt solution after expelling material onto glass slides to create direct smears. Imprint or scrape slides were made by touching or scraping the cut surface of fresh tissue (at the time of frozen section examination) to create conventional smears that were stained and interpreted along with the tissue section. Liquid-based slides were made using SurePath technology (BD Diagnostics, Burlington, NC). Slides stained using both Papanicolaou and Romanowsky (R) staining were available for 11 of 13 cases. All FNA smears were air-dried. Papanicolaou-stained slides underwent rehydration and alcohol fixation before staining. Formalin-fixed, paraffin-embedded cell block sections made from FNA cases were stained with hematoxylin and eosin.
FISH analysis used commercially available probes on formalin-fixed, paraffin-embedded cell block or whole-tissue sections. The IGH/BCL-2 probe is a dual-color, dual-fusion Vysis LSI probe (Abbott Molecular, Inc, Des Plaines, Ill) that detects translocations involving IgH at 14q32 and BCL-2 at 18q21. The expected pattern in a normal nucleus hybridized with this probe is a 2 orange, 2 green signal (2O2G). The most common pattern in a nucleus harboring a t(14;18) is a 1 orange, 1 green signal representing the normal homolog and 2 orange/green (yellow) fusion signals representing the 2 derivative chromosomes resulting from a reciprocal translocation (1O1G2F). The c-MYC probe is a dual-color, break apart rearrangement Vysis LSI probe (Abbott Molecular, Inc) comprised of 2 probes that hybridize to opposite ends of the c-MYC gene at 8q24 and encompass the vast majority of breakpoints. The normal signal is 2 orange/green (yellow) fusion (2F). A 1 orange, 1 green, 1 fusion (1O1G1F) pattern is expected when the c-MYC gene is rearranged for either t(2;8), t(8;22), or t(8;14).
Immunohistology performed on cell block material used antibodies to the proliferation marker Ki-67 (MIB-1 clone; 1:400 dilution [Dako, Carpinteria, Calif]) and to bcl-2, bcl-6, and various CD markers in cases according to routine laboratory protocol. The examining pathologist performed a semiquantitative PI by counting nuclear Ki-67 staining. Four-color flow cytometry used a standard antibody panel on needle rinses of FNA material preserved in RPMI.
We recovered 15 cytogenetically confirmed examples (having both IGH/BCL2 and c-MYC translocations using FISH) of DHL that had cytologic material (Table 1). The majority of these were only from the past 2 years because FISH testing for this set of translocations was not typically performed before the 2008 WHO recognition of the DHL category. Of these, 12 cases were FNA biopsies, 2 were pleural fluids, and 1 was a cytologic touch imprint of a bone lesion. The ages of the patients ranged from 38 years to 81 years (mean age, 64 years), and the male/female ratio was nearly equal. No patient was known to be positive for the human immunodeficiency virus. Follow-up ranged from 2 weeks to 29 months, with 3 patients being lost to follow-up. Seven FNAs were of lymph nodes and 5 (42%) were from extranodal sites. Three FNAs were diagnosed definitively as DHL using FISH probes on cell block material before surgical excision (Cases 7, 9, and 14). Direct smears from these 3 cases revealed an intermediate to large cell lymphoma of no special type, flow cytometry confirmed B-cell clonality, and the cell block demonstrated both positive bcl-2 staining and a high percentage of Ki-67-positive (> 85%) stained nuclei. This prompted us to then submit cell block sections for FISH analysis of c-MYC translocations, which, when positive, then caused us to request IGH/BCL-2 analysis by FISH. Positivity for both translocations confirmed the diagnosis of DHL in all 3 examples (Fig. 1). Remaining FNAs and non-FNA cases offered an initial diagnosis of lymphoma or variants of B-cell lymphoma before surgical biopsy in all but 2 cases (Cases 5 and 6) that were diagnosed as “atypical lymphocytosis” (Table 1). Of 14 cases tested for the bcl-2 antibody, all were positive. The estimated Ki-67 PI ranged from 50% to > 95% on cell block or surgical specimens for all cases. One case was a cytologic touch imprint from tissue sent for frozen section analysis; a specific imprint diagnosis was neither rendered nor requested. Cell blocks were available in all but 1 FNA specimens from our institution. Cases seen in consultation did not include a cell block. Flow cytometry performed in 10 of 13 cases revealed B-cell monoclonality in all of them.
Table 1. Clinical Features in 15 Cases of “Double-Hit” NHL
Cytology Specimen, Site
LDH (Nl, 100-190)
Abbreviations: +, positive; -, negative; AL, atypical lymphocytes; auto, autologous; AWD, alive with disease; BCL, B-cell lymphoma; BMT, bone marrow transplantation; CB, cell block; ChemoRx, chemotherapy; CVP, cyclophosphamide, vincristine, and prednisone; c/w, consistent with; DLBL, diffuse large B-cell lymphoma; DHL, “double-hit” lymphoma; DOD, dead of disease; Dx, diagnosis; FC, follicular cell; FISH, fluorescence in situ hybridization; FNA, fine-needle aspiration biopsy; LDH, lactate dehydrogenase; LFU, lost to follow-up; LN, lymph node; MTX, methotrexate; Nl, normal; NA, not applicable; ND, not done; NHL, non-Hodgkin lymphoma; PI, proliferation index; R-CHOP, rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone; R-CVP, rituximab plus CVP; R-EPOCH, rituximab plus etoposide, vincristine, doxorubicin, cyclophosphamide, and prednisone; R-ICE, rituximab plus ifosfamide, carboplatin, and etoposide; RP, retroperitoneal; s/p, status post.
Ki-67 PI: 50% BCL-2+
AWD, 10 mo
FNA, RP mass
BCL of FC origin
Ki-67 PI: >95% BCL-2+
DOD, 1 mo
Ki-67 PI: 90% BCL-2+
Ki-67 PI: 95% BCL-2+
DOD, 28 mo
Ki-67 PI: >95% BCL-2+
DOD, 29 mo
FNA, right neck LN
DOD, 29 mo
FNA, left neck LN
BCL-2+ Ki-67 PI: >90% (CB)
s/p BMT, recurred, now auto BMT
AWD, 6 mo
BCL, high grade
Ki-67 PI: 70%-80% BCL-2+
AWD, 4 mo
FNA, peritoneal LN
CD20+,CD10+, BCL-2+ Ki-67 PI: >85% (CB)
AWD, 2 mo
FNA, axillary LN
Ki-67 PI: 80% EBER+
DOD, 2 mo
FNA, left neck LN
Ki-67 PI: >95% BCL-2+, CD10-
FNA, left submandibular LN
Ki-67 PI: 90% CD20+, BCL-2+, BCL-6+
Touch imprint, left femoral neck mass
Ki-67 PI: >90% CD20+, CD10+, BCL-2+, BCL-6+
AWD, 3 mo
FNA, back, and left supraclavicular LN
Ki-67 PI: >95% CD20+, CD10+, BCL-2+, BCL-6+
None at time of last FU
AWD, 2 wk
Ki-67 PI: >90% BCL-2+
AWD, 6 mo
Morphologic review of slides evaluated cellularity, cell pattern, background, cell composition, and cytomorphology (Table 2). Variables were assigned scores (1+: < 25% of cells; 2+: 25%-75% of cells; and 3+: >75% of cells). All cases were hypercellular with a predominantly dissociated single-cell pattern; 3 cases also demonstrated occasional lymphoid pseudoclusters. A background of numerous (3+) lymphoglandular bodies was present in all but the single case with liquid-based slides. Smears were generally comprised of a “2-cell” population consisting of a relatively monomorphous population of intact viable lymphocytes and a second population of cells in various stages of apoptosis and/or necrosis. Tingible body macrophages ranged from 1+ to 3+, but a “starry sky pattern” as described in some FNAs of BL was absent. In fact, no case was recognizable morphologically as typical BL either initially or during retrospective review. Mitotic figures were noticeable in all cases, but were readily apparent (3+) in only 2.
Table 2. Cytomorphologic Features of “Double-Hit” NHL
Abbreviations: CB, cell block; NHL, non-Hodgkin lymphoma; Mono, monoclonal; NA, not applicable; ND, not done; P, Papanicolaou stain; R, Romanowsky stain; TBM; tingible body macrophages.
Lymphocytes demonstrated some variation in size, with the majority being intermediate to large, and having rounded to slightly irregular nuclear contours. Four cases demonstrated many excessively cleaved/notched nuclei that were best appreciated in Romanowsky-stained direct smears. One such case had rare large cells with a “clover leaf” nucleus (Fig. 2). Most nuclei had coarse chromatin and indistinct or small nucleoli; some had single small, discrete nucleoli, but macronucleoli were not noticeably present in any case. All smears had cells with minimal cytoplasm, and at least rare cytoplasmic vacuoles. These were best appreciated on Romanowsky-stained slides and were classified as 3+ in only 3 cases. Dendritic-lymphocytic aggregates were rarely seen.
Although FNA is a fast, safe, and cost-effective method of diagnosing NHL when practiced in a setting bolstered with appropriate ancillary techniques, it has been controversial and extremely limited when based largely or solely on morphology. 10 Specimen inadequacy and sampling error may lead to an uncertain diagnosis and necessitate an open biopsy for definitive classification. In addition to histopathology, the 2008 WHO classification of hematopoietic and lymphoid tissue tumors relies heavily on cytogenetic, clinical, and molecular characteristics for classification, 7 thereby creating an opportunity to increase the accuracy and precision of cytologic methods with regard to the diagnosis of hematopoietic tumors.
DHL is currently included in a novel diagnostic category of unclassifiable lymphomas that was introduced into the 2008 WHO classification of hematopoietic neoplasms as “B-cell lymphoma, unclassifiable, with features intermediate between DLBL and BL.” 7 These presumably rare lymphomas do not, as of yet, represent actual distinct disease entities according to the WHO. DHL shares morphologic and cytogenetic features of both DLBL and BL and in the past was often classified as either BLL, atypical BL (a category now eliminated in the current WHO classification), or DLBL with high PI/bcl-2 expression. Clinical features of DHL include a poorer prognosis (compared with DLBL and BL), an aggressive clinical course, rarity in patients aged < 18 years, and treatment refractoriness. A recent histologic study of 20 DHL cases found clinical presentation at a more advanced stage with a higher incidence of bone marrow and CNS involvement, higher LDH levels, and a markedly shorter median overall survival, with 70% of patients dying within 8 months of diagnosis despite intensive chemotherapy regimens. 8
Analysis of the Ki-67 (MIB-1) PI is one tool for guiding NHL subclassification. A study by Ali et al suggested cutoff percentages of MIB-1–positive cells to subdivide 91 cases of NHL cytospin preparations into “indolent, aggressive, and highly aggressive” groups based on the 2008 WHO classification. 11 They demonstrated a statistically significant increase in PI according to tumor aggressiveness, noting a high PI (70%) in their small group of “unclassifiable B-cell lymphoma with features intermediate between diffuse large B-cell lymphoma and Burkitt lymphoma.” They also recognized that the highest percentage of discrepant cases fell into this group, potentially due the overlapping nature of the morphologic features of this category. Other laboratory methods are also easily available for use on cytology specimens. Recently, 2 studies have evaluated the use of FISH studies in cytospin preparations 12 and direct smears 13 of a variety of lymphomas. In one of these studies, FISH performed from cytospin slides allowed for the subclassification of 61.6% of NHL cases by identifying specific gene rearrangements, including those rearrangements involved in DHL. 12 Similarly, FISH has been helpful in the subclassification of NHL using unstained direct smears.
Our experience suggests the importance of cell block preparation (in addition to the submission of cells for flow cytometry), particularly when coupled with bcl-2 staining to assist in the cytologic diagnosis of DHL. It allowed us to correctly recognize DHL from 3 of our 12 FNAs after immunohistochemical staining for Ki-67 revealed a high PI (> 85%) in FNAs that did not have the classic cytopathology of BL or the expected negative staining for bcl-2. 14 These 3 cases were instructive, because 2 were preliminarily interpreted as “atypical lymphoid infiltrate” and 1 as “suspicious for lymphoma” before material was sent for flow cytometry and cell block preparation. The cytologic recognition of DHL has great clinical importance, particularly in those patients in whom attempted surgical biopsy would be difficult or hazardous. We have demonstrated in these 3 examples that a cell block is useful not only for immunohistochemical staining, but also for FISH studies. Others have demonstrated this use in cytospin preparations or direct smears. 12, 13
To the best of our knowledge, the current study is among the first to specifically evaluate the cytomorphology of DHL in the hopes of making an early identification based primarily on FNA material. Unfortunately, we found no specific cytomorphologic feature or set of features that would reliably recognize possible DHL on FNA smears or exfoliative cytology specimens. In fact, features of both BL and DLBL were variably appreciated across all cases, including nuclear size, shape, and chromatin structure, or quantity of tingible body macrophages, apoptotic cells, and cytoplasmic vacuoles. Overall, the morphologic features of the 15 cases in the current study were not uniform. The only consistent features included high cellularity, a predominantly dissociated pattern, and a 2-cell population of viable and apoptotic cells. The majority of nuclei had rounded to slightly irregular nuclear borders; however, 4 cases (27%) had noticeably exaggerated nuclear clefting. Because this extreme nuclear segmentation occurred in a minority of cases, it is unclear whether it has any diagnostic significance. Nucleoli ranged from small and discrete to indistinct with coarse chromatin. The florid cytoplasmic vacuolization that is so often typical of many BL smears was only readily apparent in a subset (20%) of cases in the current study. It is interesting to note that those cases that demonstrated numerous cytoplasmic vacuoles also had more noticeable mitotic figures, apoptotic bodies, and tingible body macrophages.
Some authors have recommended the evaluation of c-MYC and IGH/BCL2 rearrangements in patients with lymphomas with certain histologic and/or clinical features. 8 These include patients with high-grade B-cell lymphomas of advanced stage; those with bone marrow and/or CNS involvement, extranodal location, and a serum LDH value that is triple the upper limit of normal; those with tumors with a BL-like morphology in patients in whom BL would be unlikely (eg, middle aged or elderly patients not infected with the acquired immunodeficiency syndrome); and those with BLL with an abnormal immunophenotype including bcl-2-positive staining and/or a Ki-67 PI of < 95%, a history of low-grade FL with recurrence into high-grade B-cell lymphoma, and DLBL not otherwise specified with a Ki-67 PI of > 80%. Immunophenotypically, in addition to bcl-2 positivity, DHL is to be suspected when an intermediate to large cell lymphoma exhibits a germinal center phenotype including coexpression of CD19, CD20, CD22, CD79a, bcl-6, and CD10 and is negative for multiple myeloma-1/interferon regulatory factor-4 (MUM1/IRF4). Although several patients in the current study presented with aggressive disease, including extranodal involvement (50%), many of these other clinical details were unknown at the time of cytologic interpretation, and would be unknown to the evaluating cytopathologist in most instances.
The results of the current study demonstrate that because no distinctive cytomorphologic feature(s) led to a specific diagnosis of DHL in this series, a high index of suspicion using a combination of bcl-2 and Ki-67 staining is extremely helpful in guiding further testing. Obtaining sufficient diagnostic material with which to make direct smears, to send cells for flow cytometry, and to create a cell block (or additional unstained cytospin preparations/smears) for immunostaining is critical in this regard. This was accomplished in our 3 definitively diagnosed DHL cases by using our normal routine of 3 to 4 FNA passes into a mass with a needle rinse into RPMI after each pass. A positive bcl-2 stain and high Ki-67 PI are important factors in recognizing DHL in FNAs (particularly those from extranodal sites) that demonstrate the cytomorphology of an intermediate to large cell lymphoma without the classic features of BL.