The cytomorphology of anaplastic large cell lymphoma (ALCL) is distinctive yet variable. To the authors' knowledge, to date only small case series have described the cytologic findings noted in patients with ALCL. The current series is the largest case series presented to date to retrospectively review the cytomorpholgic findings noted in patients with ALCL, with specific attention paid to those with anaplastic lymphoma kinase (ALK)-negative ALCL.
Over a 13-year period, the available Diff-Quik cytology smears and surgical excision specimens taken from patients with ALCL were evaluated. Different clinical and morphologic parameters were evaluated, including ALK status.
A total of 37 cases were retrieved and evaluated, 19 of which had both cytology and surgical pathology specimens available for review. ALK-negative ALCL cytology smears were found to have a high number of anaplastic cells compared with ALK-positive cases. The hallmark cells in the ALK-negative cases were not classic.
The application of molecular techniques to cytopathology is transforming the discipline. Moving beyond immunohistochemical analysis microarrays, in situ hybridization and polymerase chain reaction can be performed using cytologic material and result in a plethora of information. Much of the information gained from these exciting techniques has yet to be fully realized as it pertains to diagnosis, treatment, and prognosis. Microscopic examination remains the cornerstone of diagnosis and cytopathology continues to gain momentum as a primary diagnostic modality and economical triage procedure. Although the 21st century is upon us, many institutions do not have the capabilities for high-throughput molecular tests, or may not have a wide array of immunohistochemical stains. Furthermore, as many cytopathologists know, there are also situations in which the material limits one; either no cell block or blank smears are available, further emphasizing the need for a detailed microscopic assessmentwith diagnostic morphologic criteria. Nonetheless, in the context of cytopathology, the combination of both traditional microscopy and molecular methods provides a powerful and efficient diagnostic tool.
Systemic anaplastic large cell lymphoma (ALCL) is a rare subtype of T-cell lymphoma that constitutes <5% of all non-Hodgkin lymphoma (NHL) cases, with the majority of cases occurring in childhood and adolescence. Initially described in 1985, ALCL (prior terms include epithelioid lymphoma and Ki-1 lymphoma) was defined by its characteristic horseshoe-shaped, pleomorphic-appearing cells and CD30 immunopositivity.1 In 1994, the disease definition evolved and the anaplastic lymphoma kinase (ALK) t(2,5) translocation took its place as the molecular morphologic counterpart to the “hallmark” cell. ALK positivity was linked not only to the t(2,5) translocation but also to a favorable prognosis.2 Currently, both the cytomorphologic, molecular, and immunohistochemical features of ALCL are under discussion because the clinicopathologic features of ALK-negative ALCL may comprise a separate disease entity and require more aggressive treatment. This comprehensive retrospective study of the 13-year experience at 1 institution aims to further describe the many cytologic features of ALCL, with specific emphasis on ALK-negative tumors. Although the “hallmark” cells recognized in ALCL initially were believed to be diagnostic of the lesion, the morphologic spectrum of ALCL has proven to be vast. In the current study, the differential diagnosis of ALCL is discussed as well as the pitfalls that are commonly encountered.
MATERIALS AND METHODS
Computer-generated searches were performed in the archives of the Pathology Departments of New York University and/or Bellevue Hospital Medical Center during the period between 1994 and 2006. Patients included in the study were diagnosed with ALCL by liquid-based cytology or fine-needle aspiration cytology (FNAC) and/or had a surgical excision specimen that was diagnosed as ALCL. If the surgical excision was diagnosed as ALCL, any prior or concurrent FNA specimen of that anatomic location was reviewed. The diagnosis of ALCL was defined by the presence of anaplastic and/or hallmark cells, CD30 immunopositivity, and a T-cell immunophenotype. ALK status also was noted. Cytopathologists performed the superficial aspirations, whereas deep aspirations were performed by interventional radiologists with adequacy assessed by cytotechnologists. The aspirated material was expelled onto glass slides and smeared. The smears were fixed immediately in 95% ethanol and stained using the Papanicolaou method or air-dried and stained with modified Giemsa (Diff-Quik) (Differential Stain Kit, Newcomersupply, Middleton, Wis) and the ultrafast Papanicolaou stain (Richard Allan Scientific, Kalamazoo, Mich). When necessary, additional smears were prepared for cytochemical stains or material was collected in either RPMI-1640, Hank solution, or CytoRich Red (TriPath Imaging, Burlington, NC). The collected material was used for flow cytometry or immunocytochemistry as the case dictated. Immunochemical stains were performed on described aspirate smears and stained for CD3, CD4, CD8, CD15, CD20, CD30, CD45, CD45Ro, ALK, epithelial membrane antigen (EMA), cytokeratin AE1/3, human leukocyte antigen (HLA)-DR, and TIA-1 (Ventana Medical Systems, Tucson, Ill). Cytochemical stains were prepared for acid-fast bacilli and fungus, as the case dictated (Grocott's methenamine silver stain; Ventana Medical Systems, Tucson, Ill). Follow-up consisted of either clinical evaluation or review of a surgically excised specimen and immunohistochemical stains.
The available Diff-Quik cytology smears were evaluated for the presence or absence of atypical small lymphoid cells, plasmacytoid lymphoid cells, wreath-like multinucleated giant cells (MNGC), hallmark cells, fine cytoplasmic vacuoles, monotony, and cellular cohesion. The degree of pleomorphism among the atypical cells was graded. Mitoses and anaplastic cells were counted as the number in 10 contiguous high-power fields (×400). Hallmark cells were defined as cells with a horseshoe-shaped or donut-shaped nucleus with an eosinophilic paranuclear region.
The study complied with the Health Insurance Portability and Accountability Act (HIPAA) and a waiver of consent was obtained from the institutional review board.
FNA and liquid-based cytology specimens were retrieved for all age groups and sexes. A total of 37 patients met the search criteria as outlined earlier. Patients ranged in age from 16 to 91 years with an average age of 49 years and a median age of 47 years. Overall, there was a male predominance (25 patients). Of the 37 patients/cases studied, 32 (87%) also had a surgically excised specimen, 19 of whom had both surgical and cytopathology slides available for review. FNAC specimens from lymph node lesions comprised the majority of the specimens (23 cases). The location of the remaining specimens included the retroperitoneum (5 cases), the mediastinum (3 cases), fluids (cerebrospinal fluid and ascites [1 case each]), liver (1 case), spleen (1 case), and bone/soft tissue (2 cases). The cytologic diagnoses of the 37 patients were as follows: ALCL (13 cases); NHL, large cell type (5 cases); lymphoma (1 case); atypical lymphocytes (3 cases); T-cell lymphoma (4 cases); Hodgkin lymphoma (2 cases); reactive (1 case); malignant neoplasm (3 cases); and nondiagnostic (5 cases). All but 2 of the 32 patients for whom surgical follow-up was available were diagnosed with ALCL. The first case was diagnosed as diffuse large B-cell lymphoma (DLBCL) and the second was diagnosed as anaplastic large B-cell lymphoma. Nine patients were known to have the human immunodeficiency virus (HIV) and 2 patients had a prior history of mycosis fungoides. Table 1 details the demographic data and cytologic characteristics of the reviewed smears; all were diagnosed as ALCL on follow-up surgical biopsy.
Table 1. Demographic Data and Cytomorphologic Features of the Samples
Of the 13 cases that were diagnosed as ALCL on cytology, only 1 was rendered as the primary diagnosis; the remaining 12 cases were either recurrent or the result of the distant spread of a known primary ALCL. The next most common diagnosis was NHL, large cell type. Eleven of 20 cases contained the characteristic hallmark cells. The hallmark cells had the typical C-shaped or circular-shaped nucleus with abundant basophilic cytoplasm (Fig. 1). Some of these cells also had a conspicuous eosinophilic paranuclear globule or hof (Fig. 1a). In 8 of the 11 cases with hallmark cells, wreath-like multinucleated giant cells were observed. The presence of severe pleomorphism in the smears was found to be correlated with >5 anaplastic cells per 10 high-power fields. Anaplastic cells were large, at least 3 times the size of a resting lymphocyte nucleus, and had hyperchromatic nuclei with irregular nuclear contours (Fig. 1b). The nucleoli varied from a single prominent large nucleolus, usually irregularly shaped, to multiple small nucleoli. Many such cells also contained clear, minute cytoplasmic vacuoles (Fig. 1b). The average mitotic count per 10 high-power fields was low (0.75). Only 1 FNAC case did not have a mixed inflammatory background with small lymphocytes, neutrophils, and eosinophils. Lymphoglandular bodies were present from aspiration specimens that involved lymph nodes. However, they were not abundant and generally were absent from nonlymph node aspirates.
Monotony and a lack of pleomorphism of the lymphoid cells was not a common feature among the smears. One notable case demonstrated a monotonous population of medium-sized lymphoid cells without tingible body macrophages. The nucleus was oval with a smooth nuclear contour, a single prominent nucleolus, and a thin rim of pale cytoplasm (Fig. 2). Blank smears were not available and only a diagnosis of lymphoma was rendered. The biopsy showed a monotonous population of medium-sized to small-sized atypical lymphoid cells. However, rare pleomorphic Reed-Sternberg (RS)-like cells along with hallmark cells were found. Both the small cells and the hallmark cells were found to stain positively for CD3 and strongly for CD30, but were negative for ALK and CD20. Based on these findings, the case was diagnosed as ALK-negative ALCL.
A surgical specimen followed 6 cases, which were diagnosed as ALK-negative ALCL. Five of the 6 cases contained >5 anaplastic cells per 10 high-power fields. The remaining case was previously discussed in the previous paragraph. Four of the 6 cases contained typical hallmark cells and/or wreath-like multinucleated giant cells. However, the degree of pleomorphism varied; 4 of the aspiration specimens demonstrated anaplastic morphology. Compared with the ALK-positive cases (3 cases), the ALK-negative cases were diagnosed in older patients (average age of 59 years vs 26 years). The number of anaplastic cells per high-power field appeared to be positively correlated with age (Fig. 3). The ALK-positive tumors were found to have fewer anaplastic cells per 10 high-power fields compared with the ALK-negative tumors (average of 2 vs 9 anaplastic cells). Giant pleomorphic tumor cells with bizarre nuclei (both multinucleated and multilobated) were numerous. The cytoplasm of these anaplastic cells was moderate, with a slate gray-blue color and small vacuoles (Fig. 4). In rare cells, fine red granules were noted. The surgical excision specimen was found to contain similar cells. Five of the 6 surgical specimens stained positive for either CD3 or CD4 and for CD30. It is interesting to note that the above discussed ALK-negative ALCL tumor with a small cell morphology contained no anaplastic cells per 10 high-power fields. Three cases represented progressive disease in patients with a known history of ALCL. Two of these cases raised an interesting differential diagnosis of melanoma, germ cell tumor, and ALCL.
FNAC of a mediastinal mass in a patient aged 22 years yielded many loosely cohesive, medium-sized cells with abundant vacuolated cytoplasm and eccentrically placed nuclei. The nuclear contours were relatively smooth and contained a single prominent nucleolus. Among these plasmacytoid cells were larger cells with irregular nuclear contours, single prominent nucleoli, and multinucleated giant cells (Fig. 5). The mixed inflammatory background had a predominance of small lymphocytes. Only smears were available and a differential diagnosis was offered between a germ cell tumor and ALCL. Seminoma was favored over a nonseminomatous germ cell tumor because of the dyscohesive nature of the malignant cells. On the surgical excision specimen, the plasmacytoid and more pleomorphic cells stained strongly with CD45Ro and CD30, but were negative for ALK. The second case involved a supraclavicular lymph node and the aspirated material was described as yellow and creamy, similar to an abscess. The smear pattern was predominately neutrophils partially obscuring single large atypical cells. Some cells had a histiocytoid appearance and vaguely resembled hallmark cells with reniform nuclei. Severely pleomorphic cells also were found to be present with a high nuclear:cytoplasmic ratio and multilobated nuclei with prominent angulated basophilic nucleoli. A diagnosis of malignant neoplasm was made with a differential diagnosis of either ALCL or another anaplastic malignancy such as melanoma or anaplastic carcinoma.
Two cases, one of which occurred as a paraspinal mass in the retroperitoneum and the other as a supraclavicular lymph node, contained cohesive groups of malignant cells. Both clinical locations implied a malignant process; however, on review, the differential diagnosis was broad. In the retroperitoneal lesion, the highly anaplastic cells were singly distributed as well as grouped together enmeshed in a magenta stroma (Fig. 6). The number of anaplastic cells was high compared with the cohort (16 of 10 high-power fields). A sarcoma, anaplastic carcinoma, or lymphoma was considered. Rare anaplastic cells stained positively for CD3 but no diagnostic cells were present in deeper sections for CD30 staining. The patient had a history of mycosis fungoides and this was believed to represent involvement by his T-cell lymphoma. The surgical excision specimen demonstrated a typical pattern of ALCL with hallmark cells that stained positively for CD3, CD30, and EMA. The patient was diagnosed with ALCL.
The 2 cases in which Hodgkin lymphoma was diagnosed were morphologically distinct. The cervical lymph node aspiration specimen was taken from a patient who was positive for HIV and demonstrated a dyscohesive pleomorphic cell pattern with numerous anaplastic cells, up to 17 per 10 high-power fields (Fig. 7). These cells were binucleated and multinucleated with prominent nucleoli. They had an intermediate resemblance to both RS cells and hallmark cells. Although some large cells had irregular nuclear contours with a C-shaped or irregularly shaped nucleus and multiple prominent nucleoli and basophilic cytoplasm with minute vacuoles, other cells were binucleated or multinucleated with each nucleus having a central prominent nucleolus. The supraclavicular lymph node demonstrated a monotonous smear pattern with very rare RS-like cells. These binucleated or multinucleated cells had single prominent nucleoli, with irregular rather than smooth nuclear contours. The lymphoid cells in the background were large and monotonous and did not resemble reactive benign lymphocytes. Both cases were diagnosed as ALCL on surgical specimens with positive staining of the large anaplastic cells as well as the intermediate cells with CD3 and CD30 and negative staining for CD15. Two cases designated as lymphoma listed only Hodgkin lymphoma and ALCL in the differential diagnosis. In 1 case, a cervical lymph node aspiration specimen demonstrated many RS-like cells and a smaller subpopulation of large lymphoid cells with irregular nuclear contours, abundant cytoplasm, and cytoplasmic vacuoles, which did not resemble RS cells. Although classic hallmark cells or granulomas were not observed, wreath-like multinucleated giant cells were easily found in the smears.
The only benign diagnosis was made in a young patient who had a short history of cough, lethargy, and fever with an enlarged inguinal lymph node. The smear contained a polymorphous population of lymphocytes, tingible body macrophages, and rare enlarged atypical lymphoid cells with a prominent nucleolus resembling immunoblasts or, remotely, RS cells. Rare hallmark cells also were identified. The large “immunoblast-like” cells stained positively for CD30 only. In the context of the reactive background, this pattern was interpreted as consistent with a reactive viral phenomenon such as infectious mononucleosis.
Necrosis accounted for 4 of the 5 nondiagnostic aspirates. All 5 nondiagnostic aspirates were followed with either a surgical biopsy or a surgical excision specimen.
The diagnosis of ALCL by FNAC requires the recognition of the diverse smear pattern. The anaplastic and pleomorphic cells can be lineage-nonspecific with very few clues for classification. Other than the general dyscohesiveness of the cells, the moderate amphophilic cytoplasm, a general lack of lymphoglandular bodies, and varying clinical presentation, it can be very difficult to pinpoint this entity as a lymphoma, let alone ALCL. The hallmark cell is believed to be a clue and rather specific for the diagnosis of ALCL. Although hallmark cells were found in 11 of 20 reviewed aspiration specimens, it was not the predominant cell found in the smears. The combination of hallmark and wreath-like multinucleated giant cells appeared to be predictive of ALCL. However, 7 cases lacked both hallmark and wreath-like multinucleated giant cells but 6 of these cases did have a predominant pleomorphic cell pattern. It is in cases such as these especially that one must be diligent in obtaining adequate additional material for ancillary studies to accurately classify this anaplastic neoplasm. In the current study, NHL was diagnosed in 60% of cases, with subclassification of ALCL made in only 35% of the studied cohort. The current study highlights the need for cytopathologists to maximize the accuracy, efficiency, and effectiveness of FNA by allocating aspiration material to several modalities. This may include preparing a cell block for immunocytochemistry and polymerase chain reaction, and fresh material for cytogenetics and flow cytometry. If an anaplastic neoplasm is encountered, recommended immunostains would include CD45, CD20, CD15, CD30, ALK, EMA, CD3, cytokeratin, vimentin, placental alkaline phosphatase (PLAP), and HMB-45.
The morphologic variants of ALCL delineated by the World Health Organization classification include lymphohistiocytic, small cell, and sarcomatoid, all of which have been described by FNAC.3,4 It appears that these variants do not have a prognostic implication. However, the ALK status of patients with ALCL may impart prognostic significance. ALK chimeric protein expression and the t(2,5) translocation can be found in 72% to 85% of ALCL cases.1 ALK-negative ALCL comprises the remaining cases and is observed to have a variable and unfavorable clinical outcome. In fact, ALK-negative ALCL may be a separate entity that has yet to be fully defined.1 Recent studies have shown that lymph node ALK-negative ALCL has aberrant apoptotic mechanisms, involving Bcl-2 and caspase-3. The resulting resistance to apoptosis may account for its chemotherapy resistance and subsequent poor prognosis.5, 6 In the current study, 6 cases of ALCL were diagnosed as ALK-negative tumors. As reported in the literature, the ALK-negative cases occurred in older patients (average age of 59 years), whereas on average the majority of ALK-positive ALCL cases occur during the first 30 years of life.7 Five of these cases were found to demonstrate hallmark cells and severe anaplasia; the remaining case resembled the small-cell variant of ALCL and was monotonous. The average number of anaplastic cells in the ALK-negative cases greatly outnumbered the number found in the ALK-positive cases (11.2 vs 2). To our knowledge, the cytology of ALK-negative ALCL is not well described in the literature. The current series is small but attempted to begin to characterize the cytomorphology of ALK-negative ALCL.
If the diagnosis of ALCL is considered, the differential diagnosis should be expanded to include a limited number of entities. The most discussed differential is Hodgkin lymphoma. In the common variant of ALK-positive ALCL, the number of anaplastic or hallmark cells are in the minority, similar to the pattern of RS cells noted in classic Hodgkin lymphoma. Both Hodgkin lymphoma and ALCL have neoplastic cells that stain for CD30 antibodies and have a mixed inflammatory background. Mourad et al found that FNA smears of ALCL had significantly more abnormal cells and multilobated cells than specimens of Hodgkin lymphoma.8 A high percentage of abnormal cells, a spectrum with regard to the size and type of pleomorphic or abnormal cells, and multilobated/wreath cells were findings that favored ALCL and not Hodgkin lymphoma. The current study was able to correctly classify Hodgkin lymphoma in >95% of cases based on cytomorphology alone. It is well know that Hodgkin lymphoma may contain granulomas and multinucleated RS cells; however, wreath-like multinucleated giant cells are rarely if ever noted, in contrast to ALCL. In the current study, although we did not compare the ALCL smears with smears from Hodgkin lymphoma cases, we did find either wreath-like multinucleated giant cells or hallmark cells in 65% of the aspiration specimens. RS-like cells are often observed in ALCL; however, the nucleolus usually is irregular and basophilic rather than eosinophilic. The syncytial variant (grade 2, according to the British National Lymphomas Investigation grading system) of Hodgkin lymphoma and the lymphocyte-depleted subtype of Hodgkin lymphoma can have a predominant pattern of numerous pleomorphic, anaplastic, and RS cells. These 2 entities can most closely resemble ALCL, especially the ALK-negative variant. It is interesting to note that infection with the Epstein-Barr virus occurs in a high percentage of patients with syncytial Hodgkin lymphoma and the lymphocyte-depleted variant of Hodgkin lymphoma, a finding that is not noted in ALCL.9, 10 Immunochemistry also can aid in the diagnosis because PAX-5 and CD15 are expected to be positive in classic Hodgkin lymphoma.11 ALK expression is not observed in Hodgkin lymphoma.12 EMA expression may be helpful in the differential diagnosis of classic Hodgkin lymphoma and ALCL because it is well correlated with ALK expression in ALCL and can be used as a surrogate if ALK is not available.13 In both ALK-positive and ALK–negative cases of ALCL, clonal rearrangement of the T-cell receptor by polymerase chain reaction can be demonstrated in >90% of cases.14
Other malignant diagnostic entities that enter the differential diagnosis, based on both morphologic and immunocytochemical grounds, include diffuse large B-cell lymphoma, embryonal carcinoma, melanoma, and sarcoma. All these entities can stain positively for the CD30 antigen and have anaplastic morphology. Although seminoma may cytologically resemble ALCL, with dispersed single pleomorphic cells in a polymorphous lymphoid background, they are distinct immunocytochemical entities. Because seminoma should not stain for ALK or CD30, the admixed benign lymphocytes may stain for T-cell markers; therefore, caution should be used. Other immunohistochemical stains that are positive in seminoma include PLAP, c-KIT, and Oct-4.15 Embryonal carcinoma can exhibit poorly differentiated or anaplastic morphology and also stain positively for CD30 and EMA but be negative for ALK. The loosely cohesive anaplastic lymphoid cells of ALCL, especially the sarcomatoid variant, admixed with fibroblasts and stroma may simulate a sarcoma, especially in the retroperitoneum. Although uncommon, high-grade sarcomas occasionally may express CD30, but should be negative for ALK, except in the case of inflammatory myofibroblastic tumors (IMTs).16 IMTs exhibit a spindled morphology and are expected to be positive for ALK but negative for CD30. The cases of retroperitoneal ALCL in the current series did not have a spindle cell morphology.
ALK-positive anaplastic B-cell lymphoma has been reported in the surgical pathology literature.17 Morphologically, ALK-positive anaplastic DLBCL has an immunoblastic or plasmablastic appearance. These tumors are described as consisting of a monotonous population of large lymphoid cells with a central, prominent nucleolus. In distinguishing these cases from ALCL, CD30 and T-cell antibodies are helpful because ALK-positive anaplastic B-cell lymphoma is reportedly negative for these markers and ALCL should be negative for the plasmacytic markers (ie, CD138 and immunoglobulin light chains). EMA may be equivocal. Although ALK-positive anaplastic B-cell lymphoma exhibits a sinusoidal pattern of spread similar to ALCL, cytologically hallmark cells or cell variants have not been reported. However, RS-like cells may be observed.
Although difficult, ALCL can be accurately diagnosed by FNAC. Problems with reproducibility and accuracy are not limited to cytopathology; 1 study reported a reproducibility of 46% based on histomorphologic grounds alone.18 The importance of obtaining adequate material, keeping a wide differential diagnosis, and utilizing both immunochemistry and molecular techniques when one encounters an anaplastic neoplasm cannot be underestimated. Careful cytologic evaluation of hallmark cells and wreath-like multinucleated giant cells will provide a clue for further investigation, including the necessary staining for ALK or procuring material via FNA biopsy for cytogenetic studies. CD30 status is no longer sufficient for a diagnosis of ALCL because the prognostic implications of ALK are significant.