TO THE EDITOR Hairy cell leukemia (HCL) is a B cell disorder with distinct features. A variant form, HCL-variant, was recognised by Cawley et al1 in 1980 and was later described as prolymphocytic variant of HCL.2 HCL-variant is an uncommon disorder accounting for approximately 0.4% of chronic lymphoid malignancies evolving with lymphocytosis and representing 10% of all HCL cases. Early reports in a few patients have documented resistance to Interferon-alpha (IFN) and partial or no response to pentostatin (2 deoxycoformycin) and/or cladribine (2-chorodeoxyadenosine).3,4 There is so far little information on the natural history of this disease, details of the tissue histology and the response to therapy particularly to the purine analogues which are highly effective in typical HCL. We describe here the clinical and laboratory features on 52 patients with HCL-variant diagnosed between 1981 and 1998 and emphasise the differential diagnosis with other B cell disorders evolving with splenomegaly and the response to therapy and survival. Median age of our patients was 71 years (range, 48–92) and male/female ratio: 1.6. The most frequent presenting features were abdominal discomfort usually related to splenomegaly, those derived from cytopenias and, less often, was an incidental finding. Splenomegaly was present in 85% of patients, hepatomegaly in 19% and lymphadenopathy in only 4%. Two patients developed skin lesions during the disease course and 12% lymphadenopathy. The WBC was raised (median, 34  109/1) in 47 (90%), but all patients had atypical circulating hairy cells. The absolute numbers and proportion of monocytes were normal, anemia was detected in 29% and thrombocytopenia in 43%. The circulating cells were of medium to large size, had a prominent nucleolus and abundant villous cytoplasm with variable basophilia (Figure 1a). Bone marrow could be aspirated and showed normal cellularity with infiltration (5 to 85%) by atypical hairy cells. Histology in 29 cases showed slightly increased cellularity, good hemopoietic reserve and moderate increased reticulin. Most cases (79%) had interstitial infiltration with scattered hairy cells lying within the sinusoids and 21% a mixed pattern (nodular and interstitial). The infiltration was minor in 12 and was only apparent by immunohistochemistry. The interstitial pattern became diffuse in association with disease progression in two cases. Spacing among the cells varied and in five cases, was indistinguishable from that of typical HCL (Figure 1b). Spleen specimens in 16 cases showed expansion of the red pulp with no residual or atrophic white pulp and, as in typical HCL, lymphoid cells were lying within dilated sinusoids. Evidence of pseudolake formation was seen in three cases. Skin histology showed extensive dermal lymphoid infiltrates without evidence of epidermotropism. Lymph node histology in three cases showed diffuse infiltration by lymphoid cells. Liver histology showed lymphoid infiltration in the portal tracts and within the sinusoids. By immunohistochemistry, the majority of cells in all the tissues examined were CD20, CD79a and DBA44 positive and CD3 negative. Immunophenotypic analysis of the circulating and bone marrow cells showed the presence of a clonal B cell population with a mature phenotype. All but two cases expressed surface immunoglobulin (Smlg) and the most frequent lg heavy chain expressed was IgG (20 cases), either alone or in combination with other heavy chains. Cells from the two Smlg negative cases did not express cytoplasmic lg heavy and light chains either. Lambda light chain was more frequently expressed than kappa (kappa/lambda ratio: 0.75). Cells from >95% of cases expressed in the membrane FMC7, CD19, HLA-DR and CD22; the monoclonal antibodies (McAb) CD79b and CD24 were positive in 24% and 27% of cases, respectively, and <15% were CD10 and CD38 positive; CD23 and CD5 were, as a rule, negative. Considering the four markers characteristic of hairy cells: CD11c was strongly positive in 87% and CD103 in 60% of cases while HC2 and CD25 were negative in most of them (<10% positive cases). According to the scoring system used for distinguishing HCL from HCL-variant, 26% of cases had a score of 0, 45% scored 1, 26% score 2 and only two cases scored 3. This is in contrast to typical HCL in which the scores are 3 or 4 in 98% of cases.5 Cytogenetic analysis showed various numerical and/or structural complex translocations in nine patients including two with abnormalities involving the 14q32 chromosome breakpoint and another with a t(2;8) similar to that seen in Burkitt's lymphoma. Abnormalities of chromosome 17 were present in only two patients but fluorescence in situ hybridization (FISH) analysis showed p53 deletion in cells from all 12 cases investigated. Seven patients were not treated and three are alive at 6 to 129 months with stable disease. Two of these patients have died at 10 and 11 months of disease unrelated causes and the follow-up was lost at 8 and 26 months in the two other patients. Forty-two patients underwent one or more treatment modalities and the best responses were seen with splenectomy (Table 1). A good partial response (PR) to splenectomy was experienced in 74% lasting from 1–10+ years (median: 48 months), four did not benefit and one patient was not evaluable as he died from sepsis 1 month after surgery. Four patients received splenic irradiation and one of them had a PR, but died 11 months later. Alkylating agents (chlorambucil or cyclophosphamide) plus or minus prednisolone were used in 12 patients and only one benefited. Two patients received CHOP and achieved PR but relapsed 12 and 15 months later. IFN was used in 14 patients and only two had a PR lasting for 12 months. Fifteen patients received pentostatin, and eight had a PR. Of the eight patients treated with cladribine, four achieved PR. Three of these had previously responded to pentostatin. Fludarabine was used in three patients and one had a PR lasting for 30 months. Complete follow-up data were available in 35 patients and it was lost at 1–154 months in the remaining. Fourteen (40%) patients are alive at 7 months to 17+ years from diagnosis while the rest have died 1 month to 15 years from diagnosis. In nine patients the causes of death were unrelated to the disease and in 12, they were disease-related. In three patients, the disease underwent transformation at 6, 13 and 14 years from diagnosis. All developed abdominal masses, B symptoms and one of them skin lesions. The WBC increased rapidly to 250–300 in two of the three patients with large transformed blast cells which had identical phenotype to the cells at presentation. CT scan confirmed massive abdominal and/or retroperitoneal nodes and histology demonstrated infiltration by large cells with a CD20+, DBA44+ phenotype. The three died with progressive disease at 1 to 12 months following transformation. Figure 2 illustrates the overall survival and probability for survival in 47 patients with HCL-variant and compares survival with 188 patients with HCL. The overall median survival for HCL-variant was 9 years and 15% are alive 17 years from diagnosis, while median survival was not reached for HCL at 12 years. Our data show that HCL-variant has similarities but also differences with the typical form of HCL. The patients are older without male predominance and there is a marked lymphocytosis with no monocytopenia and neutropenia. Other cytopenias are largely due to hypersplenism rather than bone marrow infiltration. The morphology of the circulating cells is different from typical hairy cells as it is the immunophenotype. In HCL-variant, the lymphoid cells are, as a rule CD25 and HC2 negative, although like hairy cells do express CD11c and CD103. Histologically, the patterns of bone marrow and spleen infiltration are almost identical to typical HCL and thus, morphology and membrane markers are the main diagnostic tests to distinguish typical HCL from its variant form. HCL-variant should be distinguished from B-prolymphocytic leukemia (B-PLL) and splenic lymphoma with villous lymphocytes (SLVL) or marginal zone lymphoma. Cell morphology and tissue histology are helpful to distinguish HCL-variant from B-PLL and SLVL. B-prolymphocytes do not show villous projections and have less abundant cytoplasm, while SLVL cells are smaller than HCL-variant cells, have more condensed chromatin and the cytoplasm is villous but scanty and more basophilic. The bone marrow histology in SLVL does not show spacing among the cells and instead a predominantly nodular and/or intrasinusoidal pattern; the latter pattern, nevertheless, may be similar to that of HCL-variant. Spleen histology in both, B-PLL and SLVL, is clearly different form HCL-variant and shows expansion of the white pulp with a typical pattern of infiltration in SLVL.6 Immunological markers may help too in distinguishing these disorders from HCL-variant. Distinction is relevant from the clinical point of view due to the different response to therapy and outcome. There is no evidence of a recurrent chromosomal or molecular change in HCL-variant but, unlike typical HCL, chromosome 5 is rarely involved. Of interest are the few cases with a t(2;8) with a breakpoint involving c-myc. Molecular studies using Northern blot have shown that cells from HCL variant, unlike typical hairy cells, contain c-myc transcripts and it has been postulated that c-myc transcripts in HCL-variant might be one of the mechanisms by which cells are resistant to IFN as they do not modulate by 'in vitro' IFN treatment. Although abnormalities involving 17p are rare, our recent study has shown that cells from most HCL-variant carry monoallelic deletions of the p53 and that patients with higher numbers of cells with deleted p53 tend to progress and/or undergo transformation. The clinical course of HCL-variant is chronic with a long lymphocyte doubling time; the main problems are those derived from splenomegaly and hypersplenism. The disease is resistant to IFN and only a half of the cases experienced PR to either pentostatin or cladribine in contrast to the excellent response seen to these drugs in typical HCL. The lack of response of HCL-variant to cladribine has also been documented in six patients.4 One out of three patients achieved a PR with fludarabine and there has been another case reported. Our experience suggests that splenectomy is the best option because it corrects hypersplenism and removes the bulk of disease. Thus, two-thirds of the patients achieved good long lasting PR to splenectomy. Splenic irradiation may be an alternative in patients with high surgical risks but seems less effective than splenectomy. Despite the chemoresistance, the clinical course in most patients was protracted with a median survival of 9 years, but worse than in typical HCL (median survival not reached at 12 years). The continuous downward trend in survival may reflect the lack of response to purine analogues in HCL-variant and the high CR in HCL.7 HCL-variant, rarely undergoes transformation. Such event is estimated here to be around 5% which compares to that described in SLVL and CLL. In such instance, prognosis is poor with resistance to therapy and early death. A form of transformation with large hairy cells has also been described in HCL. In summary, HCL-variant has distinct morphological, immunophenotypic and histological features and pursues a chronic clinical course. At present, there is no effective therapy for HCL-variant with splenectomy providing, so far, a good palliative therapeutic measure.
|
