Classical hairy cell leukemia (HCL) coexpressing both CD25 and CD103 is highly responsive to the treatment with cladribine and pentostatin (complete response rate, 75–95%; overall response rate, 86–100%). HCL variant in the literature is characterized by HCL-like morphology, lack of CD25 coexpression, and variable immunophenotypes indistinguishable from those of splenic marginal zone B cell lymphoma (

Matutes, et al.
). Importantly, HCL variant is associated with a poor response to standard therapy. It is unclear if HCL cases lacking CD25 but otherwise phenotypically identical to classical HCL belong to HCL or HCL variant, and if the standard therapy is effective in these patients. To further delineate features of HCL and its variants, we analyzed immunophenotyping data, by flow cytometry or immunohistochemistry, or by both when feasible, in 260 consecutive patients with HCL (
Dong, et al.
Mod Pathol
). The diagnosis was established by hairy cell morphology together with coexpression of strong CD20 and CD22, bright CD11c, and CD103. Clinical data were obtained for a subset of cases in which expression of CD25 was distinctly absent. Our results were consistent with the literature in that HCL has a consistent and unique immunophenotypic profile, which allowed detection of residual HCL by flow cytometry at levels as low as 0.1% of total cells. In addition, approximately 20% and 37% of classical HCL coexpressed CD10 and BCL-1 respectively, which may significantly confuse the diagnosis in an incomplete work-up. Interestingly, there were 43 cases (20% of all CD103+ cases) that lacked CD25 but were otherwise identical to classical HCL in their uniform phenotypic profiles. Compared with classical HCL, patients with the CD25- HCL were generally older (medium age: 59yrs vs. 79yrs; p=0.001) and frequently had leukocytosis (medium WBC: 3.0x109/L vs. 24.5x109/L; p=0.014). Among 14 patients with follow up data, 7 were treated with cladribine or pentostatin. The complete response (CR) rate was 14.3% (1/7) and the overall response rate was 57.1% (4/7). Three patients had no response. One patient who had an initial partial response (PR) to pentostatin died of the disease 10 months after the diagnosis. Of others, an additional two patients achieved CR and PR upon initial treatment with fludarabine and Rituxamab respectively. Four patients were untreated and were alive with disease (follow up, 21–41 months). One patient died of the disease in 2 years and treatment for this patient was unknown. These clinical features overlapped with those of HCL variant in the literature. However, unlike the HCL variant that has significant phenotypic heterogeneity, including lack of CD11c and CD103 in a substantial number of cases, the CD25- HCL was remarkably uniform in phenotype and can be easily identified by immunophenotyping. In conclusion, these results suggest that lack of CD25 in HCL defines a clinically distinct chronic lymphoproliferative disorder, which appears to require different clinical management.