Title: Hairy cell leukaemia‐variant and splenic red pulp lymphoma: a single entity?
Abstract: In a recent correspondence article, Hockley et al (2010) reported that the hairy-cell lymphoproliferative disorders display distinct Immunoglobulin Gene Heavy Variable (IGHV) repertoires. Notably, it was suggested that the hairy cell leukaemia-variant (HCL-V) repertoire, analysed for the first time for IGHV mutational status, shares more similarities with splenic marginal zone lymphoma/splenic lymphoma with villous lymphocytes (SMZL/SLVL) than with hairy cell leukaemia (HCL) (Hockley et al, 2010). We agree with the authors position, who then called into question the origin of HCL-V, as we did previously when describing a series of splenic lymphomas cases (Traverse-Glehen et al, 2008). Those lymphomas were identified from our SMZL hospital registered cases because they presented with clearly distinct circulating cell morphology and spleen pattern. These cases, which are now categorized as unclassifiable splenic lymphomas according to the recent World Health Organization classification (Swerdlow, 2008), displayed clinical and morphological features strongly reminiscent of HCL-V and have the same IGHV pattern as the one described by Hockley et al (2010). They indeed overlap with both SMZL and HCL, but also show original features and a quite indolent clinical course. We thought at this time that classifying them as SMZL-V or HCL-V would be based upon subjective criteria and therefore proposed to consider them as a provisional entity that might be called splenic red pulp lymphoma (SRPL) with numerous basophilic villous lymphocytes, thus reflecting their characteristics and the absence of objective criteria to determine the cell of origin. In our report (Traverse-Glehen et al, 2008), 37 cases were retrieved from our SMZL collection, based on particular cytology in peripheral blood (PB), characterized by homogenous lymphoid cells with condensed chromatin, basophilic cytoplasm and well-defined polar villi that were unevenly distributed around the cells and broad based. These villi were distinct from those observed in other well-described entities and from the short villoid expansions currently observed in classic SMZL. However this cytology was close to that described in HCL-V except for the characteristic large prominent nucleolus present only in a few cases of our series. In addition, our cases shared numerous features with HCL-V, distinct from classic HCL or SMZL. Indeed they both presented at an older age and showed a lesser male predominance than HCL, a large splenomegaly and lymphocytosis without pancytopenia (Matutes et al, 2003). The immunophenotype was similar, with expression of IgM+D, IgM+G, IgM or IgG alone with an equal repartition of kappa and lambda light chains. The expression pattern of CD11c, CD76, CD103, CD123, CD25 and Annexin A1 is intermediate between SMZL and HCL. Other common features are more reminiscent of SMZL, in particular the bone marrow pattern of infiltration with a good hematopoietic reserve and a mild infiltrate, predominantly intrasinusoidal (Matutes et al, 2008). In addition, the chromosomal abnormalities described in SRPL are those recurrently described in SMZL (7q deletion, partial trisomy 3q, and complete trisomy 18) (Callet-Bauchu et al, 2005; Matutes et al, 2008). Although the spleen histology of HCL-V is not well described, it is generally considered as being similar to HCL, with primarily involvement of the red pulp, a particular feature also observed in the 12 SRPL cases of our series for which the spleen histology was available and also found in six recently diagnosed additional cases. We never observed a marginal zone enlargement with biphasic pattern, characteristic of SMZL. The clinical evolution is indolent both in SRPL and HCL-V, with a median survival between 9 and 10 years and we observed a significantly (P < 0.05) longer progression-free survival in SRPL compared with SMZL. The clinico-biological comparison between B-cell neoplasms presenting cytoplasm projections in PB and splenomegaly is summarized in Table I. In the report from Hockley et al (2010), the frequent usage of IGHV4-34 (7/41, 17.1%) with the high number of mutated cases (30/41, 73%) tallies with our result in SRPL (15.1% of IGHV4-34: 5/33; 78.7% mutated: 26/33). However the IGHV1-2 families seem more frequent in our experience of SMZL (20/70, 28.5%) than in their series (22/113, 19.4%), and we observed a lower frequency of IGHV4-34 (6/70, 85% vs. 14/113 or 123%). In an other recent series (Arons et al, 2009), IGHV4-34 was also observed more frequently in a subgroup called HCL-V (40%, 8/20 cases) than in classic HCL (10%, 6/60 cases). However the authors did not compare their results with the English group ones or our SRPL series. Interestingly, the IGHV4-34 rearrangements were more frequently associated with unmutated IGHV sequence. IGHV4-34 patients had a greater white blood count at diagnosis than cases of HCL or cases of HCL-V with another rearranged IGHV segment, a lower response rate and progression-free survival after first-line cladribine, and a shorter overall survival from diagnosis. The authors concluded that IGHV4-34 could be recurrent but not quite specific for the diagnosis and an independent prognostic factor when hairy cells are observed in PB. However, in this report, neither spleen histology nor PB blood cytology were reported. The median age at diagnosis for HCL-V cases was closer to that of classic HCL patients than what is usually described, suggesting that these cases are distinct from the initial description of HCL-V and could represent, as concluded by the author, a true variant of HCL with poor prognosis. In conclusion, the differential diagnosis of rare indolent mature B-cell neoplasms with cytoplasm expansions in PB cells includes SMZL/SLVL, SRPL, HCL, HCL-V and Japanese variant HCL (HCL-J). It is important to accurately diagnose these entities, as they have different clinical and biological features, particularly regarding their response to purine analogue-based treatment or splenectomy. While SMZL and HCL are well-described entities, no such clear definition is made for the others. Hence, we thus agree with Hockley et al (2010), that HCL-V could be more related to SMZL and closer to our description of SRPL, thus suggesting that they may represent overlapping entities.