Diagnostic Approach to CD5+/CD23+ Leukemic Non-Hodgkin Lymphomas Lacking Lymphnode Histopathology.

Lymphoproliferative disorders (LPD) may be characterized at presentation by peripheral blood (PB) and/or bone marrow (BM) involvement. These may represent the unique diagnostic sources to subtype LPD in patients who lack nodal involvement or are not eligible for aggressive diagnostic procedures. Lymphocyte morphology and immunophenotype allow to distinguish typical chronic lymphocytic leukemia (CLL) from other leukemic B-cell LPD, though they cannot provide a specific classification of the latter. CD5+/CD23+ LPD are a heterogeneous group including typical/atypical CLL and leukemic B-cell non-Hodgkin lymphomas (B-NHL). Aim of this study was to define the value of BM and PB as the only tissue available for the differential diagnosis of B-LPD, focusing on CD5+/CD23+ subgroup. Between January 2003 and December 2004, we evaluated at our institution 305 consecutive patients characterized by clonal B-lymphocytosis. Among 205 CD5+/CD23+ cases, 135 (66%) had typical features of CLL, as for PB standard morphologic/immunophenotypic criteria, 70 (34%) were provisionally defined as B-NHL/atypical CLL. 36 of them underwent a lymphnode biopsy. In the remaining 34, diagnosis was approached only by BM biopsy in 27 cases (20 lacked superficial and/or internal adenopathies and 7 with superficial adenopathy could not undergo a lymphnode biopsy), while in 7 both a BM and node biopsy were performed. BM biopsy was analyzed by histopathology and immunocytochemistry (CD20, CD79a, Ig light chain, CD5, CD23, CD43, CD10, bcl-2, cyclinD1). Median age of patients (24 males and 10 females) was 68 years (range 39–78). Spleen enlargement was detected in 10 cases. Median WBC count was 17x10⁹/l (range 8.5–90), lymphocyte count 9x10⁹/l (range 4.3–69). In all cases, PB morphology was not typical of CLL: >10% of lymphocytes were large in 9 cases, cleaved in 15, nucleolated in 2, villous in 2 and with a mixed pattern in 6. Matutes’ immunophenotypic score was 5–4 in 25 cases and 3–2 in 9. BM infiltration was diffuse in 9 cases, interstitial in 21 or nodular +/− interstitial in 4. Distinctive lymphoma infiltrates, such as an intrasinusoidal pattern or proliferation centers, were recognized in 2 and 3 cases, respectively. CyclinD1 was negative in 5/5 evaluated cases. A final diagnosis was reached by PB/BM evaluation in 29 out of 34 cases (85%): small lymphocytic lymphoma (SLL/CLL) in 24 (70%), lymphoplasmacytic lymphoma in 2 (6%), marginal zone lymphoma in 2 (6%), follicular lymphoma in 1 (3%). Five cases (15%) remained unclassified. Among the 7 patients who underwent also a lymphnode biopsy, a discordant diagnosis between BM and lymphnode was observed only in 1 case that proved a Richter syndrome. In summary, CD5+/CD23+ clonal B-LPD not fulfilling criteria for CLL diagnosis, lacking node biopsy, can in most cases be adequately classified by PB/BM morphology/immunophenotype and BM immunohistochemistry. A small proportion of cases remain unclassified. Notably, 17% of CD5+/CD23+ “non-CLL” LPD are leukemic B-NHL. BM biopsy represents a valuable source to define as B-NHL with leukemic spillover cases not fulfilling a diagnosis of CLL, even in the absence of primary tissue histopathology.