CD26 (dipeptidyl peptidase IV, DPP IV) is widely expressed by T and natural killer (NK) cells, epithelial and endothelial cells of different tissues, and it is strongly upregulated in activated B-cells; moreover it plays a regulatory role in the neoplastic transformation and progression of various types of tumours. CD26 expression was evaluated by means of flow cytometry in various peripheral B-cell lymphoid tumours: 12 follicular and 12 mantle cell lymphomas, 20 multiple myelomas (MMs), 12 hairy cell leukaemias (HCLs), 112 chronic lymphocytic leukaemias (CLLs), 20 CD5(negative) B-cell chronic lymphoproliferative diseases (CD5(neg) B-CLPDs) and 12 diffuse large cell lymphomas (DLCLs). CD26 expression was absent or barely detectable in follicular and mantle cell lymphomas, high in MMs and HCLs, and variable in CLLs, in CD5(neg) B-CLPDs and in DLCLs. CD26 significantly correlated with CD49d and CD38 expressions (p < 0.0001) in B-CLLs, and there was a significant correlation between CD26 and ZAP-70 expressions or IgVH mutational status (p < 0.0001). After a median follow-up of 36 months, 65 B-CLL patients were treated; taking 10% as the best CD26 cut-off value, Kaplan-Meier curves revealed a significantly shorter time to treatment in the CD26-positive cases (p < 0.0001). Overall, our data indicate that CD26 expression may identify subsets of B-CLL patients with an unfavourable clinical outcome in terms of therapeutic need, thus suggesting its potential role as a marker (together with CD38 and CD49d) in a future routine cytofluorimetric panel to be validated for the prognostic stratification of B-CLLs.

1259 Poster Board I-282 We describe the clinical-biological features of 63 cases of variant B chronic lymphocytic leukemia (v-B-CLL), characterized by a mantle cell lymphoma-like immunophenotype, atypical cytomorphology in absence of t(11;14)(q13;q32) in FISH analysis. A historical series of 130 B-CLL was used as comparison. The v-B-CLL were significantly different from the B-CLL in terms of the following clinico-hematological variables: age <70 yrs (p <.001), lymphocytosis <20 × 109/ (p <.001), lymphocyte doubling time < 12 months (p = .02), high serum β2-microglobulin levels (p <.001), and splenomegaly (p = .002). Considering immunophenotipic features, CD38 and CD49d expression were significantly more expressed in v-B-CLL than B-CLL (p <.001); whereas, no statistical difference was observed for ZAP-70 reactivity. Considering the IgVH mutation status, there were more patients mutated in the v-B-CLL group than in the B-CLL group (p = .001). Other significant differences were found about the frequency of the recurrent chromosome alterations, evaluated by means of FISH analysis: trisomy 12 was more frequent in v-B-CLL ( p<.001), while del13q14, considered as a single alteration, was more frequent in B-CLL (p=.008). Gene expression profiling of a panel of 9 v-B-CLL compared with 60 B-CLL samples indicated that the variant group is characterized by a specific molecular pattern of gene expression. In particular we observed the upregulation of tumor protein D52 (TPD52), and that of 6 genes (AFF1, GMPS, PICALM, JUN, REL, RAC2) known to be protooncogenes. Furthermore, we found that upregulated genes with apoptosis related functions (IL-7, HSP90B1, NOTCH2, BECN1, ANXA4, MCL1) are all negative regulators of apoptosis. Microarray analysis revealed that various genes (TRIM38, EEF1D, CASP1, MALT1, RHOH0) involved in the I-kB kinase/NF-kB cascade of the canonical NF-kB signaling pathway, are furtherly upregulated in v-B-CLL. Furthermore, among the genes found differentially expressed in SAM analysis, we observed also the upregulation of CD1c (according to the surface expression of this antigen), OSBPL3 and ITGA4. After a median follow-up of 55 months (range 4-196) and 60 months (range 6-180), 25/42 (59%) v-CLL and 55/93 (59 %) CLL pts were treated. TTT was significant different between 2 groups when the IgVH mutational status was considered (p= .006). Median OS of v-CLL subset was 112 months vs 171 months of CLL subset. When the IgVH mutational status was considered, mutated cases showed a worse OS even if a statistical difference was not observed (p= 0.062). In conclusion, our study identifies, on the basis of a defined CFM-FISH diagnostic approach, a variant form of B-CLL that shows peculiar biological and clinical features that should be considered in the future clinical and prognostic studies. The inclusion of this form in B-CLL study could alter the interpretation of results, especially related to biological markers. Disclosures No relevant conflicts of interest to declare.

In the present study, we describe the clinico-biological features of 63 cases of variant B-CLL (v-B-CLL) defined according to a previously described diagnostic algorithm for CD5+ mature B-cell leukemias, based on the combined use of Cytofluorimetric (CFM) analysis and t(11;14)(q13;q32) detection by means of fluorescence in situ hybridisation (FISH). These leukemias were characterized by SIgbright/CD23+ or CD23+/− , CD79b/CD20 bright, and negativity for t(11;14)(q13;q32). A historical series of 112 classical B-CLL was used as comparison. The mean age was 61 years (33–85) and M/F ratio 1.9. The v-B-CLL cases were significantly different from the B-CLL cases in terms of the following clinico-hematological variables: age <70 yrs, lymphocytosis <20 ×109/L (P <.0001), lymphocyte doubling time ≤ 12 months (P=.04), high serum ß2-microglobulin levels (P=.003) and presence of splenomegaly (P=.007). No differences were found in terms of sex, Hb levels, platelets count, presence of serum monoclonal component, serum LDH levels and HCV-Ab positivity. In v-B-CLL and B-CLL pts the median follow-up was respectively 56 and 84 months. The median overall survival was 80 months in v-B-CLL pts while it was not reached in B-CLL pts; 37 pts (58.7%) in the first group started chemotherapy vs 84 pts (56.4%) in the latter. The analysis of leukemic cell biological characteristics related to the prognosis was also made in part of the cases. A classic or mixed-CLL cytomorphology was seen only in 22/61 pts (36.1%). The percentage of CD38 positive cases was higher in v-B-CLL (60.3) than in B-CLL (44.6; P=.028); the IgVH mutation status, evaluated in 30/63 cases, was more frequently hypermutated in v-B-CLL group (23/30) than B-CLL (P=.03); the percentage of CFM Zap-70 positivity, evaluated in 42/63 cases, was similar in two groups (47.6 vs 57.4%, respectively). Interestingly, significant differences were found about the frequence of the recurrent chromosome alterations, evaluated in 48/63 cases, by means of FISH analysis: trisomy 12 was more frequent in v-B-CLL (37.5 vs 18.4%, p=0.02), while del13q14 was more frequent in B-CLL (14.5 vs 37.8%, P=0.007). In both groups, del 11q22.3 and del 17p13.1 were relatively rare (4 vs 8%, and 2 vs 7.7%, respectively). In conclusion, our study identifies, on the basis of a defined CFM-FISH diagnostic approach, a variant form of B-CLL that shows significant differences in terms of genetic and clinical features.

We reviewed flow cytometric immunophenotyping (FCI) data of 470 tissue suspensions suspected of being involved by lymphoma (371 lymph nodes, 263 surgical biopsies and 108 fine needle aspirations,16 spleens, 6 tonsils, and 77 other tissues) and we have compared corresponding histologic diagnosis. The screening panel of FCI (CD45, CD19, CD3, CD4, CD8, and sIgκ/sIgλ ratio) identified three main groups: 239 cases with demonstrable light chain restriction (sIgκ/sIgλ ratio ≤ 0.5 or ≥4); 37 cases with demonstrable T cell proliferation and polyclonal B cell population; 194 cases with polyclonal B cell and normal T cell populations (146 cases) or with not detectable CD45 reactivity and other leucocyte markers (48 cases). In group 1, 235 cases were then evaluated by means of the following markers: CD5, CD10, CD11a, CD11c, CD20, CD22, CD23, CD30, CD38, CD43, CD79a–b, CD103, CD138, FMC7, and, in group 2, 30 cases were evaluated by means of the following markers: CD1a, CD2, CD5, CD7, CD16, CD26, CD43, CD56, CD57, CD45RA/RO, anti-TCR-ab gd. The complete analysis identified in group 1 four main diagnostic subsets: A (26 cases) characterized by CD5+, CD23+/±, sIg dim (CLL-like); B (36 cases) characterized by CD5+, sIg bright (MCL/CLLv); C (89 cases) characterized by heterogeneous expression of Ig, CD20 and CD43 (lymphoproliferative syndromes CD5−, excluding hairy cell leukemia); D (84 cases) characterized by CD10+, CD20 bright, CD43− (follicular NHL like). In cluster B, FISH analysis for t(11;14) resulted positive in 15/27 cases; in cluster D, FISH analysis for t(14;18) resulted positive in 60/60 cases. Histologic analysis convalidated FCI data in 23 cases of subset A (88%), in 29 cases of subset B (80%) and in 60 cases of subset D (71%). Subset C included 39 cases of LCL and 15 cases of MZL. In group 2, all cases studied expressed an aberrant T phenotype but, with the exception of 1 case of NK proliferation, 1 case of likely Sezary syndrome/Micosis Fungoides (CD4+CD7−CD26−) and 1 case of T lymphoblastic lymphoma (CD3cy+TdT+), in the other 34 cases was not possible identify other peculiar subset. In this group, histologic analysis included 19 cases of NHL T (63%). Considering NHL vs no NHL, sensibility (SE) and specificity (SP) of FCI analysis resulted 88% and 92%, respectively; positive predictive value (PPV) was 96% and negative predictive value (NPV 78%). Considering B-cell NHL only, SE was 91%, SP 95%, PPV 97% and NPV 84%. Even if histology is the basis for the diagnosis of lymphoid tumors, however our study supports that FCI can play an important role mainly in B-NHL diagnosis, combining rapidity of analysis with a multiparametric analysis, also in presence of size limitated samples.

CD26 (dipeptidyl peptidase IV, DPP IV) is a multifunctional type II cell surface glycoprotein that is widely expressed on T and natural killer cells, as well as on epithelial, endothelial and acinar cells of different tissues; its expression on B cells is very low but it is greatly upregulated following activation. We evaluated, by means of flow cytometry, the expression of CD26 in various types of B-cell lymphoid tumors: Follicular Lymphoma (Fo-Ly, 12 cases), Mantle cell Lymphoma (MCL, 12 cases) Multiple Myeloma (MM, 20 cases), Hairy cell Leukemia (HCL, 12 cases), B-cell Chronic Lymphocytic Leukemia (B-CLL, 112 cases), CD5 negative B-cell Chronic Lymphoproliferative Diseases (CD5neg-B-CLPD, 20 cases) and Diffuse Large cell Lymphoma (DLCL, 12 cases). CD26 expression was absent or low of Fo-Ly and MCL, high on MM and HCL, variable on B-CLL, CD5neg-B-CLPD and DLCL. Fluorescence intensity of positive cells was dim in B-CLL and CD5neg-B-CLPD, heterogeneous in DLCL, and bright in HCL and MM. Interestingly, in CLL patients, CD26 expression was significantly correlated with CD49d and CD38 expressions (p<0.0001); moreover Spearman’s test revealed a significant correlation between CD26, CD38 or ZAP-70 expression and IgVH mutational status (p<0.0001). We also found a significant correlation between CD26 expression and the presence of FISH chromosomal abnormalities grouped on the basis of their prognostic relevance (p<0.0001). After a median follow-up of 36 months, 65/107 CLL pts were treated; taking 10% of positive cells as best cut-off, Kaplan-Meier curves showed that the time without therapy was significantly longer for CD26 negative cases compared with CD26 positive cases (P <0.0001). Moreover the concordant negativity of CD26 and CD38, or ZAP-70 and the combination of CD26 negativity with IgVH mutated status, identified subsets of CLL patients with a favourable outcome in terms of need of therapy.