BCL-2 Expression at the Time of Diagnosis Affects the Clinical Outcome of Patients with Germinal Center and Non-Germinal Center Diffuse Large B-Cell Lymphoma (DLBCL) Treated with Chemo-Immunotherpy

Abstract 3130

DLBCL has been recognized as a heterogeneous disease varying in molecular biology and clinical outcome. The use of genetic expression profiling has led to the sub-classification of DLBCL into germinal center B-cell like (GCB) and non-germinal center B cell like (non-GCB) based on the cell of origin of the neoplastic B-cell. Immunohistochemistry (IHC) algorithms had been developed and validated to identify GCB or non-GCB DLBCL. Deregulation of Bcl-2 family member of proteins plays an important role in the development, progression, and prognosis of various subtypes of B-cell neoplasms, including DLBCL. Bcl-2 protein expression is a previously known negative prognostic indicator of clinical outcome in DLBCL treated with antracycline-containing combination chemotherapy (e.g. CHOP) in the past. In the post-rituximab (R) era (e.g. use of upfront R-CHOP), the negative prognostic value of Bcl-2 protein expression needs to be reevaluated to ensure its validity. To study the prognostic value of Bcl-2 in patients with either GCB or non-GCB DLBCL, we retrospectively analyzed differences in progression-free survival (PFS) and overall survival (OS) between Bcl-2+ and Bcl-2- de novo DLBCL (GCB or non-GCB subtypes). Using the RPCI tumor registry and pharmacy database, we identified 201 DLBCL patients treated with equivalent doses of rituximab and anthracycline-based therapy (i.e. R=CHOP or R+DA-EPOCH) at RPCI between 1997 and 2007. Demographic, clinical and pathological characteristics were obtained for each patient. Patients were classified into GCB or non-GCB DLBCL according to the Han's algorithm based on the expression of CD10, Bcl-6 and MUM-1. Bcl-2 was determined by IHC and was available for 101 patients. Using the Han's algorithm, fifty-three patients (26.4%) were classified as GCB, 54 patients (26.9%) non-GCB, and 94 patients (46.8%) could not be classified due to inadequate data/sample. Bcl-2 expression was detected in 67% and 73% of the GCB- and non-GCB DLBCL subtypes respectively. Demographics and clinical characteristics were equally distributed between GCB- and non-GCB DLBCL. Patients received either R+CHOP (90%) or R+DA-EPOCH (10%). The complete response (CR) rate of the entire cohort was 82.6% and no differences were observed between GCB- or non-GCB DLBCL (78.4% vs. 75.5%, P=0.73) or by Bcl-2 expression (+:72.5% vs. -:81.3%, P=0.35]. After a median follow up period of 74 months, significant differences were observed between Bcl-2 positive or negative and GCB or non-GCB DLBCL. GCB-DLBCL had a longer 5-yr PFS and 5-yr OS than non-GCB DLBCL (58.5% vs. 37%, P=0.026; 81.1% vs. 53.7% and, P=0.002; respectively). By itself, Bcl2 over-expression, had a negative impact in PFS (P=0.002) and OS (P=0.001) in R+CHOP/R+DA-EPOCH treated de novo DLBCL. The combined prognostic value of the Han's algorithm and Bcl-2 expression was also evaluated. Bcl-2 expression in the context of both GBC and non-GCB subtypes remains an unfavorable prognostic indicator for overall survival, with a more pronounced influence in the GCB-DLBCL phenotype (See table). Our data supports the predictive value of the Han's algorithm and Bcl-2 expression in DLBCL patients undergoing front-line chemo-immunotherapy. Bcl-2 expression is associated with a poor prognosis in GCB and non-GCB DLBCL. It is possible that intrinsic biological pathways involved in lymphomagenesis and/or “resistance” of these subtypes of DLBCL may play a role in their responsiveness to rituximab-based therapies and could be influenced by the net balance between pro- and anti-apoptotic proteins. Attempts to further delineate the biological heterogeneity of DLBCL may help identify subgroups of patients at high risk of resistance to chemo-immunotherapy and lead to the development of new therapeutic strategies. In conclusion, our data analysis confirms that the DLBCL immunophenotypes based on cell of origin and Bcl-2 status continues to have predictive significance on clinical outcomes in DLBCL in the rituximab era.