Abstract
Background: Various subsets of Diffuse Large B-cell lymphoma (DLBCL) are distinguished based on molecular and Immunohistochemical features. CD5 expressing DLBCL (CD5+ DLBCL) is increasingly recognized as a subtype of DLBCL with an aggressive disease course in small elderly Asian population cohorts comprising 5-10% of de-novo DLBCL. We attempt to describe its prevalence and prognostic implications in concert with IHC expression profiles of biomarkers
Methods: A retrospective review of elderly patients (65yr-92yr) with de-novo Diffuse Large B-Cell lymphoma (DLBCL) uniformly treated with RCHOP was conducted between Jan 98 and Dec 2016. Patient demographics, clinical parameters and IHC markers were collected. Analysis using chi square, or Fisher exact where appropriate, was performed. Kaplan-Meier analysis and univariate and multivariate Cox regression analysis identified associations with overall survival. The protocol, and subsequent modifications, were reviewed and approved by the IRB at the Texas Tech University Health Sciences Center
Results: One hundred fourteen de-novo DLBCL with known CD5 IHC status, Ki67, c-MYC, BCL2, BCL6, MUM1, CD10 marker status and Han's classifiers were evaluated. Median age of the CD5 negative cohort was 75yrs ±12.8 vs 80yrs±9.3 in the CD5+ subgroup (p=0.51). Median OS 6.7 years (SD 62.3mo; Range 1-256). CD5+ expression on IHC was seen in 23/110 (21%) patients. IHC marker status showed c-myc (12%), BCL2(42%) and BCL6(39%) respectively. Compared with matched group of patients with CD5 negative DLBCL, CD5+ DLBCL were more likely to be non-GCB type (14/23 vs 9/87; p<0.001), IPI>2 (14/23 vs 23/81; p=0.009) and Stage >III (p=0.004). Three patients with CNS involvement all had CD5 expression. Kaplan-Meier analysis showed that Han's non-GCB classification (p < 0.001), and CD10 positivity (p = 0.003) had worse survival. In a multivariate Cox regression model Ki67 >70% (HR=0.33, P=0.017), BCL6 (HR=0.47, P=0.019) and CD10 (HR=0.2, P=0.003) were associated with worse outcomes. CD5+ expression was not significantly associated with OS (HR=0.76; -1.4 to 0.8; p=0.64) in multivariate model, however was significant in a univariate analysis (P<0.001).
Conclusion: CD5+ DLBCL subset is more common in this elderly western cohort than historically seen in the Asian population with a signal toward Hans non-GCB, higher R-IPI, higher Stage and perhaps higher risk of CNS involvement. More data is needed to evaluate if CD5+ de-novo DLBCL has any correlation with other antiapoptotic or proliferative markers like BCL2,6 and c-MYC and larger patient subset and prospective studies are needed to corroborate these findings
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.