[1377] Racial Differences in Prognostic Biomarkers of Diffuse Large B-Cell Lymphoma

Elizabeth C Chastain, Kevin E Fisher, Kevin Bumpers, Vishnu Reddy, Kai Fu, Christopher R Flowers, David L Jaye. Emory University, Atlanta, GA; University of Alabama, Birmingham, AL; University of Nebraska, Omaha, NE

Background: Evidence suggests that black (B) Americans have a lower incidence of diffuse large B cell lymphoma (DLBCL) than white (W) Americans, but present at an earlier age, have worse clinical risk factors, and worse survival. Our data for a large patient cohort (Emory, UAB) receiving the same standard chemotherapy concur with these findings. Importantly, biological differences inherent to DLBCLs that may underlie the observed disparities have not been described. Thus, we asked whether recognized prognostic markers linked to known biological variants [germinal center B cell (GCB) and activated B cell-like (ABC)] are differentially expressed between these groups.
Design: From 204 DLBCL cases with complete treatment, survival, staging and demographic data, 89 cases had adequate blocks of formalin fixed material for tissue microarray (TMA) construction (3 cores/case, tonsil controls). H&E slides and immunostains, including CD3, CD5, CD10, CD20, BCL2, BCL6, FOXP1, GCET1, LMO2 and MUM1, were reviewed and scored. Published “cell of origin” algorithms were employed (Meyer et al, JCO 29:200-207). Cases were excluded from final analyses if lack of evaluable cores (no tissue or DLBCL) precluded algorithm classification.
Results: As shown in Table 1, data were adequate to study 75-78 cases/algorithm. The proportion of B and W cases analyzed were representative of the general population from which patients came. Interestingly, significantly lower rates of the more favorable GCB subtype were observed for B patients using the 5 tested algorithms. However, no differences were identified for CD5, BCL2 or CD20 immunoreactivity (p = 0.23, 0.77, 0.34 respectively). For each stain, > 90% of cases had 2-3 evaluable cores. Control staining was appropriate.

Table 1

Algorithm	W (n)	W, % GCB	B (n)	B, %GCB	p value
Hans	51	64.7	26	30.8	0.01
Natkunam	52	63.5	26	30.8	0.01
Choi	49	63.3	26	38.5	0.05
Choi (mod)	49	61.2	26	34.6	0.03
Tally	51	51.0	26	26.9	0.05

Conclusions: The rate of GCB subtype of DLBCL, which has a better prognosis, is significantly lower among B compared to W patients, regardless of analytical algorithm. By contrast, no differences were observed for CD5 or BCL2 expression, other described prognostic immunohistochemical markers. These novel findings point to potentially important biological differences intrinsic to DLBCL between B and W patients that may in part explain the comparatively adverse outcomes among B patients. Moreover, such findings may have therapeutic implications with the emergence of biologic subtype-specific therapies.
Category: Hematopathology

Monday, March 19, 2012 9:30 AM

Poster Session I Stowell-Orbison/Surgical Pathology/Autopsy Awards Poster Session # 222, Monday Morning