Background: Germinal center B cell-like (GCB) DLBCL, as determined by gene expression profiling or immunohistochemistry, is more likely to be cured by initial conventional chemoimmunotherapy than non-germinal center B cell-like (non-GCB) DLBCL. For patients with relapsed or refractory chemosensitive DLBCL, high-dose chemotherapy and autologous hematopoietic stem cell (HSC) transplant is considered standard-of-care treatment, but it is unknown whether the outcome of these patients is similarly influenced by the subtype of DLBCL. We therefore explored the differences between patients with GCB and non-GCB DLBCL as regards their clinical features at relapse after first line therapy and their outcome following salvage autologous HSC transplant.

Methods: Patients undergoing BEAM conditioning and autologous HSC transplantation for relapsed or refractory chemosensitive DLBCL at Mayo Clinic, Rochester, MN between 2001 and 2006 were included. Immunohistochemical analysis was performed for CD10, BCL-6 and MUM1 allowing classification in GCB and non-GCB-like DLBCL, as well as for BCL-2. GCB and non-GCB groups were compared in terms of known prognostic factors at time of progression and outcome after HSC transplant

Results: Fifty-nine patients were included and had retrievable tumor samples to allow immunohistochemical analysis. Median follow-up of survivors was 25 months; median age at the time of transplant was 60 years (range 17–77); All patients had failed at least one previous anthracycline-based regimen (15 had refractory disease). Overall, 25/59 cases (42%) were positive for CD10, 32/58 (55%) for BCL-6, and 19/58 (32%) for MUM1. Thirty-two patients (54%) were classified as having GCB and 27 (46%) non-GCB-like DLBCL. Patients in the GCB and non-GCB group had similar time to progression (TTP) (median 12.5 months vs. 11 months, Wilcoxon P=0.81) after first line therapy and similar IPI-R scores (Chi-square, P=0.38). In univariate analysis, GCB and non-GCB did not differ in time to relapse after HSC transplant (log rank test P=0.77) or survival (log rank test P=0.48; figure). The lack of demonstrable difference in survival persisted even after correction for IPI-R and TTP, factors know to affect transplant outcome (Cox regression, RR=0.80 for GCB; P=0.28). BCL-2 was highly expressed in both GCB (81%) and non-GCB (96%) and did not correlate with outcome in the entire population nor in any of the two groups.

Conclusion: Patients with chemosensitive relapsed or refractory GCB and non-GCB-like DLBCL derive similar benefit from autologous HSC transplant.

Disclosure: No relevant conflicts of interest to declare.