Clinical Outcomes of Hematopoietic Cell Transplantation in Patients with Diffuse Large B Cell Lymphoma Transformed From Follicular Lymphoma

Background: Transformation of follicular lymphoma (FL) to diffuse large B cell lymphoma (DLBCL) occurs in 30% of cases at 10 years. Historically, survival with chemotherapy is poor.  Limited data are available on the role of autologous or allogeneic hematopoietic cell transplantation (autoHCT or alloHCT) for DLBCL transformed from FL.

 Methods: Descriptive analysis of 155 patients (age>18 years) from 89 centers with transformation of FL to DLBCL reported to the CIBMTR who had autoHCT or alloHCT from 1990-2009. Pathology reports were reviewed in all cases to confirm transformation of FL to DLBCL. This is the largest epidemiologic series analyzing outcomes of HCT in transformed DLBCL.

 Results: Patients were classified into 3 groups consisting of autoHCT (N=108), alloHCT (N=33), alloHCT (N=14) after prior autoHCT (performed for FL) with median ages of 56, 49, 51 years (yr) who received median of 3, 4, 5 lines of chemotherapy prior to HCT, respectively.  AutoHCT, alloHCT, and alloHCT with prior autoHCT groups had 5 yr relapse rates 54% (95% confidence intervals [CI] 45-64%), 33% (95% CI 17-50%), 38% (95% CI 15-65%); 1 yr treatment related mortality  8% (95% CI 3-14%), 41% (95% CI 25-58%), 62% (95% CI 35-85%); 5 yr progression free survival (PFS) 36% (95% CI 27-46%), 18% (95% CI 6-35%), 0%;  5 yr overall survival (OS)  50% (95% CI 40-60%), 22% (95% CI 8-41%), 7% (95% CI 0-26%) respectively.  The size of the cohort did not allow multivariate analysis.  Univariate survival analysis of associations between risk factors and post autoHCT and alloHCT outcomes showed age, Karnofsky performance status, rituxan use, extranodal disease and disease status at HCT (complete remission versus primary induction failure) were not significant factors.  Use of HLA identical sibling vs unrelated donors, ATG/campath vs none, did not impact PFS/ OS after alloHCT.  Chemotherapy vs total body irradiation-based conditioning for autoHCT had improved 3 yr PFS (46% [95% CI 35-57%] vs 24% [95% CI 9-43%]; p=.04); but 3 yr OS was not significantly different.  Transformation >1 yr vs <1yr from diagnosis (dx) yielded improvement in 1 yr PFS for the autoHCT (61% [95% CI 51-71%] vs 29% [95% CI 11-53%]; p=.01) and alloHCT groups (44% [95% CI 24-64%] vs 11% [95% CI 0-38%]; p=.03). Transformation > 1 yr vs <1 yr from dx also had better 2 yr OS in the alloHCT group (43% [95% CI 24-64%] vs 11% [95% CI 0-38%]; p=.03). Time from dx to transformation was not associated with OS after autoHCT.  The Kaplan Meier graph for PFS/ OS showed a plateau after 40 months only for autoHCT group.

 Conclusions: Both autoHCT and alloHCT are feasible and provide durable PFS and OS in this high risk cohort. Transformation to DLBCL > 1 year after dx of FL is associated with improved 1 yr PFS in both autoHCT and alloHCT groups, and superior 2 yr OS in the alloHCT group compared to those cases with transformation< 1year from dx of FL.   Patients with primary induction failure also benefit from consolidation with HCT.