Post-Transplant Sorafenib Improves Overall Survival in FLT3 Mutated AML: A Report from the EBMT Acute Leukemia Working Party

Rationale:FLT3 in mutated in 30% of AML and is associated with poor prognosis. Allogeneic hematopoietic cell transplantation (allo-HCT) in first complete remission (CR) is recommended in FLT-3 ITD AML. However frequent and early post-transplant relapse leads to poor outcome. Several small studies suggested the efficacy of sorafenib as prophylactic or preemptive therapy or as treatment for relapse post allo-HCT. The purpose of this study was to assess the impact of sorafenib on outcomes of FLT3 mutated AML post allo-HCT.

Patients and Methods: We identified 462 adult patients (51% males) with FLT3 mutated AML (FLT3 ITD-95%) allografted at EBMT participating centers between 2010 and 2015. Median age was 50 years (range 18-75). Patients mostly belonged to the intermediate cytogenetic risk group (82%) and 55% of those with available data had NPM1 mutation. Disease status at allo-HCT was CR1 (72%), CR2 (10%) or active disease (18%). Donors were matched related (MSD, 40%), matched unrelated (49%) or haploidentical (11%), respectively. Preparative regimens were myeloablative (MAC) in 53% and reduced intensity conditioning (RIC) in 47% of allo-HCT. Sixty two patients received post-transplant sorafenib either as prophylactic (n=19) or preemptive therapy (n=9) or as treatment for relapse (n=34). Median time from allo-HCT to initiation of maintenance sorafenib was 55 days (1-173). Median follow-up for alive patients was 39 months (range 1-87).

Results: The 2-year leukemia free survival (LFS), overall survival (OS) and GVHD relapse free survival (GRFS) was 51%, 59% and 38%, respectively. In multivariate Cox analysis, sorafenib maintenance (either prophylactic or preemptive) significantly improved OS (HR=0.36; p=0.03). OS was also positively affected by NPM1 mutation, allo-HCT in CR1, and in vivo T cell depletion, but negatively affected by the need for >1 induction. Similarly, sorafenib maintenance significantly improved GRFS (HR=0.44; p=0.02). GRFS was also positively affected by NPM1 mutation (HR=0.66; p=0.002), haploidentical donors compared to MSD (HR=0.61; p=0.04), and in vivo T cell depletion (HR=0.55; p=0.00001), but negatively affected by the need for >1 induction (HR=1.5; p=0.005) or active disease at transplant (HR=2.5; p<10^-5). We then performed pair matched analysis on 26 patients in the sorafenib group and 26 controls who engrafted and with survival without relapse and without acute GVHD grade II-IV at least ≥ to time to sorafenib initiation for each case. Other matching factors were age at transplant, NPM1 mutation, disease status at transplant (CR1 Vs CR2 Vs active disease), and conditioning (RIC Vs MAC). Two year LFS and OS were 79% and 83% in the sorafenib group versus 54% and 62% for controls (p=0.002 and 0.007 respectively). On the other hand, for 244 patients who relapsed after allo-HCT, multivariate Cox analysis with sorafenib salvage as time dependent variable, showed that treatment of relapse with sorafenib significantly improved OS (HR=0.48; p=0.006). Similarly, age, active disease at allo-HCT and RIC conditioning negatively affected OS. Finally, we performed another pair matched analysis on 25 relapsed patients treated with sorafenib and 25 controls with survival after relapse ≥ to the time from relapse to initiation of sorafenib for each case. Other matching factors were age (< Vs > than 50 years), disease status at allo-HCT (CR1 Vs CR2 Vs active disease), conditioning (RIC Vs MAC), and time from allo-HCT to relapse (+/- 1 month). One year and 2 year OS were 41% and 30% for patients treated with sorafenib versus 14% and 14% for controls (P=0.0015).

Conclusion: For AML patients with FLT3 mutation, post-transplant sorafenib whether given as prophylactic or preemptive therapy or as treatment for relapse, significantly improves OS and may be considered as standard of care in that setting.