Post-Transplant Cyclophosphamide Improves Progression Free Survival after Allogeneic Hematopoietic Stem Cell Transplantation in AML/MDS Patients with Prior CTLA-4 or PD-1 Blockade

Checkpoint inhibitors, anti-PD-1 and anti-CTLA-4 monoclonal antibodies are being increasingly used with encouraging results in patients (pts) with AML and MDS. The recipients of these agents may become candidates for allogeneic stem cell transplantation (SCT). However, a relatively high rate of grade III-IV acute graft versus host disease (GvHD) has been reported in lymphoma pts with prior PD-1 and CTLA-4 blockade using standard GvHD prophylaxis. We retrospectively analyzed 18 AML and 6 MDS pts treated with an anti-PD-1 (nivolumab) or anti-CTLA-4 antibody (ipilimumab) at some point prior to SCT at our institution over the last 3 years, and examined the impact of GvHD prophylaxis regimen on the outcome.

Median age was 55 years; 22 (85%) were treated with PD-1 blockade prior to SCT (10 in combination with azacitidine and 12 with idarubicin + cytarabine), and 4 with CTLA-4 blockade (3 in combination with azacitidine and 1 alone). The number of treatment cycles were <3 in 11, 3-4 in 11 and 5 or higher in 4 pts. Of 26, 20 pts (77%) proceeded to ASCT directly after the use of anti-PD-1 or anti-CTLA-4 monoclonal antibodies, while 6 pts received additional leukemia therapy before SCT. The median interval from anti-PD-1 or anti-CTLA-4 therapy to SCT was 55 days (range, 19-198). Disease status at SCT was complete remission in 10 pts (38%) and 15 pts (58%) received myeloablative conditioning regimens.

GvHD prophylaxis was either post-transplant cyclophosphamide in addition to tacrolimus and +/- mycophenolate mofetil (post-Cy group) or tacrolimus and methotrexate/mycophenolate mofetil (no post-Cy group). The post-Cy group included 3 matched unrelated (MUD), 3 matched related (MRD) and 4 haploidentical donors while the no post-Cy group had 12 MUDs, 2 MRD and 2 cord blood recipients. Disease and transplant characteristic were comparable between the two groups with the exception that more pts with post-Cy had ≥4 cycles of PD-1 or CTLA-4 blockade prior to SCT (60% vs. 25%, p=0.09), and a longer than 60 day interval between checkpoint inhibitor therapy and SCT (70% vs. 25%, p=0.03).

Median follow-up of survivors was 4.5 months (range, 0.5-18). There were 2 primary graft failure with an incidence of 8%. The incidences of grades II-IV and III-IV acute GvHD (aGvHD) at day 100 were 48% and 20% respectively in the whole cohort. The number of treatment cycles with PD-1 and CTLA-4 antibody prior to SCT and the time interval from last checkpoint antibody therapy to SCT did not have an impact on the rate of aGvHD. The use of post-Cy as GvHD prophylaxis (14% vs. 60%, p=0.1) decreased while the use of MUD increased (69% vs 0%, p=0.01) the incidence of grade II-IV aGvHD by day 100. However, there was a significant negative association between the use MUD and post-Cy as GvHD prophylaxis as few pts with MUD received post-Cy compared with the rest of the study cohort (20% vs. 64%, p=0.02). Despite this, there was still a trend of lower incidence of grade II-IV aGvHD by day 100 among MUD recipients with post-Cy compared with no post-Cy (33% vs. 75%, HR=0.4, p=0.3).

Other transplant outcomes of interest were also improved in post-Cy group compared with no post-Cy group (Table 1). Progression free survival was 100% in pts with post-Cy compared with 63% in no post-Cy (p=0.05) (Figure 1). No deaths were observed in post-Cy group and causes of death were disease progression in 3, aGvHD in one, infection in two, graft rejection in one patient who had no post-Cy as GVHD prophylaxis. Despite the increased rate of aGvHD; MUD recipients had comparable outcomes of PFS (64% vs. 67%, p=0.7), transplant-related mortality (15% vs. 21%, p=0.5) and overall survival (70% vs. 67%, p=0.4) compared with the rest of the cohort.

In conclusion, SCT after PD-1 or CTLA-4 blockade appears feasible. Our results indicate 1) a high rate of grade III-IV aGvHD in MUD recipients who did not receive post-Cy and 2) improved PFS with the use of post-Cy independent of the donor type after the use of PD-1 or CTLA-4 blockade. The use of post-Cy as the GvHD prophylaxis deserves further investigation to improve the tolerability of SCT after the use PD-1 or CTLA-4 blockade.

View largeDownload slide

View largeDownload slide

Close modal