Post Allogeneic Stem Cell Transplant (SCT) Cyclophosphamide Improves Progression Free Survival (PFS) in Pts with AML/MDS Treated with CTLA-4 or PD-1 Blockade Prior to SCT

Background: Encouraging results have been reported in AML/MDS patients (pts) treated with checkpoint inhibitors (CPIs) in frontline and relapsed AML/MDS, and increasing number of pts exposed to CPIs are receiving SCT. However, a high rate of grade 3-4 acute graft versus host disease (aGvHD), 23%, was reported in lymphoma pts following CPI. We performed a retrospective study of pts with AML/MDS treated with an anti-PD1 and/or anti-CTLA-4 blockade at some point prior to undergoing SCT, with the goal of describing transplant outcomes.

Methods: In this analysis, we included 30 AML and 13 MDS pts treated with a PD-1 (nivolumab, n=32) or CTLA-4 (ipilimumab, n=9) based therapies on clinical trials, who subsequently had SCT at our institution. Two patients received both nivolumab and ipilumumab. The primary outcome was aGvHD by day 100 after SCT. Secondary outcomes were progression-free survival (PFS) and disease progression. Outcome analyses were stratified by GvHD prophylaxis.

Results: Median (med) age was 57 years (yrs) (range, 35-71) (Table 1). Of the 43 pts, 33 (77%) had achieved a best response of complete remission (CR, n=22) or CR without count recovery (CRi/CRp, n=11) to PD-1 or CTLA-4 based therapies, and 2 received PD-1 therapy as maintenance in CR. Med time to response from CPI based therapy was 49 days (range, 21-161). Twelve pts (28%) had no response or lost response to PD-1 and CTLA-4 based therapies prior to proceeding to SCT but achieved a CR/CRi with subsequent therapies (AraC-based therapies in 5, hypomethylator-based therapies in 5, molecular therapies in 2) allowing them to proceed to SCT.

Med time to SCT after CPI treatment was 63 days (range, 7-386) and med CPI treatment cycles was 3 (range, 1-10). Conditioning intensity was myeloablative in 22 (51%) pts. Donors were matched related in 7 (MRD, 16%), matched unrated in 25 (MUD, 58%), haploidentical (haplo) in 8 (19%), and cord blood (CB) in 3 (7%) pts. Hematopoietic stem cell source was peripheral blood stem cells (PBSCs) in 25 (53%) and bone marrow (BM) in 17 (40%) pts.

GvHD prophylaxis consisted of post-SCT cyclophosphamide (PTCy), 50 mg/kg on days +3 and +4, tacrolimus± mycophenolate mofetil (MMF) (PTCy)(n=22) or tacrolimus and mini-methotrexate/MMF ± antithymocyte globulin (non-PTCy)(n=21). As expected, all haplos had PTCy. None of the CB recipients had PTCy. For MRD and MUD, GvHD prophylaxis was determined by the treating physician; with 57% and 40% of pts receiving PTCy, respectively.

There were no significant differences between PTCy and non-PTCy in age, disease, disease status, MRD vs. MUD, donor age, cytogenetics, conditioning intensity, hematopoietic stem cell source, hematopoietic cell transplantation-specific comorbidity index, and number of CPI doses received, except a longer interval between CPI blockade therapy and SCT in PTCy (Table 2).

Med follow-up of was 6 months (range, 1.6-23). The 100-day cumulative incidences of grade 2-4 and grade 3-4, aGvHD were 52% and 13% (Table 2). Pts in PTCy group had a trend to lower 100-day grade 3-4 (5% vs 22%, p=0.2). The interval between CPI and SCT did not impact 100-day aGvHD incidence.However, a higher incidence of grade 3-4 aGVHD was observed in patients treated with >4 treatment of PD-1 or CTLA-4 blockade prior to SCT if not given PTCY as GvHD prophylaxis (43% vs. 12%, p=0.01)

Considering the association between donor type, stem cell source and aGvHD, subgroup analysis were performed in MUD pts. In 15 MUD patients with PBSC source, there were no grade 3-4 aGvHD in 6 PTCy patients while 4 of 9 non-PTCy patients had grade 3-4 a GvHD (HR=NE due to absence of events in one group, p=0.05). On the other hand, in 10 MUD patients with BM source, no grade 3-4 aGvHD was observed by day 100, in 4 PTCy and 6 non-PTCy pts.

At 6-month post SCT, estimated PFS for all pts was 66% (Table 2). PTCy pts had improved PFS at 6 months (p=0.03) and 1-year (p=0.02) post SCT. The med RFS was 4.6 mo in non-PTCy pts while it was not reached in PTCy pts.

Conclusion: SCT after PD-1 or CTLA4 based therapies appears feasible in pts with AML/MDS. Our results indicate 1) a high rate of grade 3-4 aGvHD in MUD recipients with PBSC who did not receive PTCy, and 2) improved PFS with the use of PTCy. The use of post-Cy as GvHD prophylaxis to improve the tolerability of SCT in pts treated with PD-1 or CTLA-4 based therapies pre-transplant deserves further investigation.

View largeDownload slide

View largeDownload slide

Close modal