Graft-Versus-Leukemia Effect after Haplo-Identical Stem Cell Transplantation with Post-Transplant Cyclophosphamide in Patients with AML- No Association with Graft-Versus-Host Disease: A Study on Behalf of the Acute Leukemia Working Party of EBMT

Introduction. Allogeneic stem-cell transplantation (SCT) is a curative approach in acute myeloid leukemia (AML) by providing dose-intensive chemo-radiotherapy and enhancing a graft-versus-leukemia (GVL) effect. The GVL effect is provided by alloimmune T- cells and is usually closely associated with graft-versus-host disease (GVHD). Patients with GVHD have a lower incidence of relapse, but at high grades a higher incidence of non-relapse mortality (NRM). However, GVL can also occur independently of GVHD. The use of haplo-identical SCT is rapidly increasing over the last decade due to the introduction of non-T depleted methods, in particular with post-transplant cyclophosphamide (PTCy), with expected outcomes that are similar to other donor sources. The GVL effect after T- depleted haplo-identical SCT is mostly related to natural-killer cell alloreactivity and is not necessarily associated with GVHD. However, there is no definite data whether GVL after SCT in the non-T depleted setting is similarly associated with GVHD as in the matched donor setting.

Methods. We assessed the impact of acute and chronic GVHD on SCT outcomes in patients with AML following non-T depleted haplo-identical SCT with PTCy, by using a series of landmark analyses.

Results. The study included 605 patients with AML in CR1 (73%) or CR2 (27%) after non-T depleted haplo-identical SCT with PTCy, given during the years 2009-2016. The median age was 53 years (range, 18-76). 71% had myeloablative conditioning and 29% had reduced-intensity conditioning . The overall rate of acute GVHD grade II-IV and III-IV was 28.4% and 8.0%, respectively. The rates of chronic GVHD all grades and extensive were 32.7% and 12.3%, respectively. The rate of relapse and NRM were 21.8% and 18.2% and the rates of leukemia-free survival (LFS) and overall survival, 2 years after SCT were 59.9% and 64.1%, respectively. 509 patients were alive and leukemia-free 100 days after transplant. 366 had no prior acute GVHD at this landmark, 107 had acute GVHD grade II and 36 had grade III-IV. The overall rate of subsequent relapse was 20.3%, 18.3% and 11.9%, respectively (P=0.60). The overall rates of subsequent NRM were 10.3%, 19.0% and 35.7%, respectively (P<0.001). The rates of LFS were 69.4%, 62.6% and 52.4%, respectively (P=0.01). Multivariate analysis showed that acute GVHD grade II before day 100 was not associated with subsequent relapse (Hazard ratio (HR) 1.02, P=0.93), had borderline association with NRM (HR 1.79, P=0.09) and no association with LFS (HR 1.28, P=0.27). Acute GVHD grade III-IV before day 100 was not associated with subsequent relapse (HR 0.92, P=0.87), while it was associated with higher NRM (HR 5.23, P<0.001) and inferior LFS (HR 2.35, P=0.003). Both acute GVHD grade II and grade III-IV were associated with subsequent extensive chronic GVHD, HR 2.02 (P=0.04) and 6.93 (P<0.001), respectively. 393 patients were alive and leukemia-free 6 months after transplant. Out of them 316 had no prior chronic GVHD at this landmark, 55 had limited and 22 extensive chronic GVHD. The overall rate of subsequent relapse was 14.3%, 9.2% and 23.9%, respectively (P=0.60). The overall rates of subsequent NRM were 7.3%, 10.4% and 31.7%, respectively (P=0.003). The rates of LFS were 78.4%, 80.4% and 44.4%, respectively (P=0.01). Multivariate analysis showed that limited grade chronic GVHD occurring before 6 months was not associated with subsequent relapse (HR 0.69, P=0.44), NRM (HR 1.43, P=0.55) or LFS (HR 0.88, P=0.74). Extensive chronic GVHD was not associated with subsequent relapse (HR 1.44, P=0.56) but was associated with higher NRM (HR 5.77, P=0.004) and inferior LFS (HR 2.75, P=0.01). Similar results were seen for prior chronic GVHD at the 1 year landmark.

Conclusions. By a series of landmark analyses at 100 days, 6 months and 1 year after SCT, we did not find any association of acute GVHD grade II or III-IV or chronic GVHD (limited or extensive) with subsequent relapse. Acute GVHD grade III-IV was associated with a higher NRM and lower LFS rates after day100. All grades of acute GVHD were associated with a higher incidence of chronic GVHD. Previous extensive chronic GVHD was also associated with a higher NRM and a lower LFS. GVL is thus not associated with GVHD after non T-deleted haplo-identical SCT with PTCy. Future novel strategies for prevention of significant GVHD are warranted.