Long-Term Survival in Patients with AML or MDS Relapsed after Allogeneic Hematopoietic Cell Transplantation: Importance of Second Cell Therapy

Abstract

<h3>Introduction</h3> Patients (pts) with acute myeloid leukemia (AML) or myelodysplastic syndrome (MDS) who relapse after allogeneic hematopoietic cell transplantation (allo-HCT) have historically shown very poor post-relapse overall survival (prOS). <h3>Objective</h3> To study the factors and treatment strategies associated with the best prOS. <h3>Methods</h3> We conducted a retrospective analysis of all pts at our institution who received a first allo-HCT between 2010 and 2017 for AML or MDS but relapsed post-transplant. Bone marrow (BM) and peripheral blood stem cells (PBSC) grafts were included, either T-cell depleted (TCD) by ex-vivo CD34+ selection (Miltenyi CliniMACS) or unmodified. Univariate (UVA) Cox proportional hazards regression was used to examine the association between prOS and patient-, disease-, and transplant-related variables. Receipt of second cell therapy (sCT) after relapse was a time-dependent covariate. <h3>Results</h3> Key characteristics of the study population are listed in Table 1. Median time to relapse (TTR) was 6 months (range: 1-61) and median prOS was also 6 months (95%CI: 4.8-8.8) in the whole population. AML pts showed shorter median prOS compared with MDS pts (5.3 [95%CI: 3.9-8] vs. 9.4 [95%CI: 5.7-24] months, p=0.026). No statistically significant associations were observed between prOS and conditioning intensity (reduced vs. ablative) or graft manipulation (TCD vs. unmodified). There was a significant association between TTR and prOS (Fig. 1). Forty-five pts received sCT after relapse, either a second allo-HCT (28) or donor lymphocyte infusions (17). The hazard of death after relapse for these patients was lower (HR 0.53, 95%CI: 0.32-0.87, p=0.01) and 2-year survival after sCT was 44.9% (95%CI: 31.1-64.8, Fig. 2). We used a 4-level categorization of TTR (</≥ 12 months) and receipt of sCT. Compared to the reference (TTR < 12 months and no sCT), the most favorable combination for prOS resulted TTR ≥ 12 months and receiving sCT (HR 0.28, 95%CI: 0.13-0.6, p=0.001, Fig. 3). <h3>Conclusion</h3> Although limited by the retrospective nature and potential selection bias, we show in a large, contemporary cohort that pts who relapse after ≥ 12 months from a first allo-HCT and receive sCT have the best prOS.