From Ex-Vivo T-Cell Depletion to Post-Transplant Cyclophosphamide: Improved GvHD-Free & Relapse-Free Survival but Comparable Chronic GvHD Incidence in Haploidentical Transplantation. A 15 Years EBMT Registry Analysis on Behalf of the TCWP-EBMT

Abstract

Background: Haploidentical stem cell transplantation (Haplo HSCT) has emerged in the past three decades as an alternative curative option when an HLA match donor is not available. Over time, the use of Haplo has increased dramatically, reaching superimposable results when compared to unrelated and related HSCT strategies, confirming its validity. The widespread use of Haplo mainly relies upon technical advances, control of alloreactivity through graft-versus-host disease (GvHD) prophylaxis combined with a rapid and almost universal probability to find an Haplo donor for any candidate patient. The aim of our study was to provide a picture of acute (aGvHD) and chronic GvHD (cGvHD) incidence in Haplo HSCT across different platform in the past 15 years, where Haplo moved from ex-vivo T-cell depleted (TCD) platform to in-vivo TCD platform to the post-transplantation cyclophosphamide (PTCy). Methods: We compared the outcomes of adult patients receiving a 1st Haplo HSCT for any hematological malignancy according to GvHD prophylaxis - ex-vivo + in-vivo TCD (n=160), in-vivo only TCD (n=507) or PTCy (n=2593) - and reported to the EBMT registry in 2004-2016. Patients with missing data on disease status at last follow-up and GvHD information were excluded. Primary endpoint was GvHD-free & Relapse-free survival (GRFS) with events defined by death or relapse or grade ≥3 aGvHD or extensive (ext) cGvHD. Secondary endpoints were progression-free survival (PFS), overall survival (OS), aGvHD and cGvHD, incidence of relapse (IR) and non-relapse-mortality (NRM). Due to sample size in the first cohort of ex-vivo TCD, multivariate analysis compared only in-vivo TCD vs PTCy cohorts. Table 1 illustrates patients' characteristics. Results: Univariate analysis for 3-year outcomes are reported on table 2. PTCy provides better GRFS, OS, PFS, NRM versus ex-vivo or in-vivo. IR was not significantly different. Likewise, the 3-year CI of cGvHD and ext cGvHD were similar between PTCy, in vivo TCD and ex-vivo TCD (cGvHD 27% [25-29%], 25% [21-29%], 18% [12-25%], p 0.03; ext cGvHD 11% [10-12%], 10% [8-13%], 8% [4-13%], p 0.45). On the contrary the 100-day CI of grade ≥2 aGvHD were lower in the ex-vivo TCD vs PTCy and in-vivo TCD (19% [14-26%], 28% [26-30%], 32% [28-36%], p 0.002) while grade ≥3 aGvHD were lower in the PTCy group vs ex-vivo and in-vivo TCD (9% [8-10%], 11% [7-17%], 14% [11-18%], p <0.001). After adjustment for diagnosis, patient age, disease status, Karnofsky PS, donor/patient gender and CMV, cell source, conditioning intensity, previous auto and year of transplant, the multivariable model comparing in-vivo TCD and PTCy showed better outcome for PTCy. Compared to in-vivo TCD, the hazards for GRFS was 0.76 for PTCy (p 0.004), the HR for PFS was 0.71 (p 0.001) and the HR for OS was 0.7 (p 0.0008), the HR for NRM was 0.63 (p 0.001). Moreover, compared to in-vivo TCD, PTCy yielded similar hazards for grade≥2 aGvHD (HR: 1.02, p 0.89), grade≥3 aGvHD (HR 0.79, p 0.27), cGvHD (HR 1.17, p 0.37), ext cGvHD (HR 1.18, p 0.52) and relapse (HR 0.8, p 0.1). Variables associated with GRFS were active disease, Karnofsky PS ≥90%, diagnosis, donor/patient gender and CMV. An ancillary analysis evaluating the stem cell source effect in the PTCy cohort only, demonstrates comparable outcome endpoints (OS, PFS, NRM, IR) at 2-year between bone marrow (BM) and peripheral blood (PB) PTCy. In univariate analysis GRFS and the 2-year CI of cGvHD were not different between BM and PB (GRFS 47% [45-50%], 46% [44-49%], p 0.085; 2-year CI of cGvHD 25% [23-28%], 27% [25-30%], p 0.2) while ext cGvHD, 100-day CI of grade ≥2 aGvHD and grade ≥3 aGvHD were lower in BM PTCy vs PB PTCy (ext cGvHD 8% [7-10%], 12% [10-14%], p <0.001; grade ≥2 aGvHD 20% [18-22%], 36% [33-38%], p <0.001; grade ≥3 aGvHD 6% [5-7%], 12% [10-14%], p <0.001). Compared to BM PTCy, the HR for cGvHD was 1.55 for PB PTCy (p 0.001), the HR for ext cGvHD was 2.04 (p 0.0003), the HR for grade ≥2 aGvHD was 1,94 (p <0.0001), the HR for grade ≥3 aGvHD was 2.01 (p 0.0001). Conclusions: In the present EBMT registry study on more than 3000 patients transplanted from an Haplo donor, we report improved outcome (better GRFS - in spite of comparable chronic GvHD - OS and PFS, lower NRM) and widespread use in different diagnosis setting other than acute leukemia in PTCy platform. PTCy strategy provides a concrete progress into the field: even if cGvHD still represent a major issue, exploitation of BM PTCy seems to protect against most severe GvHD manifestation. Disclosures Mohty: Jazz Pharmaceuticals: Honoraria, Research Funding. Kröger:Sanofi-Aventis: Honoraria; Riemser: Research Funding; Novartis: Honoraria, Research Funding; Neovii: Honoraria, Research Funding; Medac: Honoraria; JAZZ: Honoraria; DKMS: Research Funding; Celgene: Honoraria, Research Funding. Basak:Celgene: Honoraria; Teva: Honoraria.