AUTHORS
Daiana Gisela Rodriguez, MD, Omar Eduardo Fernandez Vargas, MD, MSc2, Imran Ahmad, MD, MSc3, Lea Bernard, MD4, Sandra Cohen, MD5, Jean-Sebastien Delisle, MD, PhD6, Thomas Kiss, MD7, Denis-Claude Roy, MD8, Jean Roy, MD9, Guy Sauvageau, MD,PhD10, Sylvie Lachance, MD, FRCPC, DRCPC11
AFFILIATIONS
1 Université de Montréal, Montreal, Quebec, Canada
2 Université de Montreal, Montreal, Quebec, Canada
3 Université de Montreal, Montreal, Quebec, Canada
4 Université de Montreal, Montreal, Quebec, Canada
5 Université de Montreal, Montreal, Quebec, Canada
6 Université de Montreal, Montreal, Quebec, Canada
7 Université de Montreal, Montreal, Quebec, Canada
8 Université de Montreal, Montreal, Quebec, Canada
9 Université de Montreal, Montreal, Quebec, Canada
10 Université de Montreal, Montreal, Quebec, Canada
11 Université de Montréal, Montreal, Canada
ABSTRACT
Background
In 10/10 HLA-matched allogeneic hematopoietic cell transplantation (HCT), non-permissive HLA-DPB1 mismatches (DPB1-mm) have been associated with an increased risk of graft-versus-host disease (GVHD), whereas permissive mismatches may improve survival potentially through enhanced graft-versus-leukemia (GVL) activity. Although anti-thymocyte globulin (ATG)-based GVHD prophylaxis (px), effectively attenuates HLA DPB1-driven alloreactivity via in vivo T-cell depletion, its impact on the GVL activity remains unknown
Aims
To evaluate the impact of HLA-DPB1 disparities on transplantation outcomes following 10/10 HLA-matched related donor (MRD) or matched unrelated donor (MUD) HCT using universal ATG-based GVHD prophylaxis.
Methods
This retrospective single-center study included 214 consecutive patients undergoing first 10/10 HLA-matched HCT between January 2020 and May 2025 at Maisonneuve-Rosemont Hospital, Montreal, Canada. Among them, 69 patients were HLA-DPB1 matched (DPB1-m), including 44 siblings and 145 were DPB1-mm. Of the latter, 29 (20%) were non-permissive mismatches (DPB1-np). All patients received rabbit ATG-based GVHD px administered at a median dose of 2.5 mg/kg on days -3 and -2, combined with a calcineurin inhibitor initiated on day -1 and either methotrexate or mycophenolate mofetil. Remission status at HCT was classified as in remission or not in remission. Overall survival (OS), disease-free survival (DFS) and GVHD-free, relapse-free survival (GRFS), were estimated using Kaplan-Meier method. Multivariable cox and Fine-Gray or regression analysis were performed to identify predictors of outcomes.
Results
Among 214 patients included: 69 (32%) were DPB1-m, 116 (54%) DPB1-mm permissive (DPB1-p), and 29 (14%) DPB1-np. Median recipient age was 53.2 years, with a median follow-up of 38.3 months. Baseline characteristics were generally balanced between groups. Acute leukemia (AML 40%, ALL 18%) was the leading indication for HCT in both groups. High DRI accounted for 14.5% and 11% of DPB1-m and DPB1-mm pts (p=0.47). Non-Caucasian ethnicity was more prevalent among DPB1-mm recipients (20% versus 6% (p=0.002). More DPB1-mm patients underwent transplantation with active disease (22% (n=32) vs 10% (n=10) for DPB1-m; p=0.04). Conditioning intensity and total body irradiation exposure were similar between groups. MUD transplants represented the majority of the DPB1-mm cohort and donors were significantly younger (25-y) than matched related donors (40-y) (p<0.001).
The estimated 36-month DFS was 71.5%. Relapse incidence was comparable between DPB1-m and DPB1-mm recipients at 2 years (24%). Despite a higher rate of active disease at transplantation, DPB1-mm recipients achieved DFS comparable to that of DPB1-m patients (71,3% vs. 72%:p=0.4). As expected, multivariable analysis showed a trend toward higher relapse risk for pts not in remission (aHR=1.9, p=0.06). and a non-significant lower relapse risk in DPB1-mm (aHR=0.7, p=0.19). Non-relapse mortality plateaued at 4.6% after month 12 and remained nearly identical across cohorts. The cumulative incidence of acute GVHD was 29.4% at day 100 and 33.2% at day 180 not significantly different between DPB1-m and DPB1-mm recipients (22% vs. 33% p=0.18, and 25% vs. 37%, p=0.14), as well as between permissive and non-permissive mismatch groups. Non-permissive DPB1-mm in the GVH vector showed a trend toward higher early acute GVHD (HR p=0.058). Grade III-IV acute GVHD occurred in 11% of patients and was not significantly influenced by DPB1 matching status (p=0.28), or mismatch permissiveness (p=0.6).
Chronic GVHD incidence remained similar between groups being 10.4% at 6 mo (11% DPB1-mm vs. 8.8% DPB1-m; p=0.6) and 23.4% at 12 mo (24.2% DPB1-mm vs. 21.8% in DPB1-m; p=0.54). Severe chronic GVHD occurred in only 6,9%. GVHD-free, relapse-free survival (GRFS) reached 50.8% at 24 months (46.3% for DPB1-mm vs 53% for DPB1-m; p=0.4). Active disease at transplantation was the only independent predictor of GRFS failure (p=0.006).
Conclusion
In 10/10 HLA-matched allogeneic HCT, universal ATG-based GVHD prophylaxis appears to minimize the adverse immunologic effects traditionally associated with HLA-DPB1 disparities on GVHD incidence and severity while preserving the graft-versus-leukemia effect. These findings translated into excellent 3-year DFS, low non-relapse mortality and favorable GVHD-free, relapse-free survival irrespective of HLA-DPB1 matching status suggesting
DISCLOSURES
Daiana Gisela Rodriguez, MD: No relevant conflicts of interest to declare.