Synergistic effect of HLA evolutionary divergence and post-transplant biomarkers on non-relapse mortality prediction in adult haploidentical hematopoietic stem cell transplantation

Abstract Background: In hematopoietic stem cell transplantation, post-transplant biomarkers are recognized as reliable prognostic indicators for acute graft-versus-host disease (aGVHD), infections, and organ dysfunction to enable risk stratification for non-relapse mortality (NRM). HLA Evolutionary Divergence (HED), quantifying evolutionary distance between donor and recipient HLA alleles, recently has emerged as a new parameter to predict immune-related post-transplant complications and prognosis. However, the synergistic value between HED and existing biomarkers in predicting post-transplant risk such as NRM among leukemia patients undergoing haploidentical transplantation remains unknown. Methods: A multicenter study was conducted, enrolling 535 adult patients (aged ≥ 16 years) who underwent haploidentical HSCT following the GIAC/ATG-based protocol between April 2019 and May 2023. 188 patients were stratified into low and moderate-high NRM risk groups using a validated model based on serum sST2 and TNFr1 levels (Shao et al. ASH 2024). HED for HLA class I (A, B, C) and class II (DRB1, DQB1) alleles within donors, recipients, and donor-recipient pairs was calculated using Grantham distances (PMID: 39229273). Stepwise regression modeling was used to identify independent risk factors. Cumulative incidence of NRM was calculated using cumulative incidence functions. Gray's test was used to assess differences between two groups, and a p-value < 0.05 was considered statistically significant. Furthermore, an independent multicenter validation cohort (n = 347) was utilized to validate the findings. Results: Based on post-transplant biomarker concentrations, 77 (41%) patients were classified as moderate-high risk and 111 (59%) as low risk corresponding to 1-year NRM cumulative incidences of 21.79% and 6.33% respectively (p<0.001). Higher HED of the patient's Class I HLA was associated with an increased NRM rate. The 1-year cumulative incidences of NRM were 18.15% vs. 7.06% (p=0.029) in the high HED group (≥72.8327) vs. low HED group (<72.8327). This finding was reproducible in an independent multicenter validation cohort (14.08% vs 7.91%, p=0.009). When combining HED risk with post-transplant biomarker risk, in the moderate-high biomarker risk subgroup, the 1-year NRM cumulative incidences were 34.29% vs. 12.11% (p=0.046) in the high and low HED groups, whereas no significant difference was observed in the low biomarker risk subgroup. Multivariate analysis identified HLA-I HED of patients as a new independent risk factor for NRM. Conclusion: Our finding indicates that in a severe inflammatory milieu, the risks of aGVHD and NRM elevate when patients exhibit greater HLA-I diversity (higher HLA-I HED). This can be explained by the increased diversity of antigen presentation by the patient's HLA-I proteins, which can enhance the probability of recognition of host antigens by donor T cells and, augment the risks of aGVHD and NRM. The finding of this study may facilitate more precise patient risk stratification at early stage and support preventive treatment with selected patient group with high NRM risk.