BACKGROUND:We aimed to develop multiple single-nucleotide polymorphism (SNP)-based risk models associated with the risk of transplant outcomes including graft-versus-host disease (GVHD).
METHODS:The study evaluated 259 SNPs in 53 genes in 394 pairs of donors and recipients. In a discovery set (n=307 receiving related donor transplantation), overall survival, relapse-free survival (RFS), nonrelapse mortality, and acute or chronic GVHD were evaluated.
RESULTS:Eight recipients' SNPs of IL2, IL6R, FAS, EDN1, TGFB1, and NFKBIA genes and 12 donors' SNPs of NOS1, IL1B, TGFB2, NOD2/CARD15, TNFRII, IL1R1, and FCGR2A genes were identified in univariate analyses. Risk models were generated using significant clinical variables and genetic SNP markers after filtering out through multivariate analyses. Then, we divided patients into four quartiles (25%, Q) according to their risks. The final models stratified patients into low-risk (Q1), moderate-risk (Q2, Q3), and high-risk (Q4) groups in terms of overall survival (P<0.0001), RFS (P<0.0001), nonrelapse mortality (P=0.0043), and acute GVHD (P<0.0001), but not for chronic GVHD (P=0.763). External validation was performed in 87 transplant pairs that received matched unrelated donor transplantation, especially for RFS (P=0.016) and acute GVHD (P=0.027).
CONCLUSION:Risk models can improve prognostic stratification of patients according to their risk for transplant outcome.