Abstract
The Predicted Indirectly Recognizable HLA Epitopes (PIRCHE) score quantifies the number of PIRCHEs in patient-donor pairs and represents an in silico measure of indirect alloreactivity. This biologic process is defined as T cell recognition of epitopes derived from mismatched, allogeneic HLA peptides that are subsequently presented by shared HLA molecules. Its association with clinical outcome has not been examined in haploidentical hematopoietic cell transplantation (haplo-HCT) with post-transplantation cyclophosphamide (PTCy). We hypothesized that the PIRCHE score (PS) would correlate with indirect alloreactivity and predict graft-versus-host disease (GVHD) risk and the incidence of relapse after haplo-HCT with PTCy. We retrospectively analyzed 148 patients who underwent peripheral blood stem cell T cell-replete haplo-HCT with PTCy at a single center between 2009 and 2016. For each patient-donor pair, the PS was calculated using the PIRCHE online matching tool. PSs were categorized by class and vector. The median class I graft-versus-host (GVH) PS was 11 (range, 0 to 56), and the median class I host-versus-graft (HVG) PS was 10 (range, 0 to 51). Class I GVH PS was associated with increased risk of grade II-IV acute GVHD (adjusted hazard ratio, 1.03 per PS unit increase; 95% confidence interval, 1.01 to 1.05; P= .008) but not of chronic GVHD or relapse. Our data show that use of the PS is a novel strategy for predicting clinical outcome in haplo-HCT; further studies using registry data and prospective cohorts are warranted to validate these findings.