Predictive value of molecular matching tools for the development of donor specific HLA‐antibodies in patients undergoing lung transplantation

Abstract

Molecular matching is a new approach for virtual histocompatibility testing in organ transplantation. The aim of our study was to analyze whether the risk for de novo donor‐specific HLA antibodies (dnDSA) after lung transplantation (LTX) can be predicted by molecular matching algorithms (MMA) and their combination. In this retrospective study we included 183 patients undergoing LTX at our center from 2012–2020. We monitored dnDSA development for 1 year. Eplet mismatches (epMM) using HLAMatchmaker were calculated and highly immunogenic eplets based on their ElliPro scores were identified. PIRCHE‐II scores were calculated using PIRCHE‐II algorithm (5‐ and 11‐loci). We compared epMM and PIRCHE‐II scores between patients with and without dnDSA using t‐test and used ROC‐curves to determine optimal cut‐off values to categorize patients into four groups. We used logistic regression with AIC to compare the predictive value of PIRCHE‐II, epMM, and their combination. In total 28.4% of patients developed dnDSA (n = 52), 12.5% class I dnDSA (n = 23), 24.6% class II dnDSA (n = 45), and 8.7% both class II and II dnDSA (n = 16). Mean epMMs (p‐value = 0.005), mean highly immunogenic epMMs (p‐value = 0.003), and PIRCHE‐II (11‐loci) (p = 0.01) were higher in patients with compared to without class II dnDSA. Patients with highly immunogenic epMMs above 30.5 and PIRCHE‐II 11‐loci above 560.0 were more likely to develop dnDSA (31.1% vs. 14.8%, p‐value = 0.03). The logistic regression model including the grouping variable showed the best predictive value. MMA can support clinicians to identify patients at higher or lower risk for developing class II dnDSA and might be helpful tools for immunological risk assessment in LTX patients.