Biopsy-Based Transcriptomics Support Rejection Monitoring Through Repeated Kidney Allograft Biopsies
Lukas Weidmann, Dušan Harmáček, Ariana Gaspert, Birgit Helmchen, Britta George, Kerstin Hübel, Seraina von Moos, Elena Rho, Thomas Schachtner
Abstract
Introduction Biopsy-based transcriptomics may detect subthreshold signals suspicious for rejection in histologically "rejection-free" biopsies, reflect antirejection treatment responses, and indicate gradual phenotyping of rejection in kidney allograft biopsies. Methods We investigated 80 "biopsy series" (baseline and corresponding follow-up biopsies) from 2018 to 2025, assessed by histopathology (Banff classification) and the Molecular Microscope Diagnostic System (MMDx). Results Baseline biopsies showed histological rejection, including partial antibody-mediated rejection (AMR) and borderline T-cell–mediated rejection (TCMR) in 55 of 80 cases (69%), with 55% molecular rejection confirmation. After a median of 9 months (interquartile range: 4–18), follow-up biopsies detected histological rejection in 9 of 25 (36%) previously "rejection-free" cases. Corresponding baseline biopsies had higher rejection (Rejection prob ; median 0.23 vs. 0.02, P = 0.008) and AMR (AMR prob ; 0.08 vs. 0.03, P = 0.002) classifier scores than those without follow-up rejection ( n = 16). Histological interstitial inflammation (i) + tubulitis (t) and glomerulitis (g) + peritubular capillaritis (ptc) scores were similar ( P = 0.411, P = 0.602). In molecular TCMR treated conventionally, 4 of 6 (67%) had TCMR prob < 0.1 at follow-up, with significant reductions in i+t scores (4.5–1.5, P = 0.031). Two of 6 (33%) progressed to mixed phenotypes. Among treated molecular AMR cases, AMR prob decreased (0.76–0.51, P = 0.047), but only 1 of 7 (14%) reached AMR prob < 0.2, whereas g + ptc scores remained unchanged ( P > 0.99). In treated mixed molecular AMR/TCMR, molecular and histological scores improved. Four of 9 (44%) showed rejection resolution, 3 of 9 (33%) shifted to molecular AMR, 1 of 9 (11%) to TCMR, and 1 of 9 (11%) remained mixed molecular phenotypes. Conclusions Biopsy-based transcriptomics differentiated suspicious molecular signals among histologically "rejection-free" biopsies progressing to rejection and provided a monitoring tool after antirejection treatment interventions.