We previously reported a microarray-based test for T cell-mediated rejection (TCMR) in biopsies from renal transplants. To validate this test, we undertook a prospective international multicenter study of central reading of 300 kidney transplant biopsies (INT300) from 264 patients.

Our goals were to compare the histology-DSA assessments in INTERCOM with the previous BFC403 population; correlate the TCMR score with TCMR lesions in INTERCOM biopsies; estimate the AUC for the molecular measurements compared to the histology-DSA diagnoses assigned locally; define the relationship of the TCMR score to subsequent death-censored graft survival. We used only the previously derived molecular measurement, the TCMR score.

The TCMR score correlated with i-score (r=0.52, p=<2.2e-16) and t-score (r=0.48, p=<2.2e-16), and to a lesser extent v-score (r=0.25, p=2.07e-05), across all biopsies and within the biopsies with histologic TCMR, but not with the antibody-mediated rejection (ABMR) scores (g-, cg-scores). The TCMR score indicated that 24% of borderline biopsies and 61% of BK biopsies actually had TCMR that was not diagnosed by histology. ROC analysis as prediction of TCMR and Mixed (TCMR+ABMR) histologic diagnosis demonstrated AUCs of 0.84 and 0.81 for INT300 and BFC403 respectively. The agreement with histology was about 85% but much of the disagreement was attributable to known histology inaccuracies – e.g. unrecognized TCMR in BK nephropathy; isolated v-lesions incorrectly called TCMR; and inability to diagnose TCMR in borderline biopsies. TCMR had no impact on graft survival.

Molecular test confirms previous conclusions, and shows that central molecular testing of biopsies for the presence of TCMR with predefined equations can correct many of the errors in histology assessment.

« Back to 2013 American Transplant Congress