Building a tissue-based molecular diagnostic system in heart transplant rejection: The heart Molecular Microscope Diagnostic (MMDx) System.

Abstract
Halloran PF, Potena L, Van Huyen JD, Bruneval P, Leone O, Kim DH, Jouven X, Reeve J, Loupy A.

The emergence of molecular systems offers opportunities for improving the assessment of rejection in heart transplant biopsy specimens. The present study developed a microarray-based system for assessing heart transplant endomyocardial biopsy (EMB) specimens.
We analyzed 331 protocol or for-cause EMB specimens from 221 subjects in 3 centers (Edmonton, Bologna, and Paris). Unsupervised principal component analysis (PCA) and archetype analysis used rejection-associated transcripts (RATs) shown in kidney transplants to be associated with antibody-mediated rejection (ABMR) or T cell-mediated rejection (TCMR), or both. To compare EMB specimens to kidney biopsy specimens, rejection status in both was simplified to TCMR, ABMR, or no rejection.
The pattern of RAT expression was similar in EMB and kidney specimens, permitting use of RATs to assign scores and group ("cluster") membership to each EMB, independent of histology. Three clusters emerged in EMB specimens, similar to kidney specimens: TCMR, ABMR, and no rejection. This permitted each EMB specimen to be given 3 scores and assigned to 1 cluster by its highest score. There was significant agreement between molecular phenotype-archetype scores or clusters-and both histologic diagnoses and donor-specific antibody. Area under curve estimates for predicting histologic TCMR, ABMR, and no rejection by molecular assessment were lower in EMB specimens than in kidney specimens, reflecting more uncertainty in EMB specimens, particularly in histologic diagnosis of TCMR.
Rejection-associated transcripts can be used to estimate the probability of TCMR and ABMR in heart transplant specimens, providing a new dimension to improve the accuracy of diagnoses and an independent system for recalibrating the histology guidelines.

J Heart Lung Transplant. 2017 Nov;36(11):1192-1200
PMID:28662985