*Purpose: Traditional microscopic histopathologic assessment (HDx) can be prone to error due to inter-observer variability as well as the non-specificity of some histologic lesions. Donor-derived cell free DNA (ddcfDNA), a plasma test to detect rejection, was validated using histopathology. The seminal DART study showed that dd-cfDNA (Allosure; CareDx) had a robust sensitivity for antibody-mediated rejection (AbMR) but less so for early T-cell mediated rejection (TCMR). This may partly be due to the noise in HDx itself where early TCMR maybe over-diagnosed. The Molecular microscope (MMDx, ATAGC) may provide increased precision to histopathology. At our center, we have incorporated simultaneous measurements of ddcfDNA, MMDx and HDx in to our standard of care. Here we present our experience, focusing attention on the contrast between HDx and MMDx for the calibration of ddcfDNA for rejection diagnoses.

*Methods: We analyzed 175 kidney transplant biopsies done at our center from January 2018-October 2019 using both HDx and MMDx. AUC Curves as well as sensitivity and specificity were calculated using the previously published ddcfDNA cut-offs of < 0.21% to rule-out rejection and >1% to rule-in rejection.

*Results: Of 175 biopsies done at a median of 5.4 months post-transplant, 90 (51%) were done for cause, while the remainder (85; 49%) for surveillance of high-immunologic risk patients or after treatment of rejection. There were significant discrepancies between HDx and MMDx diagnoses, with MMDx (76; 43%) identifying a higher number of rejection cases vs HDx (64; 37%). While MMDx identified a significantly higher number of AbMR cases (56; 32%) than HDx (41; 23%); the opposite was true for TCMR cases [HDx: 27 (15%); MMDx: 15 (9%)]. Most of TCMR discrepancies were seen for borderline or TCMR 1A. AUC Curves for ddcfDNA concentration and prediction of any rejection diagnoses were more robustly correlated with MMDx (AUC=0.83; p<0.001) than with HDx (AUC=0.78; p<0.001). Using MMDx for calibration; ddcfDNA<0.21% had a sensitivity of 95% and a specificity of 37% to rule out rejection. To rule-in rejection; ddcfDNA>1% had a sensitivity of 59% and a specificity of 91%. Based upon the Youden Index though, the best sensitivity (67%) and specificity (89%) was seen with a ddcfDNA cut-off criterion of >0.82%.

*Conclusions: In this large single-center study that used MMDx for calibration, we confirm that ddcfDNA can be used as a screening test to rule-out or rule-in rejection at the two diagnostics thresholds of <0.21% and 1% with a very high likelihood. Most of the discrepancies noted in traditional histopathology vs MMDx in our study were due to either an over-diagnosis of early TCMR or an under-diagnosis of subclinical AbMR possibly due to boundary error.

« Back to 2020 American Transplant Congress