Donor‐Derived Cell‐Free DNA Levels Predict Renal Allograft Biopsy Findings in a UK Single‐Centre Study. Results of the KORAD Study

Prior studies have shown that donor-derived cell-free DNA (dd-cfDNA) predicts renal biopsy-defined transplant rejection with a high negative predictive value (NPV). However, these studies were not in UK-based cohorts. The KORAD study aimed to assess relationships between dd-cfDNA levels and biopsy-defined rejection in a UK kidney transplant cohort, along with relationships with DSA and creatinine levels. dd-cfDNA was determined using AlloSeq cfDNA (CareDx) and DSA was identified with LABScreen single-antigen beads (Thermo Fisher Scientific). Eighty samples from 78 transplant patients had concomitant dd-cfDNA, DSA and creatinine results. The median percentage of dd-cfDNA in plasma samples with concomitant DSA negative results was significantly less than in DSA positive samples (0.37% vs. 3.15%, p = 0.002). There was no significant difference in the median percentage of dd-cfDNA in plasma samples with concomitant samples with elevated creatinine levels (0.41% vs. 0.36%, p = 0.779). Thirty-nine of these plasma samples had concomitant renal biopsy results. Thirty-two biopsies were negative for rejection and seven were positive for rejection. A receiver operating characteristic curve analysis showed that dd-cfDNA levels distinguished biopsy-defined rejection from no rejection with an acceptable AUC of 0.737 (95% CI, 0.485-0.987). The %dd-cfDNA threshold maximising overall correct classification was 0.7%. Using this threshold might have prevented 26 patients from undergoing renal biopsy unnecessarily, with two patients directed not to have a biopsy incorrectly (26 dd-cfDNA values being true negatives and two false negatives; NPV: 93% (95% CI, 83%-100%)), which outperformed the NPV for DSA: 88% (95% CI, 77%-99%) and creatinine 85% (95% CI, 69%-100%). We showed that the optimal dd-cfDNA threshold (0.7%) in this study was lower than the 1% threshold identified in non-UK studies. AlloSeq cfDNA (CareDx) with the optimal threshold showed an acceptable level of discrimination of biopsy-proven rejection and a potential to reduce the number of biopsies performed.