Donor-derived cell-free DNA as a new biomarker for acute rejection in heart transplant recipients: results from the freedna car study

Abstract Background There is an increasing body of evidence supporting the potential utility of donor-derived cell-free DNA as a non-invasive biomarker that could allow surveillance of rejection without the need of repeated endomyocardial biopsies (EMB) in heart transplant (HT) recipients. Methods Free-DNA CAR was a multicenter, observational, prospective study, performed between 2019 and 2022 (NCT 04973943). All patients underwent per-protocol surveillance EMB at 0.5, 1, 2, 3, 4, 6 and 12 months post-HT. Dd-cfDNA levels were determined prior to each EMB, using Next Generation Sequencing technology. Primary end-point was the association between dd-cfDNA levels and the presence of clinically relevant rejection defined as ACR grade ≥2R and/or pAMR≥1. Results The cohort included 206 patients from 12 HT centers (mean age 54 ± 12 years, 74% male), with a total of 1006 pairs of EMB-dd-cfDNA determinations. ACR ≥2 was present in 44 EMB (4.4%) and AMR ≥1 in 17 EMB (1.7%). Two samples had simultaneous ACR≥2 and pAMR≥1. Patients with clinically relevant rejection had dd-cfDNA levels significantly higher than those without, 0.189% (0.058-0.6) vs 0.097% (0.04-0.24), p=0.0049 (Figure 1A, with %dd-cfDNA logtransformed as log2(x+0.01) in order to reduce the skewness). Median (IQR) dd-cfDNA values for each group are shown in figure 1B. Using GEE (generalized estimating equation) models in order to account for repeated individual measures, a trend to significant association between %dd-cfDNA and rejection was found (OR 1.25, IC95% 0.95-1.64) (Figure 2), even though it did not reach statistical significance (p = 0.109). In the subgroup of samples evaluated for ACR (n =991, after excluding 15 samples that did not have ACR but had AMR ≥1), there was also a significant difference between those with and without rejection: 0.205% (0.06-0.65) in ACR≥2 vs 0.097 (0.04-0.24) in ACR 0-1, p = 0.0097. Area under the ROC curve for the composite endpoint was 0,61. A 0.15% dd-cfDNA threshold showed a negative predictive value of 96%, which could allow to safely eliminate 61% of EMB. Conclusions Dd-cfDNA appears to be an excellent biomarker to rule out ACR and AMR. Its high negative predictive value could allow to eliminate a significant percentage of surveillance EMB currently performed in HT recipients.Median (IQR) dd-cfDNA values