Background: Rejection is the major cause of graft failure after pancreas transplantation, but pancreas graft biopsies are a challenging and risky technique. Molecular biomarkers, such as donor-derived cell-free DNA (dd-cfDNA) and gene expression profile (GEP), have emerged as an alternative to monitor graft rejection in other solid organ transplants. We hypothesized whether these techniques could complement the noninvasive diagnostic performance of currently available biomarkers of pancreas graft dysfunction.

Methods: We performed a longitudinal analysis of stored plasma samples from 38 pancreas transplant recipients during the first years. Plasma samples were collected in PAXgene tubes pre-transplant, and afterwards at 1h, 24h and 7 days post-transplant, and at time of pancreas biopsy – either per protocol at 3 weeks and 12 months, or for clinical indication. The Allonext assay (Eurofins Genoma Group) was used to report the dd-cfDNA as a percentage of the donor-derived fraction as compared with total cfDNA by using panels of more than 500 SNPs . GEP was analyzed with the TruGrafâ algorithm—a gene expression algorithm analyzing differential expression of 120 genes. GEP results were provided as a probability score normalized on a 0–100 scale. The TruGrafâ assay (EurofinsTransplant Genomics) has a previously defined probability threshold of 50 to differentiate the TX (no rejection) from the not TX phenotype in kidney transplantation.

Results: Dd-cfDNA increased significantly during the first hour after pancreas transplantation (4.30±2.53%), and reduced progressively (at 24h 2.41±1.79%, at 1 week 0.65±0.47%, at 3 weeks 0.47±0.28%). Patients with biopsy-proven acute rejection (P-BPAR) episodes presented higher values of dd-cfDNA (2.1±2.46% vs 0.50±0.88%; p<0.027), most prominently in those with acute rejection diagnosed beyond 30 days from transplant.

In patients with pancreas acute rejection, GEP diagnosed the rejection episode in 58.8%, presenting a sensitivity of 74.1% (95% CI 53.7-88.9) and a specificity of 83.3% (95% CI 59.8-92.5) for the diagnosis of acute rejection, with negative predictive value of 86.7% (95% CI 77.3-92.6). Sub-clinical rejection (lipase <3xs normal) was diagnosed by GEP in 57.1%, with a specificity of 86.1% (95% CI 71-95). The combination of both parameters presented a negative predictive value of 82.3% (67-91%), with thirteen patients with P-BPAR (72.2%) presenting either test positive.

Conclusions: Dd-cfDNA and GEP can improve the detection of pancreas acute rejection episodes. GEP may provide non-invasive monitoring for sub-clinical rejection of the pancreas graft.