Conclusion
These results suggest that the clinical implementation of the monitoring of this noninvasive plasmatic score may allow the prediction and diagnosis of rejection and identify patients with graft dysfunction due to rejection, helping with a more efficient guide for immunosuppressive therapy adjustment. This finding warrants the development of prospective biomarker-guided clinical trials.
Methods
A prospective, observational study was conducted with a cohort of 79 patients followed during the first year after LT. Plasma samples were collected at predetermined time points for the analysis of miRNAs and the CXCL-10. Patients with LFTs abnormalities were submitted to a LB to rule out rejection, assessing previous and concurrent expression of the biomarkers to evaluate their predictive and diagnostic ability. Information from 86 patients included in a previous study was collected and used as a validation cohort.
Results
Twenty-four rejection episodes were diagnosed in 22 patients. Plasmatic CXCL-10 concentration and the expression of the three miRNAs were significantly elevated prior to and at the moment of the diagnosis of rejection. We developed a logistic model for rejection prediction and diagnosis, which included CXCL-10, miR-155-5p and miR-181a-5p. The area under the ROC curve (AUROC) for rejection prediction was 0.975 (79.6% sensitivity, 99.1% specificity, 90,7% PPV; 97.7% NPV; 97.1% correctly classified) and 0.99 for diagnosis (87.5% sensitivity, 99.5% specificity, 91.3% PPV; 99.3% NPV; 98.9% correctly classified). In the validation cohort (n=86; 14 rejections), the same cut-off points were used obtaining AUROCs for rejection prediction and diagnosis of 0.89 and 0.92 respectively. In patients with graft dysfunction in both cohorts the score could discriminate those with rejection regarding other causes with an AUROC of 0.98 (97.3% sensitivity, 94.1%specificity).