*Purpose: kSORT, a 17 gene blood biomarker, has been retrospectively validated for detection of biopsy proven acute rejection (AR) and immune quiescence in kidney transplantation (tx). This is the first assessment of the predictive accuracy of pre and serial post-tx kSORT in a prospective clinical trial of high immunologic risk tx patients.

*Methods: 113 kidney tx recipients, with cPRA of >50% (median 97%) were enrolled pre-tx and followed for 12 months post-tx in the PRISM (Prediction of Rejection In Sensitized saMples) trial. A protocol bx was done at 6 months and/or at graft dysfunction. kSORT was run pre-tx and 5 times post-tx (0.5, 1, 2, 3, 6 months post-tx). Customized software kSAS generated actionable immune risk scores as High-(HR) or Low-(LR) risk for AR. All patients had induction with Thymoglobulin and maintained on TAC, MMF and prednisone. Statistical analysis used R and Fisher’s exact test and kSORT reads were correlated with biopsy histological results.

*Results: 98 evaluable patients had kSORT analyzed across 560 serial blood samples and results correlated with 93 bx (34 for cause/indication), with 12 AR and 20 borderline AR (bAR) cases. The cumulative incidence of biopsy confirmed rejection was 25% at one year (25/98 patients). 15.7% of blood samples processed for kSORT had a null call of Indeterminate-Risk (IR). The overall predictive accuracy of the pre-tx kSORT (available in 54 pts) was 90.3% for no-rejection post-tx. 18% of pre-tx samples were HR and did not correlate with cause of sensitization. For post-tx blood samples paired with bx, 51 patients had no-rejection and 47 were correctly classified as LR (specificity 92.2%). Among the 20 BL-AR, kSORT was HR in 93% with definitive kSORT scores (HR in 14; LR in 1, IR in 5). Among the 12 AR, 78% were HR. Combining the AR and BL-AR groups, the accuracy of kSORT for matching HR results with a biopsy rejection call had 84% accuracy. A diagnostic odds ratio of a kSORT+ (HR) call in the AR and BL-AR group, compared to the odds of a LR call in the same group was >27-fold (dOR=27.4), confirming that the kSORT test can discriminate between BPAR outcomes (p<0.0001). After treatment of AR, 42% of patients remained kSORT positive.

*Conclusions: kSORT can be applied both pre-tx and at serial times post-tx to monitor patients at risk of rejection in highly sensitized patients. A LR kSORT score has 90% accuracy for predicting freedom from rejection, either before or after tx. Pre and post-tx kSORT assessment is an important adjunct measure for instituting precision medicine in the management of sensitized renal tx recipients and optimizing their outcomes.

« Back to 2019 American Transplant Congress