Interim results from the KidneyIntelX™ expanded validation study have been announced in a press release from Renalytix AI. The KidneyIntelX artificial intelligence-enabled in vitro diagnostic uses a machine learning algorithm to examine results from proprietary blood biomarkers in combination with information from a patient’s electronic health record (EHR) to generate a rapid kidney function decline (RKFD) score. RKFD is defined as a change in kidney function (glomerular filtration rate) of at least 5 mL/min/1.73 m2 per year.
The expanded validation study includes stored patient plasma samples and corresponding EHRs from three academic medical centers: Emory University, the Ichan School of Medicine at Mount Sinai, and the University of Pennsylvania.
In the multi-center cohort of patients with type 2 diabetes, performance targets were met or exceeded by the KidneyIntelX algorithm. The algorithm identifies patients experiencing rapid decline in kidney function as well as patients who will progress to kidney failure and/or dialysis. The data will support the ongoing regulatory progress with the US Food and Drug Administration under Breakthrough Device designation, announced in May 2019.
Initial performance targets reported in April demonstrated that the ability to predict which patients went on to experience RKFD was significantly increased with the KidneyIntelX machine learning algorithm compared with currently used diagnostic methods. In the expanded validation study population, the positive predictive value of KidneyIntelX for RKFD in patients with type 2 diabetes exceeded 50% in patients who were in the highest 15% of the risk distribution, as well as a negative predictive value of >95% for patients unlikely to develop RKFD.