KidneyIntelX sifts through thousands of datapoints such as large electronic health records, blood-based biomarkers and other genomic information to help identify people at risk of kidney failure.
During the trial, the algorithm “significantly increased” the ability to predict which patients, all of whom had Type 2 diabetes, would go on to experience rapid kidney function decline (RKFD).
Impressive negative predictive value results
It was also able to better predict which patients would eventually progress to kidney failure and/ or dialysis.
When KidneyIntelX deemed a patient not to be at risk, it was right an impressive 95% of the time, compared with just 30% using current methods.
When it thought a patient was at risk of kidney problems, the algorithm was correct in more than 50% of cases.
“We have now completed the complex process of integrating electronic health record data and our proprietary blood biomarker measurements from over 3,000 patients with Type 2 diabetes from three independent, diverse population groups,” said chief technology officer Fergus Fleming.
“KidneyIntelX continues to demonstrate significant performance improvements over current diagnostic standards and now in a multi-centre environment.”
Commercial launch ‘later this year’
A final cohort of patients is expected to test KidneyIntelX which will form the basis for consideration in the ongoing regulatory approval process, an update on which is expected in the “near term”.
Renalytix expects to commercially launch KidneyIntelX in the US later this year.
Shares were up 4.3% to 339p on Tuesday morning, valuing the company at £182mln.