AI Model Predicts Kidney Stone Type More Accurately
Mayo Clinic researchers report their model identified kidney stone composition correctly in 73% of patients, compared with 33% for a standard calculation.
Updated on Nov. 12, 2026 in Artificial Intelligence

Live Poll
Would you trust an AI tool to help inform your own medical care?
Researchers at Mayo Clinic developed an artificial intelligence model that predicts kidney stone composition more accurately than a conventional calculation. The findings were published in Mayo Clinic Proceedings.
Why it matters
Kidney stones affect a large share of Americans and recur often, and many patients never have a stone available for laboratory analysis. A tool that predicts stone type from urine test data could help guide prevention and treatment without one.
The AI model was developed on data from more than 13,000 patients, including age, sex and 24-hour urine test results, and evaluated on 1,653 further patients, where it outperformed the conventional calculation at 73% accuracy versus 33%. Whether this accuracy holds in broader clinical use has not been established by the reported study.
The players
Mayo Clinic
Mayo Clinic is a nonprofit academic medical center based in Rochester, Minnesota, known for its research and patient care.
The details
The model analyzes multiple measurements from 24-hour urine tests together to identify patterns and estimate the most likely stone type. It performed better than the conventional calculation alone, including across different age and sex groups. Most stones are calcium oxalate, though calcium phosphate and uric acid stones are also common.
Timeline
The article on the study was published November 12, 2026.
Big Picture
The study extends Mayo Clinic's kidney stone research by applying artificial intelligence to the 24-hour urine test that clinicians already use to classify stones. It moves the field toward predicting stone composition without requiring a physical stone sample.
People who have had kidney stones or face recurrence risk of 60% to 80% could eventually benefit from a test that predicts stone type from urine data. The tool is still at the research stage, so no change to current testing or treatment is reported.
The takeaway
Predicting kidney stone type from routine urine data may reduce reliance on stone samples that many patients never provide. Patients with recurrent stones should continue to follow their clinician's current testing and prevention guidance.
Further reading
Explore more Artificial Intelligence coverage in the United States.
Source note: This article includes information reported by KIMT-TV 3 Mason City.
Live Poll
Would you trust an AI tool to help inform your own medical care?






