Cleerly reported to the FDA that its AI analysis of a 67-year-old patient's coronary CT scan showed moderate (50-69%) stenoses; months later the patient had acute chest pain, was found to have two 98% blockages and had open-heart surgery the next day. Cleerly said it is unknown whether its analysis underreported the stenoses or the disease progressed, and filed the report because an unconfirmed inaccuracy may have contributed; the patient had not pursued the follow-up testing recommended.