Ultromics’s EchoGo did not meet primary endpoint in heart disease trial

An AI tool developed by Ultromics showed lower accuracy compared to traditional clinical decision-making when identifying patients for further follow-ups, although it may still be helpful for less experienced clinicians. This finding was part of a presentation at the European Society of Cardiology (ESC) Congress 2024 in London, UK.

The PROTEUS trial (NCT05028179) tested the AI’s ability to assist in recommending invasive coronary angiograms based on stress echocardiography results. Stress echocardiography, a common diagnostic test for coronary artery disease, relies on clinician expertise, which can impact accuracy.

Oxford University developed EchoGo Pro, the AI software, to help interpret scan images. Despite being FDA-approved for diagnosing heart failure, the AI tool underperformed, with the standard clinical approach sending 27 of 36 patients for invasive procedures compared to 34 of 49 for the AI-augmented method. However, Ultromics suggested the AI could support less experienced clinicians and improve outcomes for difficult-to-interpret cases.