The FDA has cleared Heartvue.Proton, a machine-learning platform for cardiac MRI analysis that comes with permission to keep improving itself without returning to the agency for each tweak.
The 510(k) clearance covers software from Heartvue.ai, a company founded by cardiologists in Brentwood, Tennessee, with regulatory support from Austin-based consultancy Innolitics. The tool automates time-consuming measurements across three areas: two-dimensional linear dimensions, ventricular volumes and ejection fraction, and blood flow quantification. Every result remains under physician review.
Reading a single cardiac MRI by hand can stretch past an hour, which the company says keeps quantitative cardiac care out of reach for many patients. Heartvue.Proton integrates with hospital systems and is built to slot into existing imaging workflows.
The notable feature of the clearance is an FDA-authorized predetermined change control plan, or PCCP. The regulatory tool lets companies pre-clear certain changes, meaning Heartvue’s machine-learning models can improve within defined bounds as clinical data grows without a fresh submission for each update. The company says the capability keeps AI performance current over time.
Founder and CEO Jeffrey Dendy, a practicing cardiologist and assistant professor at Vanderbilt University Medical Center, said the team cleared the FDA on its first submission in about five and a half months. Heartvue.ai now begins commercial rollout to US hospitals and imaging centers.
