AI moves into cardiology
Regulators and researchers are backing AI tools for heart care — Philips won FDA clearance for DeviceGuide to provide real‑time visualization in minimally invasive valve repairs, a Penn Medicine AI model now reads cardiac MRIs at near‑expert accuracy, and a Ventripoint PR highlights decentralized AI unlocking cardiac diagnostics. The combo is accelerating predictive, image‑driven cardiology and remote surveillance. (globenewswire.com; itnonline.com; prnewswire.com)
The U.S. Food and Drug Administration issued 510(k) clearance for EchoNavigator R5.0 with DeviceGuide on March 26, 2026, formalizing regulatory authorization for the AI-enabled software to assist image guidance in mitral transcatheter edge-to-edge repair workflows. (philips.com) DeviceGuide is built on Philips’ EchoNavigator echo‑fluoro fusion and integrates with the Azurion image‑guided therapy platform to render a virtual 3D representation of repair devices by combining live echocardiography and X‑ray streams in real time. (philips.com) Philips developed DeviceGuide in close collaboration with Edwards Lifesciences and the cleared software was described by industry reporting as initially compatible with Edwards’ Pascal Ace clip for M‑TEER procedures, targeting direct embedding into cath‑lab workflows. (fiercebiotech.com) Penn Medicine’s team published a Nature Biomedical Engineering paper on March 25, 2026, describing a self‑supervised, contrastive‑learning “foundational” vision system trained across data from four large U.S. academic centers and validated on the UK Biobank plus external datasets. (nature.com) The Nature study reported training on more than 300,000 cine MRI video clips from roughly 20,000 patients and demonstrated the model’s ability to regress left‑ventricular ejection fraction and classify 39 distinct cardiac conditions, including cardiac amyloidosis and hypertrophic cardiomyopathy. ( ) VentriPoint’s March 27, 2026 PR announced a partnership with First Light Health to deploy its VMS+ hub‑and‑spoke model across British Columbia, Yukon, Alberta, Manitoba and New Brunswick, converting 2D echocardiograms into 3D volumetric images the company says can approach cardiac MRI accuracy for decentralized screening. (prnewswire.com) The same VentriPoint release named VentriPoint, HeartBeam, HeartFlow, Hyperfine and GE Healthcare as focal players in a funding shift toward AI cardiac diagnostics and cited the FDA’s authorized registry surpassing 1,200 AI medical devices with cardiovascular applications the second‑largest category. (prnewswire.com)
