FDA clears DeepHealth AI that finds and characterizes breast ultrasound lesions
RadNet subsidiary DeepHealth received FDA 510(k) clearance on July 30, 2026 for DeepHealth Breast Ultrasound, software that automates lesion detection, characterization and reporting. In a multi-reader study with 16 U.S. board-certified radiologists, the company reports more than 98% lesion localization accuracy, an 8% gain in cancer-detection sensitivity and a 37% cut in interpretation time. It is commercially available in the U.S., and DeepHealth expects to roll it out across RadNet's imaging centers by year end.
Why it's notable: Breast ultrasound is labor-intensive and highly operator-dependent, and it is used more often for women with dense breasts. A cleared tool that drafts findings and reports could ease radiologist workload in breast imaging.