DeepHealth, a wholly owned subsidiary of RadNet, has received FDA 510(k) clearance for its AI-powered Breast Ultrasound platform, expanding its portfolio of breast imaging technologies. The software is designed to automate lesion detection, characterize findings according to the American College of Radiology's BI-RADS standards and generate structured draft reports, helping standardize breast ultrasound interpretation and reduce variability across imaging centers. RadNet plans to deploy the platform throughout its nationwide outpatient imaging network, where it is expected to support more than 700,000 breast ultrasound examinations annually.
The AI platform was validated in a multi-reader, multi-case clinical study involving 16 board-certified U.S. radiologists. According to the company, the software achieved lesion localization accuracy exceeding 98%, increased overall cancer detection sensitivity by approximately 8% and reduced radiologist interpretation time by 37%. The platform analyzes ultrasound characteristics such as lesion shape, margins, echogenicity and posterior acoustic features before automatically incorporating measurements and findings into draft reports for physician review, reducing manual documentation and supporting more consistent reporting.
Beyond workflow improvements, the technology also offers healthcare providers a reimbursement pathway through an existing Category III CPT code covering quantitative ultrasound tissue characterization. The platform will become part of DeepHealth OS, the company's cloud-based imaging ecosystem that already includes AI tools for mammography detection, breast density assessment, arterial calcification analysis and breast cancer risk prediction. By integrating multiple AI applications into a single operating environment, DeepHealth aims to improve efficiency, diagnostic consistency and scalability across breast imaging services.
Click here to read the original news story.
