DeepHealth Wins FDA Nod for AI Breast Ultrasound Tool

DeepHealth, a subsidiary of RadNet, has received FDA 510(k) clearance for an artificial intelligence tool that reads breast ultrasound images and generates draft radiology reports. Announced on 4 August 2026, the tool automates lesion detection, characterisation and reporting from breast ultrasound scans, while leaving final clinical assessment with the radiologist. RadNet plans to deploy the tool across its US network of imaging centres by the end of 2026.

The tool helps clinicians localise suspicious soft tissue lesions, characterise details such as shape, orientation and margin, and generate a draft report with key findings and impressions. In study data submitted to the FDA as part of the 510(k) clearance, DeepHealth reported that the breast ultrasound feature improved sensitivity for breast cancer detection by 8%, while reducing radiologist interpretation times by 37%. The validation study involved 16 radiologists at imaging centres and hospitals, with the full results not yet published. The technology sits within DeepHealth’s broader breast suite, which also includes reporting and analysis tools for mammography.

The technology came into DeepHealth’s portfolio via RadNet’s 2025 acquisition of AI ultrasound firm See-Mode Technologies. RadNet, which operates more than 400 outpatient imaging centres, has estimated that more than 700,000 breast ultrasound studies annually across its network may be eligible for reimbursement under an existing CPT code for quantitative ultrasound tissue characterisation. The clearance comes as the FDA is evaluating other devices combining image analysis with report drafting, with Aidoc and Cognita both having received the FDA’s breakthrough device designation for AI tools that read chest X-rays and generate draft reports for radiologist review.

The commercial signal is the AI radiology reporting category consolidating around image analysis plus automated report drafting, with the DeepHealth/RadNet model uniquely combining clearance, national imaging network infrastructure and established reimbursement pathway. Expect Aidoc, Cognita, Rad AI, iCAD, ScreenPoint Medical, Volpara Health and Lunit to sharpen their own workflow automation positioning through the rest of 2026 and 2027.

For the latest updates and in-depth insights into the world of Medical Devices, including breakthrough treatments, industry trends, and regulatory news, contact Albert Baines today!