On July 30, 2026, DeepHealth, Inc., a subsidiary of RadNet, announced that the FDA had granted 510(k) clearance, number K260303, for DeepHealth Breast Ultrasound. The cleared device is formally the See-Mode Augmented Reporting Tool, Breast (SMART-B), manufactured by See-Mode Technologies Pty Ltd and distributed by DeepHealth. It automates three steps of the ultrasound workflow that are normally manual: lesion detection, lesion characterization, and report generation.
The clearance rests on a multi-reader, multi-case study with 16 U.S. board-certified radiologists. The company reports greater than 98 percent accuracy in localizing breast lesions, an 8 percent improvement in sensitivity for breast cancer detection, and a 37 percent reduction in radiologist interpretation time. The time figure is the one that tends to decide whether a tool is actually adopted, because reading volume, not diagnostic ambition, is what constrains most breast imaging practices.
The reimbursement path is already in place, which is unusual for a newly cleared AI device. DeepHealth says the tool is billable under an existing Category III CPT code for quantitative ultrasound tissue characterization, and RadNet estimates that more than 700,000 annual breast ultrasound studies across its network may be eligible for reimbursement by the end of 2026.
Breast ultrasound is a good test case for clinical AI economics. It is high volume, operator dependent, and used heavily as a supplemental screen for dense breast tissue, so variability in detection and reporting is a real problem rather than a theoretical one. A cleared device that improves sensitivity while cutting reading time, inside a network that owns both the imaging volume and the software, is a much cleaner demonstration of return on investment than most published AI-in-medicine results.