RadNet Inc. highlighted its rapid expansion in digital health at the 2026 Jefferies Healthcare Services and Technology Conference, with its DeepHealth unit reporting a 90% year-over-year revenue growth in the second quarter. The company’s core imaging business generated over $2 billion annually, while its digital health segment—split evenly between clinical AI and operational software—guided revenue to $135 million to $145 million for 2026. Revenue from external customers accounted for 65% of total digital health income, with RadNet contributing 35%.
The imaging division’s same-store growth remained robust, with MRI procedures up over 10%, while PET and CT scans grew by 8% to 9%. Self-pay mammography revenue exceeded $15 million annually, nearing $20 million, and breast ultrasound procedures approached 1 million—comparable to England’s annual mammogram volume. Thyroid ultrasound exams totaled about 250,000 per year, generating roughly $60 per exam through reimbursements from over half of payers.
RadNet operates 442 imaging centers, with 36% in joint ventures, aiming to increase this share toward 50%. The company’s leverage ratio improved to below 2x, down from historical levels of around 6x.
AI-driven innovations underpin its growth. RadNet has 27–28 FDA-cleared clinical AI products, including a chest X-ray tool using visual language models instead of convolutional neural networks. Thyroid ultrasound AI achieved a 94%–95% radiologist acceptance rate, reducing exam times from 30 to 20 minutes and generating $60 per exam. Breast ultrasound AI automates report generation, while acquisitions like Gleamer expanded X-ray and fracture-detection capabilities.
The company’s partnerships include Trinity Health in Idaho, Cleveland Clinic, and three mammography centers inside Walmart locations. Greg Sorensen, RadNet’s Chief Strategy Officer, emphasized workflow efficiency, noting that AI’s value lies in augmenting radiologist workflows rather than cancer detection alone. He also highlighted a potential Jevons paradox—declining radiology costs while expanding demand.
RadNet’s digital health unit employs 600 people, up from just 12 upon its 2020 acquisition of DeepHealth. The company targets over 12 acquisitions annually, with internal tools replacing legacy systems like PowerScribe through AI-assisted speech recognition. Future AI submissions include breast cancer risk prediction, with a decision expected by year-end.
