RadNet to acquire ‘leading’ artificial intelligence firm founded by former Siemens CEO

Imaging industry giant RadNet announced Wednesday morning that it’s acquiring “leading” radiology artificial intelligence vendor DeepHealth Inc.  

The AI and machine learning firm is based in Cambridge, Massachusetts, and is primarily focused on solutions to aid in the interpretation of mammography exams. A former Siemens CEO cofounded DeepHealth in 2015, and the firm made a splash recently as the top-performing team in the image crowd-sourcing DREAM Challenge.

RadNet plans to acquire the AI vendor for 1 million shares of its common stock at the April 1 closing date, and an additional payment of 1.5 million shares, contingent on several milestones. At $17.46 per share as of Wednesday morning, the deal would be valued at more than $43.6 million.

“We are more certain today than ever before that artificial intelligence will transform the diagnostic imaging and radiology industry,” Howard Berger, MD, chairman and CEO of Los Angeles-based RadNet, said in a statement. “Machine learning, big data applications and automation algorithms will allow us to deliver our services more cost-effectively, efficiently and accurately. We have been tracking DeepHealth’s progress and evaluating its algorithms for some time.”

Berger added that RadNet spends nearly 20% of its net revenue on radiologists’ interpretation of images. In 2019 alone, the company completed about 1.4 million mammography exams, representing 17% of its procedural volume and 15% of revenue. Adding the knowledge of DeepHealth to its toolbox will improve efficiency and accuracy in this and other modalities, while also “materially” benefitting shareholders, he said.

Board-certified neuroradiologist Gregory Sorensen, MD, cofounded the AI company after previously serving as president and CEO of Siemens Healthcare North America up until 2015. He’ll now pilot all of RadNet’s AI efforts, including acquisitions, partnerships and product development.

In the announcement, Sorensen noted that RadNet operates across some 335 facilities, with 700 affiliated radiologists performing 8 million exams per year. For DeepHealth’s team of AI experts, “its database of stored images and corresponding radiologist reports is as broad and as deep as any of which we have identified anywhere in the world.”

“It has unparalleled relationships with referring physicians, health plans and capitated payors,” Sorensen said of RadNet. “I cannot think of a better positioned company to capitalize on AI opportunities and to develop products and services that can transform the entire industry.”

Closing of the deal is contingent on several factors, including approval from the FDA for mammography and prostate MRI algorithms, and “demonstrating certain features, functionality and radiologist productivity.”

RadNet and DeepHealth plan to host a conference call Friday morning to discuss the deal, in addition to the former's fourth quarter earnings call slated for Thursday.