- Google’s health division has expanded its agreement to integrate its artificial intelligence technology into medtech company iCAD’s 2D mammography suite, about nine months after the tech giant first announced the agreement bringing its AI into real-world clinical practice.
- The amendment expands the previous licensing agreement from five to 20 years and could allow for iCAD’s ProFound Breast Health Suite to serve as second reader in breast cancer screenings, pending regulatory clearance, according to a press release. Google and iCAD partnered in November to validate and incorporate Google’s mammography AI with iCAD products.
- iCAD is currently working on a submission for regulatory approval in the European Union once the commercial product is complete, president and CEO Dana Brown told Healthcare Dive in a statement. Though dates are subject to change, the technology could be available in certain countries next year.
But a looming radiologist shortage complicates the picture, particularly in low- and middle- income countries. Some countries also use double reading, a practice where mammograms are assessed by two different radiologists, which iCAD argues can strain the radiologist workforce.
Proponents of AI argue the technology could analyze imaging results and flag potential concerns for radiologist review, lessening workloads and potentially catching signs of cancer that could be missed. But the technology didn’t outperform human radiologists in qualifying exams, suggesting AI could currently be most useful as an adjunct, according to a study published in the BMJ.
Google Health announced the first version of the agreement late last year. It was the first commercial agreement from the tech giant involving its mammography AI research model, which had been in development for several years.
Google has published research on its mammography AI system, including a 2020 report that found it reduced the workload of a second radiologist reader by 88%.
In 2021, Google, Imperial College London and three NHS trusts announced they had received government funding to build on that research and evaluate the system’s ability to work as a second reader.
Radiology is a common therapeutic area for the use of AI-enabled medical devices. The number of AI-backed devices has soared over the past five years, and nearly three-quarters of submissions authorized by the Food and Drug Administration between 1995 and 2022 were for radiology, in part because of the wealth of imaging and data available, according to an analysis by MedTech Dive.