Google Health licenses its AI breast cancer screening tool to a medical biz
iCAD will have five years to bring a product to market
Google has licensed its AI breast cancer screening model to a commercial medical technology company, paving the way for the system developed by researchers to be tested in real clinical settings for the first time.
The partnership with iCAD, a maker of cancer detection and radiation therapy equipment headquartered in New Hampshire, was announced on Monday. The company will begin integrating Google's computer vision system, designed to detect breast cancer in mammography scans, to build a product that will eventually be made commercially available to healthcare providers.
Under the agreement, iCAD will use Google Cloud services to develop infrastructure for storing data securely. A spokesperson from iCAD told The Register it had agreed to license Google's breast cancer screening model for five years.
"There will be a regulatory approval process that follows the completion of the commercial product. While these dates are difficult to predict and subject to change, we are estimating the availability of an iCAD cloud offering by the end of 2023 and the AI solution with Google AI in the first half of 2024, depending on the time required to achieve regulatory clearance," the spokesperson added.
Google has been building an AI model to aid doctors in diagnosing breast cancer from mammography scans more accurately for years. In 2020, a team of computer scientists – from Google Health, DeepMind, Verily Sciences, and universities across the UK and US, including Northwestern University, Imperial College London, and the University of Cambridge – published a paper in Nature claiming AI could outperform professional radiologists in detecting cancerous breast tissue.
The system reportedly had lower rates of false positives and false negatives, compared to six radiologists. It was touted as a way to reduce unnecessary patient follow-ups, allowing doctors to prioritize women who are at highest risk of having the disease, speeding up diagnosis.
Google Health said the model had been trained on datasets of mammograms from more than 76,000 women in the UK and more than 15,000 women in the US. It led to a 5.7 percent reduction of false positives in the US and a 1.2 percent reduction in the UK, as well a 9.4 percent reduction in false negatives in the US, and a 2.7 per cent reduction in the UK in experiments with the experts.
- Can AI transformer models help design drugs and treat incurable diseases?
- Doctors call for greater scrutiny of bidders for platform that pools UK's health info
- US drug watchdog green-lights first prostate-cancer-predicting AI software
- UK's National Health Service will roll existing Palantir work into patient data platform
Doctors and researchers previously criticized the company for not sharing the model's code, to allow others to replicate and validate the results, months after the paper's publication. Now, the code is being shared as a commercial license to companies that can deploy it in clinical settings for real patients.
"Google Health's AI tech could be used to make healthcare more available, more accessible, more accurate," Greg Corrado, Head of Health AI at Google, said in a statement.
"But effecting change like this will only be possible if we work closely with forward-looking partners, those with a deep tradition of pioneering innovation and the market experience and wherewithal to put innovations into real workflows."
"Google Health working with iCAD is a great example of two organizations coming together to leverage our mutual strengths, technological capabilities, and resources to improve breast cancer screening worldwide, with the ultimate goal of improving health outcomes of individuals and communities."
The Register has asked Google for comment. ®