Google teams up with medical technology companies to develop AI breast cancer screening tool

Google announced today that it has licensed its artificial intelligence research model for breast cancer screening to medical technology company iCAD. This is the first time Google has licensed the technology, which it hopes will eventually lead to more accurate breast cancer detection and risk assessment. The companies aim to deploy the technology in a real-world clinical setting by 2024, however, commercial deployment still depends on the success of research and testing.

The partnership builds on Google's previous work to improve breast cancer detection. Back in 2020, Google researchers published a paper in the journal Nature, finding that its artificial intelligence system outperformed several radiologists at identifying signs of breast cancer. Across the thousands of mammograms studied, the model reduced the false negative rate by 9.4 percent and the false positive rate by 5.7 percent.

iCAD plans to incorporate Google's mammography AI research model into iCAD's existing tools. The first is its ProFound AI tool, which analyzes images from digital breast tomosynthesis (DBT), an advanced imaging technique sometimes called 3D mammography. This tool scans DBT images looking for malignant soft tissue density and calcifications. iCAD also plans to use Google's model for its risk assessment tool, which the company says provides a personalized assessment of breast cancer risk for each individual.

It is hoped that artificial intelligence may become a tool to help radiologists and their patients. In general, medical experts are cautious about artificial intelligence. In Google's 2020 study, there were cases where radiologists spotted cancers that the model didn't see initially. Beyond that, there's no definitive gold standard for diagnosing cancer, which can make it difficult to establish a good baseline when training an algorithm. Thus, such AI tools may be improved by offering more than two options to account for diagnostic gray areas rather than sticking to a binary outcome of cancer or no cancer. Relying too much on AI without bringing in doctors to assess all the nuances of a patient's health, especially when it comes to trying to detect early-stage cancer, could increase the risk of overdiagnosis.

Separately, Google said it was also working with the UK's National Health Service (NHS) and Imperial College London to see if its artificial intelligence technology could play a role as a "second independent observer" in double-reading mammograms, allowing radiologists to focus on high-priority cases while improving exam consistency and quality.