Google’s AI System Can Detect Breast Cancer Better Than Doctors

An artificial intelligence (AI) system developed by Google reportedly surpasses expert radiologists in accurately interpreting mammograms from screening programs, according to a new study published in Nature.

In the study, which tested the accuracy of the system and developed through a collaboration between Google and cancer researchers, the authors report that the AI system outperformed both the historical decisions made by the radiologists who initially assessed the mammograms, and the decisions of six expert radiologists who interpreted 500 randomly selected cases in a controlled study.

“The use of AI in healthcare is something clinicians have long anticipated-and even welcome,” says Jay Anders, MD, chief medical officer for Medicomp Systems, a provider of physician-driven point-of-care solutions that fix EHRs using clinical AI.

According to the study, mammograms for 25,856 women in the United Kingdom and 3,097 women in the United States were used to train the AI system. The system was then used to identify the presence of breast cancer in mammograms of women who were known to have had either biopsy-proven breast cancer or normal follow-up imaging results at least 365 days later. 

Related: Consumers Embrace AI

Using AI technology also resulted in fewer false positives, where test results suggest cancer is present when it isn't, and false negatives, where an existing cancer goes undetected, according to a press release. Compared to human experts, the program reduced false positives by 5.7% for U.S. patients and 1.2% for U.K. patients. It reduced false negatives by 9.4% for U.S. patients and 2.7% for U.K. patients.

“AI that aids clinical decision support is advantageous, not only to the physician, but also the patient,” Anders says. “As noted in the [study], we have a shortage of physicians in many areas across the country, as well as many overburdened physicians. AI that can help physicians be more efficient and improve clinical decision making is especially valuable when physician resources are limited.” 

AI can play a role of clinical decision support and make clinicians’ jobs easier-but it is not intended to replace clinicians, according to Anders.

“AI cannot substitute the wealth of real-world knowledge that a clinician gains from years of training and hands-on information-gathering,” he says. “The best clinical tool in the doctor’s office will always be the one between the clinician’s ears.”