AI Outperforms Human Radiologists in Breast Cancer Screening

Google and medical centres in the U.S. and UK have reported in the journal Nature that Artificial Intelligence can help doctors do a better job of finding breast cancer on mammograms. A study conducted by Google worked with researchers from Northwestern University in Chicago, Cancer Research Imperial Centre and Royal Surrey County Hospital, using a new system for reading mammograms which are X-rays of the breast. Computers were trained to recognize patterns and interpret images using mammograms from about 76,000 women in Britain and 15,000 in the United States, whose diagnoses were already known. The computers were then tested on images from about 25,000 other women in the UK and 3,000 in the U.S. and compared the system’s performance with that of radiologists who had originally read the X-rays. Tested on these images where the diagnosis was already known, the new system performed better than radiologists. On scans from the U.S., the system produced a 9.4% reduction in false negatives, in which a mammogram is mistakenly read as normal and a cancer is missed. It also provided a lowering of 5.7% in false positives, where the scan is incorrectly judged abnormal but there is no cancer. The next step in the research is to have radiologists try using the tool as part of their routine practice in reading mammograms.