Researchers are using artificial intelligence to create an automated system for screening suspicious mammograms for breast cancer risk. The software can translate a chart into a risk assessment at 30 times the speed of a human with 99 percent accuracy, its developers claim.
The research team, led by Stephen T. Wong, PhD, and chair of the Department of Systems Medicine and Bioengineering at Houston Methodist Research Institute and Jenny C. Chang, MD and director of the Houston Methodist Cancer Center, have evaluated the mammograms of 500 breast cancer patients. The software scanned charts, collected diagnostic features, and correlated mammogram findings with breast cancer subtype.
Fifty percent of mammogram sreenings yield false positive results, says the American Cancer Society.
Currently, mammograms that fall into a "suspicious" category result in biopsy for risk assessment. With the AI system, however, these suspicious mammograms could be further assessed, quickly and accurately, to drastically lower the number of false positives.
With about 20 percent of breast biopsies being unnecessary in this country, there is definitely room for improvement says Wong and his team. The software could quickly review a patient's charts to determine if biopsy is required.