Reducing False Diagnoses

Diagnosing breast cancer can be a worrisomely old-school process. For all of the 21st century tech that can image tumors, the ultimate diagnosis still rests with the radiologist’s eyes and judgment. With 30 million breast-cancer screenings in the U.S. each year, that can lead to a lot of false negatives and positives. QuantX backstops radiologists with AI-enabled software that analyzes MRIs to confirm or challenge their diagnosis. The system was approved by the FDA in 2017 when a clinical study showed a 39% reduction in missed cancers and a 20% improvement in overall accuracy. QuantX is currently being tried out at the University of Chicago and the University of Texas MD Anderson Cancer Center, with a bigger rollout planned in the coming months. —Jeffrey Kluger