FRIDAY, Sept. 24, 2021 (HealthDay News) — Deep learning does a better job than clinical risk factors in distinguishing between the mammograms of women who will later develop breast cancer and those who will not, according to a study published online Sept. 7 in Radiology.
Xun Zhu, Ph.D., from the University of Hawaii in Honolulu, and colleagues examined the ability of deep learning models to estimate the risk for interval and screening-detected breast cancers with and without clinical risk factors. The analysis included 25,096 digital screening mammograms (January 2006 to December 2013) from 6,369 women without breast cancer, 1,609 of whom developed screening-detected breast cancer and 351 of whom developed interval invasive breast cancer.
The researchers found that the deep learning model outperformed other models when comparing patients with screening-detected cancer versus matched controls (deep learning: odds ratio [OR], 1.25; clinical risk factors with the Breast Imaging Reporting and Data System [BI-RADS] density model: OR, 2.14; combined deep learning and clinical risk factors model: OR, 1.21). Deep learning underperformed for interval cancer risk versus controls (deep learning: OR, 1.26; clinical risk factors with BI-RADS density model: OR, 7.25; combined model: OR, 1.10).
“By ranking mammograms in terms of the probability of seeing cancer in the image, [artificial intelligence is going to be a powerful second reading tool to help categorize mammograms,” a coauthor said in a statement.
Copyright © 2021 HealthDay. All rights reserved.