Predicting Risk of Breast Cancer: Automated Versus Clinical Assessment of Breast Density
Posted: Wednesday, May 16, 2018
Denser-than-average breast tissue can mask a tumor’s presence in mammography images, as well as increase a tumor’s aggressiveness, making density is an important factor when considering cancer screening options, noted Karla Kerlikowske, MD, of the University of California San Francisco, co-lead author of new research—the largest of its kind—published in the Annals of Internal Medicine.
“These findings demonstrate that breast-density evaluation can be done with equal accuracy by either a radiologist or an automated system,” said Dr. Kerlikowske in a UCSF news release.
The team first undertook to ascertain whether a computer algorithm in the standard Breast Imaging Reporting and Data System (BI-RADS) is as accurate as human radiologists in categorizing a woman’s breast tissue: fatty; scattered fibroglandular densities; heterogeneously dense; or extremely dense. The algorithm and the radiologists were, they found, equally adept. (In 30 states, women are required by law to receive notification of their breast density.)
The researchers then evaluated data from 6,369 women: 4,409 controls, 1,609 whose cancer was detected through screening, and 351 with “interval invasive” cancer—meaning diagnosed within 1 year of a negative mammography result. Compared with women in the most common density category (scattered fibroglandular), those who had extremely dense tissue had a higher risk—5.65-fold—of interval invasive cancer and a 1.43-fold higher risk of screening-detected cancer.
Since women with denser tissue “are at higher risk of aggressive tumors, [they are] thus more likely to be candidates for supplemental screening,” said Dr. Kerlikowske in a UCSF news release. Such methods include ultrasound and MRI.