Wednesday, November 29 | 3:40 p.m.-3:50 p.m. | W7-SSBR09-5 | Room S406B
In this scientific session, researchers will describe how a deep learning image-based risk model can identify women in their 40s who have a higher five-year future breast cancer risk.
In his presentation, Ray Mayo, MD, from the MD Anderson Cancer Center in Houston will show results from his team's model, which found that this method can support more informed decision-making for these women and their providers on starting screening.
While breast cancer screening is regular for women in their 50s and 60s, guidelines vary on screening younger women in their 40s. The Mayo team wanted to find out whether a deep-learning model could identify younger women at five-year breast cancer risk similar to that of older women. It included data from 30,212 consecutive bilateral 2D full-field digital screening mammograms collected between 2011 and 2016.
The researchers found that within each age group, the model determined that for women in their 40s, cancer rates were 0.7%, 3.3%, and 5.1% in average, intermediate, and high-risk subgroups, respectively. They also found that 37% of exams for patients in their 40s were identified as intermediate risk, with a five-year cancer rate of 3.3%. In comparison, the cancer rate was 3.2% for exams in patients ages 50-69.
"For programs that do not offer routine screening to patients aged 40-49 and for those with variable engagement of these patients in screening, a deep learning image-based model has the potential to support enhanced screening of this key age group," the study authors wrote.
Discover what else the team found by attending this session.
