Tuesday, December 2 | 3:10 p.m.-3:20 p.m. | T7-SSBR07-2 | Room N228
An AI-based MRI protocol can increase the accessibility and cost-effectiveness of breast cancer screening, according to research to be shared on Tuesday afternoon.
How? By combining “the efficiency of abbreviated MRI with the diagnostic depth of full protocols by determining, at the scanner, when full imaging is necessary,” wrote presenter Ritse Mann, MD, PhD, of Radboud University Medical Centre in Nijmegen, the Netherlands, and colleagues.
Although multiparametric breast MRI is highly sensitive for breast cancer detection, its use for screening is limited by high cost and long scan times, the team noted. Mann’s group developed an AI detection algorithm that incorporated data from 3,272 abbreviated breast MRI exams procured between 2004 and 2020. The investigators then applied this algorithm to an independent dataset of 1,277 screening cases.
Overall, the investigators found that the AI-based protocol reduced average scan time — which allowed for more efficient screening. They did find, however, that shorter protocol times led to higher recall rates.
Learn more about their research by attending this session.