AI Model Detects Breast Cancer Risk Up to 5 Years in Advance

AsymMirai, a new AI model, predicts breast cancer risk with up to 88% accuracy up to five years in advance. By providing transparency in its decision-making, it aims to improve trust, reduce unnecessary tests, and enhance personalized screening, reports Medscape.

Developed by Jon Donnelly, PhD Student, Duke University,?AsymMirai?improves on earlier AI systems by focusing on differences between the right and left breasts, offering a clear understanding of how risk is assessed. This approach contrasts with traditional AI models, which often operate as "black boxes," where decision-making processes are not transparent.

“One in eight women will develop invasive breast cancer, and one in 39 will die from it,” Donnelly stated. Traditional mammograms miss about 20% of breast cancers, underscoring the need for improved predictive tools. AsymMirai addresses this gap by analyzing subtle asymmetries between breast tissues, aiming to reduce unnecessary testing and lower healthcare costs.

The model was tested on over 200,000 mammograms from nearly 82,000 patients. Results showed that AsymMirai assigns a higher risk score to individuals who later develop cancer 66% of the time, compared to 71% with the earlier Mirai model. When consistent asymmetries were detected over several years, the model demonstrated an 88% accuracy rate in predicting future cancer cases.

“This new model makes AI more tangible and understandable, thereby improving its potential for acceptance,” Fides Schwartz, Radiologist, Brigham and Women’s Hospital, noted. Similarly, Vivianne Freitas, Radiologist, University of Toronto, highlighted that AsymMirai could help in developing personalized screening strategies. Patients with higher risk scores might receive more frequent screenings or supplementary MRI scans, while those with lower scores might avoid unnecessary procedures.

In?Mexico, breast cancer accounts for 25% of all cancer cases, with prevention and detection challenges exacerbated by widespread misinformation and high R&D costs, according to Scielo. To address these issues, GE Healthcare is striving to improve access to preventive healthcare in the country. Its latest innovation, the?MyBreastAI Suite,?represents a significant advancement in personalized breast cancer care. This AI-powered platform enhances mammography by providing clinicians with cutting-edge tools for more accurate detection and diagnosis, aiming to improve patient outcomes and streamline screenings, as highlighted by the company to?MBN.?

Robert Smith, PhD, American Cancer Society, expressed cautious optimism about AI’s potential in medical imaging, nonetheless. While acknowledging advancements, he cautioned against relying solely on algorithms for screening decisions due to current limitations and the need for further validation.