AI Outperforms Expectations in Breast Cancer Prediction

A study published in Radiology highlights the superior predictive capabilities of artificial intelligence (AI) algorithms over traditional clinical risk models for assessing five-year breast cancer risk. Researchers, led by Dr. Vignesh A. Arasu, analyzed data from over 324,000 women screened via mammograms in 2016, selecting a sub-cohort of 13,628 for detailed evaluation. The study revealed that AI models could use mammographic data to generate risk scores more effectively than the commonly employed Breast Cancer Surveillance Consortium (BCSC) model, which relies on various clinical factors. Notably, AI algorithms significantly improved predictions for interval cancers and future cancer risks, identifying missed cancers and crucial tissue features linked to cancer development. In instances of high-risk patients, AI detected up to 28% of cancers, compared to 21% by BCSC. The integration of AI in mammography could lead to personalized patient care, automating risk score generation in seconds to enhance care accuracy. Dr. Arasu emphasizes the chance for precision medicine on a national scale, with AI serves as a valuable tool to improve breast cancer risk assessments. The full study examined five AI algorithms relative to the BCSC model, presenting compelling evidence for AI's role in future breast cancer risk prediction.

https://instegro.net/ai-outperforms-expectations-in-breast-cancer-prediction/?fsp_sid=75923