AI Misses Nearly One-Third of Breast Cancers; MRI Technique Offers Potential Solution

A recent study reveals that AI-based computer-aided diagnosis (AI-CAD) systems missed nearly one-third (30.7%) of breast cancers in a review of 414 women diagnosed with the disease. The study identified dense breast tissue and small tumor size (≤2 cm) as major factors contributing to missed detections. However, the research also explored a potential solution: diffusion-weighted imaging (DWI) MRI. When radiologists reviewed only the DWI portion of the MRI scans, they were able to detect a significant portion of the cancers missed by the AI system, with one radiologist detecting 83.5% of the missed lesions. The findings suggest that supplementing AI-based screening with DWI could serve as an effective safety net, particularly for women with dense breasts. While promising, the authors caution that the study was conducted at a single institution and included only women already diagnosed with cancer. They call for prospective, multicenter trials to confirm whether DWI can reliably enhance AI-assisted breast cancer detection in broader screening populations. This research underscores the importance of recognizing the limitations of AI in medical imaging and exploring complementary techniques to improve diagnostic accuracy.