AI Boosts Early Breast Cancer Detection Between Screens | Mirage News

A recent study by UCLA Health Jonsson Comprehensive Cancer Center suggests that artificial intelligence (AI) can significantly improve early breast cancer detection. The research, published in the Journal of the National Cancer Institute, found that AI can identify mammographically-visible cancers that develop between routine screenings, potentially reducing the number of these cancers by 30%.

The study focused on interval cancers, which are often difficult to detect because they have subtle signs or are missed by radiologists during initial screenings. The study analyzed nearly 185,000 mammograms from 2010 to 2019, using a commercially available AI software called Transpara.

The AI tool scored each mammogram for cancer risk, flagging those with a score of 8 or higher as potentially concerning. The results showed that the AI successfully flagged 76% of mammograms later linked to interval breast cancer and 90% of cases where radiologists missed visible signs of cancer.

It also performed well in identifying cancers with subtle signs, which could reasonably be acted upon. Despite the promising results, the researchers also acknowledged the limitations of the AI. The tool wasn't always accurate in pinpointing the exact location of the cancer, and it flagged some areas as suspicious that weren't visible to the human eye.

The study highlights the need for larger prospective studies to understand how radiologists would incorporate AI into their practice and address these inaccuracies. The researchers emphasize that AI is not a perfect solution and shouldn't be used in isolation. However, they believe that AI has the potential to enhance early detection, especially for cancers that are challenging to identify.

This could provide radiologists with better tools and improve patient outcomes by enabling earlier treatment and potentially saving lives.