A new study shows that artificial intelligence (AI) can help find breast cancers in mammograms that are often missed by doctors. These are cancers that grow between regular screenings and are usually harder to detect.
The study, published in the Journal of the National Cancer Institute, found that AI can catch tumors that appear on mammograms but go unnoticed by radiologists. AI can even spot very faint signs of cancer that are hard to see with the human eye.
“This is important because catching these cancers earlier can make them easier to treat,” said Dr. Tiffany Yu, the lead researcher and assistant professor of radiology at UCLA. “Early detection often means patients need less aggressive treatment and have better chances of recovery.”
The researchers looked at data from nearly 185,000 mammograms taken between 2010 and 2019. They focused on 148 cases where breast cancer was found between screenings (called interval cancers). They wanted to understand why these cancers weren’t found earlier and used an AI tool called Transpara to see if it could help.
Here’s what they found:
- 76% of mammograms that were called “normal” ended up being linked to interval breast cancer.
- In 90% of those cases, the cancer was visible on the mammogram but missed by the doctor.
- In 89% of cases, the cancer showed very subtle signs that AI could detect.
- 69% of cancers were not visible at all on the mammogram.
However, the AI wasn’t good at spotting cancers that weren’t present during the original screening.
Dr. Hannah Milch, a senior researcher, said, “While the results are exciting, we also saw that AI still makes mistakes. We need more studies to test how it works in real-life clinics.”
The researchers say that larger studies are needed before AI can be used more widely in breast cancer screening.
Source: jang
