It’s taking detection to the next level.
Artificial intelligence (AI) is showing potential in identifying breast cancers that are typically missed during routine mammogram screenings, according to new research. A study published in the Journal of the National Cancer Institute suggests that using AI technology in mammography could help reduce the number of interval breast cancers—those that appear between scheduled screenings—by nearly 30%.
Researchers found that AI systems could detect tumors that are visible on mammograms but are often overlooked by trained radiologists. In some cases, AI was able to recognize extremely subtle indicators that might be too faint for the human eye to notice. Dr. Tiffany Yu, one of the study’s lead researchers from UCLA’s David Geffen School of Medicine, emphasized the significance of catching cancer in its early stages, as it could lead to less invasive treatments and better patient outcomes.
The study reviewed nearly 185,000 mammograms conducted between 2010 and 2019, identifying 148 cases of interval breast cancer. Researchers then applied an AI program called Transpara to evaluate whether it could have detected the cancers that radiologists missed. The AI system successfully flagged 76% of the cases that were originally assessed as normal but later linked to cancer. It also identified 90% of the cancers that had been visible on previous images but were not recognized at the time, along with 89% of cases involving extremely subtle visual signs.
Interestingly, the AI tool even flagged 69% of cancers that were considered invisible on initial mammograms. However, the tool’s accuracy varied when pinpointing the exact cancer location, correctly identifying the affected area only 22% of the time. Dr. Hannah Milch, another researcher involved in the study, noted that although the AI showed promise, further research is needed to refine its accuracy and understand its performance in everyday clinical use.
Overall, the researchers concluded that while AI isn’t flawless and should not replace radiologists, it could serve as a valuable supplement. Acting as a second set of eyes, AI might improve early cancer detection rates and provide critical support for radiologists, ultimately helping to save more lives through earlier diagnosis and intervention.
Discussion about this post