The Australian-built predictive BRAIx tool is trained to not only detect breast cancer in Australian women, but also predict their likelihood of developing cancer over the next four years by evaluating mammogram images.

Currently, mammogram images are examined by two radiologists and sent to a third if findings differ.

AI tools use algorithms to evaluate the same images and detect signs or risks of cancer.

Researchers say this will not replace the role of radiologists, but will help improve the accuracy and efficiency of breast screening.

St Vincent’s Clinical Director of Breast Screening, Dr Helen Fraser, said the algorithm was trained using real mammogram images from Australia’s population screening program to detect early cancer signs and risk patterns that are “invisible to the human eye”.

“We were able to train the algorithm on millions of mammogram images, which is far more than a radiologist can see in a lifetime,” Fraser said.

“We use simple image classification algorithms to detect signs of cancer and cancer risk on mammograms.

“We can create a highly accurate tool trained on local Australian populations to improve population screening programs for Australian women.”

The BRAIx detection tool is currently being used in the Breast Screen Victoria randomized controlled trial and will soon be rolled out in Breast Screen SA.

This technology has been expanded to generate a risk score, which predicts the likelihood that a patient will develop breast cancer within the next four years.

Fraser said risk prediction tools are under research and development.

“This is Australia’s first major decision-making in the medical field using AI,” she said.

“This provides the evidence needed to trust AI tools to make important medical decisions.”

Doctors now analyze your age, family history, and breast density to determine your risk of developing breast cancer.

However, the study found that the BRAIx tool could predict cancer with “much higher accuracy” than previously possible.

“One in 10 women with a BRAIx risk score in the top 2% of the population cohort we tested went on to develop breast cancer over the next four years,” Fraser said.

“And they were given all the permissions.” [mammogram] screening. “

“This is a truly groundbreaking discovery…The future of AI-powered risk detection is incredibly exciting.”

The tool could represent the most significant “quantum shift” in breast cancer deaths in Australia since the national breast cancer screening program was introduced for women over 50 more than 30 years ago.

Despite the introduction of screening programs, 90,000 women are expected to die from breast cancer in Australia over the next 25 years.

Fraser hopes Australian-made tools can put a dent in this grim statistic.

“What we have learned from screening programs is that if cancer is detected early, while it is still in the breast and has not spread to lymph nodes or distant organs, survival is almost complete. [rate] With modern treatments,” she said.

“If we can make it even more accurate, detect cancer even earlier, and actively invite women as young as 40, we have the potential to detect cancer early in all women, perhaps at even younger ages in the future, and potentially save all women’s lives.”

Never miss a story: Follow us on all platforms to be the first to know about breaking news and special stories.