New Delhi: Artificial intelligence (AI) algorithms outperformed standard clinical risk model to predict the risk at five years of breast cancer, according to a large study of thousands of mammograms. A woman’s risk of breast cancer is normally calculated using clinical models such as the Breast Cancer Surveillance Consortium (BCSC) risk model. This model uses self-reported information and other information about the patient, including age, family history of disease, and whether she has given birth, to calculate a risk score.
“Clinical risk models rely on collecting information from different sources, which is not always available or collected,” said lead investigator Vignesh A. Arasu, from Kaiser Permanente Northern California, USA.
“Recent advances in AI deep learning give us the ability to extract hundreds to thousands of additional mammographic features,” Arasu said in a statement.
The study, published in the journal Radiologyused data associated with negative results showing no visible evidence from 2D cancer screening mammograms performed at Kaiser Permanente Northern California in 2016.
Of the 324,009 women evaluated in 2016 who met the eligibility criteria, a random group of 13,628 women was selected for analysis.
In addition, the 4,584 patients in the eligibility group who were diagnosed with cancer within five years of the original 2016 mammogram were also studied. All women were followed up until 2021.
“We selected from the entire year of screening mammograms performed in 2016, so our study population is representative of Northern California communities,” Arasu said.
The researchers divided the five-year study period into three time periods: interval cancer risk or incident cancers diagnosed between 0 and 1 years, future cancer risk or incident cancers diagnosed between 1 and 5 years, and all cancer risk or incident cancers cancers diagnosed between 0 and 5 years of age.
Using the 2016 screening mammograms, five AI algorithms generated risk scores for breast cancer over the five-year period, including two academic algorithms used by investigators and three commercially available algorithms. The risk scores were then compared with each other and with the BCSC clinical risk score.
“All five AI algorithms performed better than the BCSC risk model in predicting 0- to 5-year breast cancer risk,” Arasu said.
“This strong predictive performance over the five-year period suggests that AI is identifying both missed cancers and breast tissue characteristics that help predict future cancer development. Something about mammograms allows us to track breast cancer risk This is the ‘black box’ of AI,” the researcher added.
Some of the AI algorithms excelled in predicting patients at high risk for interval cancer, which is often aggressive and may require a second mammogram reading, follow-up exams, or short-interval imaging.
Looking at women with the highest 10 percent risk as an example, the AI predicted up to 28 percent of cancers compared to 21 percent predicted by BCSC.
Even AI algorithms trained for short time horizons, as low as 3 months, were able to predict future cancer risk for up to five years when no cancer was clinically detected by screening mammography.
When used in combination, the AI and BCSC risk models further improved cancer prediction.
“We are looking for an accurate, efficient, and scalable means of understanding women’s breast cancer risk,” Arasu said.
“Mammography-based AI risk models provide practical advantages over traditional clinical risk models because they use a single data source: the mammogram itself,” he added.
