This study heralds the potential of AI as a critical tool for implementing personalised, precision medicine on a national scale.

The emerging role of artificial intelligence (AI) in the healthcare sector continues to push the boundaries of traditional medical practices.

A groundbreaking study, recently published in the journal Radiology, showcases AI’s potential in improving the accuracy of predicting breast cancer risk.

Breast cancer, a pervasive and life-threatening disease, impacts women globally. Presently, prediction models such as the Breast Cancer Surveillance Consortium (BCSC) clinical risk score, which heavily relies on patient-reported data such as age, family history, and other personal factors, are used to calculate a woman’s risk of developing breast cancer.

However, this method presents certain limitations, as explained by the study’s lead researcher, Dr. Vignesh A. Arasu, a research scientist and practicing radiologist at Kaiser Permanente Northern California.

Dr. Arasu noted that the prevailing clinical risk models depend on diverse sources of information, which might not always be readily available or thoroughly compiled.

By contrast, recent advances in AI deep learning offer a compelling alternative, with the capability to extract a much broader array of mammography features.

During their research, the team retrospectively examined thousands of mammograms. They used five different AI algorithms to compute risk scores for breast cancer over a five-year span. These AI-generated scores were subsequently compared with the traditional BCSC clinical risk score.

“The AI algorithms surpassed the BCSC risk model in predicting breast cancer risk within a 0 to 5 year timeframe,” reported Dr. Arasu.

He also emphasised the dual strength of the AI in identifying overlooked cancers and distinguishing crucial features in breast tissue that may be indicative of future cancer development.

Although AI’s utility in cancer detection through mammograms is currently in use by some institutions, this pioneering research suggests that AI could also be instrumental in determining patients’ future risk scores, a task that AI can accomplish in mere seconds.

“AI presents us with the opportunity to individualise each woman’s care, a capability not systematically available currently,” Dr Arasu stated.