The method combines genetics, lifestyle and social determinants using artificial intelligence, leading to better understanding of outcome disparities.

A team of researchers at UC San Diego’s School of Medicine have developed a new approach for identifying people with skin cancer using artificial intelligence.

Combining genetic ancestry, lifestyle and social determinants of health, the new method has shown to be more accurate than existing approaches and has helped the researchers better characterize disparities in skin cancer risk and outcomes.

The study, published recently in Nature Communications, analyzed data from more than 400,000 participants in the National Institutes of Health’s All of Us Research Program, which aims to build a diverse database of patient data to inform more inclusive studies on a variety of health conditions.

Skin cancer is among the most common cancers in the US, with more than 9,500 new cases diagnosed every day and about two deaths from skin cancer occurring every hour.

Risk prediction, or using technology and patient information to decide which individuals should be prioritized for cancer screening, is critical to reducing skin cancer’s rise. Traditional risk prediction tools, however, have historically performed best in people of European ancestry due to an abundance of data about those groups.

This leaves significant gaps in early detection for other populations, particularly those with darker skin, who are less likely to be of European ancestry.

These groups, then, are frequently diagnosed at later stages, leading to worse overall outcomes.

The new study ensured the team had substantial data representation from African, Hispanic/Latino, Asian and mixed-ancestry populations.

The model included a variety of determinants to determine skin cancer likelihood, achieving 89% accuracy in identifying individuals with skin cancer across all populations.

Genetic ancestry remained a strong predictor of risk, with those individuals of European ancestry at least eight times more likely to be diagnosed with skin cancer.

Using the new model may enable earlier diagnosis in people with darker skin tones, potentially easing current disparities in skin cancer outcomes.