AI model powers skin cancer detection in diverse populations

Researchers at the University of California San Diego School of Medicine have developed a new approach to identifying people with skin cancer that combines genetic ancestry, lifestyle and social determinants of health using a machine learning model. Their model, more accurate than existing approaches, also helped researchers better characterize disparities in skin cancer risk and outcomes.

The research is published in the journal Natural communications.

Skin cancer is one of the most common cancers in the United States, with more than 9,500 new cases diagnosed every day and approximately two deaths from skin cancer every hour. An important part of reducing the burden of skin cancer is risk prediction, which uses technology and patient information to help doctors decide which people should be prioritized for cancer screening.

Traditional risk prediction tools, such as risk calculators based on family history, skin type and sun exposure, have historically performed better among people of European ancestry because they are more represented in the data used to develop these models. This leaves significant gaps in early detection for other populations, particularly those with darker skin, who are less likely to be of European ancestry.

As a result, skin cancer in people of non-European ancestry is often diagnosed at later stages, when it is more difficult to treat. Due to later-stage detection, people of non-European ancestry also tend to have worse overall skin cancer outcomes.

To help correct this disparity, researchers analyzed data from more than 400,000 participants in the National Institutes of Health’s All of Us Research Program, a national initiative to create a diverse database of patient data to inform new, more inclusive studies on a variety of health issues. By leveraging All of Us program participants, researchers were able to ensure that the data they used was largely represented by African, Hispanic/Latino, Asian, and mixed-ancestry populations.

The main findings of the study include:

• The new model includes genetic and non-genetic determinants, including lifestyle choices, socioeconomic variables and medication use, to stratify individuals based on their likelihood of having skin cancer.
• The model achieved 89% accuracy in identifying people with skin cancer across all populations, with 90% accuracy for people of European ancestry and 81% for people of non-European ancestry.
• In a subset of participants who had genetic data but lacked data on lifestyle and social determinants of health, the model still maintained 87% accuracy.
• Genetic ancestry, particularly the proportion of European ancestry, was a strong predictor of risk; Individuals of European ancestry were at least 8 times more likely to be diagnosed with skin cancer.

The new model is best designed as a clinical case finding aid, meaning it can help identify people who should have full-body skin exams by a dermatologist. This could help enable earlier diagnosis in people with darker skin, potentially reducing current disparities in skin cancer outcomes. Additionally, their model could be adaptable to other diseases, paving the way for more equitable and personalized medicine for all.

The study was led by Matteo D’Antonio, Ph.D., assistant professor in the Department of Medicine, and Kelly A. Frazer, Ph.D., professor in the Department of Pediatrics at the UC San Diego School of Medicine. Frazer is also a member of the Moores Cancer Center at UC San Diego.