Artificial intelligence can predict individuals at significantly higher risk of developing melanoma within five years, according to a new study. Researchers at the University of Gothenburg used AI to analyze clinical data from six million Swedish adults between 2005 and 2014.
The AI models incorporated medication history and other diagnoses, in addition to age and gender. The most advanced model accurately distinguished individuals who later developed melanoma in approximately 73% of cases, a significant improvement over traditional methods using only age and gender (64%).
By combining diagnoses, medication, and sociodemographic data, researchers identified small, high-risk groups with a melanoma development risk around 33% within five years. Early detection is critical for melanoma, the sixth most common cancer in the EU, as survival rates drop significantly once it spreads.
This AI-driven approach could enable clinicians to prioritize follow-up and invite high-risk individuals for screening, leading to more accurate monitoring and efficient healthcare resource use. However, further research and policy decisions are needed before implementation in routine healthcare.