AI Identifies Melanoma Risk in Sweden

A study led by the University of Gothenburg, in collaboration with Chalmers University of Technology, has demonstrated that artificial intelligence can identify early risk patterns for melanoma skin cancer. Utilizing routinely collected registry data from Sweden's adult population, including age, sex, diagnoses, medications, and socioeconomic status, researchers developed AI models. The most advanced model achieved 73% accuracy in distinguishing individuals who developed melanoma within five years. This method can identify small, high-risk groups with a 33% probability of developing melanoma, suggesting that selective screening could lead to more accurate monitoring and efficient use of healthcare resources. Martin Gillstedt and Sam Polesie were key figures in the analysis and leadership of the study, respectively. While further research and policy decisions are needed, the findings indicate a significant step towards personalized risk assessments and future screening strategies for melanoma.