AI Predicts Melanoma Risk Up to Five Years in Advance Using Health Data

Researchers at the University of Gothenburg in Sweden have made a significant breakthrough in the fight against melanoma. They developed advanced artificial intelligence models that predict the risk of this skin cancer. The technology uses routine health data to anticipate which individuals are most likely to develop it. This finding allows for early detection up to five years sooner. The study, involving over 6 million people, will optimize prevention within healthcare systems.

The evaluated artificial intelligence models significantly outperformed traditional methods. These conventional approaches rely solely on an individual's age and sex. The AI analysis incorporated more complex variables. It included previous diagnoses, medication usage, and socioeconomic status. The most advanced model correctly distinguished between individuals with and without future melanoma risk in 73% of cases. Basic methods achieved only 64% accuracy.

The research was conducted in collaboration with Chalmers University of Technology. The team combined diverse population data sources to enhance the models' precision. Findings showed a significant leap in prediction capabilities. By adding sociodemographic data, diagnoses, and prior treatments, AI systems identified population segments. These groups had up to a 33% probability of developing melanoma within the next five years. This greatly facilitates the development of tools for effective risk stratification.

Existing health data allows us to identify individuals with elevated risk factors.

Statistician Martin Gillstedt, from the Sahlgrenska Academy and Sahlgrenska University Hospital, highlighted the importance of these results. "Existing health data allows us to identify individuals with elevated risk factors," he stated. While the tool is not yet implemented in daily clinical practice, the results demonstrate its utility. They show the value of population registries for anticipating public health scenarios.

The study encompassed 6,036,186 individuals in Sweden. Over a five-year follow-up period, 38,582 individuals were documented as developing melanoma. This robust quantitative evidence supports the effectiveness of the AI-based approach. Artificial intelligence can interpret routine data and predict complex diseases with remarkable accuracy.

The Swedish research team pointed out the immense potential of artificial intelligence. It enables the definition of high-risk groups for implementing targeted and frequent medical check-ups. The goal is to optimize available healthcare resources. Furthermore, it aims to generate more precise and focused prevention strategies. Associate Professor Sam Polesie, a dermatologist and co-director of the study, emphasized the importance of this personalized medicine. It brings us closer to a future where prevention is more proactive and effective.