Contribution of AI in the Assessment of Environmental Cancer Risks

Accurate assessment of environmental cancer risks is critical for public health policymaking. However, epidemiological data are often incomplete and costly. This presentation explores how artificial intelligence (AI), specifically convolutional neural networks pretrained on ImageNet, can leverage existing data such as high-resolution satellite imagery from major American cities. By linking geographic data with cancer prevalence information, this study effectively predicts local cancer risk. It also addresses the transferability of predictive models between cities, highlighting technical and scientific challenges within the GeoCancer 2023-2027 project.