Development of artificial intelligence-based predictive models for individual risk stratification and the simulation of disease progression trajectories in lung cancer screening programs

The activity focuses on the development, validation and interpretation of predictive models for individual risk stratification within WP4 of the Try-A-Lung project. A post-test model will be designed to integrate clinical data and quantitative features extracted from low-dose CT images, including nodule morphology, emphysema and risk scores for lung cancer, cardiovascular and respiratory diseases. Machine learning and deep learning models will estimate short- and long-term outcomes and will be evaluated using discrimination and calibration metrics, with comparisons against established models and advanced artificial intelligence methods. Explainable AI techniques will be used to analyse individual trajectories and identify clinically relevant progression profiles. The activities will also include the co-supervision of theses and participation in seminars and conferences.