A GAN based approach for inferring progression trajectories of costal cartilage calcification from cross-sectional data at image level

•Costal cartilage calcification is seen in populations by single observations in time.•We built a GAN to infer longitudinal data from cross-sectional data using CT images.•Batch-based training with extra loss terms favored calcification increase over time.•Synthetic images show stepwise growth of calcification sites in realistic patterns.•Trajectories of growth can inform biomechanics of the aging chest wall.