Longitudinal NSCLC Treatment Progression via Multimodal Generative Models
The Virtual Treatment (VT) framework models Non-Small Cell Lung Cancer (NSCLC) progression as a dose-aware image translation problem. It uses multimodal clinical data and radiation dose to synthesize realistic follow-up CT scans. Diffusion models outperformed GANs by producing more stable tumor evolution trajectories, making them better for in-silico treatment monitoring.