High-performance simulation enabled personalized vascular-targeted PDT planning for port wine stains

Port wine stains (PWS) are vascular malformations often treated with vascular-targeted photodynamic therapy (V-PDT). To enhance treatment personalization, we developed a real-time V-PDT simulation method that models light, photosensitizer, and oxygen interactions in tissue. Using optical coherence tomography angiography (OCTA), we visualize PWS vasculature and apply high-precision segmentation for accurate tissue modelling. The system integrates tissue optics, light distribution, photosensitizer diffusion, and photodynamic reactions to predict singlet oxygen dynamics. GPU tensor computation ensures high-speed performance, while gradient fields and ray tracing improve photon transport accuracy. Validated against digital phantoms and fiber measurements, the method rapidly simulates complex tissues. Additionally, a GPU-based 3D diffusion algorithm tracks singlet oxygen concentration changes over time, optimizing clinical treatment planning. This approach will improve precision and efficiency in personalized V-PDT for PWS.