Monte Carlo fluence simulation for prospective evaluation of interstitial photodynamic therapy treatment plans

摘要

Photodynamic therapy (PDT) delivers a localized cytotoxic dose that is a function of tissue oxygen availability, photosensitive drug concentration, and light fluence. Providing safe and effective PDT requires an understanding of all three elements and the physiological response to the radicals generated. Interstitial PDT (IPDT) for solid tumours poses particular challenges due to complex organ geometries and the associated limitations for diffusion theory based fluence rate prediction, in addition to restricted access for light delivery and dose monitoring. As a first step towards enabling a complete prospective IPDT treatment-planning platform, we demonstrate use of our previously developed PullMonte tetrahedral Monte Carlo simulation engine for modelling of the interstitial fluence field due to intravesicular insertion of fibre light sources. The goal is to enable a complete treatment planning and monitoring workflow analogous to that used in ionizing radiation therapy, including plan evaluation through dose-volume histograms and algorithmic treatment plan optimization. FullMonte is to our knowledge the fastest open-source tetrahedral MC light propagation software. Using custom hardware acceleration, we achieve 4x faster computing with 67x better power efficiency for limited-size meshes compared to the software. Ongoing work will improve the performance advantage to 16x with unlimited mesh size, enabling algorithmic plan optimization in reasonable time. Using FullMonte, we demonstrate significant new plan-evaluation capabilities including fluence field visualization, generation of organ dose-volume histograms, and rendering of isofluence surfaces for a representative bladder cancer mesh from a real patient. We also discuss the advantages of MC simulations for dose-volume histogram generation and the need for online personalized fluence-rate monitoring.