Photodynamic therapy has long struggled with a fundamental challenge: light does not travel uniformly through living tissue, and what works on average may be dangerously underdosed or toxic for an individual patient. A computational framework that can predict, patient by patient, exactly how light energy distributes through the complex architecture of the lung could redefine the precision ceiling for this cancer treatment modality.
Published in PNAS, this work presents an anatomy-resolved digital twin framework that integrates CT-derived patient geometry with computational light transport modeling to simulate photon propagation and energy deposition across individualized lung structures. Rather than relying on population-average optical tissue parameters, the system constructs a virtual replica of each patient's bronchial and parenchymal anatomy, allowing clinicians to model how light from fiber-optic delivery systems will scatter and absorb across heterogeneous tissue compartments. The framework can resolve dosimetric variation at the level of anatomical subregions, providing a mechanism to optimize fiber placement and irradiance before a single photon is delivered to a real patient.
Photodynamic therapy for lung malignancies — including endobronchial tumors and malignant pleural mesothelioma — has seen intermittent clinical interest for decades, but dosimetry has remained largely empirical. This digital twin approach connects to a broader wave of computational medicine tools that seek to personalize treatment planning the way radiation oncology dose-volume histograms transformed external beam radiotherapy. The analogy is instructive: standardized light dosing in PDT today resembles radiotherapy before treatment planning systems existed. The key limitations at this stage are the computational validation pathway — translating simulated dose distributions into confirmed clinical outcomes requires prospective trials — and the extent to which ex-vivo or phantom-derived optical properties accurately represent living, perfused lung tissue. This is an incremental but potentially foundational advance, most significant if validated in prospective clinical cohorts.