Abstract
Radiation-induced pulmonary fibrosis (RIPF) is a severe late-stage complication of radiotherapy (RT) to the chest area, typically used in lung cancer treatment. This condition is characterized by the gradual and irreversible replacement of healthy lung tissue with fibrous scar tissue, leading to decreased lung function, reduced oxygen exchange and critical respiratory deficiencies. Currently, predicting and managing lung fibrosis post-RT remains challenging, with limited preventive and treatment options. Accurate prediction of fibrosis onset and progression is therefore clinically crucial. We present a personalized in silico model for pulmonary fibrosis that encompasses tumour regression, fibrosis development and lung tissue remodelling post-radiation. Our continuum-based model was developed using data from 12 RT-treated lung cancer patients and integrates computed tomography (CT) and dosimetry data to simulate the spatio-temporal evolution of fibrosis. We demonstrate the ability of the in silico model to capture the extent of fibrosis in the entire cohort with a less than 1% deviation from clinical observations, in addition to providing quantitative metrics of spatial similarity. These findings underscore the potential of the model to improve treatment planning and risk assessment, paving the way for more personalized and effective management of RIPF.