ESTRO 2025 - Abstract Book

S3853

Physics - Radiomics, functional and biological imaging and outcome prediction

ESTRO 2025

Purpose/Objective: The microvasculature shapes the microenvironment, directly impacting oxygen delivery and the consequent radiotherapy outcome. In the general effort to tailor radiation therapy, we still have limited access to data on microvasculature and its evolution during treatment. In this work, we propose a digital twin (DT) workflow to model the microvasculature radio-induced evolution for the single-patient. Material/Methods: We collected sublingual microvascular data from 63 head-and-neck cancer patients using the non-invasive GlycoCheck™ system. We repeated acquisitions during the radiotherapy treatment until side effects allowed a measure. We built the patient-specific baseline DT by generating 2D synthetic microvascular networks using Voronoi based models and calibrating them against the single-patient data. To describe blood flow and capillary density variation with the dose, we modeled a general non-linear function that minimizes mean-squared-errors coupled with random effects. We used the obtained relationships to build a personalized computational model that describes the microvasculature alterations as a function of dose. We computed synthetic microvascular networks matching the patient’s data at each dose point, and we obtained functional information by solving for blood flow and oxygen delivery in the microenvironment [1]. We quantified the radiotherapy effect on tumour cells considering the oxygen effect by an oxygen-modified Linear-Quadratic (LQ) model[2]. Results: From our dataset, we observed that blood flow increases during radiotherapy and capillary density decreases, suggesting damage to blood vessels and an increment in blood flow to boost the tissue repair process (Figure 1).