ESTRO 2025 - Abstract Book

S3828

Physics - Radiomics, functional and biological imaging and outcome prediction

ESTRO 2025

Netherlands. 3 Department of Radiotherapy, Erasmus MC Cancer Institute, University Medical Center Rotterdam, Rotterdam, Netherlands. 4 Medical Delta, LDE, Delft, Netherlands

Purpose/Objective: Glioblastoma is notorious for its extensive tumor infiltration, which is often undetectable on conventional MRI. 1,2 Non-enhancing tissue abnormalities are assumed to partially reflect tumor infiltration, but are typically excluded from the gross tumor volume (GTV), as they are difficult to distinguish from edema. 3 The 2023 ESTRO-EANO guidelines 3 highlight that perfusion MRI may help identify areas of tumor infiltration; however, its integration for target delineation remains ill-defined. We aim to integrate perfusion MRI into target delineation and investigate its impact on the GTV. Material/Methods: Patients with glioblastoma, recruited in two prospective studies on advanced MRI and scheduled for radiotherapy treatment with 60 or 40.05 Gray, underwent an extended treatment planning MRI-scan that included the clinical brain tumor protocol plus pseudo-continuous arterial spin labeling (ASL). ASL-based cerebral blood flow (CBF) maps were generated to quantify perfusion. Construction of the conventional target volumes adhered to clinical guidelines. Based on the contralateral normal-appearing gray matter (cNAGM), a threshold for elevated CBF (S CBF,thr ) was calculated for each patient through: This threshold was used to automatically delineate a biological tumor volume (BTV), defined as elevated CBF within residual contrast-enhancement or non-enhancing tissue abnormalities. Thereafter, the BTV was added to the conventional GTV to generate a GTV that incorporates CBF information (GTV CBF ). Hence, the GTV CBF comprises structural information from conventional MRI and biological information from ASL. A Sign test compared the volumes of the conventional GTV and GTV CBF , and the dice similarity coefficient (DSC) was calculated. Results: Between June 2023 and July 2024, twenty-four patients were included. Two patients were excluded post-hoc, because they did not proceed with radiotherapy due to clinical deterioration. For another patient, ASL was not acquired due to technical issues. The median S CBF,thr was 72.4 mL/100g/min (IQR: 62.0–79.5 mL/100g/min). There was no significant difference between the volumes of the conventional GTV and the GTV CBF ( p = 0.996); both had a median volume of 52.63 mL (IQR: 35.96–68.18 mL). The median DSC GTV-GTV,CBF was 1.00 (IQR: 1.00-1.00). Figure 1 shows the delineations of an exemplary patient. Table 1 provides the demographics and target volumes of each patient.