S2483
Physics - Radiomics, functional and biological imaging, and outcome prediction
ESTRO 2026
(PC). This study aims to evaluate in vivo and in vitro ADC as a robust quantitative imaging biomarker for treatment response assessment in PC. Material/Methods: Two diffusion – weighted (DWI) sequences, REF and STUDY (b=0, 150, 500 s/mm ² [1]) were assessed on a 1.5T Unity® MR-Linac (Elekta, Sweden), with REF sequence derived from the MR-Linac Consortium protocol, and STUDY(in-house) sequence providing higher spatial resolution but requiring longer acquisition time. In vitro, evaluation of REF and STUDY was performed on the QIBA ice-water phantom (0 °C) , along with the usual diffusion -weighted sequence on a 1.5T diagnostic MRI sequence (“DIAG”, b=50, 800 s/mm ² ), following QIBA metrics : ADC bias, Signal-to- Noise ratio (SNR), linearity, repeatability, reproducibility [2]. In vivo, diffusion-weighted images were acquired at simulation and at each treatment fraction in PC patients using the REF and STUDY sequences. GTVs were manually contoured by a physician using MIM® software. Fourteen patients (REF) and twelve patients (STUDY) treated with SBRT (50 Gy in 5 fractions) were included in the analysis of mean ADC repeatability within the GTV, assessed using the repeatability coefficient (RC). An additional evaluation was performed to investigate the presence of significant bias between the REF and STUDY sequences during the simulation. A preliminary analysis on ten patients compared ADC changes between fractions 1 and 5 to the RC of each sequence, using the Wilcoxon signed-rank test to evaluate significant differences (P < 0.05). Results: In vitro, the three sequences demonstrated good performance according to QIBA standards (Table 1). Reproducibility relative to the diagnostic MRI was 0.74% for STUDY/DIAG and 0.89% for REF/DIAG. In vivo, ADC repeatability within the GTV was 22.62% for STUDY and 47.98% for REF. A significant bias of 0.10x10 ⁻ ³ mm/s was observed for simulation (p=0.039) between sequences. ADC changes during treatment did not exceed the RC for any patient, except for one patient in the STUDY sequence (+25.4%), and no consistent upward or downward trend was observed (STUDY: p=0.922, REF: p=0.125).
Conclusion: Both DWI sequences on the MR-Linac were technically robust and QIBA-compliant, with the STUDY sequence showing superior repeatability for PC monitoring. However, further optimization of the MRI sequence is required to improve repeatability and to potentially detect meaningful changes in a larger patient cohort. References: 1. Kooreman ES, van Houdt PJ, Keesman R, Pos FJ, van Pelt VWJ, Nowee ME, Wetscherek A, Tijssen RHN, Philippens MEP, Thorwarth D, Wang J, Shukla-Dave A,
Hall WA, Paulson ES, van der Heide UA. ADC measurements on the Unity MR-Linac: A
recommendation on behalf of the Elekta Unity MR- Linac consortium. Radiother Oncol. 2020; 153:106– 113. doi: 10.1016/j.radonc.2020.09.0462. Quantitative Imaging Biomarkers Alliance (QIBA). Profile: Magnetic Resonance Diffusion-Weighted Imaging (DWI) of the Apparent Diffusion Coefficient (ADC). RSNA; version n.d. Available at: https://qibawiki.rsna.org Keywords: MR-Linac, repeatability, pancreatic tumor Digital Poster 4086 Re-irradiation in pediatric brain tumors - An exploratory analysis. Parag Mude 1 , Tejpal Gupta 1 , Abhishek Chatterjee 1 , Rahat Malhotra 1 , Archya Dasgupta 1 , Girish Chinnaswamy 2 , Arpita Sahu 3 , Jeevanshu Jain 4 , Akash Pawar 5 1 Radiation Oncology, Tata Memorial Hospital, Mumbai, India. 2 Pediatric Oncology, Tata Memorial Hospital, Mumbai, India. 3 Radiodiagnosis, Tata Memorial Hospital, Mumbai, India. 4 Medical Physics, Tata Memorial Hospital, Mumbai, India. 5 Biostatistics, Tata Memorial Hospital, Mambai, India Purpose/Objective: Recurrent Medulloblastoma (MB) and Ependymoma (EPN) represent challenging treatment scenarios in pediatric neuro-oncology. While Re-irradiation (Re-RT) offers potential for disease control, the risk of cumulative radiation toxicity, particularly Radiation Necrosis (RN), remains a significant concern. This
Made with FlippingBook - Share PDF online