S2118
Physics - Inter-fraction motion management and daily adaptive radiotherapy
ESTRO 2026
Digital Poster 1054
Dosimetric impact of inter-observer contouring variability in online adaptive MR-Guided SBRT for Pancreatic Cancer Sara Poeta 1 , Zelda Paquier 1 , Akos Gulyban 1 , Madeline Michel 2 , Elisa Bodson 2 , Nicolas Jullian 2 , Robbe Vandenbegin 2 , Christelle Bouchart 2 1 Medical Physics Department, Institut Jules Bordet, Brussels, Belgium. 2 Radiation Oncology Department, Institut Jules Bordet, Brussels, Belgium Purpose/Objective: Online adaptive MR-guided stereotactic body radiotherapy (MRgSBRT) for pancreatic cancer enables daily plan adaptation to account for proximity to radiosensitive organs-at-risk (OARs). However, inter- observer variability in daily contouring may introduce uncertainties affecting target coverage and OAR sparing. This study evaluates the dosimetric impact of such variability by comparing delivered doses based on online versus expert reference offline contours. Material/Methods: A retrospective analysis was conducted on ten patients with pancreatic cancer treated with MRgSBRT isotoxic dose prescription of 50 Gy in five fractions using a 1.5 T MR-Linac between July 2023 and October 2024. For each of the 50 sessions, the target and adjacent OARs were delineated online by multiple observers (“online”) and re-contoured offline by a single expert (“ref”). OARs were evaluated within a 3 cm region-of-interest (ROI) surrounding the target. Dosimetric parameters were assessed based on the delivered plan and the two contour representations, to estimate the potential impact of contour variability on delivered dose to targets and OARs. Statistical analysis was performed using the Student’s t-test with a significance level of p < 0.05, implemented in the SciPy Python package (version 1.5.2). Results: Figure 1 compares target coverage and OAR doses between online (green) and reference (orange) contours. Target coverage (V50 Gy, V35 Gy) was consistent across fractions, with no significant differences. However, total GI dose was significantly higher for online contours (p < 0.05), while bile duct and great vessel doses remained within tolerance. GI constraints were exceeded in 24% of fractions, mainly involving the duodenum. However, these were evenly distributed across fractions, such that the accumulated dose remained within GI constraints, except for one patient for whom no GI toxicity was reported.Figure 2 shows that these discrepancies came from difficulty in distinguishing adjacent GI structures, particularly where the duodenal wall was in contact with the target; in one case, bile duct misidentification led to a localized dose increase.
Made with FlippingBook - Share PDF online