S2037
Physics - Dose prediction/calculation, optimisation and applications for photon and electron planning
ESTRO 2026
Gamma passing rates exceeded 95% (3%/3 mm) , >94% (2%/2 mm) (Fig2) , and >85% (1%/1 mm) for both Complex-ED and Simple-ED tables. DVH variations were statistically significant only in GI and thoracic organs but remained within 0.5 Gy , confirming clinical insignificance. Figure 1.
Proffered Paper 5207
Defining the Precision Needed for Electron Density Override: A Consensus Framework for MR-Only and Direct-to-Treatment Radiotherapy Across Platforms Mu-han Lin 1 , Haozhao Zhang 1 , Christopher Kabat 1 , Justin Visak 1 , Ruiqi Li 1 , Jinyong Lin 2 , Jinzhong Yang 2 1 Department of radiation oncology, University of Texas Southwestern Medical Center, Dallas, USA. 2 Department of radiation oncology, MD Anderson Cancer Center, Houston, USA Purpose/Objective Electron density (ED) override provides a practical and versatile solution for dose calculation in both MR-only and direct-to-treatment radiotherapy workflows. In MR-only planning, ED override serves as an alternative to synthetic CT (sCT) generation, enabling accurate dose computation directly from MR images. In direct- to-treatment workflows—where diagnostic MR, diagnostic CT, or CBCT datasets are repurposed for planning—ED override is essential to compensate for the absence of simulation-quality CT and to ensure reliable tissue density mapping across platforms. This study aimed to establish a comprehensive, consensus- based ED override look-up-table (LUT) using a large multi-institutional dataset to enhance consistency, accuracy, and efficiency in MR-only and direct-to- treatment workflows. Material/Methods Retrospective treatment planning data were collected from 711 patients at UT Southwestern and 454 patients at MD Anderson Cancer Center , both employing CT-derived ED assignments in MR-guided workflows. A custom script extracted ED values from treated plans based on a standardized list of 50 anatomical structures defined in Monaco. Two consensus ED tables were developed: (1) a structure- specific (Complex-ED) table with per-structure mean ED values, and (2) a category-based (Simple-ED) table with five generalized tissue groups—air cavity, free- breathing lung, soft tissue, soft bone, and hard bone. To evaluate clinical impact, 100 representative plans across CNS, head and neck, lung, pancreas, liver, pelvis, and prostate sites were recalculated using both consensus LUTs. Comparisons were performed against institution-specific overrides using 3D gamma analysis (3%/3 mm, 2%/2 mm, 1%/1 mm for dose >10%) and DVH metrics (D_mean, D_max, D_0.5cc) for each structure. Results Inter-institutional ED values showed excellent correlation ( R² = 0.98 ) with a mean difference of 0% ± 3% (Fig1) . The largest deviation occurred in the mandible (1.32 vs. 1.47 g/cm ³ ), yet corresponding dosimetric differences were minimal (<0.3 Gy).
Figure 2.
Conclusion A robust, consensus-based ED override LUT was developed from a large, multi-institutional dataset, demonstrating strong inter-center agreement and minimal dosimetric variability. Both the structure- specific and simplified category-based LUTs maintained high fidelity across sites, supporting standardized ED override adoption in MR- only and direct-to-treatment workflows. This harmonized approach provides a practical, validated reference to improve planning efficiency, reproducibility, and safety across MR-Linac and conventional linac platforms.
Made with FlippingBook - Share PDF online