ESTRO 2025 - Abstract Book

S3162

Physics - Inter-fraction motion management and offline adaptive radiotherapy

ESTRO 2025

References: [1]

ExacTrac Dynamic [online] Brainlab AG. [Consultation: 22 jun 2024] <

https://www.brainlab.com/es/productos-de-radiocirugia/exactrac/ >. [2] https://www.brainlab.com/es/productos-de-radiocirugia/rt-elements/ >.

Elements software applications [online] Brainlab AG. [Consultation: 22 jun 2024] <

3133

Digital Poster AI-Enhanced Optimization for kV-CBCT Guided Online Adaptive Radiotherapy in Hypofractionated Whole Breast Treatment Chang-Shiun Lin, Mu-Han Lin, Justin Visak, Narine Wandrey, Mona Arbab, Prasanna Alluri, Asal Rahimi, David Parsons Radiation Oncology, UT Southwestern Medical Center, Dallas, USA Purpose/Objective: The enhanced Intelligence-Optimization-Engine 2.0 (IOE2.0) introduces significant advancements in AI-guided planning, addressing key limitations of IOE1.0 in hypofractionated whole breast radiotherapy (26 Gy in 5 fractions). IOE1.0 often required extensive revisions due to highly patient-specific IMRT beam arrangements, which were heavily planner-dependent. While VMAT offered potential for streamlining, it was impractically slow. Furthermore, IOE1.0’s two-level importance system lacked flexibility, leading to challenges such as heart dose violations in adapted plans. This study evaluates IOE2.0’s improved VMAT capabilities and its four-level priority ranking system for impact on plan quality, reproducibility, and efficiency. Material/Methods: In IOE2.0, planners assign priority ranks (1=most important, 4=least important) to clinical goals, enabling precise control over competing objectives. To minimize planner dependency and optimize workflows, a standardized partial arc VMAT template was developed based on clinical experience at our institute. The ‘most-important’ priorities included target coverage, global dose limits, and sparing critical OARs, while lower priority levels were assigned to contralateral organs and additional OARs (e.g., liver, stomach). Plans were normalized to ensure 90% of the PTV received 100% of the prescribed dose. Retrospective replanning was performed on ten challenging cases, including left- and right-sided targets, to assess dosimetric quality, reproducibility, and computational efficiency. Results: Partial-arc VMAT plans created using IOE2.0 met all clinical goals, achieving consistent PTV V95% coverage (>97%) and maintaining global maximum doses below 110%. Lung (V8.0 Gy) and heart (V7.0 Gy) constraints were satisfied, even in complex cases with axillary lymph node involvement. The four-level priority system provided better customization, addressing limitations of IOE1.0 and reducing heart dose violations. High-quality plans were achieved with no more than two template revisions, and VMAT computation times averaged under 500 seconds — a marked improvement over IOE1.0, regardless of case complexity.