ESTRO 2025 - Abstract Book

S4395

RTT - Treatment planning, OAR and target definitions

ESTRO 2025

Conclusion: AI-generated contours significantly reduced inter- and intra-observer variability across all observers and for all OARs. Most notably, AI-based contouring appears to reduce the impact of variation in OAR delineation experience between observers. Further work is on-going to optimise the integration of AI-based OAR delineation in radiotherapy planning.

Keywords: Variability, OARs, Autosegmentation

3987

Proffered Paper Automated palliative IMRT plan generation for Halcyon delivery using Eclipse scripting Jonathan Paine 1 , Dualta McQuaid 1 , Donna Rickard 1 , Charles Self 2 , Elizabeth West 3 , Hayley ` Dommett 3 , Marco Yau 3 , Elizabeth Adams 1 1 Radiotherapy Physics, Royal Surrey Foundation Trust, Guildford, United Kingdom. 2 Treatment Planning, Royal Surrey Foundation Trust, Guildford, United Kingdom. 3 Radiotherapy, Royal Surrey Foundation Trust, Guildford, United Kingdom Purpose/Objective: Palliative treatments are typically performed using single or opposed fields based on Virtual Simulation (VS), with calculations performed by staff who are not trained in complex planning, and short pathways to ensure rapid access to treatment. The use of simple fields are more challenging for the ring-based Varian Halcyon due to the relatively low 6FFF beam energy, which can give rise to increased hotspots near the patient surface. Auto-feathering solutions for longer targets are available for robust IMRT plans but not for simple fields. Two scripts were developed to: i) generate bony site target volumes from an applied field and ii) quickly produce IMRT treatment plans with minimal interaction.