ESTRO 2025 - Abstract Book

S2746

Physics - Dose prediction, optimisation and applications of photon and electron planning

ESTRO 2025

1014

Digital Poster Plan Complexity Evaluation Dashboard for Breast and Prostate VMAT Based on Outstanding Plan Database Distributions Ruben O Farias, Marina Gomes Medical Physics, Genesiscare, Chelmsford, United Kingdom Purpose/Objective: Modern radiotherapy technology enables highly modulated dose distributions, which rely on combining numerous machine parameters during plan optimization. As a result, similar dose distributions can be achieved even when plan complexity varies significantly. We developed a predictive model to assess plan quality and deliverability using complexity metrics from previous, high-quality plans. This model focuses on two high-demand VMAT treatment sites involving multiple planners: CHHIP prostate and breast with nodes. The tool provides early detection of potentially problematic plan modulations within the clinical planning workflow and includes quantitative metrics to support standardized plan quality and safe delivery. Material/Methods: Clinical VMAT plans for breast with nodes and CHHIP prostate cases were exported from Pinnacle TPS. Dosimetric quality was assessed using custom scorecards in ProKnow®, while in-house Python code analysed plan complexity and deliverability. The analysis incorporated five key plan complexity metrics and the ArcCheck-specific quality assurance (QA) gamma index ratio results. Metrics used to establish baseline historical distributions included the inverse Modulation Complexity Score (iMCS), MLC Leaf Travel Index (LTi), gantry speed, MLC average speed, and modulation index per control point. Baseline distributions were derived from plans with ProKnow scores over 75% and gamma index ratios (3%/2mm criteria) above 90%. For each treatment site, fitted distributions were generated for iMCS and LTi. Remaining parameters were processed to obtain the mean, 25th, and 75th percentile values relative to gantry angle. A simple Python-based dashboard was developed for visual evaluation of new plans. Results: Baseline data included 163 VMAT breast plans and 93 CHHIP prostate plans that met the ProKnow and ArcCheck criteria. Breast plans were further stratified by treatment side for gantry angle-dependent complexity metrics. Figure 1 displays a dashboard report for a left-breast plan with an ArcCheck gamma index ratio below 90%.