ESTRO 2025 - Abstract Book

S2752

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

ESTRO 2025

clinical plans), plans generated using RP (RPP = RapidPlan plans), and plans created using RP with additional manual optimization by a physicist (CRPP = clinical RapidPlan plans).

Material/Methods: A cohort of 25 left-sided breast cancer patients underwent a retrospective comparative dosimetric analysis. Patients were selected for having optimal plans reflecting the best possible current planning knowledge and representing typical treatments. These patients received treatment with four partial-arc (with a case-specific anterolateral avoidance sector) 6 MV photon VMAT plans generated with Eclipse software (v.16.1, Varian). The RP model was trained based on these 25 clinically treated plans. For each patient RPP and CRPP plans were generated. PTV coverage, mean doses for OARs (heart, left lung, right breast and right lung), the number of monitor units, and optimization time were compared. Paired-samples t-test was used. Results: Comparison of CP and RPP plans indicated marginally inferior PTV [D95%] coverage (95.9±0.8 % vs 95.5±0.6 %, p=0.04), significantly better OAR sparing (see figures 1 and 2a for results) and comparable number of MUs (656±74 MUs vs 666±86 MUs, p=0.53) for the RPP plans. Comparison of RPP and CRPP plans showed that CRPP plans achieved significantly better PTV coverage (96.2±0.2 %, p<0.01), marginally inferior OAR sparing (see figure 1 and 2b for results) and moderately higher MUs (692±75 MUs, p=0.07). RPP plans had significantly shorter optimization time (445±63 s [RPP] vs 1850±831 s [CRPP], p<0.01). For CP optimization time was not recorded.