Abstract
INTRODUCTION: Online adaptive radiotherapy aims to improve treatment quality by accounting for inter-fractional variation in anatomy. This study presents a quantitative comparison between adapted and non-adapted scheduled plans with identical margins in a real-world clinical setting. METHODS: We retrospectively analyzed 422 fractions from 43 patients with prostate cancer treated with the Varian Ethos system. All patients received hypofractionated treatment with 3 Gy per fraction up to a cumulative dose of 60 Gy. For each fraction, the scheduled plan (planned on planning CT, calculated on synthetic CT derived from daily cone beam CT) was compared to the adapted plan (planned and calculated on actual daily anatomy) by means of several dose-volume metrics. Comparative statistics regarding dose-volume metrics were performed using Wilcoxon signed-rank test for paired data with a two-sided hypothesis. RESULTS: Adapted plans delivered significantly better target coverage, conformality, and homo-geneity than scheduled plans. The constraints D95% ≥ 95% and V95% ≥ 95% were met in 418 out of 422 fractions with the adapted plan, compared to only 41%-84% of fractions with the scheduled plan. Median absolute improvements for these metrics ranged between 1.5 and 6.0 percentage points. Most organ-at-risk metrics remained unchanged or showed only minor differences. Interquartile ranges decreased across all metrics. CONCLUSIONS: Adaptation significantly improved target dose metrics compared to non-adapted plans, without compromising organs-at-risk sparing. Interquartile ranges were reduced for all metrics evidencing better repeatability of adapted plans.