# Are treatment plans optimized on the basis of acuros XB dose calculation robust against anatomic changes during online adaptive radiotherapy for lung cancer regarding dose homogeneity?

## TL;DR

This study examines how well lung cancer radiotherapy plans remain effective when patients' anatomy changes during treatment, focusing on dose distribution in the bronchial wall.

## Contribution

The study introduces a predictive model linking dose changes to anatomic variations and air lumen water override in lung cancer adaptive radiotherapy.

## Key findings

- Dose homogeneity in the central bronchial wall (CBW) significantly increased between planning CT and initial CBCT.
- Reducing PTV overlap with the central bronchial air lumen decreased dose increases in CBW.
- Adaptive plans showed smaller dose changes than scheduled plans due to reduced intra-fractional deformations.

## Abstract

The Acuros XB dose calculation algorithm implements advanced modelling of lateral electron transport, making dose distributions sensitive to density changes between source and subsequent CT. The aim of this study was to analyse the robustness of dose distribution in the central bronchial wall (CBW) of treatment plans from lung cancer patients treated with adaptive radiotherapy.

IMRT or VMAT plans from patients with locally advanced lung cancer from a prospective registry cohort were analysed, who received definitive radiotherapy in surface-guided inspiratory breath-hold on the Ethos™ closed-bore linac, equipped with the HyperSight™ cone beam CT (CBCT). Dose homogeneity of the scheduled plans, optimized on planning CT (CTplan), was verified on the initial CBCT of a dose fraction (CBCT1). The adaptive plans were verified on a subsequent post-adaptation CBCT (CBCT2) of the same dose fraction. A predictive model was built for maximum dose (Dmax) in CBW in dependence on plan sensitivity in the central bronchial air lumen overlapping the planning target volume (CBALPTV) to water override (WOR) of the air lumen.

Ninety-one dose-fractions from 10 patients were analysed. Dmax values in the CBW of the scheduled plans showed over all significant inter-fractional increases from CTplan to subsequent CBCT1 (p < 0.0001, Wilcoxon test, stratified by patient) with significant heterogeneity between patients (p < 0.0001, Kruskal-Wallis Test). The median Dmax increase per dose fraction was 2.15% (-3.15 − 19.30%). Reducing the PTV overlap of scheduled plans with CBAL led to lower inter-fractional Dmax increases in CBW (p < 0.0001, signed rank test). Dose accumulation showed, that Dmax and D1cc values in CBW over the treatment course stayed in all patients below 110.5% and 107.5% and that the equivalent uniform dose in CTV around the CBW stayed > 95% for scheduled plans. A predictive model showed the dependence of inter-fractional Dmax increases in CBW of scheduled plans on an interaction between plan sensitivity on CTplan to WOR in CBALPTV and density change at the Dmax point in CBCT1 between CTplan and CBCT1 (p < 0.0001, t-test). Intra-fractional Dmax increases of adaptive plans in CBW amounted to only 20% +/- 1.1% of the inter-fractional increases of scheduled plans, as intra-fractional deformations were smaller than inter-fractional (p < 0.0001, signed rank test).

Dose homogeneity in CBW of Ethos plans were found sufficiently robust against intra-fractional deformations during course of online adaptive radiotherapy. Plan sensitivity to anatomic changes can be detected and controlled on the planning CT by the WOR of air in CBALPTV.

The online version contains supplementary material available at 10.1186/s13014-025-02656-1.