Multi‐institutional study on the commissioning and clinical implementation of the TrueBeam enhanced leaf model in the Eclipse treatment planning system

TL;DR

This study evaluated the accuracy and clinical feasibility of a new leaf modeling technique for radiation therapy planning across multiple institutions.

Contribution

The study introduces a machine-averaged ELM approach to streamline clinical implementation and improve dose calculation accuracy.

Findings

Machine-specific ELM achieved clinically acceptable accuracy with 96.3% of results within tolerance.

Tuned ELMAve showed dose differences within ±2% and gamma pass rates above 96%.

ELMAve meets all clinical criteria and offers a practical solution for multi-institutional use.

Abstract

Recently, the Enhanced Leaf Modeling (ELM) was developed to improve the multileaf collimator (MLC) model in the Eclipse treatment planning system (TPS). This multi‐institutional study evaluated the dose calculation accuracy and clinical feasibility of the ELM for TrueBeam linear accelerators (linacs) equipped with Millennium 120 and High‐Definition 120 MLCs. In addition, to facilitate broader clinical adoption, the feasibility of using machine‐averaged ELMs (ELMAve) for both MLC types was evaluated.

Twelve TrueBeam linacs from nine institutions were included. Machine‐specific ELM (ELMMS) parameters, including Leaf Transmission (LT) and the Leaf Gap (LG), were derived from vendor MLC sequences. Three volumetric modulated arc therapy (VMAT) plans for prostate cancer, head and neck cancer, and spinal metastasis were calculated using Acuros XB with each ELMMS and verified using an ionization chamber and either Delta4 or ArcCHECK systems, based on relative dose differences and gamma pass rates (3%/2 mm). Verification criteria were set as tolerance and action levels of ≤2% and ≤3% for relative dose difference and ≥95% and ≥90% for gamma pass rate. To further improve the ELM dose calculation accuracy for VMAT cases, the LT and LG parameters were systematically adjusted based on the verification results and then averaged across all linacs to obtain the ELMAve, which was subsequently verified using the same procedure.

For the ELMMS dose calculations, the relative dose differences across the three VMAT cases averaged −1.13%, with a two‐standard deviation (2 SD) of 1.04%. The mean gamma pass rate for all dose profiles was 99.4% (2 SD = 2.6%). The ELMMS achieved clinically acceptable accuracy, with 96.3% of the verification results within clinical tolerance and no cases exceeding the action level. After parameter tuning, the ELMAve calculations exhibited dose differences within ±2% and gamma pass rates above 96% across all verification cases investigated in this study.

The ELMMS demonstrated reliable VMAT dose calculation accuracy across multiple institutions and MLC types. Tuning the LT and LG parameters can further improve ELM performance in various VMAT plans. The tuned ELMAve satisfied all clinical tolerance criteria in all verification cases and may serve as a practical approach for efficient commissioning and objective evaluation of ELM parameters across institutions.

Figures