Investigation of TG-43 dose calculation algorithm for volumetric HDR Brachytherapy using Monte Carlo simulation in TOPAS.
TL;DR
The accuracy of TG-43 algorithm against Monte Carlo (MC) simulation is investigated, confirming its adequacy in homogeneous phantom, and the necessity of adopting Model-based calculation algorithms such as MC to improve the dosimetric precision and treatment accuracy in HDR brachytherapy is emphasized.
Abstract
The accuracy of dose calculation algorithms significantly influences brachytherapy treatment outcomes. TG-43 algorithm has been widely adopted due to its simplicity, which has standardized contemporary brachytherapy dose calculations worldwide. However, TG-43's homogeneity assumption potentially compromises clinical treatment accuracy. The present study aimed to investigate TG-43 accuracy against Monte Carlo (MC) simulation. Varian GammaMed Plus 192Ir High-dose-rate (HDR) brachytherapy source was modeled using TOPAS3.9 MC simulation-framework. Three well-defined test scenarios were examined: Homogeneous water phantom, TG-186 generic-applicator, and computed-tomography (CT) based intracavitary brachytherapy (ICBT) patient plan. Corresponding TG-43 based treatment plans were generated using Eclipse BrachyVision and imported into TOPAS for MC-simulation. Both TG-43 and MC dose distributions were evaluated using point-dose assessment, dose-volume metrics, dose-volume histogram, and γ-analysis with 2% Dose-difference and 1 mm Distance-to-agreement. In a homogeneous water phantom, TG-43 and MC exhibited close agreement, with dose-difference (Δ%) at the reference point and for High-risk clinical target volume (HRCTV) found to be below 1-3% and γ-passing rates ≥95%. However, in the presence of a TG-186 applicator, TG-43 substantially overestimated doses by 60-93% on shielded side and 4-13% on unshielded side. For CT-based ICBT patient plan, TG-43 overestimated HRCTV and organ-at-risk (OAR) doses relative to MC. The mean γ-passing rate in clinical ICBT plan was 90%, indicating moderate spatial agreement. Present study validated the accuracy of TG-43 algorithm against MC simulation confirming its adequacy in homogeneous phantom. However, significant deviations were observed in the presence of an applicator, and tissue heterogeneity revealed TG-43 limitations. TG-43 consistently overestimates dose owing to the homogeneity assumption. These findings emphasize the necessity of adopting Model-based calculation algorithms such as MC to improve the dosimetric precision and treatment accuracy in HDR brachytherapy.