Monte Carlo intercomparison of TG-43 dosimetric parameters for clinical 192Ir and 125I brachytherapy sources

Abstract

Objective. This study presents a rigorous, simultaneous Monte Carlo (MC) intercomparison of three general-purpose computational codes, MCNP6.2, TOPAS 3.9, and PHITS 3.341, for the calculation and validation of TG-43 dosimetric parameters for the clinical microSelectron-HDR v2 192Ir and Theragenics-AgX100 125I brachytherapy sources under realistic, unsimplified photon emission spectra. Approach. Simulations were performed in a spherical liquid water phantom with a radius of 18.5 cm to ensure full-scatter conditions. Dose was scored in small, distance-dependent spherical water scoring cells with varying radii. Discrete, unsimplified photon emission spectra from ICRP Publication 107 were implemented in all three codes. To address gaps in the literature, a computationally efficient simplified geometry previously adopted for the 192Ir source was evaluated, while the 125I source was modeled with high geometric fidelity. The radial dose and anisotropy functions were calculated, the radial dose data were fitted using fifth-degree polynomials, and the resulting datasets were benchmarked against established TG-43 consensus reference data. Main results. The three MC codes demonstrated good overall agreement. For the 192Ir source, the simplified geometry reproduced the consensus reference data within the ±2% criterion over the clinically relevant radial range, demonstrating its dosimetric suitability within the TG-43 framework. For the low-energy 125I source, larger deviations, reaching approximately 21%, were observed at short radial distances ( r ⩽ 0.5 cm), consistent with the greater sensitivity of low-energy photon transport to attenuation and shielding effects associated with the source encapsulation and internal components. At distances beyond approximately 1.0 cm, the results showed substantially improved agreement, with deviations generally remaining within ±3%. Fifth-degree polynomial fitting accurately reproduced the radial dose function, with deviations predominantly remaining below the recommended 2% criterion within the clinically relevant radial intervals. Significance. This work establishes a robust, cross-validated computational baseline for brachytherapy dosimetry using general-purpose MC codes under realistic photon emission spectra. The validation of the simplified 192Ir geometry supports its use in computationally efficient simulations, while the characterization of the low-energy 125I source provides relevant information for high-precision dosimetry and future investigations of low-energy radiation interactions in radiosensitization studies.