Development of a low‐cost 3D‐printed phantom for end‐to‐end testing of model‐based dose calculation algorithms (MBDCAs) in high‐dose‐rate (HDR) brachytherapy

BACKGROUND: Model-Based Dose Calculation Algorithms (MBDCAs), such as AcurosBV, have improved dosimetric accuracy in high-dose-rate (HDR) brachytherapy by accounting for medium heterogeneities and variations in scatter conditions. However, their safe and effective clinical implementation requires thorough commissioning, of which end-to-end (E2E) validation using physical phantoms is an important component. Commercial anthropomorphic phantoms can be costly and may offer limited flexibility for customized testing configurations. PURPOSE: This study aims to design, fabricate, and evaluate a novel low-cost, modular, 3D-printed phantom for E2E validation of MBDCAs in HDR brachytherapy. The phantom was developed to provide a practical, measurement-based tool for commissioning and quality assurance under controlled heterogeneous conditions, particularly in resource-limited settings. METHODS: A modular phantom incorporating heterogeneity plugs (moderate-density, air cavity, and calcium carbonate) was 3D printed using Polylactic acid (PLA) on a Bambu X1 Carbon printer. EBT4 films were scanned on an Epson 10000XL flatbed scanner, calibrated and used for planar dose measurements. The phantom was scanned with a Siemens CT simulator. Treatment plans were generated using BrachyVision in Eclipse v16.1, with dose distributions calculated using both TG-43 and AcurosBV algorithms. The plan was delivered with a GammaMed Plus iX afterloader. Film measurements were compared against calculated dose maps using gamma analysis and dose profile comparisons. RESULTS: The 3D-printed phantom demonstrated high dimensional fidelity (± 0.15 mm) and allowed for reproducible catheter placement and film positioning. AcurosBV calculations showed close agreement with film measurements, with gamma passing rates of 95% at 2%/2 mm criterion and 94.2% at 3%/1 mm criterion. Mean dose differences between AcurosBV and film were 1.7% ± 1.9% for the vertical profile and -0.5% ± 1.7% for the horizontal profile. Under the controlled heterogeneous configuration of the 3D-printed phantom, TG-43 predicted higher doses than film by 3.6% ± 1.2% and 8.2% ± 1.5% for the vertical and horizontal profiles, respectively. These differences reflect the combined influence of the non-water-equivalent phantom body and the incorporated heterogeneous inserts. CONCLUSIONS: This low-cost, modular, 3D-printed phantom provides a practical platform for E2E validation of MBDCAs under controlled heterogeneous conditions. Its customizable design may complement existing commissioning methods, particularly in resource-limited settings.