Comparative Dosimetric Impact of Dental Prostheses and Metal Artifact Reduction In Head-and-Neck Radiotherapy: Pinnacle Versus Monaco
Abstract
Purpose
Dental prostheses in head-and-neck (H&N) cancer can introduce CT artifacts and dose uncertainties. The dosimetric impact of metal artifact reduction (MAR) depends on prosthesis material and treatment planning system (TPS) algorithms. This study evaluated dosimetric effects of metal and ceramic prostheses with and without MAR.
Methods
Dosimetric validation used two phantoms: an agar-based phantom with opposed-field geometry for baseline dose perturbation assessment, and an anthropomorphic PH-47 phantom for VMAT situation. Prostheses were placed at clinically relevant locations. Images were performed on a GE Healthcare CT simulator both with and w/o MAR. Dose distributions were computed using Pinnacle and Monaco TPS. Ground-truth reference doses were defined from artifact-free CT images with manual density override in prostheses regions. Dosimetric accuracy was evaluated using mean dose deviation analysis, lateral dose profile comparison, and planar dose verification using OCTAVIUS 2D ionization chamber arrays with gamma criteria of 3%/3 mm.
Results
Under 2D irradiation, dosimetric perturbations varied across calculation algorithms. Without MAR, metal prostheses caused systematic target dose underestimation of −2.7% (Monaco) and −3.8% (Pinnacle), with magnitude increasing proportionally to field size. MAR application yielded divergent algorithmic responses: Pinnacle showed near-zero mean dose error, whereas Monaco exhibited systematic overestimation (~+2.3%), indicating MAR-induced over-compensation in Monte Carlo-based dose calculation. Ceramic caused smaller, more spatially uniform dose reductions with improved post-MAR accuracy. In VMAT, metal-induced local dose deficits up to −10%, predominantly at profile peripheries. While MAR improved mean dose accuracy, TPS-specific differences in boundary behavior persisted, with Monaco exhibiting superior edge conformity relative to Pinnacle.
Conclusion
Prostheses induce dose perturbations, primarily at dose profile boundaries rather than mean dose metrics. While MAR improves overall accuracy, its efficacy is algorithm-dependent: Monaco demonstrates potential over-correction for metal, whereas Pinnacle shows residual boundary underestimation. DVH-based evaluation alone proves insufficient; detailed dose analysis is essential for comprehensive dosimetric verification.