# Dental metal artifacts in magnetic resonance-based synthetic computed tomography for brain radiotherapy: Impact on dose, patient setup, and geometric distortion

## TL;DR

This study shows that synthetic CT from MRI is safe for brain radiotherapy, even with dental metal implants, as it provides accurate dose and positioning.

## Contribution

The study evaluates the clinical impact of dental metal artifacts on synthetic CT for brain radiotherapy, providing evidence for its safe use.

## Key findings

- Median dose differences were ±0.3%, except near metal implants.
- Geometric distortion was less than 1 mm, supporting safe sCT use.
- sCT and CT provided equivalent positioning accuracy for radiotherapy.

## Abstract

•Dental metal implants cause artifacts on synthetic computed tomography.•Similarity index shows no differences between synthetic and real computed tomography.•Median dose differences were ±0.3%, except in regions adjacent to metal implants.•Synthetic and real computed tomography provide equivalent positioning accuracy.•Geometric distortion was <1 mm, supporting safe synthetic computed tomography use.

Dental metal implants cause artifacts on synthetic computed tomography.

Similarity index shows no differences between synthetic and real computed tomography.

Median dose differences were ±0.3%, except in regions adjacent to metal implants.

Synthetic and real computed tomography provide equivalent positioning accuracy.

Geometric distortion was <1 mm, supporting safe synthetic computed tomography use.

Magnetic resonance (MR)-based synthetic computed tomography (sCT) can improve target and organs-of-interest delineation in brain radiotherapy. However, metallic implants may cause anatomical inconsistencies, dose inaccuracies, and distortions. This study evaluates the impact of dental metal artifacts on MR-only workflow.

Ninety-six patients underwent MR and CT for radiotherapy planning; fifty-two with dental implants undergoing standard or stereotactic treatments were eligible. Dose-volume metrics (Dmean, D2%, and D98%) and γ-index analyses (1%/1 mm and 3%/3 mm) were used to compare dose for targets and organs. For a subgroup of 15 patients, further analyses were performed. Image quality of sCT was evaluated against planning CT using Mean Absolute Error (MAE) of Hounsfield units (HU) within targets and organs. Offline sCT-cone beam CT (CBCT) registrations were compared with online CT-CBCT matches. Sequences for B0-map, the map of main static magnetic field, quantified geometric distortion.

Median target dose differences remained within ±0.3%, with maximum of 3.7% for D98% due to artifact-induced body contour changes. Median organ dose deviations were ±0.3%. Passing rates for γ-index were 97.8% and 100% for 1%/1 mm and 3%/3 mm, respectively. Average MAE was below 10 HU for brain, reaching 40 HU in bone-target regions. No differences were observed between CT-CBCT and sCT-CBCT registration. Geometric distortions remained within 1 mm, satisfying radiotherapy requirements.

Despite dental metal artifacts, sCT demonstrate dose accuracy comparable to standard CT. The results support clinical use of sCT in brain radiotherapy, with caution when artifacts alter the body near targets.