Accuracy of anterior cranial base surfaces acquired from computed tomography imaging.

Abstract

Understanding computed tomography scan accuracy is crucial for reliable diagnoses and minimal radiation exposure. This study assessed the accuracy of 3D surface models of the anterior cranial base created using different imaging devices and settings. Ten human skulls were scanned with two cone beam computed tomography (CBCT) scanners, including a low-radiation protocol, a CT scanner, and an optical scanner that provided highly accurate reference models. Water-filled head shells were used to simulate soft tissue during imaging. Reproducibility of the 3D models was evaluated through repeated segmentations, while accuracy was assessed by comparing the segmented models to the reference scans. The primary metric used was the mean absolute distance (MAD) between the 3D surface-approximated models. Results demonstrated high consistency across repeated segmentations, with only minor differences observed (< 0.09 mm). The low-radiation CBCT scans produced 3D models with accuracy comparable to conventional CT scans (median: 0.12 mm, IQR: 0.07 mm). No considerable differences were found between the imaging devices or protocols. These findings confirm that low-radiation CBCT protocols can reliably produce 3D anterior cranial base models comparable to standard CT scans, offering a safer alternative. This supports lower-dose imaging for craniofacial morphology assessment in clinical and research settings, ensuring accuracy while prioritizing patient safety.