The effect of cone beam computed tomography voxel size on age estimation: A three-dimensional surface area analysis study.

Abstract

Accurate three-dimensional (3D) age estimation relies on cone beam computed tomography (CBCT) image quality. This study investigated the effect of voxel size on age estimation using 3D surface area analysis (3D SAA) of developing tooth apices from different CBCT systems. CBCT scans of 128 developing maxillary second premolars from children aged 7-14 years of Malay and Chinese ethnicities were retrospectively collected using a voxel size of 76 μm. The surface area of the developing apex was measured using the 3D SAA method. A previously developed regression formula, based on 300 μm voxel size data, was applied to calculate the mean absolute error (MAE) for the 76 μm voxel data. A strong negative correlation (r = 0.960) was found between chronological age (CA) and root surface area of apex (RSAA). An MAE value of 0.51 was obtained, demonstrating the high accuracy of the existing regression model. RSAA and root development status significantly contributed to age estimation (p < 0.05), except for sex (p = 0.37) and ethnicity (p = 0.49). No significant differences in the correlation strength between CA and RSAA (p = 0.09), the correlation coefficient (p = 0.08), and the MAE values (p = 0.13) were observed when tested against the previous 300 μm findings. This study demonstrated that the 3D SAA-based age estimation model performs consistently across both voxel sizes. As smaller voxel sizes did not significantly improve accuracy, the use of larger voxel sizes may be preferred to reduce radiation exposure, especially in pediatric populations.