Effect of Different Voxel Sizes on Accuracy of 2D and 3D Mandibular Linear Measurements Using CRANEX® 3D Cone Beam Computed Tomography Machine: Preclinical Study

Abstract

Aim To assess the accuracy of linear measurement obtained from CBCT images in 3D volumetric rendering and multiplanar slices with different voxel sizes. Methodology: 9 mandibles with twelve radioopaque markers fixed on each one were imaged using CRANEX® 3D CBCT machine. Each mandible was imaged twice with voxel sizes 250 μm and 350 μm. 2D and 3D linear measurements from CBCT images were compared to the real physical measurements taken by a digital caliper. Software measurements were recorded twice by the principal investigator and once by the supervising professor. Intra- and inter-observer agreements were assessed using Intra-Class Correlation Coefficient (ICC). SPSS®IBM Version 24 and MedCalc® version 12.1 for Windows were used for statistical analysis. Results: CBCT 2D measurements with voxel sizes of 250μ and 350μ showed a mean difference= 0 mm both with 95% of difference being -0.6 mm above or 0.7mm below and-0.5 mm above or 0.6 mm below the Gold Standard respectively. CBCT 3D measurements with voxel sizes of 250μ and 350μ showed a mean difference= 0 mm both with 95% of difference being -0.5 mm below or 0.5 mm above and -0.6 mm above or 0.7 mm below the Gold Standard. Intra-observer and inter-observer reliability regarding all measurements showed excellent agreement. Conclusions: 2D and 3D measurements driven from CBCT images with voxel size of 350 μ are accurate as measurements driven from CBCT images with voxel size of 250 μ with less radiation delivered to the patient. 2D and 3D CBCT measurements driven from CBCT are reliable.