Accuracy and reliability of landmark-based, surface-based and voxel-based 3D cone-beam computed tomography...

Area of Science:

• Dentistry
• Radiology
• Orthodontics

Background:

• Three-dimensional cone-beam computed tomography (3D CBCT) is crucial for orthodontic diagnosis and treatment planning.
• Accurate superimposition of serial 3D CBCT scans is essential for evaluating treatment outcomes and changes over time.

Purpose of the Study:

• To compare the accuracy and reliability of three superimposition methods: landmark-based, surface-based, and voxel-based.
• To determine the most suitable method for analyzing changes in the craniofacial complex using 3D CBCT.

Main Methods:

• 20 subjects (mean age 11 years) with pre- and post-orthodontic treatment 3D CBCT scans (T1 and T2).
• Landmark-based superimposition using seven points on the zygomatic arch and supraorbital region.
• Surface-based and voxel-based superimpositions using the anterior cranial base as reference.
• Accuracy assessed by superimposing duplicated T1 scans; reliability assessed by repeated superimpositions and intraclass correlation coefficients (ICCs).

Main Results:

• Surface-based and voxel-based methods showed no significant differences when superimposing duplicated T1 scans, indicating high accuracy.
• Landmark-based superimposition accuracy showed statistically significant differences in several parameters.
• All methods demonstrated high reliability (ICCs > 0.90), with minor exceptions for specific landmarks in landmark-based and voxel-based methods.

Conclusions:

• Surface-based and voxel-based superimposition methods, referencing the anterior cranial base, are accurate and reliable for detecting positional changes.
• The landmark-based superimposition method is reliable but less accurate compared to surface-based and voxel-based techniques for 3D CBCT analysis.