Reproducibility of Three-Dimensional Facial Surface Models Generated from Repeated CT and CBCT Scans: An Ex Vivo...

Area of Science:

• Medical Imaging
• Radiology
• Craniofacial Anatomy

Background:

• Accurate 3D facial skeletal surface models are crucial for diagnosing and treating craniofacial abnormalities.
• Evaluating the reproducibility of models generated from different imaging modalities is essential for clinical application.

Purpose of the Study:

• To assess the reproducibility of 3D facial skeletal surface models derived from repeated computed tomography (CT) and cone-beam computed tomography (CBCT) scans.
• To compare the consistency of models generated by CT versus CBCT, including low-radiation CBCT protocols.

Main Methods:

• Four dry skulls with simulated soft tissues were scanned twice using CT; eight skulls were scanned twice using two CBCT scanners.
• Facial skeletal surfaces were segmented using an optimal threshold, and 3D models from repeated scans were superimposed.
• Deviations were quantified using distance measurements and color-coded maps; segmentation and image generation inaccuracies were analyzed.

Main Results:

• Repeated threshold determinations showed no significant differences, indicating segmentation consistency.
• Computed tomography (CT) yielded more consistent 3D models (Median Mean Absolute Distance [MAD]: 0.016 mm) compared to cone-beam CT (CBCT) (MAD: 0.059 mm).
• CBCT scanners demonstrated high reproducibility, with a low-radiation protocol achieving comparable results to higher-dose settings.

Conclusions:

• Computed tomography (CT) provides slightly superior reproducibility for 3D facial skeletal surface models.
• Cone-beam CT (CBCT) scanners, including those utilizing low-radiation protocols, exhibit high reproducibility, supporting their clinical utility.
• The findings reinforce the reliability of both CT and CBCT for diagnosing and planning treatments related to facial morphology.