Soft tissue coverage on the segmentation accuracy of the 3D surface-rendered model from cone-beam CT

Area of Science:

• Radiology and Imaging Science
• Biomedical Engineering
• Oral and Maxillofacial Surgery

Background:

• Accurate 3D hard tissue models from cone-beam CT (CBCT) are vital for dental and medical applications.
• Applications include implant planning, virtual surgical planning, orthognathic, and navigational surgeries.
• The influence of surrounding soft tissues on segmentation accuracy requires investigation.

Purpose of the Study:

• To evaluate the impact of soft tissue presence on the segmentation accuracy of 3D hard tissue models derived from CBCT data.
• To quantify discrepancies in linear and angular measurements and surface deviations.

Main Methods:

• Utilized seven pairs of CBCT DICOM datasets from human cadaver heads and dry skulls.
• Performed linear and angular measurements on segmented models.
• Employed superimposition and color mapping of surface discrepancies after midsagittal plane splitting of the maxillo-facial complex.

Main Results:

• Significant differences observed in posterior transversal measurements of the mandible (p < 0.01).
• Mean root-mean-square error (RMSE) for the maxillo-facial complex decreased insignificantly from 0.936 to 0.922 mm (p > 0.05).
• RMSE for the mandible significantly decreased from 1.240 to 0.981 mm after splitting (p < 0.01).

Conclusions:

• Soft tissue presence affects CBCT-derived 3D hard tissue model accuracy, though often below the 1mm clinical significance threshold.
• The observed accuracy may be insufficient for applications demanding high precision.
• Consideration of soft tissue effects during segmentation is crucial for high-precision applications.