Related Experiment Video
Updated: Jun 2, 2026
Reliability of Artificial Intelligence-Based Cone Beam Computed Tomography Integration with Digital Dental Images
Published on: February 23, 2024
Reliability and reproducibility of linear mandible measurements with the use of a cone-beam computed tomography and
C Tomasi1, E Bressan, B Corazza
1Department of Periodontology, Institute of Odontology, The Sahlgrenska Academy, University of Gothenburg, Sweden. cristiano.tomasi@odontologi.gu.se
Objective:
The aim of this study was to assess the influence of inclination of the object on the reliability and reproducibility of linear measurements of anatomic structures of the mandible on images obtained using cone-beam CT (CBCT).
Methods:
Ten linear dimensions between anatomical landmarks were measured in a dry mandible. The measurements were performed with a manual calliper three times by three observers. The mandible was scanned with Planmeca Promax 3D cone-beam CT (Planmeca Oy, Helsinki, Finland) with the base of the mandible parallel as well as tilted 45° to the horizontal plane. Computer measurements of the linear dimension were performed by three observers. The radiographic measurements were performed four times for each experimental setting. A total of 240 measurements were performed. Reproducibility was evaluated through comparison of standard deviation (SD) and estimation of intraclass correlation coefficient (ICC). The error was estimated as the absolute difference between the radiographic measurements and the mean manual calliper measurements.
Results:
The mean SD for the radiographic measurements was 0.36 mm for the horizontally positioned mandible and 0.48 mm for the inclined mandible. The ICC between examiners was 0.996 mm, between sessions was 0.990 mm and between CBCT measurements and calliper was 0.992 mm. The overall absolute mean measurement error was 0.40 mm (SD 0.39 mm). The percentage of errors that exceeded 1 mm was 6.7%.
Conclusion:
The results revealed high reliability of measurements performed on CBCT images independently from object position, examiner's experience and high reproducibility in repeated measurements settings.
