Related Experiment Video
Updated: May 31, 2026
05:49
Reliability of Artificial Intelligence-Based Cone Beam Computed Tomography Integration with Digital Dental Images
Published on: February 23, 2024
Comparison of linear and angular cephalometric measurements in CBCT using semi-automated software and artificial
Kaan Orhan1,2,3, Ismet Ersalıcı4, Nora Saif5,6,7
1Department of Dentomaxillofacial Radiology, Faculty of Dentistry, Ankara University, Ankara, Turkey.
BMC Oral Health
|May 29, 2026
Summary
This study evaluated the reliability of cephalometric measurements using cone-beam computed tomography (CBCT) software. While most measurements were consistent, artificial intelligence platforms showed lower repeatability for specific metrics, indicating areas for AI algorithm enhancement.
Area of Science:
- Dentistry
- Radiology
- Medical Imaging
Background:
- Assessing the reliability of cephalometric measurements is crucial for diagnostic accuracy.
- Cone-beam computed tomography (CBCT) is increasingly used for dental imaging.
- Evaluating semi-automated software and artificial intelligence (AI) platforms for cephalometry is essential.
Purpose of the Study:
- To assess the reliability of cephalometric measurements derived from CBCT scans.
- To compare the performance of two semi-automated software programs (InVivoDental, Romexis) and an AI-based platform (Diagnocat).
- To evaluate intraobserver and interobserver agreement for various cephalometric measurements.
Main Methods:
- A cross-sectional reliability study analyzing 29 CBCT scans.
- Eleven cephalometric measurements assessing vertical and anteroposterior relationships were performed.
- Measurements were taken using InVivoDental, Romexis, and an AI-based platform.
Main Results:
- High intraobserver reliability was observed across most measurements.
- Excellent interobserver agreement (ICC > 0.90) was found for InVivoDental and Romexis.
- Most measurements showed no significant differences between semi-automated software and AI, except for Wits Appraisal and U1-SN angle, which were lower with AI.
Conclusions:
- Cephalometric measurement reliability on CBCTs needs careful evaluation, especially with semi-automated and AI tools.
- Consistent intra-observer repeatability was noted for most measurements.
- AI-generated cephalograms showed reduced repeatability for certain metrics, highlighting potential areas for AI algorithm improvement.
