Artificial Intelligence-Assisted Cephalometric Analysis: A Comparative Study Between WebCeph and NemoCeph

Area of Science:

• Orthodontics and Dental Imaging
• Artificial Intelligence in Healthcare
• Medical Data Analysis

Background:

• Artificial intelligence (AI) is increasingly used in orthodontics for automated cephalometric analysis, enhancing efficiency and reducing human error.
• Validation of AI-driven systems against established digital analysis software is crucial for clinical adoption.
• NemoCeph is considered a reference standard for manual cephalometric tracing.

Purpose of the Study:

• To evaluate the accuracy and consistency of AI-assisted cephalometric analysis using WebCeph.
• To compare WebCeph's AI-driven measurements with manual tracings performed using NemoCeph, the reference standard.

Main Methods:

• Cephalometric radiographs from 50 adult patients were analyzed.
• Linear and angular parameters were measured using both NemoCeph (manual) and WebCeph (AI-assisted).
• Statistical comparisons, including Shapiro-Wilk test for normality and paired t-tests for discrepancies, were performed.

Main Results:

• WebCeph demonstrated dependable outcomes for linear dental evaluations.
• Significant discrepancies were observed between WebCeph and NemoCeph for angular parameters (e.g., sella, gonial, SNA, SNB, ANB) and vertical skeletal dimensions.
• Differences were also significant in determining facial typology, with WebCeph sometimes overestimating or underestimating values.

Conclusions:

• AI-based cephalometric tools like WebCeph provide reliable linear dental measurements but show marked discrepancies in angular and vertical skeletal assessments.
• Clinical interpretation of AI-driven cephalometric data requires caution.
• AI tools should be considered supplementary to, rather than replacements for, expert orthodontic assessments.