Artificial Intelligence Methods in Cephalometric Image Analysis-A Systematic Narrative Review

Area of Science:

• Orthodontics
• Medical Imaging
• Artificial Intelligence

Background:

• Information technology advancements, including AI, are crucial for orthodontic diagnostics.
• Cephalometric images are vital for orthodontic treatment planning, providing data on craniofacial growth.
• Machine learning and automation are increasingly applied to analyze these images.

Purpose of the Study:

• To review current AI applications in cephalometric image analysis.
• Focus on studies published between 2020-2025 from Scopus and Web of Science databases.

Main Methods:

• Systematic review of 20 key studies.
• Analysis of AI models including CNN, YOLO, BCNN, ANN, and others.
• Evaluation of performance in landmark detection and classification tasks.

Main Results:

• AI landmark detection errors are 1-2 mm, within clinical limits.
• CNN and YOLO models match orthodontist accuracy; BCNN provides uncertainty estimates.
• ANN models achieve 95% accuracy for cervical vertebral maturity (CVM); screening tools reach 96.3% agreement.

Conclusions:

• AI tools demonstrate clinically acceptable accuracy in cephalometric analysis.
• AI improves repeatability and reduces analysis time in semi-automated workflows.
• AI shows potential as a reliable tool for CVM and surgical eligibility assessment, pending broader validation.