Individualizing ANB angle based on Arab and German orthodontic cephalometric analysis: a cross-sectional multi-center...

Area of Science:

• Orthodontics
• Cephalometric Analysis
• Machine Learning in Healthcare

Background:

• The individualized ANB (ANB_indv.) is a standard method for classifying skeletal malocclusion.
• The traditional ANB_indv. equation, established in 1977, has limitations due to a small sample size and single-ethnic group basis.
• Enhancing the accuracy of ANB_indv. prediction is crucial for precise orthodontic diagnosis.

Purpose of the Study:

• To develop a novel, more accurate equation for predicting the individualized ANB (ANB_indv.) angle.
• To improve upon the traditional Panagiotidis and Witt equation using a larger, multi-ethnic sample.
• To leverage machine learning regression models for enhanced cephalometric analysis.

Main Methods:

• Collected lateral cephalogram images from 300 Arab and 623 German patients with skeletal class I occlusion.
• Applied machine-learning regression models to predict ANB_indv. using angles like SNA, ML-NSL, and SN-Pg.
• Validated the new equations through sensitivity analysis of Calculated_ANB (ANB-ANB_indv.).

Main Results:

• Individual ethnic group formulas showed varying correlation coefficients (R² Arab=0.59, R² German=0.73) compared to the traditional method (r=0.80).
• Inclusion of the SN-Pg angle significantly improved prediction accuracy (R² Arab=0.78, R² German=0.88).
• The combined multi-ethnic equation achieved a sensitivity of 89.85% for Calculated_ANB, outperforming individual ethnic equations.

Conclusions:

• A larger, multi-ethnic sample size significantly enhances the predictive capability for the ANB angle.
• The newly developed equation, incorporating SNA, ML-NSL, and SN-Pg angles, offers superior accuracy for individualized ANB classification.
• This research provides a more robust tool for diagnosing skeletal malocclusion across diverse populations.