Development And Validation of a Regional Regression Model for Predicting Unerupted Tooth Size in Mixed Dentition: A South Asian Perspective

Abstract

Background: Accurate prediction of unerupted tooth size is critical for orthodontic diagnosis in mixed dentition. While widely used, Moyers and Tanaka–Johnston methods are based on Caucasian populations and may not be reliable for South Asian children.
Objective: To evaluate the accuracy of a new regression model with tried-and-true techniques for forecasting the mesiodistal widths of South Asian children's unerupted canines and premolars.
Methods: 150 children of South Asian (Marathi and Telugu) ages 10 to 16 had their dental casts examined (75 males and 75 females). Digital callipers were used to measure the mesiodistal widths of the premolars, canines, and mandibular incisors. A new regression formula was created and contrasted with Tanaka-Johnston and Moyers' approaches. Bland-Altman plots, Pearson's correlation, and paired t-tests were used to evaluate predictive accuracy.
Results: The new regression model showed a statistically significant improvement in predictive accuracy (RMSE: 0.61 mm) over Moyers (RMSE: 0.87 mm) and Tanaka–Johnston (RMSE: 0.92 mm). The new model was more reliable in both males and females across ethnic subgroups.
Conclusion: The newly developed regression model provides a more accurate and population-specific method for mixed dentition analysis in South Asian children. It offers a clinically relevant alternative to traditional methods.