Validity of Moyers and Tanaka-Johnston Prediction Methods for Mixed Dentition Analysis in the Pediatric Population of Central Maharashtra: A Cross-Sectional Validation Study

Background Accurate prediction of the mesiodistal widths of unerupted permanent canines and premolars is essential for mixed-dentition space analysis and orthodontic planning. The Moyers' probability tables and the Tanaka-Johnston regression equations are widely used, yet their accuracy across different Indian subpopulations remains uncertain. Aim To assess and compare the validity of Moyers' and Tanaka-Johnston's prediction methods for estimating the combined mesiodistal widths of unerupted canines and premolars in children from the Central Maharashtra population. Methods This cross-sectional validation study included 40 children aged 6-11 years in mixed dentition. Mesiodistal tooth widths were measured directly on dental casts using a pointed divider and millimeter scale, while the arch perimeter was recorded using 0.5-mm stainless-steel ligature wire. Predicted values were obtained using Moyers' percentile tables (95th percentile for both sexes) and the Tanaka-Johnston equations. The primary outcome was the mean bias (Available - Predicted, mm). Secondary outcomes included the SD of bias, Bland-Altman limits of agreement, the proportion of predictions within ±1 mm and ±2 mm, and Cohen's dz. Statistical analyses used paired t-tests and Bland-Altman plots (p < 0.05). Results Both methods demonstrated close agreement with actual measurements at the group level. Moyers' method slightly underestimated mandibular widths by a mean difference of 1.18 mm and nearly matched maxillary values (-0.15 mm). Tanaka-Johnston, on the other hand, showed minimal bias in both arches (≤0.27 mm). Despite non-significant mean differences (all p > 0.05), Bland-Altman limits indicated substantial individual variability. Conclusion Moyers' and Tanaka-Johnston's analyses provide acceptable mean accuracy for Central Maharashtra children but limited reliability for individual predictions. Region-specific regression models are recommended to improve orthodontic diagnostic precision.