Correlation coefficients for predicting canine diameters from premolar and molar sizes

Abstract

Background/purpose: The permanent canine is the most widely used tooth for sex estimation because it is the tooth with the highest degree of sexual dimorphism; however, there are several factors that can limit the analysis (e.g., pathologies, postmortem loss). The aim of this work was to analyse the correlation between the dimensions of the canines and those of the premolars and molars, and evaluate the correspondence of real and predicted canine dimensions by applying the equations developed. Materials and methods: The sample was composed of digital models of 80 adult individuals from Temuco, Chile. The buccolingual and mesiodistal diameters of the canine, premolar and molar dental crowns were measured and Pearson's linear regression analysis was performed in order to determine the correlation between the diameters. The equations obtained from the reference subsample were applied to a validation subsample to predict canine dimensions from the dimensions of the postcanine teeth. Results: Four regression equations were obtained, all for prediction of the size of the lower canines, whose correlation coefficient ranged from 0.701 to 0.738. The regression equations developed with the reference sample were tested on the validation sample using a Student's t-test for paired samples and the intraclass correlation coefficient. The differences between actual dental size and that predicted by the equations were not significantly different, and concordance analysis showed a moderate degree (0.485-0.585). Conclusion: There is a limited correlation of canine dimensions with respect to premolars and molars. The correspondence between the actual and predicted canines dimensions is moderate.