Predictability of Craniofacial Skeletal Age with Geometric Morphometrics

Abstract

This study aims to estimate skeletal age of craniofacial shape obtained from cone beam computed tomography (CBCT)-defined facial and basicranial landmarks using geometric morphometrics method in a random sample of growing patients, and explore the correlation between craniofacial shape and skeletal age as determined from hand and wrist radiograph. Generalized Procrustes analysis (GPA) of craniofacial shape with estimation of centroid size was performed on CBCTs of 48 growing patients (mean age 11.7 ± 1.5 years). Greulich and Pyle method for skeletal age assessment were used for correlation with centroid size. Correlation among the variables relied on Pearson's coefficient and its 95% confidence interval was estimated. The model's R2 was calculated, (Cook's distances, Mahalanobis distances, leverage values, and studentized residuals) and multiple regression analysis performed using the Statistical Package for the Social Sciences (SPSS) version 22. Mean skeletal age was 11.9 ± 2.4 years. Centroid size (151.5 ± 7.2) was significantly correlated with chronological age (R = 0.616, 95% CI 0.355-0.789, p < 0.01) and skeletal age (R = 0.605, 95 % CI 0.331-0.794, p < 0.01). A new equation for determining craniofacial skeletal age was developed, using the centroid size of the craniofacial frame, gender, and the known chronological age. A CBCT may be used for skeletal age assessment without additional hand wrist radiograph.