A Statistical Procedure for Exploring a Skeletal Age-Explicative Tool for Growing Patients

Background: Skeletal age estimation plays a fundamental role in orthopedic treatments. Since the most reliable methods are based on ionizing radiation, this study aimed to use machine learning techniques to explore a skeletal age assessment method not based on additional radiographies. Methods: Patients aged between 6 and 16 years old whose clinical records included orthopantomography, radiographs of the second phalanx of the third finger, and biometric data were enrolled for the study. The radiographs were analyzed to estimate the maturation degree of the left lower first premolars, the midpalatal suture, and the second phalanx of the third finger. Both an explicative data analysis and a multivariate analysis were performed. Results: The sample comprised 111 subjects. The multivariate analysis revealed an explanatory role for sex (p < 0.01) and chronological age (p < 0.01). The ordinal tool showed how the use of height (p = 0.02) and weight (p = 0.03) was explicative of skeletal age against a loss of statistical significance corresponding to the use of body mass index (p = 0.6). The median palatine suture (p = 0.01) was explicative. Conclusions: The combined evaluation of weight, height, sex, chronological age, and grade of maturation of the midpalate suture provides an explicative tool for assessing skeletal age without additional radiographic exams, besides a routine orthopantomography.