From Molars to Milestones: Predicting Growth Spurts via AI and Panoramic Imaging
Prosthesis · 2025-09-10 · 7 (5) · 116
DOI: 10.3390/prosthesis7050116 ↗Abstract
Background: A promising improvement in orthodontic diagnostics is the use of convolutional neural networks (CNNs) to predict skeletal maturity using dental radiographic data, namely the calcification phases of the lower canine. Objectives: The purpose of this study is to improve the prediction of skeletal growth maturation through the use of sophisticated deep learning techniques, particularly CNNs, in the analysis of orthopantomography developmental markers. Methods: The CNN was trained and validated using the suggested model, which enables precise assessment of skeletal maturity across data collected from patients undergoing orthodontic and dental evaluations. By using a multiclass classification framework to classify the various stages of skeletal maturation. Results: CNN model predicting the development of the lower canine from the second molar provided a test accuracy of 97.50%, the model made it possible to automatically interpret radiographic features that were previously evaluated manually. Conclusions: CNN models can be trained to correctly identify the lower canine calcification stage, this helped clinicians with treatment planning and timing, especially with regard to growth modification, implant, prosthodontics approach and its clinical applicability. It guarantees ethical imaging procedures and optimizes clinical workflows by doing away with the necessity for further imaging.
EUVIMED