Comparison of individualized facial growth prediction models using artificial intelligence and partial least squares based on the Mathews growth collection.

Abstract

Objectives: To develop facial growth prediction models using artificial intelligence (AI) under various conditions, and to compare performance of these models with each other as well as with the partial least squares (PLS) growth prediction model.

Materials and methods: Longitudinal lateral cephalograms from 33 subjects in the Mathews growth collection were utilized. A total of 1257 pairs of before and after growth lateral cephalograms were included. In each image, 46 hard and 32 soft tissue landmarks were manually identified. Growth prediction models were constructed using a deep learning method based on TabNet deep neural network and partial least squares (PLS) method. Prediction accuracies of the two methods were compared.

Results: On average, artificial intelligence (AI) showed 0.61 mm less prediction error than PLS. Among the 77 predicted landmarks, AI was more accurate than PLS in 60 landmarks. When comparing AI models with varying numbers of training epochs, those with higher epochs yielded more accurate predictions. Overall, PLS and AI exhibited greater prediction errors for soft tissue and mandibular landmarks compared to hard tissue and maxillary landmarks. However, AI showed a smaller increase in prediction error in areas with greater variability.

Conclusions: AI proved to be a valuable growth prediction method, with clinically acceptable prediction errors averaging 1.49 mm for 45 hard tissue landmarks and 1.71 mm for 32 soft tissue landmarks. PLS accurately predicted landmarks with low variability. However, AI generally outperformed PLS, particularly for landmarks in the lower part of the craniofacial structure and soft tissue, where uncertainty is considerable.