Determination of growth and developmental stages in hand–wrist radiographs: Can fractal analysis in combination with artificial intelligence be used? Ermittlung von Wachstums- und Entwicklungsstadien in Handwurzel-Röntgenaufnahmen: Kann die Fraktalanalyse in Kombination mit künstlicher Intelligenz eingesetzt werden?


Gonca M., Sert M. F., Gunacar D. N., Kose T. E., Beser B.

Journal of Orofacial Orthopedics, 2024 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Publication Date: 2024
  • Doi Number: 10.1007/s00056-023-00510-1
  • Journal Name: Journal of Orofacial Orthopedics
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, MEDLINE
  • Keywords: Artificial intelligence, Fractal analysis, Machine algorithms, Machine learning, Orthodontics
  • Recep Tayyip Erdoğan University Affiliated: Yes

Abstract

Purpose: The goal of this work was to assess the classification of maturation stage using artificial intelligence (AI) classifiers. Methods: Hand–wrist radiographs (HWRs) from 1067 individuals aged between 7 and 18 years were included. Fifteen regions of interest were selected for fractal dimension (FD) analysis. Five predictive models with different inputs were created (model 1: only FD; model 2: FD and Chapman sesamoid stage; model 3: FD, age, and sex; model 4: FD, Chapman sesamoid stage, age, and sex; model 5: Chapman sesamoid stage, age, and sex). The target diagnoses were accelerating growth velocity, very high growth velocity, and decreasing growth velocity. Four AI algorithms were applied: multilayer perceptron (MLP), support vector machine (SVM), gradient boosting machine (GBM) and C 5.0 decision tree classifier. Results: All AI algorithms except for C 5.0 yielded similar overall predictive accuracies for the five models. In order from lowest to highest, the predictive accuracies of the models were as follows: model 1 < model 3 < model 2 < model 5 < model 4. The highest overall F1 score, which was used instead of accuracy especially for models with unbalanced data, was obtained for models 1, 2, and 3 based on SVM, for model 4 based on MLP, and for model 5 based on C 5.0. Adding Chapman sesamoid stage, chronologic age, and sex as additional inputs to the FD values significantly increased the F1 score. Conclusion: Applying FD analysis to HWRs is not sufficient to predict maturation stage in growing patients but can be considered a growth rate prediction method if combined with the Chapman sesamoid stage, age, and sex.