Head & Face Medicine (Aug 2024)

Sex determination through maxillary dental arch and skeletal base measurements using machine learning

  • Cristiano Miranda de Araujo,
  • Pedro Felipe de Jesus Freitas,
  • Aline Xavier Ferraz,
  • Isabella Christina Costa Quadras,
  • Bianca Simone Zeigelboim,
  • Sidnei Priolo Filho,
  • Svenja Beisel-Memmert,
  • Angela Graciela Deliga Schroder,
  • Elisa Souza Camargo,
  • Erika Calvano Küchler

DOI
https://doi.org/10.1186/s13005-024-00446-w
Journal volume & issue
Vol. 20, no. 1
pp. 1 – 10

Abstract

Read online

Abstract Background Cranial, facial, nasal, and maxillary widths have been shown to be significantly affected by the individual’s sex. The present study aims to use measurements of dental arch and maxillary skeletal base to determine sex, employing supervised machine learning. Materials and methods Maxillary and mandibular tomographic examinations from 100 patients were analyzed to investigate the inter-premolar width, inter-molar width, maxillary width, inter-pterygoid width, nasal cavity width, nostril width, and maxillary length, obtained through Cone Beam Computed Tomography scans. The following machine learning algorithms were used to build the predictive models: Logistic Regression, Gradient Boosting Classifier, K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Multi-Layer Perceptron Classifier (MLP), Decision Tree, and Random Forest Classifier. A 10-fold cross-validation approach was adopted to validate each model. Metrics such as area under the curve (AUC), accuracy, recall, precision, and F1 Score were calculated for each model, and Receiver Operating Characteristic (ROC) curves were constructed. Results Univariate analysis showed statistical significance (p < 0.10) for all skeletal and dental variables. Nostril width showed greater importance in two models, while Inter-molar width stood out among dental measurements. The models achieved accuracy values ranging from 0.75 to 0.85 on the test data. Logistic Regression, Random Forest, Decision Tree, and SVM models had the highest AUC values, with SVM showing the smallest disparity between cross-validation and test data for accuracy metrics. Conclusion Transverse dental arch and maxillary skeletal base measurements exhibited strong predictive capability, achieving high accuracy with machine learning methods. Among the evaluated models, the SVM algorithm exhibited the best performance. This indicates potential usefulness in forensic sex determination.

Keywords