EUVIMEDEUVIMED entdecken
Crossref

Quantifying cross-site effect in AI-based dental age estimation: evidence from Brazilian panoramic radiographs

Willian Oliveira, Matheus L Oliveira iD, Francisco Haiter Neto iD, Mariana Albuquerque Santos iD, Maria Luiza Dos Anjos Pontual iD, Ricardo V Beltrão, Andrea A Pontual iD, Flávia Maria M Ramos-Perez iD, Deborah Queiroz Freitas iD, Cleber Zanchettin iD

Dentomaxillofacial Radiology · 01.07.2026 · 55 (5) · 477-486

DOI: 10.1093/dmfr/twag014 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Dentomaxillofacial Radiology, online erschienen am 19.02.2026. Verfasst von Willian Oliveira und 9 weiteren Autor:innen. Beteiligte Einrichtung: Centro de Informática, Universidade Federal de Pernambuco, Recife, Pernambuco 50740-560, Brazil. Hauptthemen laut MEDLINE: Radiography, Panoramic, Age Determination by Teeth, Artificial Intelligence. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 18 Literaturangaben. Gefördert von: FACEPE and FAPESP.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Objectives

To quantify cross-regional generalization of dental age-estimation models, identify practical strategies to improve their performance, and report uncertainty in a transparent manner.

Methods

A total of 21 722 panoramic radiographs from 2 Brazilian regions were acquired using distinct equipment. A robust Inception-v4 model was evaluated under 4 scenarios: (1) training on Northeast data and testing on Southeast data; (2) fine-tuning using Southeast data only or both regions; (3) training from scratch on pooled data; (4) pooled training with augmentation. Model performance was assessed using mean absolute error (MAE), mean signed error (bias), R2, Bland-Altman analysis, and calibration metrics.

Results

A marked performance drop was observed when the model trained on Northeast data was applied to Southeast radiographs (MAE 4.97 years vs 3.10 years in-region), with negative bias and wider Bland-Altman limits at older ages. Training with pooled regions and modest fine-tuning improved accuracy and calibration across both cohorts (MAE 3.24-3.69; R2 0.93-0.95). Data augmentation yielded only small additional improvements and did not eliminate large residual errors. Heatmaps highlighted clinically relevant anatomical structures commonly used by dental experts for age estimation.

Conclusions

Cross-site and domain shifts significantly impact the performance of AI models for dental age estimation. Multi-regional training combined with light model adaptation provides robust, well-calibrated, and interpretable results across regions, whereas data augmentation alone has limited effectiveness. This study offers a 2-region benchmark, code, and data-access protocols to support reproducible evaluation and guide clinical deployment.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Radiography, PanoramicAge Determination by TeethArtificial IntelligenceHumansBrazilChild

Stichwörter der Autor:innen

age estimationdeep neural networkforensic sciencesradiological methods

Autor:innen und Institutionen

  • Centro de Informática, Universidade Federal de Pernambuco, Recife, Pernambuco 50740-560, Brazil.
  • Faculdade de Odontologia de Piracicaba, Universidade Estadual de Campinas, Piracicaba, Sao Paulo 13414-903, Brazil.
  • Departamento de Clínica e Odontologia Preventiva, Universidade Federal de Pernambuco, Recife, Pernambuco 50670-901, Brazil.
  • Department of Clinical and Social Dentistry, Centro de Ciências da Saúde, Universidade Federal da Paraíba, Jardim Universitário, Joao Pessoa, Paraiba 58051-900, Brazil.

ORCID-Profile: Matheus L Oliveira, Francisco Haiter Neto, Mariana Albuquerque Santos, Maria Luiza Dos Anjos Pontual, Andrea A Pontual, Flávia Maria M Ramos-Perez, Deborah Queiroz Freitas, Cleber Zanchettin

Förderung

  • FACEPE and FAPESP

Zitieren

APA

Oliveira, W., Oliveira, M. L., Haiter Neto, F., Albuquerque Santos, M., Dos Anjos Pontual, M. L., Beltrão, R. V., Pontual, A. A., Ramos-Perez, F. M. M., Freitas, D. Q., & Zanchettin, C. (2026). Quantifying cross-site effect in AI-based dental age estimation: evidence from Brazilian panoramic radiographs. Dentomaxillofacial Radiology, 55(5), 477-486. https://doi.org/10.1093/dmfr/twag014

Vancouver

Oliveira W, Oliveira ML, Haiter Neto F, Albuquerque Santos M, Dos Anjos Pontual ML, Beltrão RV, et al. Quantifying cross-site effect in AI-based dental age estimation: evidence from Brazilian panoramic radiographs. Dentomaxillofacial Radiology. 2026;55(5):477-486. doi:10.1093/dmfr/twag014

BibTeX
@article{Oliveira2026,
  author = {Oliveira, Willian and Oliveira, Matheus L and Haiter Neto, Francisco and Albuquerque Santos, Mariana and Dos Anjos Pontual, Maria Luiza and Beltrão, Ricardo V and Pontual, Andrea A and Ramos-Perez, Flávia Maria M and Freitas, Deborah Queiroz and Zanchettin, Cleber},
  title = {Quantifying cross-site effect in AI-based dental age estimation: evidence from Brazilian panoramic radiographs},
  journal = {Dentomaxillofacial Radiology},
  year = {2026},
  volume = {55},
  number = {5},
  pages = {477--486},
  doi = {10.1093/dmfr/twag014}
}
Kostenlos starten →