Scholarlink

Background <p>Accurate age estimation is essential in forensic identification. While traditional skeletal methods are often limited by preservation status and individual variability, teeth offer greater taphonomic stability. The first molar undergoes continuous age-related pulp chamber reduction due to secondary dentin deposition, making it a valuable target for adult age estimation. This study aimed to develop and compare curvilinear regression and machine learning models for age estimation based on three-dimensional pulp chamber volume (PCV) of first molars in a Han Chinese population.</p> Results <p>A retrospective analysis was performed on 1,857 right first molars (maxillary tooth 16 and mandibular tooth 46, FDI notation) from Han Chinese adults aged 18–65 years using cone-beam computed tomography (CBCT) in a single-center setting. PCV showed significant differences by sex and maxillary/mandibular position. Strong negative correlations were observed between PCV and age (r = –0.88 to –0.81), with the strongest correlation in female maxillary first molars (r = –0.88). The best cubic regression model (female, maxillary) achieved a mean absolute error (MAE) of 4.95 years. Machine learning models demonstrated superior performance. The sex- and position-specific XGBoost model for female maxillary molars achieved an MAE of 3.14 years (95% CI: 2.92–3.37) and R² = 0.87 (95% CI: 0.84–0.89), representing a 36.5% error reduction compared to the best regression model.</p> Conclusions <p>The combination of first molar pulp chamber volume with machine learning, particularly sex- and position-specific XGBoost models, provides a precise and reproducible method for adult age estimation in forensic practice. This approach demonstrates substantial improvements over traditional regression models and offers a promising, high-precision tool for adult age estimation in forensic practice, particularly when sex- and tooth position-specific data are available.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Adult age estimation from first molar pulp chamber volume in a Han Chinese population using a machine learning approach

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月22日

查看全文
  • Yanjie Ding,
  • Tao Zhong,
  • Yuxin He,
  • Wei Wang,
  • Shilin Zhang,
  • Xiao Zhang,
  • Wenli Shi,
  • Bo Jin

摘要

Background

Accurate age estimation is essential in forensic identification. While traditional skeletal methods are often limited by preservation status and individual variability, teeth offer greater taphonomic stability. The first molar undergoes continuous age-related pulp chamber reduction due to secondary dentin deposition, making it a valuable target for adult age estimation. This study aimed to develop and compare curvilinear regression and machine learning models for age estimation based on three-dimensional pulp chamber volume (PCV) of first molars in a Han Chinese population.

Results

A retrospective analysis was performed on 1,857 right first molars (maxillary tooth 16 and mandibular tooth 46, FDI notation) from Han Chinese adults aged 18–65 years using cone-beam computed tomography (CBCT) in a single-center setting. PCV showed significant differences by sex and maxillary/mandibular position. Strong negative correlations were observed between PCV and age (r = –0.88 to –0.81), with the strongest correlation in female maxillary first molars (r = –0.88). The best cubic regression model (female, maxillary) achieved a mean absolute error (MAE) of 4.95 years. Machine learning models demonstrated superior performance. The sex- and position-specific XGBoost model for female maxillary molars achieved an MAE of 3.14 years (95% CI: 2.92–3.37) and R² = 0.87 (95% CI: 0.84–0.89), representing a 36.5% error reduction compared to the best regression model.

Conclusions

The combination of first molar pulp chamber volume with machine learning, particularly sex- and position-specific XGBoost models, provides a precise and reproducible method for adult age estimation in forensic practice. This approach demonstrates substantial improvements over traditional regression models and offers a promising, high-precision tool for adult age estimation in forensic practice, particularly when sex- and tooth position-specific data are available.