Age estimation from pubic symphysis based on cinematic volume rendering: comparison between Suchey-Brooks staging and...

Area of Science:

• Forensic Anthropology
• Medical Imaging
• Artificial Intelligence

Background:

• Accurate adult age estimation is crucial in forensic anthropology.
• Conventional methods like the Suchey-Brooks (SB) method have limitations.
• Advancements in imaging and AI offer new possibilities for age estimation.

Purpose of the Study:

• To compare the performance of the conventional Suchey-Brooks (SB) method with a deep learning (DL) approach for adult age estimation.
• To evaluate age estimation accuracy using standardized cinematic volume rendering (cVR) from pelvic CT scans.
• To analyze age-dependent biases and the focus of DL models in age estimation.

Main Methods:

• Analysis of 1,359 pelvic CT scans from a Chinese cohort.
• Application of the conventional Suchey-Brooks (SB) method with cubic regression.
• Implementation of a deep learning (DL) model for age estimation.
• Utilized gradient-weighted class activation mapping (Grad-CAM) for DL model visualization.

Main Results:

• SB method achieved Mean Absolute Errors (MAE) of 5.94 years (males) and 6.06 years (females).
• DL model produced MAEs of 6.64 years (males) and 7.03 years (females).
• No significant accuracy difference was found between SB and DL methods.
• Both methods showed age-dependent bias: overestimation in younger adults and underestimation in older individuals.
• DL model visualization revealed focus on pubic symphyseal surface morphology.

Conclusions:

• Cinematic volume rendering (cVR)-based SB phase assignment and deep learning regression achieve compatible precision for adult age estimation.
• Deep learning offers a scalable, automated, first-line approach for objective age estimation in forensic anthropology.
• Both methods demonstrate age-related biases that warrant further investigation.