A novel region-growing based semi-automatic segmentation protocol for three-dimensional condylar reconstruction using...

Area of Science:

• Dentistry
• Medical Imaging
• Biomedical Engineering

Background:

• Accurate 3D reconstruction of mandibular condyles is crucial for diagnosing and treating temporomandibular joint disorders.
• Current segmentation methods can be time-consuming and subjective, potentially affecting the accuracy of 3D models.

Purpose of the Study:

• To introduce and validate a novel semi-automatic segmentation protocol for precise 3D reconstruction of mandibular condyles.
• To assess the accuracy and reproducibility of this protocol using cone beam computed tomography (CBCT) data.

Main Methods:

• The protocol combines 3D region-growing and local thresholding algorithms.
• Twenty mandibular condyles from ten CBCT datasets were segmented by two independent observers.
• Accuracy and reproducibility were evaluated using Dice-coefficient and distance map calculations.

Main Results:

• The mean inter-observer Dice-coefficient was 0.98, indicating high agreement between observers.
• An average 90th percentile distance of 0.32 mm demonstrated excellent inter-observer similarity.
• No systematic errors were detected, confirming the protocol's reliability.

Conclusions:

• The developed semi-automated protocol is a highly accurate and reproducible tool for segmenting and 3D rendering mandibular condyles.
• This method enables CBCT to serve as a quantitative imaging modality for analyzing condylar morphology in clinical practice.