3D dento-maxillary osteolytic lesion and active contour segmentation pilot study in CBCT: semi-automatic vs manual...

Area of Science:

• Oral and Maxillofacial Radiology
• Medical Imaging Analysis
• Quantitative Dento-maxillary Imaging

Background:

• Accurate segmentation of dento-maxillary osteolytic lesions is crucial for diagnosis and treatment planning.
• Cone-beam computed tomography (CBCT) is widely used for visualizing these lesions.
• Manual segmentation, while precise, is time-consuming and operator-dependent.

Purpose of the Study:

• To assess the reliability of semi-automatic segmentation tools for dento-maxillary osteolytic lesions.
• To compare the accuracy of semi-automatic methods against manual segmentation in CBCT scans.
• To identify limitations and potential improvements for automated image analysis in dentistry.

Main Methods:

• Five CBCT scans with periapical radiolucencies were analyzed.
• Manual segmentation was performed by two clinicians.
• Three semi-automatic segmentation procedures were applied by four operators.
• Volume measurements and statistical analyses (ANOVA, t-tests) were conducted.

Main Results:

• Manual segmentation showed low variability (2.5-3.5% coefficient of variation) and no operator difference.
• Semi-automatic procedures exhibited higher dispersion, varying significantly between operators and cases.
• ANOVA confirmed significant operator effects on semi-automatic segmentation accuracy.

Conclusions:

• Region-based semi-automatic segmentation is reliable only for circumscribed osteolytic lesions.
• Current semi-automatic methods are limited by complex surface structures.
• Future research should explore hybrid methodologies combining manual and semi-automatic strengths for improved dento-maxillary image analysis.