Extraction of pulmonary Trachea by dynamic tubular edge contour algorithm

Area of Science:

• Medical imaging
• Computer vision
• Pulmonary medicine

Background:

• Accurate extraction of the pulmonary trachea from lung CT images is a significant challenge in computer vision and medical imaging.
• Existing bronchial extraction methods suffer from bronchial loss and leakage, failing to extract higher-level structures needed for guiding lung procedures.

Purpose of the Study:

• To develop a novel algorithm for precise pulmonary trachea extraction from CT images.
• To overcome limitations of current methods, specifically bronchial loss and leakage.
• To provide a tool that can guide lung puncture procedures effectively.

Main Methods:

• A dynamic tubular edge contour algorithm is proposed, leveraging the tubular structure of the pulmonary trachea in CT scans.
• The algorithm extracts skeletal lines, performs elliptical fitting, and obtains 3D point cloud data in axial, coronal, and sagittal views.
• Point cloud data is fused to generate a complete 3D pulmonary trachea model.

Main Results:

• The algorithm successfully extracts the pulmonary trachea to the 10-11 level, as validated on the "EXACT09" dataset.
• It effectively resolves the issues of tracheal leakage and loss encountered with previous methods.
• A complete 3D model of the pulmonary trachea was obtained.

Conclusions:

• A novel pulmonary trachea extraction algorithm has been developed with significant theoretical and practical implications.
• The algorithm provides guidance for selecting puncture paths, helping to avoid major tracheal structures.
• This advancement is crucial for reducing puncture complications and lowering mortality rates in lung procedures.