Functional Region Annotation of Liver CT Image Based on Vascular Tree.

Area of Science:

• Medical Imaging
• Radiology
• Computational Anatomy

Background:

• Accurate liver region annotation is crucial for diagnosing and treating liver diseases.
• Current methods face challenges in functionally segmenting the liver from Computed Tomography (CT) images.

Purpose of the Study:

• To propose a vessel-tree-based method for precise liver annotation in CT scans.
• To develop a system that automatically divides the liver into eight segments based on Couinaud classification.

Main Methods:

• Extraction of the liver region, including vessels and tumors, from CT scans.
• Application of a 3D thinning algorithm to derive the spatial skeleton and geometric structure of liver vessels.
• Formulation of portal vein topology using a directed acyclic graph with geometrical attributes.
• Construction of a hierarchical vascular tree for liver segmentation.

Main Results:

• The proposed method successfully extracts liver structures and generates a topological graph of the portal vein system.
• The hierarchical vascular tree effectively divides the liver into eight functional segments according to Couinaud classification.
• Experimental results confirm the method's effectiveness in precise liver annotation.

Conclusions:

• The vessel-tree-based liver annotation method provides accurate functional region identification in CT images.
• This approach supports improved diagnosis and treatment planning for liver diseases.
• The method demonstrates significant potential for clinical application in liver imaging analysis.