Development of an automated 3D segmentation program for volume quantification of body fat distribution using CT

Area of Science:

• Medical Imaging
• Radiology
• Computational Anatomy

Background:

• Accurate quantification of body fat distribution is crucial for assessing obesity and related metabolic syndrome risks.
• Current methods for body fat segmentation on CT images can be time-consuming and subjective.
• Volumetric analysis of visceral and subcutaneous adipose tissue requires precise segmentation techniques.

Purpose of the Study:

• To develop a fully automated computing tool for segmenting body fat distributions on volumetric CT images.
• To enable quantitative evaluation of abdominal obesity and obesity-related metabolic syndrome.
• To enhance the efficiency and accuracy of body fat analysis using 3D visualization.

Main Methods:

• Development of an algorithm for automatic identification of body perimeter and visceral-subcutaneous fat contours.
• Model-based segmentation for extracting diaphragmatic surfaces to define the upper abdominal limit.
• Implementation of quantitative evaluation functions on a prototype 3D image processing workstation.
• Calculation of volumetric ratios of visceral to total fat and visceral to subcutaneous fat.

Main Results:

• Successful development of a fully automated segmentation tool for body fat on CT images.
• Accurate identification of body perimeter and visceral-subcutaneous fat layers.
• Quantitative calculation of visceral and subcutaneous fat volumes and their ratios.
• 3D surface display with color intensity mapping effectively visualizes fat distribution and obesity risk.

Conclusions:

• The developed computing tool provides an efficient and accurate method for full-automatic segmentation of body fat distributions.
• The tool facilitates quantitative evaluation of abdominal obesity and associated metabolic syndrome risks through 3D visualization.
• Preliminary results demonstrate the utility of the tool in medical checkups for improved obesity assessment.