Two-dimensional echocardiographic image texture analysis: reduction of regional variability using polar coordinates

Area of Science:

• Medical Imaging
• Ultrasound Technology
• Image Analysis

Background:

• Quantitative ultrasound (QUS) texture analysis is sensitive to imaging system artifacts.
• Variations in texture measures can obscure true tissue structural changes.
• Standard image acquisition and analysis methods may introduce biases.

Purpose of the Study:

• To evaluate the impact of image data acquisition and analysis methods on quantitative texture variations.
• To determine if polar coordinate analysis can mitigate position-dependent texture measure differences.
• To assess the effectiveness of different coordinate systems in reducing regional variations in QUS texture.

Main Methods:

• Analysis of texture measures from regions of interest (ROIs) within a uniform tissue-equivalent phantom.
• Comparison of conventional scan-converted (rectangular) data analysis with pseudo-polar and true polar coordinate analysis.
• Statistical assessment of significant differences in texture measures based on ROI position (range and azimuth).

Main Results:

• Conventional rectangular coordinate analysis revealed significant texture measure differences in over 50% of cases due to ROI position.
• Pseudo-polar analysis reduced texture measure variations by 80% (p < 0.01).
• True polar coordinate analysis reduced variations by 74% (p < 0.01) and completely eliminated azimuth-dependent differences.

Conclusions:

• Image acquisition and analysis methods significantly influence QUS texture measures.
• Polar coordinate analysis, particularly true polar, substantially reduces position-dependent artifacts.
• Implementing polar coordinate analysis for tissue texture data can enhance the definitive identification of abnormal tissues.