Quantification of coronary arterial stenoses in CTA using fuzzy distance transform

Area of Science:

• Medical Imaging
• Cardiovascular Disease Diagnosis
• Image Processing

Background:

• Coronary arterial stenosis quantification is crucial for diagnosing coronary heart diseases.
• Computed tomographic angiography (CTA) is a common, noninvasive imaging modality for monitoring disease.
• Existing methods for stenosis quantification often involve binarization, which can introduce errors.

Purpose of the Study:

• To present a novel method for detecting and quantifying coronary arterial stenosis in CTA.
• To improve the accuracy of stenosis quantification by eliminating the need for binarization.
• To enhance the analysis of coronary arterial stenosis using fuzzy distance transform (FDT) and coherence analysis.

Main Methods:

• Developed a new method utilizing fuzzy distance transform (FDT) for detecting and quantifying coronary arterial stenosis in CTA.
• Employed a new skeletal pruning algorithm to improve medial axis computation for enhanced accuracy.
• Introduced a coherence analysis of observed diameters to estimate the expected local diameter along arterial branches.

Main Results:

• The FDT-based method achieved significantly higher accuracy on a phantom dataset (0.53% error) compared to binary distance transform (2.11%) and conventional binary (3.71%) methods.
• In clinical CTA datasets, the FDT-based method showed high agreement with expert assessments (kappa coefficient = 87.9%), outperforming other methods (75.2% and 69.5%).
• The method successfully detects and quantifies stenosis severity by analyzing FDT values along the arterial tree's skeleton.

Conclusions:

• The proposed FDT-based method offers a more accurate and reliable approach for quantifying coronary arterial stenosis in CTA.
• This novel technique demonstrates superior performance compared to existing binary methods, showing high concordance with expert evaluation.
• The method holds significant potential for improving the diagnosis and management of coronary heart diseases.