Quantifying stenosis in renal arteriograms: a fuzzy syntactic analysis

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Cardiovascular Medicine

Background:

• Digital subtraction angiography (DSA) is crucial for visualizing renal artery stenosis.
• Automated quantification systems aim to improve accuracy and efficiency in stenosis assessment.
• Existing non-fuzzy systems struggle with inherent variability in image processing and reference diameter knowledge.

Purpose of the Study:

• To introduce fuzzy logic into a renal artery lesion quantification system.
• To improve the representation of reference diameter using fuzzy quantities.
• To compare the performance of a fuzzy logic-based system against a non-fuzzy system.

Main Methods:

• A two-step approach was employed, starting with non-fuzzy syntactic analysis.
• Fuzzy logic was integrated to represent the reference diameter and stenosis.
• Statistical comparison using a two-tailed paired t-test was performed on an index of variability.

Main Results:

• The fuzzy logic approach provided a description of stenosis in fuzzy quantities.
• The index of variability was significantly reduced using the fuzzy approach (t = 2.37; p = 0.025).
• The fuzzy system demonstrated statistically significant improvements over the non-fuzzy system.

Conclusions:

• Fuzzy logic integration significantly enhances the automatic quantification of renal artery lesions.
• The fuzzy approach effectively handles the uncertainty and variability in image analysis and reference diameter.
• This improved system offers a more robust and accurate method for assessing renal artery stenosis.