Resolution and SNR effects on carotid plaque classification

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Cardiovascular Research

Background:

• Multicontrast-weighted MRI and automatic classification algorithms show promise for assessing plaque composition.
• A gap exists in understanding how imaging conditions affect segmentation accuracy.

Purpose of the Study:

• To investigate the impact of image resolution and signal-to-noise ratio (SNR) on the accuracy of automatic plaque classification.
• To determine the robustness of multicontrast MRI-based plaque analysis under varying imaging conditions.

Main Methods:

• Carotid endarterectomy samples (n=13) were imaged using proton density, T1, T2, and diffusion weightings at high resolution and SNR.
• Maximum likelihood (ML) algorithm classified plaque components using contrast weighting intensities.
• Image resolution and SNR were systematically degraded to assess classification accuracy.

Main Results:

• Classification accuracy was independent of in-plane resolution from 156 microm to 1250 microm.
• Accuracy was dependent on SNR, decreasing by less than 10% until SNR reached 25% of original values, then dropping sharply.
• Automatic classifiers demonstrated robustness across a range of imaging conditions.

Conclusions:

• Multicontrast MRI combined with automatic classification is a viable tool for plaque composition analysis.
• The method is robust to resolution variations and moderately robust to SNR degradation, supporting its use in standard in vivo carotid imaging.