Deep Learning-Based Automated Coronary Plaque Quantification with Ultra-high Resolution Photon-Counting Detector CT...

Area of Science:

• Medical Imaging
• Cardiovascular Imaging
• Photon-Counting Detector CT

Background:

• Coronary artery disease (CAD) diagnosis relies on accurate imaging.
• Photon-counting detector CT (PCD-CT) offers potential for improved spatial resolution.
• Assessing the impact of varying spatial resolutions on image quality and deep learning (DL) quantification is critical.

Purpose of the Study:

• To evaluate how different spatial resolutions from PCD-CT affect image quality.
• To assess the impact of spatial resolution on DL-based quantification of coronary stenosis and plaque volumes.
• To determine the optimal spatial resolution for reliable CAD assessment.

Main Methods:

• Prospective study of 70 patients with suspected CAD undergoing ultra-high resolution CT angiography (UHR-CCTA) on a PCD-CT system.
• CCTA datasets reconstructed at standard (0.6 mm), high (0.4 mm), and ultra-high (0.2 mm) resolutions.
• Subjective image quality assessed by radiologists; DL tool quantified stenosis and plaque volumes.

Main Results:

• Higher spatial resolution improved vessel sharpness and plaque visualization but increased perceived noise.
• DL-based stenosis quantification showed a decreasing trend with increasing resolution for calcified plaques (P=0.011).
• Median plaque volume significantly decreased with increasing spatial resolution (P<0.001).

Conclusions:

• Ultra-high resolution PCD-CT enhances image quality and reduces DL-based overestimation of stenosis and plaque volume in calcified lesions.
• Resolution-dependent variability in quantification necessitates careful consideration for CAD risk stratification.
• Superior spatial resolution of UHR PCD-CT aids in more accurate assessment of coronary lesions.