Evaluating coronary artery calcification with low-dose chest CT reconstructed by different kernels

Area of Science:

• Radiology
• Cardiovascular Imaging
• Medical Imaging Analysis

Background:

• Coronary artery calcification (CAC) is a key indicator of cardiovascular disease.
• Standard cardiac CT (CaCT) is effective but involves higher radiation doses.
• Low-dose chest CT (LDCT) is increasingly used for lung screening, raising interest in its potential for incidental CAC detection.

Purpose of the Study:

• To assess the reliability of LDCT for CAC detection and quantification.
• To compare the performance of different LDCT reconstruction kernels against CaCT.
• To evaluate image quality, radiation dose, and diagnostic accuracy.

Main Methods:

• Retrospective analysis of patients undergoing both CaCT and LDCT.
• LDCT images reconstructed using smooth, standard, and sharp kernels.
• Comparison of CAC scores (volume, mass, Agatston) and cardiovascular risk classification.
• Evaluation of image quality metrics (noise, SNR, CNR) and radiation dose.

Main Results:

• LDCT demonstrated high sensitivity (98.5%) for CAC detection across all kernels.
• Smooth kernel reconstruction (Group B1) showed the highest agreement with CaCT for CAC scores and risk classification (ICC > 0.90, Kappa = 0.757).
• Smooth kernels yielded improved image quality (lower noise, higher SNR/CNR) and a 41.2% lower radiation dose compared to CaCT.

Conclusions:

• LDCT with smooth kernel reconstruction is a reliable method for detecting and quantifying CAC.
• This approach may enable incidental CAC detection during lung screening, expanding LDCT's clinical utility.
• Smooth kernel reconstruction optimizes LDCT for cardiovascular risk assessment with reduced radiation exposure.