The objective of this study was to assess the potential of grating-based phase-contrast computed tomography (gb-PCCT) for the detection and characterization of human coronary artery disease in an experimental ex vivo validation study.
Materials and Methods
The study was approved by the institutional review board, and informed consent was obtained from all patients. Specimens were examined using a conventional low-coherence x-ray tube (40 kV) and a Talbot-Lau grating interferometer. Histopathologic assessment was used as the standard of reference. Signal characteristics of calcified, fibrous (FIB), and lipid-rich (LIP) tissue were visually and quantitatively assessed by phase-contrast Hounsfield units (HU). Conventional absorption-based HU values were also measured. Conservative measurements of diagnostic accuracy for the detection and differentiation of plaque components as well as quantitative measurements of vessel dimensions were obtained, and receiver operating characteristic curve analysis for plaque differentiation was performed.
A total of 15 coronary arteries from 5 subjects were available for analysis (386 sections). Calcified, FIB, and LIP displayed distinct gb-PCCT signal criteria. The diagnostic accuracy of gb-PCCT was high with sensitivity, specificity, and negative and positive predictive values of 0.89 or greater for all plaque components with good interrater agreement (к ≥ 0.88). In addition, quantitative measurements of vessel dimensions in gb-PCCT were strongly correlated with measurements obtained from histopathology (Pearson R ≥ 0.86). Finally, phase-contrast Hounsfield units were superior to conventional HU in differentiating FIB and LIP (receiver operating characteristic analysis, 0.86 vs. 0.77, respectively; P < 0.05).
In an ex vivo setting, gb-PCCT provides improved differentiation and quantification of coronary atherosclerotic plaque and may thus serve as a tool for nondestructive histopathology.