Quantitative evaluation of extracellular volume in patients with heart failure using dual-layer spectral detector CT: a feasibility study / 中华放射学杂志

ABSTRACT

Objective:To explore the feasibility in evaluating segmental extracellular volume (ECV) using dual-layer spectral detector CT in patients with heart failure with various percentages of delayed hyper-enhanced volume, using MRI as a reference.Methods:Twenty-seven patients with heart failure in Peking Union Medical College Hospital were prospectively enrolled in this study from July 2019 to January 2021. All the patients underwent both CT late iodine enhancement and MR late gadolinium enhancement imaging within a week. According to percentages of hyper-enhanced volume on cardiac MR image, myocardial segments were classified into 3 groups: 0-4% (group A), 5%-49% (group B) and 50%-100% (group C). ECV images were independently observed by 2 experienced radiologists. The correlation and agreement between CT-ECV and MRI-ECV for the whole myocardial assessment and for the segmental assessment were evaluated using Spearman correlation and Bland-Altman analysis. Inter-observer agreement was assessed using the ICC.Results:Of the 27 patients, 432 segments were analyzed, with 199 segments in group A, 151 segments in group B and 82 segments in group C. In the whole myocardial analysis, the CT-ECV and MRI-ECV were 32.25% (28.04%, 34.28%) and 32.27% (28.35%, 34.77%), respectively. A good correlation was found between CT-ECV and MRI-ECV ( r=0.969, P<0.001). Bland-Altman analysis between CT-ECV and MRI-ECV showed a small bias (0.4%), with 95% limits of agreement of -6.9% to 7.8%. as for the segmental myocardial assessment, the CT-ECV was 28.23% (26.18%, 29.92%) for group A, 33.29% (30.16%, 35.96%) for group B and 40.22% (34.06%, 46.70%) for group C. There was statistically significant difference in ECV among three groups (all P<0.001). Good correlations between CT-ECV and MRI-ECV were found in group A ( r=0.614, P<0.001), group B ( r=0.852, P<0.001) and group C ( r=0.953, P<0.001). Bland-Altman analysis between CT-ECV and MRI-ECV showed a small bias 2.0% (95% limits of agreement:-17.6%-21.6%) in group A, a small bias -0.4% (95% limits of agreement: -13.8%-12.9%) in group B and a small bias -1.8% (95% limits of agreement: -17.4%-13.9%) in group C. The ICC values between 2 observers were 0.877 and 0.945 on CT-ECV and MRI-ECV, respectively. Conclusions:The larger percentage of hyper-enhanced volume on cardiac MR image, the stronger correlation are between the CT-ECV and MRI-ECV. CT-ECV may be taken as an alternative tool for quantitative evaluation of myocardial tissue in patients with heart failure.