Associate variables of coronary computed tomographic angiography-derived extracellular volume fraction and cardiac magnetic resonance-derived late-gadolinium enhancement

Abstract

Abstract Background Coronary computed tomographic angiography (CCTA) with late-iodine enhancement enables the segmental quantification of extracellular volume (ECV) fraction which represents myocardial fibrosis. However, its association with cardiac magnetic resonance (CMR)-derived late gadolinium enhancement (LGE) has not been fully elucidated. Purpose We aimed to assess the concordance and discordance between CCTA-derived ECV and CMR-derived LGE and to investigate associated variables of each index. Methods This retrospective analysis included a total of 123 segments from 41 patients (anterior, inferior and lateral segments per patient) with known or suspected coronary artery disease (CAD) who underwent both clinically indicated CCTA and CMR. Patients with cardiomyopathy such as amyloidosis were excluded. Myocardial fibrosis was evaluated by CCTA-derived segmental ECV and by CMR-LGE with ischemic pattern on a per-segment basis (AHA 16 segments). Clinical and echocardiographic variables associated with CCTA-derived segmental ECV and CMR-LGE were studied by univariable linear and logistic regression analyses, respectively. Results Among the 123 segments, 21 segments showed ischemic CMR-LGE pattern. CCTA-ECV was significantly higher in segments with versus without CMR-LGE (42 [39, 47] % versus 34 [29, 39] %, P<0.01). Receiver-operating characteristic curve analysis showed that the best cut-off of CCTA-ECV to predict CMR-LGE was 38.9% (area under the curve 0.82 [95% confidence interval 0.74, 0.90], positive predictive value: 40.0%, negative predictive value: 96.2%). Univariable analyses revealed that male sex, prior history of revascularization, usage of statin, aspirin, angiotensin-converting enzyme inhibitor and beta-blockade, higher C-reactive protein, NT-pro BNP, lower left ventricular ejection fraction, greater left ventricular chamber size, shorter deceleration time, and larger segmental and global longitudinal strain were associated with both higher CCTA-ECV and the presence of CMR-LGE. Lower lipid profiles, higher high-sense troponin I, and greater left atrial volume index were associated with higher CCTA-ECV, while not with the presence of CMR-LGE. The frequency of CMR-LGE were significantly different according to CCTA-ECV tertiles. (P<0.05) Statin use and higher NT-proBNP were determinants of the highest ECV tertile without LGE. Conclusion We found a strong interrelationship between variables associated with ECV and the presence of LGE, although high ECV may not necessarily indicate the presence of LGE in patients with known or suspected CAD. Clinical factors associated with discordance, such as high-ECV without CMR-LGE, should be taken into account in the assessment of ECV and LGE. Further studies are needed to elucidate clinical significance of the concordance and discordance between ECV and LGE.