Quantitative analysis of parameters of three-dimensional coronary angiography and coronary artery calculation correlation analysis of myocardial fractional blood flow reserve measured by computed tomography

Abstract

<strong>Objective</strong> This study evaluates the feasibility of combining parameters derived from three-dimensional quantitative coronary angiography (3D-QCA) with fractional flow reserve (CT-FFR) calculated by computed tomography based on an artificial intelligence deep learning algorithm and to further assess the relevance of coronary artery stenosis lesions and myocardial functional lesions. <br/> <br/> <strong>Methods</strong> This study retrospectively included 595 patients suspected of coronary heart disease who received coronary computed tomography angiography (CTA) or invasive coronary angiography (no coronary artery therapy has been performed in the past, coronary CTA examination was completed after admission followed by coronary angiography, and the interval was no more than 14 days) during hospitalization from December 2017 to December 2020. From these patients, a total of 1004 diseased blood vessels have been studied. Subsequent 3D-QCA was completed, followed by CT-FFR calculations, and the correlation analysis of the relevant parameters was carried out. According to the CT-FFR value, the patients were divided into non-functional lesion group with CT-FFR &lt; 0.80 (n = 633) and functional lesion group with CT-FFR≤0.80 (n = 371). The diseased vessels included 443 anterior descending (LAD), 283 left circumflex (LCX), and 278 right coronary (RCA) vessels. ROC curves were drawn to analyze the sensitivity and specificity of 3D-QCA parameters for predicting non-functional lesions. <br/> <br/> <strong>Outcome</strong> Statistical analysis confirms that the differences in coronary artery diameter stenosis rate [(28.83±13.25)% and (39.29±15.39)%, P&lt;0.001], cross-sectional area stenosis rate [(47.75±17.64)% and (60.89±17.12)%, P&lt;0.001 ], volume stenosis rate [(22.20±8.85) % and (29.73±8.66) %, P＜0.001] of non-functional lesion group and functional lesion group were significant. While above parameters was found to be negatively correlated with CT-FFR values (the correlation coefficients were r=-0.359, P&lt;0.001, r=-0.361, P&lt;0.001, r=-0.382, P&lt;0.001), coronary artery stenosis diameter [( 1.75 ± 0.50) mm vs ( 0.83 ± 0.39) mm, P＜0.001], stenosis cross-sectional area [(2.59 ± 1.45) mm^2 vs (1.58 ± 1.11) mm^2, P &lt;0.001], plaque volume [(39.95±26.24) mm^3 vs. (45.86±27.73) mm^3, P&lt;0.001] were found to be positively correlated with the CT-FFR value with statistical significance (the correlation coefficients were r = 0.433, P &lt; 0.001, r = 0.413, P &lt; 0.001, r = -0.076, P = 0.016). The stenosis diameter greater than 1.455mm predicted non-functional ischemia with a sensitivity of 71.7% and a specificity of 64.2%. The stenosis cross-sectional area greater than 1.585 mm^2 predicted nonfunctional ischemia with a sensitivity of 73.5% and a specificity of 60.1% <br/> <br/> <strong>Conclusions</strong> There is a specific correlation between the diameter stenosis rate, area stenosis rate, and volume stenosis rate of coronary artery in the relevant parameters of quantitative analysis of three-dimensional coronary angiography and CTFFR value. However, the sensitivity and specificity of predicting whether the lesions are myocardial ischemia-related lesions are not high. Besides, since CT-FFR has certain characteristics in the functional evaluation of related diseased blood vessels, we recommended popularizing and promoting CT-FFR measurement technology for patients with suspected coronary heart disease before coronary angiography.