Accuracy of 4D cardiac magnetic resonance for flow measurement

Background 4D Flow CMR has emerged as an innovative imaging modality providing flow assessment at any selected region of interest. The wide range of dynamic properties and velocity of cardiac flows are a commonly admitted technical limit. We aim to assess if this technique is robust enough to validate the conservation of mass principle applied to vessels and valves of the whole heart. Our working hypothesis was that 4D CMR would prove accurate compared to 2D CMR. Methods Thirty healthy patients underwent both 2D CMR and 4D CMR using a 1.5-T machine. Ventricular volumes were assessed and cardiac output was calculated and used as a reference value. We mapped using 4D Flow analysis (Arterys), cardiac flows from the vena cava to the aorta. The 4D flows through the cardiac valves, and vessels were independently measured by two readers. Agreement between flows measured using 4D CMR and CMR was assessed by linear regression analysis and Bland-Altman plots. Results Mean net flows were 5.5 ± 1.2 L/min through venae cavae, 5.4 ± 1.1 L/min through the tricuspid valve, 5.5 ± 1.1 L/min through the pulmonary valve, 5.4 ± 1.4 L/min through the pulmonary artery, 5.3 ± 1.1 L/min through the pulmonary veins, 5.5 ± 1.0 L/min through the mitral valve, 5.5 ± 1.4 L/min through the aortic valve and 5.5 ± 1.4 L/min through the aorta compared to 5.6 ± 1.1 with 2D CMR. Correlation coefficients between 4D flows measurements and 2D CMR ranged from 0.60 to 0.86. Agreement was strong between cardiac outputs with a mean bias ranging from 0.01 to 0.33 with no significant bias. Interobserver reproducibility of 4D CMR net flow measurements was good (interclass coefficient ranging from 0.89–0.98; and coefficient of variation of 3.7%–6.3%). Conclusion 4D CMR flow measurements correlated with cardiac output measured using CMR in healthy volunteers in all valves and vessels, respecting the conservation of mass principle. 4D CMR seems robust enough to obtain precise hemodynamic features of cardiac flow.