Comparison Of Different Methods Of Measuring LVEF By Magnetic Resonance

Objectie:To measure Left ventricular ejection fraction(LVEF) using the methods in post-processing software including SAX, LAX, Triplane and Biplane. The consistency and reproducibility of these methods, as well as the difference and consistency between these methods and the broadest possible M-mode ultrasonography at present, were compared. Methods:A total of 20 healthy young volunteers, including 11 males and 9 females, were recruited. MR cine sequence scanning was used to obtain images on horizontal long axis, vertical long axis and standard short axis. By positioning above images, the images on multilayer short axis and multi-angle long axis were acquired. Based on the semi-automatic depiction of endocardium, the LVEF was then calculated using the four methods respectively. The methods were then compared in pairs using Bland-Altman analysis and ICC to determine the consistency of results and intra and inter-observer error. T-test(P>0.05) and Bland-Altman analysis were performed on the results in comparison with M-mode ultrasonography. Results:All ICC was greater than 0.75(>0.75), except that the intra-observer ICC of SAX was smaller than 0.75(<0.75). Bland-Altman graph showed good consistency. Comprehensive analysis found that the consistency of the four methods and the intra and inter-observer consistency were all good; t-test was performed between the four methods and the most widely used M-mode ultrasonography, and the results showed that P>0.05 and Bland-Altman graph showed good consistency with M-mode ultrasonography. Conclusion:The four MR analytical methods show good consistency. All of them can be applied flexibly to measure the indicators of cardiac function clinically based on their different characteristics. The LAX method showed the best stability.