Two wrongs sometimes do make a right: errors in aortic valve stenosis assessment by same-day Doppler echocardiography and 4D flow MRI.

Abstract

This study aims to systematically verify if the simplified geometry and flow profile of the left ventricular outflow tract (LVOT) assumed in 2D echocardiography is appropriate while examining the utility of 4D flow MRI to assess valvular disease. This prospective study obtained same-day Doppler echocardiography and 4D flow MRI in 37 healthy volunteers (age: 51.9 ± 18.2, 20 females) and 7 aortic stenosis (AS) patients (age: 64.2 ± 9.6, 1 female). Two critical assumptions made in echocardiography for aortic valve area assessment were examined, i.e. the assumption of (1) a circular LVOT shape and (2) a flat velocity profile through the LVOT. 3D velocity and shape information obtained with 4D flow MRI was used as comparison. It was found that the LVOT area was lower (by 26.5% and 24.5%) and the velocity time integral (VTI) was higher (by 28.5% and 30.2%) with echo in the healthy and AS group, respectively. These competing errors largely cancelled out when examining individual and cohort averaged LVOT stroke volume. The LVOT area, VTI and stroke volume measured by echo and 4D flow MRI were 3.6 ± 0.7 vs. 4.9 ± 1.0 cm2 (p < 0.001), 21.2 ± 3.0 vs 15.2 ± 2.8cm (p < 0.001), and 75.6 ± 15.6 vs 72.8 ± 14.1ml (p = 0.3376), respectively. In the ensemble average of LVOT area and VTI, under- and over-estimation seem to compensate each other to result in a 'realistic' stroke volume. However, it is important to understand that this compensation may fail. 4D flow MRI provides a unique insight into this phenomenon.