Slice positioning in phase-contrast MRI impacts aortic stenosis assessment

Abstract

AimsTo determine the phase-contrast cardiovascular magnetic resonance imaging (PC-CMR) slice-position above aortic leaflet-attachment-plane (LAP) that provides flow-velocity, –volume and aortic valve area (AVA) measurements with best agreement to invasive and echocardiographic measurements in aortic stenosis (AS). Methods and resultsFifty-five patients with moderate/severe AS underwent cardiac catheterization, transthoracic echocardiography (TTE) and CMR. Overall, 171 image-planes parallel to LAP were measured via PC-CMR between 22 mm below and 24 mm above LAP. AVA via PC-CMR was calculated as flow-volume divided by peak-velocity during systole. Stroke volume (SV) and AVA were compared to volumetric SV and invasive AVA via the Gorlin-formula, respectively.Above LAP, SV by PC-CMR showed no significant dependence on image-plane-position and correlated strongly with volumetry (rho: 0.633, p < 0.001, marginal-mean-difference (MMD): 1 ml, 95 % confidence-interval (CI): −4 to 6). AVA assessed in image-planes 0–10 mm above LAP differed significantly from invasive measurement (MMD: −0.14 cm2, 95 %CI: 0.08–0.21). In contrast, AVA-values by PC-CMR measured 10–20 mm above LAP showed good agreement with invasive determination without significant MMD (0.003 cm2, 95 %CI: −0.09 to 0.09). Within these measurements, a plane 15 mm above LAP resulted in the lowest bias (MMD: 0.02 cm2, 95 %CI:-0.29 to 0.33). SV and AVA via TTE correlated moderately with volumetry (rho: 0.461, p < 0.001; bias: 15 ml, p < 0.001) and cardiac catheterization (rho: 0.486, p < 0.001, bias: −0.13 cm2, p < 0.001), respectively. ConclusionPC-CMR measurements at 0–10 mm above LAP should be avoided due to significant AVA-overestimation compared to invasive determination. AVA-assessment by PC-CMR between 10 and 20 mm above LAP did not differ from invasive measurements, with the lowest intermethodical bias measured 15 mm above LAP.