Abstract 390 Positioning of the Image Plane in Phase Contrast MRI impacts Aortic Stenosis Assessment

Background. Phase-contrast cardiovascular magnetic resonance imaging (PC-CMR) has become an emerging tool in characterizing aortic stenosis (AS). While it is increasingly used in quantifying stenosis degree, there is still uncertainty which PC-CMR image plane within the aortic root yields the most reliable results.

Purpose. To determine the PC-CMR imaging level above aortic leaflet attachment plane (LAP) that generates the most valid measures of flow velocity and volume compared to invasive catheterization.

Methods. Fifty-one patients with moderate to severe AS underwent cardiac catheterization, transthoracic echocardiography (TTE) and CMR including cine-imaging and PC-CMR. A total of 243 imaging planes parallel to the aortic annular plane were measured via PC-CMR, at levels ranging from 22mm below to 24mm above LAP at end-diastole. Aortic valve area (AVA) via PC-CMR was calculated as mean flow volume divided by peak velocity over the whole systole. In all layers, stroke volume (SV) and AVA were determined as well as compared to volumetric SV and invasive AVA via the Gorlin-formula, respectively.

Results. Above LAP, SV by PC-CMR compared to cine-volumetry showed no significant differences depending on image plane position and correlated strongly with volumetric measurements (r: 0.647, p<0.001, bias: 1ml, p=0.540). AVA assessment in layers from 0-10mm above LAP differed significantly between PC-CMR and invasive measurement (0.85±0.30cm² vs. 0.71±0.24cm², bias: 0.14cm², p<0.001). In contrast, AVA values by PC-CMR measured 10-20mm above LA showed good agreement with invasive determination without any significant bias (0.70±0.25cm² vs. 0.70±0.25cm², bias: 0.001cm², p=0.974). Within these measurements, the layer at 15mm above LA displayed the lowest bias ( 0.05cm², p=0.516). By contrast, SV and AVA via TTE correlated moderately (r: 0.492, p<0.001; bias: -14ml, p<0.001) and weakly (r: 0.356, p<0.001, bias: 0.17cm², p<0.001), respectively.

Conclusion. PC-CMR measurements at 0-10mm above LA should be avoided due to significant AVA overestimation compared to invasive determination. AVA assessment by PC-CMR between 10 and 20mm above LA did not differ from invasive measurements, with the lowest intermethodical bias measured 15mm above LA.

Clin Res Cardiol (2021). 10.1007/s00392-021-01933-9
Positioning of the Image Plane in Phase Contrast MRI impacts Aortic Stenosis Assessment
F. Troger¹, M. Reindl², I. Lechner³, C. Tiller², M. Holzknecht³, M. Pamminger¹, P. Poskaite¹, C. Kremser¹, S. J. Reinstadler², A. Bauer², B. Metzler³, G. Klug³, A. Mayr¹
¹Universitätsklinik für Radiologie, Medizinische Universität Innsbruck, Innsbruck, AT; ²Kardiologie und Angiologie, Tirol Kliniken GmbH, Innsbruck, AT; ³Department für Innere Medizin III - Kardiologie und Angiologie, Medizinische Universität Innsbruck, Innsbruck, AT;
Background. Phase-contrast cardiovascular magnetic resonance imaging (PC-CMR) has become an emerging tool in characterizing aortic stenosis (AS). While it is increasingly used in quantifying stenosis degree, there is still uncertainty which PC-CMR image plane within the aortic root yields the most reliable results. Purpose. To determine the PC-CMR imaging level above aortic leaflet attachment plane (LAP) that generates the most valid measures of flow velocity and volume compared to invasive catheterization. Methods. Fifty-one patients with moderate to severe AS underwent cardiac catheterization, transthoracic echocardiography (TTE) and CMR including cine-imaging and PC-CMR. A total of 243 imaging planes parallel to the aortic annular plane were measured via PC-CMR, at levels ranging from 22mm below to 24mm above LAP at end-diastole. Aortic valve area (AVA) via PC-CMR was calculated as mean flow volume divided by peak velocity over the whole systole. In all layers, stroke volume (SV) and AVA were determined as well as compared to volumetric SV and invasive AVA via the Gorlin-formula, respectively. Results. Above LAP, SV by PC-CMR compared to cine-volumetry showed no significant differences depending on image plane position and correlated strongly with volumetric measurements (r: 0.647, p<0.001, bias: 1ml, p=0.540). AVA assessment in layers from 0-10mm above LAP differed significantly between PC-CMR and invasive measurement (0.85±0.30cm² vs. 0.71±0.24cm², bias: 0.14cm², p<0.001). In contrast, AVA values by PC-CMR measured 10-20mm above LA showed good agreement with invasive determination without any significant bias (0.70±0.25cm² vs. 0.70±0.25cm², bias: 0.001cm², p=0.974). Within these measurements, the layer at 15mm above LA displayed the lowest bias ( 0.05cm², p=0.516). By contrast, SV and AVA via TTE correlated moderately (r: 0.492, p<0.001; bias: -14ml, p<0.001) and weakly (r: 0.356, p<0.001, bias: 0.17cm², p<0.001), respectively. Conclusion. PC-CMR measurements at 0-10mm above LA should be avoided due to significant AVA overestimation compared to invasive determination. AVA assessment by PC-CMR between 10 and 20mm above LA did not differ from invasive measurements, with the lowest intermethodical bias measured 15mm above LA.
https://dgk.org/kongress_programme/ht2021/P531.htm