https://doi.org/10.1007/s00392-025-02625-4
1Herzzentrum Dresden GmbH an der TU Dresden Klinik für Innere Medizin, Kardiologie und Intensivmedizin Dresden, Deutschland; 2Medizinische Fakultät Carl Gustav Carus der TU Dresden Dresden, Deutschland; 3Universitätsklinikum Carl Gustav Carus Dresden Radiologie Dresden, Deutschland
The computed tomography (CT) calcium score of the aortic valve is recommended by current clinical practice guidelines as an additional parameter for aortic stenosis severity assessment. However, CT exposes the patient to radiation and may cause harm to the patient. Therefore, we aimed to assess and validate an echocardiographic method for the quantification of aortic valve calcification.
Methods
47 patients with moderate to severe aortic stenosis of the native aortic valve were evaluated for aortic valve replacement. For severity assessment, transesophageal echocardiography (TEE) as well as a CT scan were performed. In TEE long- (LA) and short-axis (SA) views we performed triplicate measurements of the echo density of individually aligned regions of interest (ROI) of the aortic valve left coronary (LCC), non-coronary (NCC) and right coronary cusp (RCC). A ROI placed within the left ventricular septum served as reference value to correct for differences in gain strength. All other device-related parameters were kept constant. CT aortic valve calcification was quantified using a modified Agatston method considering the aortic valve cusps exclusively. Normalized aortic cusp echo densities were compared to CT calcium score using correlation analyses, Bland- Altman plots, intraobserver and interobserver variability.
Results
Normalized echo densities of the LCC in long axis (r=0.86, p<0.001) and in short axis (r=0.79, p<0.001) showed strong correlation to the CT derived calcium score. RCC in long-axis (r=0.68, p<0.001) and short axis (r=0.64, p<0.001), NCC in short axis (r=0.45, p<0.001) as well as the combined echo density of all cusps in short axis (r=0.72, p<0.001) exhibited lower correlation strength compared to the reference method.
The Bland-Altman comparison between LCC in long axis and LCC in short axis showed no significant difference of the parameter between both measurements (mean difference: -0.002; LoA: 0.037; p<0.001).
For all methods we observed very low intra- and interobserver variabilities as shown by intraobserver interclass coefficient (ICC) and interrater reliability coefficient (IRR) (values for LA LCC, e.g.: ICC = 0.985 [CI 0.971; 0.993]; IRR = 0.956 [CI 0.821; 0.989], each p<0.001).
Discussion
Echo density assessment of the LCC showed the best correlation with CT calcium score and excellent concordance between measurements in long and short axis. The superiority of LCC assessment is likely the result of less acoustic attenuation because of the shortest distance to the echocardiography probe in TEE and superior cusp delineation when compared to the other cusps. Assessment of LCC echo density in TEE is a simple, fast and radiation free method for the assessment of aortic valve calcification. Further studies are needed to validate this method in a larger cohort with a broader range of aortic stenosis severities.