Precision of forty slice spiral computed tomography for quantifying aortic valve stenosis: comparison with...

We evaluated the precision of multislice spiral computed tomography (MSCT) for the quantification of aortic valve stenosis in comparison with echocardiography and cardiac catheterization.

An electrocardiogram-gated MSCT scan (detector collimation 40 x 6.25 mm, gantry rotation time 420 milliseconds, pitch 0.2, tube voltage 120 KV, tube current 333 mA) was performed in 32 patients with known aortic valve stenosis. In each patient the aortic valve orifice area (AVA) was determined by planimetry on MSCT and compared with the results obtained from transthoracic Doppler echocardiography (using the continuity equation) and cardiac catheterization (using the Gorlin formula).

Planimetry of the AVA on MSCT was feasible in all cases. The AVA on MSCT (1.11 +/- 0.49 cm2) was significantly larger compared with echocardiography (0.81 +/- 0.37 cm2, P < 0.001) and cardiac catheterization (0.87 +/- 0.45 cm2, P < 0.001). The correlations between MSCT and echocardiography (r = 0.86, limits of agreement +/-0.52 cm2) and also between MSCT and cardiac catheterization (r = 0.90, limits of agreement +/-0.44 cm2) were good, but inferior to the correlation between echocardiography and cardiac catheterization (r = 0.94, limits of agreement +/-0.32 cm2). Using an AVA of 1.0 cm at cardiac catheterization as reference standard, the best cut-off level for detecting severe-to-critical stenosis at MSCT was an AVA of 1.20 cm, resulting in a sensitivity, specificity, and accuracy of 91%, 100%, and 94%, respectively.

AVA determined by MSCT correlates well with echocardiography and cardiac catheterization. However, AVA derived from MSCT is consistently larger, requiring an adjustment of cut-off values for the classification of stenosis severity and therapeutic decision making.