Skip to main content

Table 3 Effect of temporal resolution

From: Estimation of aortic pulse wave transit time in cardiovascular magnetic resonance using complex wavelet cross-spectrum analysis

 

Avg by 2

Avg by 3

Avg by 4

archPWVWU

Linear fit slope

1.01

0.93

0.86

Bland-Altman mean bias [limits] (m/s)

0.2 [-1.45 to 1.87]

0.12 [-2.06 to 2.30]

0.10 [-2.50 to 2.71]

Correlation r(p)

0.95 (<0.001)

0.91 (<0.001)

0.87 (<0.001)

archPWVTU

Linear fit slope

0.99

0.60

0.30

Bland-Altman mean bias [limits] (m/s)

0.35 [-1.72 to 2.40]

-0.07 [-4.28 to 4.22]

-0.03 [-7.81 to 7.71]

Correlation r(p)

0.88 (<0.001)

0.72 (<0.001)

0.50 (<0.001)

archPWVF

Linear fit slope

1.03

1.03

0.89

Bland-Altman mean bias [limits] (m/s)

0.16 [-2.51 to 2.8]

0.42 [-3.29 to 4.12]

0.32 [-4.67 to 5.32]

Correlation r(p)

0.97 (<0.001)

0.94 (<0.001)

0.89 (<0.001)

  1. For the three methods, aortic pulse wave velocity (archPWV) estimated from the original ascending and descending aortic flow curves, were used as reference for comparisons against archPWV estimated from curves with lower temporal resolution obtained by averaging time points by blocks of 2 (Avg by 2), 3 (Avg by 3) and 4 (Avg by 4). For these associations the linear fit slope, Bland-Altman mean bias and limits of agreement as well as linear correlation coefficients were provided. Wavelet-based approach is archPWVWU, time-based approach is archPWVTU and Fourier-based approach is archPWVF