Effects of inaccuracies in arterial path length measurement on differences in MRI and tonometry measured pulse wave velocity

Abstract

BackgroundCarotid-femoral pulse wave velocity (cf-PWV) and aortic PWV measured using MRI (MRI-PWV) show good correlation, but with a significant and consistent bias across studies. The aim of the current study was to evaluate whether the differences between cf.-PWV and MRI-PWV can be accounted for by inaccuracies of currently used distance measurements.MethodsOne hundred fourteen study participants were recruited into one of 4 groups: Type 2 diabetes melltus (T2DM) with cardiovascular disease (CVD) (n = 23), T2DM without CVD (n = 41), CVD without T2DM (n = 25) and a control group (n = 25). All participants underwent cf.-PWV, cardiac MRI and whole body MR angiography(WB-MRA). 90 study participants also underwent aortic PWV using MRI. cf.-PWVEXT was performed using a SphygmoCor device (Atcor Medical, West Ryde, Australia). The true intra-arterial pathlength was measured using the WB-MRA and then used to recalculate the cf.-PWVEXT to give a cf.-PWVMRA.ResultsDistance measurements were significantly lower on WB-MRA than on external tape measure (mean diff = −85.4 ± 54.0 mm,p < 0.001). MRI-PWV was significantly lower than cf.-PWVEXT (MRI-PWV = 8.1 ± 2.9 vs. cf.-PWVEXT = 10.9 ± 2.7 ms−1,p < 0.001). When cf.-PWV was recalculated using the inter-arterial distance from WB-MRA, this difference was significantly reduced but not lost (MRI-PWV = 8.1 ± 2.9 ms−1 vs. cf.-PWVMRA 9.1 ± 2.1 ms−1, mean diff = −0.96 ± 2.52 ms−1,p = 0.001). Recalculation of the PWV increased correlation with age and pulse pressure.ConclusionDifferences in cf.-PWV and MRI PWV can be predominantly but not entirely explained by inaccuracies introduced by the use of simple surface measurements to represent the convoluted arterial path between the carotid and femoral arteries.

Development and Validation of a Path Length Calculation for Carotid–Femoral Pulse Wave Velocity Measurement

“…This was not the case in the current study with the only variables significantly impacted by the new formula being BMI and waist circumference—both of which are known to affect the accurate path length measurement. 14 , 20 We found no significant shift in the magnitude of differences in PWV between those with and those without cardiovascular disease and no substantial change in the hazard ratios for PWV for prediction of future major adverse cardiovascular events and mortality.…”

“…Multiple studies have documented significant differences between carotid–femoral PWV and central aortic PWV, 15 , 16 with this difference being markedly reduced when the true arterial path lengths are used in the carotid–femoral PWV calculation rather than the external distance measurements. 14 , 17 We are not the first to attempt to derive a formula for the calculation of the true arterial path length for use in PWV calculation. Weber et al 6 produced a calculation for distance being equal to the (height/4+7.28) from a population of 135 being investigated for coronary artery disease, but, however, used the distance from the ascending aorta to the aortic bifurcation as the comparator which is not representative of the path traveled by the pulse wave between the carotid and femoral artery.…”

“…One hundred forty-eight participants of the SUMMIT study underwent WB-MRA using the same acquisition parameters and with the intra-arterial path length measured using the same technique as in the TASCFORCE cohort. 14 This was then compared with the distance generated using the formula from the TASCFORCE cohort. PWV in this cohort was calculated using the 3 distances: the external distance (PWV SUMM ), the true distance on MRI (PWV TD ), and the formula distance (PWV MRA ).…”

Artificial Intelligence Estimation of Carotid-Femoral Pulse Wave Velocity using Carotid Waveform

“…We can observe, from Figs 2 and 3 , that as PWV goes above 12 m / sec the error starts to increase. It is both because of the inherent measurement error 37 – 40 and also sparsity of data in that region. For smaller values of PWV , this behavior is less pronounced as data concentration is higher and also the measurement errors are smaller.…”

Acute phase proteins as prospective risk markers for arterial stiffness: The Malmö Diet and Cancer cohort

“…Even though c-f PWV is a valid measure of the arterial stiffness in the aortic and aorto-iliac vascular tree, it should be acknowledged that c-f PWV does not reflect the heterogeneity which occurs in various primary and secondary branches. A recent study by Weir-McCall et al demonstrated differences in carotid-femoral measured PWV and MRI measured aortic PWV, which can be for the most part be explained by use of simple surface measurements [ 55 ]. Nevertheless, distance estimation has been performed in a standardized way in this study.…”