Variability in Analysis of Volumetric Flow using a Flow Phantom

Abstract

Abstract: : Purpose: In a vessel of constant diameter, flow is a linear function of velocity. To evaluate the variability in the analysis of volumetric flow in ATL HDI-5000 cineloops, error in the flow/velocity ratio was quantified. Methods: A phantom flow model was constructed using agarose gel to mimic fatty soft tissue. 1.57mm and 2.36mm diameter lumens were created in the gel. A blood-mimicking fluid was pumped through each tube at low, medium, and high rates. The ATL HDI-5000 measured the velocity and volumetric flow in the phantom model using cineloops (a rapidly acquired sequence of CDI images). Measurements were performed with the probe in four different positions: 1) 45° angle, parallel to the flow, 2) 45° angle, offset to the flow, 3) 75° angle, parallel to the flow, 4) 75° angle, offset to the flow. Twenty five measurements were obtained at each of the four probe positions for all three flow settings in both lumens (n=600). The flow values obtained were correlated with the velocity measurements at each setting. Results: All correlations were statistically significant (p<0.001). In the pooled data set, velocity was significantly correlated to flow (R²= 0.408, n=600). The large lumen (R² = 0.637, n=300) had less error than the small (R²= 0.537, n=300). Obtaining measurements with the probe at 75° (R² = 0.836, n=300) had less error than at 45° (R² = 0.073, n=300). Using the large lumen with the probe held at 75°, obtaining measurements offset to the flow (R² = 0.862, n=75) had less error than when parallel to the flow (R² = 0.785, n=75). Conclusions: Based on agreement between flow and velocity data, the calculation of volumetric flow from cineloops contain the least amount of noise when the probe is placed at a 75° angle to the body surface with the probe offset to the direction of blood flow.

Keywords: