Abstract No. 84 Comparison of qDSA with a time-of-arrival–based method for determining velocity in a porcine iliac artery model

Abstract

Quantitative digital subtraction angiography (qDSA) is a recently developed 2D-DSA post-processing tool that can be used to calculate velocities in an artery-of-interest using multiple spatial and temporal points from an entire contrast injection. qDSA has been previously shown to correlate with 4D Flow MRI and intravascular Doppler. While time-of-arrival (TOA)-based methods are commercially and clinically available, they rely on a single spatial point at a particular time. This study aimed to compare velocities calculated using qDSA with a TOA-based method in a porcine iliac artery model. Thirty-three 2D-DSA acquisitions (30fps) were performed in 7 female domestic swine (∼54kg) with a 4 French catheter placed in the common iliac artery using 1.5-3 mL/s injection rates for 8 seconds. The vessel length divided by the difference in TOA of contrast was used to derive blood velocities in the common, external, and internal iliac arteries. Blood velocities were also calculated with qDSA using a shifted least squares method on the contrast-attenuation curves. The two methods were then compared with descriptive statistics, paired t-test, Pearson correlation, and Bland-Altman. Of 99 data points, one was excluded because TOA difference could not be detected. Mean blood velocities with the TOA-method versus qDSA were 39.7 ± 23.3 versus 45.6 ± 20.5 cm/s in the common (n = 33), 26.5 ± 12.2 versus 23.9 ± 9.5 cm/s in external (n = 32), and 27.6 ± 12.0 versus 34.5 ± 12.7 cm/s in internal iliac (n = 33) arteries. The two methods provided significantly different velocities based on the paired t-test (P< 0.05) and were moderately correlating (Pearson r = 0.42, P< 0.05). A Bland-Altman plot showed the TOA method to have a bias of -3.45 cm/s (95% LOA: -40.2 to 33.3), underestimating higher velocities ( >50 cm/s) with a higher degree of uncertainty. In a porcine iliac artery model, TOA-based blood velocities moderately correlated with those calculated using qDSA while TOA method underestimated higher velocities.