Test-retest multisite reproducibility of neurovascular 4D flow MRI

Background 4D flow MRI shows great potential in neurovascular disorders such as stenosis, atherosclerotic disease, aneurysms, and vascular malformations. Its widespread application in the neurovascular system requires evidence of good test-retest multicenter reproducibility. Purpose To assess the multicenter reproducibility, test-retest reliability and interobserver dependence of 4D flow MRI in measurements of cerebral blood flow/velocity in main intracranial vessels. Study Type Prospective study. Subjects Ten healthy subjects underwent 4D flow scans at three different centers. All subjects were scanned twice at 2 different days at each center. Field Strength/Sequence 3.0 T; 4D flow sequence. Assessment Multicenter reproducibility, test-retest reliability and interobserver agreement for measurements of the blood flow and peak velocity from five regions of interest were assessed (bilateral internal carotid arteries, bilateral medial cerebral arteries, and sagittal sinus). Statistical Test A Shapiro-Wilks test was conducted to assess normality of measurements in each scan. Coefficient of variances (CVs) was computed to evaluate intra- and intersite variances of all measurements. The multicenter reproducibility was assessed by two-way mixed intraclass correlation coefficient (ICC). A Bland-Altman plot and Pearson correlation were used to evaluate test-retest reliability. ICC was calculated to assess interobserver agreements. Results All P-values for Shapiro-Wilks tests were greater than 0.05, which indicated the normality of all measurements. Both intra- and intersite CVs were lower than 12%. There was good test-retest reliability for both blood flow and peak velocity of all ROIs (r = 0.75-0.94). In addition, high multicenter reproducibility was detected (ICC = 0.77-0.96, all P < 0.001). The results of these measurements also showed great interobserver agreement (all ICC > 0.9 and all P < 0.001). Data Conclusion High multicenter reproducibility and test-retest reliability was shown for 4D flow in the measurements of blood flow and peak velocity of intracranial vessels. In addition, these measurements showed great interobserver agreement.