Repeatability of cerebral arteriovenous pulse wave propagation in flow-related enhancement MRI

Flow Related Enhancement (FREE) MRI is a non-contrast technique for temporally resolved cerebral pulse-wave analysis. Clinical implementation requires proven repeatability. We aimed to evaluate the intra- and inter-scan repeatability of FREE-MRI measurements in healthy volunteers. Twenty-four healthy volunteers were scanned on a 3T MRI using a balanced steady-state free precession (bSSFP) sequence. A test-retest protocol was performed: two scans, a break with repositioning, and two more scans. From the resulting pulse-wave delay maps, the arteriovenous delay (AVD) was computed for the anterior (ACA), middle (MCA), and posterior (PCA) cerebral arteries. Repeatability was assessed using Bland-Altman analysis and Spearman correlation. Mean AVDs were 366 ± 51 ms (ACA), 371 ± 55 ms (MCA), and 376 ± 53 ms (PCA). Intra-scan repeatability was variable; the first session (Run 1 vs. 2) showed no significant differences, whereas the second session (Run 3 vs. 4, post-repositioning) showed significant deviation. However, inter-scan repeatability (comparing pooled pre- vs. post-break acquisitions) showed no significant differences after Bonferroni correction. Bland-Altman analysis confirmed that averaging measurements (Before vs. After) narrowed the limits of agreement compared to single-run comparisons, indicating improved stability. FREE-MRI provides quantitative assessments of cerebral pulse-wave dynamics, though single-acquisition precision is sensitive to physiological state. While intra-scan repeatability was high at rest, immediate post-repositioning scans showed significant variability, highlighting the need for a settling period. Crucially, while averaging measurements across sessions (Before vs. After) helps mitigate this noise, inter-scan reproducibility remains limited and protocol-dependent. Despite these limitations, this study provides a foundational step toward establishing a practical protocol for future clinical and longitudinal applications.