Assessment of arterial pulsatility of cerebral perforating arteries using 7T high-resolution dual-VENC phase-contrast MRI

Abstract

Purpose: Directly imaging the function of cerebral perforating arteries could provide valuable insight into the pathology of cerebral small vessel diseases (cSVD). Arterial pulsatility has been identified as a useful biomarker for assessing vascular dysfunction. In this study, we investigate the feasibility and reliability of using dual velocity encoding (VENC) phase-contrast MRI (PC-MRI) to measure the pulsatility of cerebral perforating arteries at 7 T.

Methods: Twenty participants, including 12 young volunteers and 8 elder adults, underwent high-resolution 2D PC-MRI scans with VENCs of 20 cm/s and 40 cm/s at 7T. The sensitivity of perforator detection and the reliability of pulsatility measurement of cerebral perforating arteries using dual-VENC PC-MRI were evaluated by comparison with the single-VENC data. The effects of temporal resolution in the PC-MRI acquisition and aging on the pulsatility measurements were investigated.

Results: Compared to the single VENCs, dual-VENC PC-MRI provided improved sensitivity of perforator detection and more reliable pulsatility measurements. Temporal resolution impacted the pulsatility measurements, as decreasing temporal resolution led to an underestimation of pulsatility. Elderly adults had elevated pulsatility in cerebral perforating arteries compared to young adults, but there was no difference in the number of detected perforators between the two age groups.

Conclusion: Dual-VENC PC-MRI is a reliable imaging method for the assessment of pulsatility of cerebral perforating arteries, which could be useful as a potential imaging biomarker of aging and cSVD.

Keywords: arterial pulsatility; cerebral perforating artery; dual‐VENC; phase‐contrast MRI; pulsatility index (PI); ultra‐high field 7 T.

Figure 1

a. TOF MIP image as a reference to position the imaging slice of PC-MRI (yellow line) which is perpendicular to most of the perforating arteries; b,c. The PC-MRI magnitude and structural image in which the perforating arteries in the basal ganglia region are clearly detected highlighted in a red box; d. The averaged normalized velocity curve of detected perforators from the PC-MRI data. PI is calculated by dividing the difference between the maximum and minimum velocities by the mean velocity across a cardiac cycle.

Figure 2

a and b show an example of the zoomed-in PC-MRI magnitude images in the basal ganglia region with VENC = 40 cm/s (a) and VENC = 20 cm/s (b), respectively, from a representative subject. The detected perforators are circled in green. c. The comparison of the number of detected perforators using VENC of 40 cm/s and 20 cm/s across subjects. The colored markers show measurements from individual subjects, and each black line shows the measurements from a single subject.

Figure 3

a and b show the individual comparisons of PI and V_mean measurements, respectively, with dual VENC, high VENC of 40cm/s, and low VENC of 20cm/s. The colored markers show measurements from individual subjects, and each black line shows the measurements from a single subject. c and d show the group comparisons of the PI and V_mean measurements using box plots, respectively, with dual VENC, high VENC of 40cm/s, and low VENC of 20cm/s.

Figure 4

Bland-Altman plots of test-retest measurements of the detected number of perforators with high VENC of 40cm/s (a) and low VENC of 20cm/s (b), respectively.

Figure 5

Bland-Altman plots of the test-retest PI measurements using low-VENC of 20 cm/s (a), dual-VENC (b), and high VENC of 40 cm/s (c), respectively.

Figure 6

a. Averaged velocity curves of detected perforators across subjects using dual-VENC PC-MRI from the test and retest data. The shaded areas show the standard deviation of the velocity measurements at each data point. b. Bland-Altman plot of the test-retest V_mean values measured using dual-VENC PC-MRI.

Figure 7

Bland-Altman plots of the inter-reader reproducibility of N_perforator (a) and PI (b) measurements.

Figure 8

a. Averaged normalized velocity curves of detected perforators across subjects using dual-VENC PC-MRI at the temporal resolution of 63ms, 95ms, and 127ms, respectively. The shaded areas show the standard deviation of the velocity measurements at each data point. b and c show box plots of the number of detected perforators (b) and PI measured with dual-VENC PC-MRI at the temporal resolution of 63ms, 95ms, and 127ms, respectively.

Figure 9

Comparison of the number of detected perforators (a), PI (b), and V_mean (c) between the young and elderly participants measured using dual-VENC PC-MRI, respectively.