ESMRMB 2020 Online, 37th Annual Scientific Meeting, September 30-October 2: Abstracts, Friday
- PDF / 9,544,826 Bytes
- 21 Pages / 595.276 x 790.866 pts Page_size
- 8 Downloads / 188 Views
ESMRMB Congress (2020) 33 (Suppl 1): S47–S67 was calculated, comparing the mean CVR of the standardized ROIs (Fig. 2a).
S06 Scientific Session
10:30–12:30
Room B
Brain & fMRI S06.01 Measuring cerebrovascular reactivity in patients with Moyamoya disease: comparison of resting-state BOLD fMRI to breath-hold BOLD fMRI L. Zerweck1, T.-K. Hauser1, C. Roder2, U. Klose1 1 University Hospital Tuebingen, Diagnostic and Interventional Neuroradiology, Tuebingen, GERMANY, 2University Hospital Tuebingen, Neurosurgery, Tuebingen, GERMANY
Introduction: Patients with Moyamoya Disease (MMD) require the estimation of remaining cerebrovascular reactivity (CVR) to evaluate vascular territories at risk of stroke [1]. Positron emission tomography (PET) with acetazolamide (ACZ) challenge is the diagnostic gold standard to assess the CVR. Another, more readily available approach seems to be hypercapnia-triggered BOLD fMRI, for example by breath-holding (bh), where BOLD signal increases are expected in brain tissue with physiological CVR [2]. Recent findings suggest the use of resting-state (rs) BOLD fMRI to estimate the CVR [3]. This approach requires minimum patient compliance and no complex equipment. The aim of this study was to compare rs fMRI to bh fMRI in patients with MMD. Subjects/methods: The rs fMRI and bh BOLD fMRI data sets of seven patients with angiographically proven MMD were analyzed. All images were acquired on a 3 T MR Scanner. During the rs task the patients were instructed to close their eyes without performing any task. The bh task involved five repetitive cycles, each consisting of 9 s end-expiratory breath-holding and 60 s of regular breathing. The images of both modalities were realigned, normalized, segmented into 6 standardized ROIs [4] and spatially smoothed by a Gaussian kernel of 12 mm FWHM. The bh images were additionally slice-timing corrected. The rs data was temporally filtered with a band-pass filter of 0.02–0.04 Hz, because previous findings suggest that global rs BOLD signal fluctuations within this frequency range correlate best with end-tidal CO2 fluctuations [3]. The rs CVR maps (Fig. 1a) were calculated by performing a linear regression analysis in which the cerebellar reference time-course was the independent variable and individual voxels signal time-courses were the dependent variable. The bh CVR maps (Fig. 1b) were calculated by voxel-wise integrating the signal time-course averaged over the five cycles.
The correlation between the normalized rs and bh maps of all patients
123
Results/discussion: The present analysis indicates that rs fMRI leads to similar estimation of the CVR to bh fMRI (correlation coefficient = 0.80, p \ 0.001, Fig. 2b). It seems to be a promising method for hemodynamic evaluation because it is widely available and easily implementable in a high number of patients and therefore merits further research. References 1. Roder C et al. (2018) World Neurosurg 117:e75–e81. 2. Hauser TK et al. (2019) Neuroimage Clin 22:101713. 3. Liu P et al. (2017) Neuroimage 146:320–326. 4
Data Loading...
