GRAPPA-based susceptibility-weighted imaging of normal volunteers and patients with brain tumor at 7 T

Abstract

Susceptibility-weighted imaging (SWI) is a valuable technique for high-resolution imaging of brain vasculature that greatly benefits from the emergence of higher field strength MR scanners. Autocalibrating partially parallel imaging techniques can be employed to reduce lengthy acquisition times as long as the decrease in signal-to-noise ratio does not significantly affect the contrast between vessels and brain parenchyma. This study assessed the feasibility of a Generalized Autocalibrating Partially Parallel Acquisition (GRAPPA)-based SWI technique at 7 T in both healthy volunteers and brain tumor patients. GRAPPA-based SWI allowed a twofold or more reduction in scan time without compromising vessel contrast and small vessel detection. Postprocessing parameters for the SWI needed to be modified for patients where the tumor causes high-frequency phase wrap artifacts but did not adversely affect vessel contrast. GRAPPA-based SWI at 7 T revealed regions of microvascularity, hemorrhage and calcification within heterogeneous brain tumors that may aid in characterizing active or necrotic tumor and monitoring treatment effects.

FIG 1

Depiction of brain regions used to generate contrast ratios. In (a) is a representative large vessel mask in green, while in (b), a region is enlarged to visualize a small vessel mask and neighboring white matter region, delineated in pink.

FIG 2

Effects of coil combination method on vessel contrast for the fully sampled and reduced sampled images. The top row shows uncorrected combined magnitude images with greatest uniformity when weighting solely by the noise variance of each coil. In the lower two rows, the enlarged images demonstrate the reduced small vessel contrast observed when using a weighted sum of coil sensitivity profiles for coil combination.

FIG 3

Acquired normal volunteer SWI with a minimum intensity projection of 15 mm for the (a) fully sampled, (b) 2-fold reduction (R=2), and (c) 3-fold reduction (R=3) acquisitions. The bottom row displays magnified SWI images for the visualization of small vessels.

FIG 4

Variation of CNR and stability of CR across the head in a representative normal volunteer acquired with the R=2 acquisition. In A, small vessel (blue) and adjacent white matter (red) regions in four different brain locations. In B, the regions from A are magnified to show the apparent similarity in contrast despite the variation in CNR values.

FIG 5

Illustration of fully sampled and GRAPPA-based SWI of brain tumors compared to corresponding 3T anatomical images. In left panel, a recurrent glioma patient post-therapy with (a) pre-Gd T1 weighted SPGR, (b) post-Gd T1 weighted SPGR, and (c) T2 weighted FLAIR imaging at 3T, and (d) fully sampled SWI at 7T. In right panel, a newly-diagnosed glioma patient pre-therapy with (e) pre-Gd T1 weighted SPGR, (f) post-Gd T1 weighted SPGR, and (g) T2 weighted FLAIR imaging at 3T, and (h) GRAPPA-based SWI (R=2) at 7T.

FIG 6

Depiction of (a) blood products, (b) radiation effects, and (c) abnormal vasculature in treated brain tumors using SWI at 7T. The two dark areas at the brain's surface in (c) are due to artifacts from surgical excision.