Clinical feasibility study of 3D intracranial magnetic resonance angiography using compressed sensing.
Zhiyong Lin,Xiaodong Zhang,Li Guo,Ke Wang,Yuan Jiang,Xiaoyu Hu,Yong Huang,Juan Wei,Shuai Ma,Yi Liu,Lina Zhu,Zhizheng Zhuo,Jing Liu,Xiaoying Wang
DOI: https://doi.org/10.1002/jmri.26752
IF: 4.4
2019-01-01
Journal of Magnetic Resonance Imaging
Abstract:Background Compressed sensing (CS) has been widely used to improve the speed of MRI, but the feasibility of application in 3D intracranial MR angiography (MRA) needs to be evaluated in clinical practice. Purpose To evaluate the clinical feasibility of CS-MRA in comparison with conventional 3D-MRA (Con-MRA). Study Type Retrospective. Subjects Forty-nine consecutive patients with suspected intracranial arterial disease. Field Strength/Sequence 3T MRI. 3D time-of-flight (TOF) MRA using a CS algorithm and conventional 3D TOF MRA scan. Assessment Three radiologists (4, 11, and 12 years of experience in neuroradiology) independently assessed the image quality, vascular lesions, and variations of intracranial arteries of both CS-MRA and Con-MRA, respectively. Statistical Tests The Kendall W test was performed to assess the interobserver agreement of image quality and intracranial arterial stenosis. A nonparametric test (Wilcoxon test) was used for comparison of the image quality and definition of the external carotid artery (ECA). Weighted kappa analysis was performed for the interstudy agreement of intracranial arterial stenosis. The aneurysm, decreased branches, congenital hypoplasia, absence, and variant branching of intracranial arteries were observed and evaluated for interobserver agreement and interstudy agreement by kappa analysis. Paired-t-tests for signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) were conducted. Results Image quality is better for CS-MRA compared with Con-MRA with significance (Z = -3.710 to -2.673, with P < 0.01). The interstudy agreement of lesion and variation of intracranial arteries assessment for each observer was excellent. The SNR and CNR were significantly higher in CS-MRA compared with Con-MRA (P < 0.001). The definition of ECA of CS-MRA was significantly better (Z = -4.9, P < 0.001). Data Conclusion CS-MRA showed significantly higher image quality with less blur, comparable image diagnostic performance of intracranial arteries, and better display of ECA than Con-MRA. Technical Efficacy Stage: 2 J. Magn. Reson. Imaging 2019;50:1843-1851.