Objectives: The aim of this study was to evaluate the clinical feasibility of accelerated time-of-flight (TOF) magnetic resonance angiography with sparse undersampling and iterative reconstruction (sparse TOF).
Materials and Methods: The local institutional review board approved the study protocols. Twenty healthy volunteers were recruited (mean age, 31.2 years; age range, 22-52 years; 14 men, 6 women). Both sparse TOF and parallel imaging (PI) TOF were obtained on a 3 T scanner. Acceleration factors were 3, 4, 5, 6, and 8 for sparse TOF (Sp 3×, Sp 4×, Sp 5×, Sp 6×, and Sp 8×, respectively) and 2, 3, 4, and 6 for PI TOF (PI 2×, PI 3×, PI 4×, and PI 6×, respectively). Images were reconstructed on the scanner, and maximum intensity projection images were subjected to visual evaluation, wherein each segment of the major brain arteries was independently evaluated by 2 radiologists on a 4-point scale (1, poor; 2, limited; 3, moderate/good quality for diagnosis; and 4, excellent). As a quantitative evaluation, the apparent contrast-to-background deviation (apparent CBD) was calculated at the level of the basilar artery and the pons.
Results: A total number of 1800 segments were subjectively evaluated. There was substantial agreement regarding vessel visualization (κ = 0.759). Sparse TOF received scores above 3 (good for diagnosis) at any acceleration factor up to the third segments of major arteries. The middle and distal segments of PI 4× and PI 6× were graded below 3 (limited or poor diagnostic value). Sp 3×, 4×, 5×, and 6× retained diagnostic information (graded above 3), even at distal segments. The apparent CBD of sparse TOF at any acceleration factor was equivalent to that of PI 2×, whereas the apparent CBD of PI 3×, PI 4×, and PI 6× attenuated with the acceleration factor.
Conclusions: Sparse TOF can achieve better image quality relative to PI TOF at higher acceleration factors. The diagnostic quality of distal branches (A2/3, M4, P4) was maintained with Sp 6×, which achieved a shorter acquisition time less than half of PI 2×.
From the *Department of Diagnostic Imaging and Nuclear Medicine, Graduate School of Medicine, Kyoto University, Kyoto, Japan; †Siemens Healthcare, Erlangen, Germany; and ‡Siemens Medical Solutions USA, Inc, Los Angeles, CA.
Received for publication June 11, 2015; and accepted for publication, after revision, August 26, 2015.
Supported in part by the Grant-in-Aid for Scientific Research on Innovative Areas “Initiative for High-Dimensional Data-Driven Science through Deepening of Sparse Modeling” (MEXT grant numbers 25120002).
Conflicts of interest and sources of funding: K.F., T.O., Y.F., and K.T. are currently receiving a grant (#25120002) from the Ministry of Education, Culture, Sports, Science, and Technology, Japan. A.F.S. and M.S. are employees of Siemens Healthcare, Erlangen, Germany. Y.N. is an employee of Siemens Medical Solutions USA, Inc.
Correspondence to: Koji Fujimoto, MD, PhD, Department of Diagnostic Imaging and Nuclear Medicine, Graduate School of Medicine, Kyoto University, 54 Shogoin Kawaharacho, Sakyo-ku, Kyoto City, Kyoto, Japan, 606-8507. E-mail: email@example.com.