Rapid multi-contrast magnetic resonance imaging and time-of-flight angiography

Abstract

Magnetic resonance imaging (MRI) is a frequently used imaging modality for examining soft tissue structures. Long scanning time is the most crucial constraint that limits the use of MRI in the clinics. Partial Fourier (PF), parallel imaging (PI), and compressed sensing (CS) methods have been proposed to accelerate acquisitions by undersampling the data in k-space. However, further increase in acceleration factor as well as image quality are needed in certain applications of MRI, such as T2-weighted imaging and time-of- flight (TOF) angiography. Previous studies have adopted SPIRiT, a popular CS method, to the problem of multi-contrast image reconstruction. However, the mutual information across different contrast images were not utilized in these studies. In this thesis, a new method is proposed to benefit from the correlated structural information among the images by emphasizing high-spatial frequencies during joint reconstruction. The results obtained from in vivo brain scans and numerical phantom show that the proposed method is more robust against parameter selection when compared to conventional methods. For TOF angiography images, the goal of this thesis is to increase the signal-to-noise-ratio and shorten the scanning time, simultaneously. For this purpose, a combination of 2D acceleration in the phase-encode dimensions via CS and 1D PF data acquisition in the frequency-encode dimension to reduce echo time is proposed. Following this data acquisition, a joint reconstruction that iteratively alternates between CS and PF is introduced. In vivo angiography results in the brain show that the proposed time-efficient TOF method improves the vessel-background contrast, while decreasing the scanning time.