MRI Scan Time Reduction through K-Space Data Sharing in Combo Acquisitions with a Spin Echo Sequence

We propose a technique to reduce scan time for magnetic resonance imaging (MRI) through sharing of k-space data between images. As a proof of concept, we ran simulations of MRI experiments based on Bloch equations using a spin echo sequence. We generated images of a realistic brain phantom containing the tissues: white matter, gray matter, and cerebrospinal fluid. A set of k-space data was acquired while varying two acquisition parameters: repetition time (TR) and echo time (TE). This data set was then used to reconstruct multiple images of different contrast. Customized variation of TR and TE allowed us to obtain different contrast weightings of signal values. We present results for 2-contrast and 3-contrast "combo" acquisitions and compare them with images from acquisitions with fixed TR and TE. Scan time reductions of 30%-52% were achieved. Artifacts stemming from non-uniform and tissue-dependent data weighting in the Fourier domain were minimized through systematic optimization of the order of phase encoding and of variation schemes for TR and TE. No obvious degradation of image quality and resolution was observed. In addition, we quantitatively analyzed preservation of contrast, image profiles of sharp tissue boundaries, and signal-to-noise-ratio