Novel Methods For Spatiotemporally Encoded Single-shot MRI And Chemical Exchange Saturation Transfer MRI

Spatiotemporally encoded MRI(SPEN MRI)is a recently proposed ultrafast technique.Compared to EPI,SPEN MRI remains the outstanding performance in imaging speed,while possesses much better resistance to inhomogeneous B0 field and chemical shift effects.The advantages of SPEN MRI show enormous potential in clinical applications.Limited by the quadratic phase encoding,the data of SPEN MRI cannot be directly reconstructed through conventional fast Fourier transform.Super-resolved reconstruction(SR)is a data post-processing method designed for SPEN MRI aiming at improving spatial resolution without additional acquisition.To date,many SR algorithms have been proposed,such as conjugate gradient,partial Fourier transform,de-convolution algorithm and compressed sensing based reconstruction.Though these algorithms can improve the spatial resolution to some extent,the images reconstructed by these algorithms do not reach the optimal resolution in theory.In this thesis,we proposed an optimal SR algorithm SEED(super-resolved enhancing and edge deghosting).SEED algorithm can retrieve the optimal resolution in theory by fully exploiting the quadratic phase encoding information.Besides,SEED can also be applied to multi-shot SPEN MRI,which can reduce half of acquisition points without obvious degradation in image quality.Compared to Cartesian sampling scheme,non-Cartesian sampling has a great advantage in sampling efficiency and feasibility,resistance to aliasing artifacts induced by undersampling.However,few studies relating to non-Cartesian SPEN MRI have been carried out so far.In this thesis,we firstly proposed a method to design the sampling trajectory and decoding gradient for non-Cartesian SPEN MRI.A simple and efficient non-Cartesian SR algorithm was also proposed for non-Cartesian SPEN MRI,which can greatly improve the image resolution.Compared to Cartesian SPEN MRI,non-Cartesian SPEN MRI can yield better image quality in regard of spatial resolution and aliasing artifacts.In this thesis,we also carried out a feasibility study on referenceless geometric distortion correction and T2*relaxation correction for non-Cartesian SPEN MRI,which will benefit the image quality of non-Cartesian SPEN MRI.Chemical exchange saturation transfer(CEST)MRI,as a novel molecular imaging technique,has received great attention in recent years because of its potential to detect low concentration proteins and metabolites in vivo.The technique has been successfully applied to detect pathological chemical changes in many diseases,such as neurological disorders in human brain,animal models of cerebral ischemia and cancer in both animal models and humans.Though there are many successful applications,CEST MRI suffers from the lack of specificity in the Z-spectrum of endogenous exchanging protons.In this thesis,we proposed a new CEST technique based on off-resonance variable delay multi-pulse.Combined with the fitting based on Bloch equation,this novel method can be used to separate out the fast-and slow-exchanging components in the Z-spectrum.Besides,we proposed a localized lorentzian line-shape fitting method to extract the creatine signal in the Z-spectrum under 11.7 T.Combined with in vivo MRS,this method can be used to quantify the creatine signal and obtain the creatine concentration.A high-resolution creatine map was obtained via this method.