Total: 1
Purpose: This study aimed to develop a reliable whole-brain multi-parameter CEST imaging sequence at 3T. By overcoming the limitations of existing imaging techniques, such as low SNR, image distortion, and magnetic susceptibility artifacts, this research intended to facilitate clinical research on brain diseases. Methods: A whole-brain single-shot CEST sequence with True FISP readout,also called bSSFP was designed. The sequence included a pre-saturation module followed by a fast True FISP readout. The four-angle method was used to acquire B0, rB1, and T1 maps for CEST data correction. MRI experiments were carried out on five healthy volunteers using a 3T whole-body MRI system. Data processing involved motion correction, deep-learning denoising, B0 correction, neural network B1 correction, and four-pool Lorentz fitting. One participant underwent three scans over three days to calculate the coefficient of variation of CEST metrics in different brain regions and nuclei. Results: The CEST contrast of MTRLD and AREX with B1 correction, incorporating APT, NOE, and MT effects, was obtained within 9 minutes. Neural network B1 correction not only reduced the relative error of CEST images but also eliminated the inhomogeneous spatial distribution related to the B1 field. The coefficient of variation of CEST metrics in most brain regions was below 10%. Notably, no banding artifacts or magnetic susceptibility artifacts were observed, and the SAR value was within an acceptable range. Conclusion: Homogeneous, unbiased, multi-parameter whole-brain CEST imaging can be achieved within 9 minutes at 3T using a single-shot True FISP readout. This sequence enables rapid acquisition of high-SNR CEST images free from banding artifacts and magnetic susceptibility artifacts, making it suitable for clinical multi-parameter CEST imaging applications.