Accelerating whole brain vessel wall imaging of isotropic 0.4 mm3 on 5T by 10-fold using deep learning reconstruction

Motivation: Whole brain vessel wall imaging (VWI) of isotropic 0.4 mm3 on 3T can’t utilize higher than 5-fold acceleration to reduce the scan time due to insufficient signal-to-noise. Goal(s): To achieve 10-fold accelerated whole brain VWI of isotropic 0.4 mm3 on the 5T scanner with a 48-channel transmit receive head coil. Approach: Deep learning (DL) reconstruction equipped with 3D convolution neural network was developed to alleviate the nonuniform noise amplified by SPIRiT reconstruction and the B1 inhomogeneity of 5T scanner. Results: The proposed DL SPIRiT reconstruction achieves 10-fold accelerated intracranial VWI scan on 5T in 6 minutes and give better VWI quality than 3T. Impact: This work develops a 10-fold accelerated whole brain vessel wall imaging of isotropic 0.4 mm3 in 6 minutes using deep learning (DL) unrolled SPIRiT reconstruction on the 5T scanner equipped with a 48-channel transmit receive head coil.