Fast 3D chemical exchange saturation transfer imaging with variably-accelerated sensitivity encoding (vSENSE)

Magn Reson Med. 2019 Dec;82(6):2046-2061. doi: 10.1002/mrm.27881. Epub 2019 Jul 1.

Abstract

Purpose: To extend the variably-accelerated sensitivity encoding (vSENSE) method from 2D to 3D for fast chemical exchange saturation transfer (CEST) imaging, and prospectively implement it for clinical MRI.

Methods: The CEST scans were acquired from 7 normal volunteers and 15 brain tumor patients using a 3T clinical scanner. The 2D and 3D "artifact suppression" (AS) vSENSE algorithms were applied to generate sensitivity maps from a first scan acquired with conventional SENSE-accelerated 2D and 3D CEST data. The AS sensitivity maps were then applied to reconstruct the other CEST frames at higher acceleration factors. Both retrospective and prospective acceleration in phase-encoding and slice-encoding dimensions were implemented.

Results: Applying the 2D AS vSENSE algorithm to a 2-fold undersampled 3.5-ppm CEST frame halved the scan time of conventional SENSE, while generating essentially identical reconstruction errors (p ≈ 1.0). The 3D AS vSENSE algorithm permitted prospective acceleration by up to 8-fold, in total, from phase-encoding and slice-encoding directions for individual source CEST images, and an overall speed-up in scan time of 5-fold. The resulting vSENSE-accelerated amide proton transfer-weighted images agreed with conventional 2-fold-accelerated SENSE CEST results in brain tumor patients and healthy volunteers. Importantly, the vSENSE method eliminated unfolding artifacts in the slice-encoding direction that compromised conventional SENSE CEST scans.

Conclusion: The vSENSE method can be extended to 3D CEST imaging to provide higher acceleration factors than conventional SENSE without compromising accuracy.

Keywords: amide proton transfer (APT); artifact suppression (AS); chemical exchange saturation transfer (CEST); fast imaging; variably accelerated sensitivity encoding (vSENSE).

Publication types

  • Research Support, N.I.H., Extramural
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Algorithms
  • Artifacts
  • Brain / diagnostic imaging*
  • Brain Neoplasms / diagnostic imaging*
  • Humans
  • Image Interpretation, Computer-Assisted / methods
  • Image Processing, Computer-Assisted / methods*
  • Imaging, Three-Dimensional / methods*
  • Magnetic Resonance Imaging*
  • Magnetic Resonance Spectroscopy
  • Reproducibility of Results
  • Sensitivity and Specificity
  • Signal Processing, Computer-Assisted