Real-time motion and retrospective coil sensitivity correction for CEST using volumetric navigators (vNavs) at 7T.

Esau Poblador Rodriguez,Sami Auno,Wolfgang Bogner,Korbinian Eckstein,Andre Van Der Kouwe,Barbara Dymerska,Stephan Gruber,Philipp Moser,Siegfried Trattnig

doi:10.1002/mrm.28555

Abstract

PurposeTo explore the impact of temporal motion‐induced coil sensitivity changes on CEST‐MRI at 7T and its correction using interleaved volumetric EPI navigators, which are applied for real‐time motion correction.MethodsFive healthy volunteers were scanned via CEST. A 4‐fold correction pipeline allowed the mitigation of (1) motion, (2) motion‐induced coil sensitivity variations, ΔB1‐, (3) motion‐induced static magnetic field inhomogeneities, ΔB0, and (4) spatially varying transmit RF field fluctuations, ΔB1+. Four CEST measurements were performed per session. For the first 2, motion correction was turned OFF and then ON in absence of voluntary motion, whereas in the other 2 controlled head rotations were performed. During post‐processing ΔB1‐ was removed additionally for the motion‐corrected cases, resulting in a total of 6 scenarios to be compared. In all cases, retrospective ∆B0 and ‐ΔB1+ corrections were performed to compute artifact‐free magnetization transfer ratio maps with asymmetric analysis (MTRasym).ResultsDynamic ΔB1‐ correction successfully mitigated signal deviations caused by head motion. In 2 frontal lobe regions of volunteer 4, induced relative signal errors of 10.9% and 3.9% were reduced to 1.1% and 1.0% after correction. In the right frontal lobe, the motion‐corrected MTRasym contrast deviated 0.92%, 1.21%, and 2.97% relative to the static case for Δω = 1, 2, 3 ± 0.25 ppm. The additional application of ΔB1‐ correction reduced these deviations to 0.10%, 0.14%, and 0.42%. The fully corrected MTRasym values were highly consistent between measurements with and without intended head rotations.ConclusionTemporal ΔB1‐ cause significant CEST quantification bias. The presented correction pipeline including the proposed retrospective ΔB1‐ correction significantly reduced motion‐related artifacts on CEST‐MRI.

Highlights

Chemical exchange-sensitive MRI techniques such as CEST1-3 and chemical exchange based on T1ρ or T2 relaxation enhancement provide an alternative chemically specific MRI contrast to MR spectroscopy, based on indirect detection of endogenous and exogenous molecules.[4,5,6,7,8,9,10] In contrast to MRSI, the encoding of the spectral dimension is slow
To explore the impact of temporal motion-induced coil sensitivity changes on CEST-MRI at 7T and its correction using interleaved volumetric EPI navigators, which are applied for real-time motion correction
Previous volumetric EPI navigators (vNavs) studies tackled the mitigation of subject motion simultaneously with a third type of motion-induced artifacts,[27,28,29,33,54] whereas we opted for a self-corrected dynamic ΔB0 method directly from the phase of the pre-saturated gradient echo (GRE) images.[43]

Summary

Introduction

Chemical exchange-sensitive MRI techniques such as CEST1-3 and chemical exchange based on T1ρ or T2 relaxation enhancement (ie, chemical exchange sensitive spin-lock, CESL) provide an alternative chemically specific MRI contrast to MR spectroscopy, based on indirect detection of endogenous and exogenous molecules.[4,5,6,7,8,9,10] In contrast to MRSI, the encoding of the spectral (chemically specific) dimension is slow. CEST and similar techniques rely on the comparison between images (ie, representing different spectral points) acquired at different time points. They reported motion-induced artifacts in the same order of magnitude as the measured CEST effects (ie, 1% for every 0.6 mm of translation and every 7 Hz of shift in the static magnetic field; B0)

Objectives

Methods

Findings

Discussion

Conclusion