English
 
Help Privacy Policy Disclaimer
  Advanced SearchBrowse

Item

ITEM ACTIONSEXPORT
  DeepCEST: fast mapping of 7T CEST MRI parameters with uncertainty quantification

Hunger, L., Rajput, J., Fabian, M., Mennecke, A., Glang, F., Schmitt, M., et al. (2022). DeepCEST: fast mapping of 7T CEST MRI parameters with uncertainty quantification. In 24. Jahrestagung der Deutschen Sektion der ISMRM (DS-ISMRM 2022) (pp. 10-11).

Item is

Basic

show hide
Genre: Meeting Abstract

Files

show Files

Creators

show
hide
 Creators:
Hunger, L, Author
Rajput, J, Author
Fabian, MS, Author
Mennecke, AB, Author
Glang, FM1, Author                 
Schmitt, M, Author
Dörfler, A, Author
Maier, A, Author
Zaiss, M1, Author                 
Affiliations:
1Department High-Field Magnetic Resonance, Max Planck Institute for Biological Cybernetics, Max Planck Society, ou_1497796              

Content

show
hide
Free keywords: -
 Abstract: Introduction
To make 7T CEST MRI more available for radiologists, we developed a deepCEST pipeline for 7T MRI that predicts CEST
contrasts from just one scan with robustness against B1 inhomogeneities. The pipeline includes an uncertainty quantification
and a confidence map to evaluate the quality of the predictions. The proposed approach results in a reduction of 50% of the
measurement time and delivers the predicted CEST contrast with in 1 sec.
Methods
The input data for a neural network (NN) consisted of 7T in vivo raw Z-spectra of a single B1 level, and a B1 map. The 7T raw
data was acquired using the 3D snapshot GRE MIMOSA CEST3 at a Siemens MAGNETOM 7T scanner. These inputs were
mapped voxel-wise on target data consisting of Lorentzian amplitudes conventionally generated by 5-pool-Lorentzian fitting
performed on normalized, denoised, B0- and B1-corrected Z-spectra. The network consisted of two fully connected hidden
layers with RELU activation and was trained with Gaussian negative log likelihood loss. The output layer consisted of 10
nodes with linear activation to obtain the amplitudes and uncertainty of the 5-pool Lorentzian fit.
Results
Figure 1a, b, d and e shows the Lorentzian fit and the prediction of the amide and rNOE contrast in a tumor patient. Figure
1c, f shows the segmented uncertainty map over all contrasts with a threshold of 10%. The first row shows the predictions
and uncertainty for the measurement made with a bad shim. Such a strong B0 shift was not part of the training distribution.
Therefore, the predictions and fits do not match. Consequently, the NN outputs a high uncertainty for these voxels (Fig. 1c).
In the second row of (Fig. 1), the Z-spectra of the patient was centered, resulting in NN predictions that agree well with the
fit, and only a low uncertainty is yielded.
Discussion
The deepCEST approach has already shown very promising results for 3T, the clear advantage of 7T data is the better SNR
and higher spectral resolution. The 7T deepCEST approach uses only one B1 level, this saves about 50% of scan time (now
6:42 min), but still predicts accurately with low uncertainty (Fig. 2) and provides both B0- and B1-corrected homogeneous
CEST contrast.

Details

show
hide
Language(s):
 Dates: 2022-09
 Publication Status: Published online
 Pages: -
 Publishing info: -
 Table of Contents: -
 Rev. Type: -
 Identifiers: -
 Degree: -

Event

show
hide
Title: 24. Jahrestagung der Deutschen Sektion der ISMRM (DS-ISMRM 2022)
Place of Event: Aachen, Germany
Start-/End Date: 2022-09-21 - 2022-09-24

Legal Case

show

Project information

show

Source 1

show
hide
Title: 24. Jahrestagung der Deutschen Sektion der ISMRM (DS-ISMRM 2022)
Source Genre: Proceedings
 Creator(s):
Affiliations:
Publ. Info: -
Pages: - Volume / Issue: - Sequence Number: V004 Start / End Page: 10 - 11 Identifier: ISBN: 978-3-948023-28-7