0754

Mapping the myelin bilayer with short-T2 MRI: Validation studies
Emily Louise Baadsvik1, Markus Weiger1, Romain Froidevaux1, Wolfgang Faigle2, Benjamin Victor Ineichen3, and Klaas Paul Pruessmann1
1Institute for Biomedical Engineering, ETH Zurich and University of Zurich, Zurich, Switzerland, 2Neuroimmunology and MS Research Section, Neurology Clinic, University of Zurich, University Hospital Zurich, Zurich, Switzerland, 3Department of Neuroradiology, Clinical Neuroscience Center, University Hospital Zurich, University of Zurich, Zurich, Switzerland

Synopsis

Various techniques for myelin mapping based on signal from the lipid-protein bilayer have been proposed, and a common way of validating these techniques is using ex-vivo tissue samples. However, it is still unclear to what extent experimental factors such as tissue storage conditions and processing affect MR signal. In this work, we investigate how long-term deep-frozen storage impacts tissue signal, and evaluate whether signal component mapping is feasible in non-D2O-exchanged samples. We also determine whether animal tissue can act as a substitute for human tissue and investigate signal differences between white and grey matter.

Introduction

Several diseases such as multiple sclerosis critically disrupt the myelin sheath. Hence, noninvasively tracking myelin content by means of imaging poses an urgent medical need. However, despite several attempts at mapping the myelin bilayer directly using short-T2 MRI techniques1-7, it is not yet clear how best to interpret the observed signal in-vivo. Experimental approaches to signal analysis frequently involve the use of animal2,5,6,8 or human9 tissue samples where the background water signal is reduced by D2O exchange5,6,8,9, but this raises additional questions related to the preparation and storage of samples. Common practice is to freeze the tissue, but it is not yet known if this alters the MR signal like with formalin fixation10. Similarly, it is unclear how well signal properties of animal and human tissue match, and whether the employed analysis approaches would work in samples without D2O exchange.

In this work, we investigate the signal components observable with state-of-the-art short-T2 imaging techniques in healthy human brain samples to address how different experimental procedures affect tissue signal. Additionally, we investigate signal differences between white matter (WM) and grey matter (GM).

Methods

Samples
Eight tissue samples of approximately 5x25x18 mm3 from two regions of the cerebrum of a 56-year-old male were studied. The patient died of a cardiovascular event and neuropathological workup did not suggest gross brain pathology. Tissue preparation included D2O exchange and frozen storage for certain samples. A schematic of the full procedure is given in Figure 1.

Imaging
A 3T Philips Achieva system equipped with a high-performance gradient11, fast transmit/receive switches12 and a proton-free loop coil of 40 mm diameter were used. Two imaging protocols were applied, one using single-point imaging (SPI)13 and one using the zero-TE variant HYFI14. The SPI protocol consisted of 14 images at TE between 33-2000 µs with an isotropic resolution of 1.56 mm, which are used for model fitting. The HYFI images have an isotropic resolution of 0.39 mm. Details on the protocols can be found in Ref.9.

Analysis
The signal behaviour of the SPI series was analysed via a fitting procedure based on that used in Ref.6, consisting of the following steps:

  1. ROIs for WM were drawn in each sample, and in D2O samples also for GM.
  2. The average signal in each ROI was calculated and used for the fits in steps 3 and 4.
  3. The 3-component fitting model described in Ref.6 was applied to all D2O samples. This model splits the non-aqueous signal into an ultrashort (U) and a short (S) component of super-Lorentzian lineshape representing the myelin bilayer and residual non-aqueous content, respectively, and the aqueous signal is represented by one component (W) with a T2 of 50 ms and chemical shift (δ) of 4.7 ppm.
  4. The component parameters of the U- and S-components (T2,min and δ) were fixed at the average values found for WM in step 3, and a second fit was run on all samples to determine the amplitudes of each signal component in the different ROIs.
  5. The fixed-component fitting procedure from step 4 was applied on a single-voxel basis for all samples, yielding amplitude maps of the signal components.
For H2O samples, the two longest-TE data points were discarded for fitting because of artefacts related to the presence of a large water pool (TEmax 827 µs).

Interpolation was applied to all images. Amplitude maps for the non-aqueous components in H2O samples were masked based on the sample outline in the W-component map.

Histology
The tissue samples were cryosectioned (20 µm) and underwent myelin staining by immunohistochemistry for myelin oligodendrocyte glycoprotein (MOG, 1:50).

Results

SPI data points and fitted curves (analysis step 4) for several ROIs in tissue block 2 are shown in Figure 2.

The fitted values of T2,min and δ from step 3 are shown in Table 1. The amplitudes found in step 4 are given in Table 2.

Figure 3 shows photographs (both sides), MOG stains, reference HYFI images and the fitted amplitude maps from step 5 for all samples from tissue block 2.

Discussion and Conclusions

Based on data presented in Table 2 and Figure 2, we conclude that storing samples under frozen conditions appears to affect the T1 of the different signal components, which alters the steady-state signal magnitudes and the amplitude relationship of the U- and S-components. However, the component parameters remain largely unchanged. Thus, frozen samples are at least qualitatively equivalent to fresh samples.

From Table 1 we see that GM has different component parameters than WM, which reflects the different composition of the two tissue types. However, as seen in Figure 3, fixing fit components based on WM values produces reasonable amplitude maps also in GM.

The component parameters found in this study for human samples are similar to those found previously in animal brain tissue2,6. Therefore, animal brain tissue can be a useful phantom depending on the particular purpose of the study.

As seen in Figure 3, component mapping works well also in samples without D2O exchange, bringing model-fitting approaches significantly closer to in-vivo applications.

Overall, results obtained for the experimental procedures investigated here can act as a foundation for signal interpretation in in-vivo studies.

Acknowledgements

The authors would like to thank Karl Frontzek from the Institute of Neuropathology, University Hospital Zurich, Switzerland for providing the brain samples.

References

1. Nayak KS, Pauly JM, Gold GE, Nishimura DG. Imaging ultra-short T2 species in the brain. In Proceedings of the 8th Annual Meeting of ISMRM, Denver, USA, 2000. p.509.

2. Wilhelm MJ, Ong HH, Wehrli SL et al. Direct magnetic resonance detection of myelin and prospects for quantitative imaging of myelin density. Proc Natl Acad Sci. 2012;109:9605-9610.

3. Du J, Ma G, Li S et al. Ultrashort echo time (UTE) magnetic resonance imaging of the short T2 components in white matter of the brain using a clinical 3T scanner. NeuroImage. 2014;87:32-41.

4. Sheth V, Shao H, Chen J et al. Magnetic resonance imaging of myelin using ultrashort echo time (UTE) pulse sequences: Phantom, specimen, volunteer and multiple sclerosis patient studies. NeuroImage. 2016;136:37-44.

5. Seifert AC, Li C, Wilhelm MJ, Wehrli SL, Wehrli FW. Towards quantification of myelin by solid-state MRI of the lipid matrix protons. NeuroImage. 2017;163:358-367.

6. Weiger M, Froidevaux R, Baadsvik EL, Brunner DO, Rösler MB, Pruessmann KP. Advances in MRI of the myelin bilayer. NeuroImage. 2020;217:116888.

7. Waldman A, Rees J, Brock C, Robson M, Gatehouse P, Bydder G. MRI of the brain with ultra-short echo-time pulse sequences. Neuroradiology. 2003;45:887-892.

8. Fan SJ, Ma Y, Zhu Y et al. Yet more evidence that myelin protons can be directly imaged with UTE sequences on a clinical 3T scanner: Bicomponent T2* analysis of native and deuterated ovine brain specimens. Magn Reson Med. 2018;80:538-547.

9. Baadsvik EL, Weiger M, Froidevaux R, Faigle W, Ineichen BV, Pruessmann KP. Mapping myelin content in ex-vivo MS brain tissue using short-T2 MRI of the lipid-protein bilayer. In Proceedings of the 29th Annual Meeting of the ISMRM, Vancouver, Canada, 2021. p.2827.

10. Seifert AC, Umphlett M, Hefti M, Fowkes M, Xu J. Formalin tissue fixation biases myelin-sensitive MRI. Magn Reson Med. 2019;82:1504-1517.

11. Weiger M, Overweg J, Rösler MB et al. A high-performance gradient insert for rapid and short-T2 imaging at full duty cycle. Magn Reson Med. 2018;79:3256-3266.

12. Brunner DO, Furrer L, Weiger M et al. Symmetrically biased T/R switches for NMR and MRI with microsecond dead time. J Magn Reson. 2016;263:147-155.

13. Balcom BJ, MacGregor RP, Beyea SD, Green DP, Armstrong RL, Bremner TW. Single-point ramped imaging with T1 enhancement (SPRITE). J Magn Reson Ser A. 1996;123:131-134.

14. Froidevaux R, Weiger M, Rosler MB, Brunner DO, Pruessmann KP. HYFI: Hybrid filling of the dead-time gap for faster zero echo time imaging. NMR in biomedicine. 2021;34:e4493.

Figures

Figure 1: Flow chart of the processing steps for each sample. Tissue blocks from two regions (1 + 2) on the cerebrum were sliced into four samples each, half of which were imaged directly and half of which were stored at -80°C for 5 months. Of the four samples that were prepared either fresh (s) or frozen-thawed (z), one from each tissue block underwent a 2-step D2O exchange (D) while the other was placed in H2O (H). After imaging, the samples were cryosectioned for subsequent myelin immunohistochemistry.

Figure 2: Magnitude and phase of the SPI data and fitting results with fixed components for the samples from tissue block 2. Data points represent an average over large WM ROIs, except for the GM plot (red) for the fresh D2O sample. H2O and D2O samples are clearly distinguishable in both plots, with the phase curves exhibiting a characteristic shape for low water content. Frozen samples have reduced magnitude w.r.t. their fresh counterparts. GM and WM signal behaviour is similar but not identical.

Table 1: Fitted signal component values T2,min (μs) and chemical shift δ (ppm) in WM and GM ROIs for all D2O samples, as well as the average results per tissue type. The T2,min and δ values are generally larger in GM than WM, but still on the same order. In WM and partly in GM, there is more variation between the two tissue blocks than between fresh and frozen samples. This indicates that anatomical location affects the detectable signal components to some degree while freezing has little impact.

Table 2: Normalised fitted amplitudes for each signal component in all samples, where the T2,min and chemical shift values are fixed at the average WM results given in Table 1. Relative water content is much higher in H2O than D2O samples and higher in GM than WM. The U/S distribution is around 90/10 % of the non-aqueous signal in WM and 80/20 % in GM. Differences between fresh and frozen is only significant in H2O samples, where freezing increases relative water content and changes U/S in WM to 80/20 %.

Figure 3: Comparison of photographs, MOG immunohistochemistry, high-resolution HYFI images and fitted component amplitude maps for each brain sample from tissue block 2 (ordered according to Figure 1). Good correspondence in terms of sample geometry and WM/GM areas (arrowheads) is achieved for all images. The component maps are of similar quality for H2O and D2O samples, bringing the method significantly closer to in-vivo applications.

Proc. Intl. Soc. Mag. Reson. Med. 30 (2022)
0754
DOI: https://doi.org/10.58530/2022/0754