1118

Advances in direct myelin imaging

Markus Weiger¹, Romain Nicolas Froidevaux¹, David Otto Brunner¹, Manuela Barbara Rösler¹, and Klaas Paul Pruessmann¹

¹Institute for Biomedical Engineering, ETH and University of Zurich, Zurich, Switzerland

Synopsis

Direct imaging of myelin is of great interest for improved diagnosis of neurodegenerative diseases. However, MR signals from myelin exhibit ultra-short T2 values in a range of 8 μs - 1 ms with a large fraction below 100 μs. Due to restrictions in sequence timing, current short-T2 imaging approaches cannot sufficiently capture these very short signals. In the present work, advanced short-T2 methodology and hardware are employed to actually image the majority of the ultra-short-T2 components in the brain. The abilities of the approach are proven in excised brain tissue and applied in vivo.

Introduction

Depiction and quantification of myelin is of great interest for the diagnosis of various neurodegenerative diseases. With MRI this is conventionally approached by using indirect techniques based on myelin water or magnetisation transfer¹. However, due to limited specificity of such surrogate measures, direct detection of the myelin signal itself would be preferable². The challenge for this task arises from the restricted mobility of the main constituents – lipids and proteins – of the myelin membrane. Their proton signals exhibit ultra-short T2 (uT2) values below 1 ms^3-5 with a large fraction below 100 μs, including contributions as short as 8 μs⁶. Therefore, dedicated short-T2 techniques have been employed to directly image the uT2 signals of myelin^2,6-9.

However, in all current approaches the timing of the MR sequence is not fully suited to sufficiently capture and spatially encode the large portion of signals with T2s below 100 μs. In particular, either extended excitation pulses, echo times (TE), gradient ramping, or readout times lead to considerable loss or blurring of such signals in the image^2,10.

In the present work, advanced short-T2 methodology and hardware are employed to actually image the majority of the uT2 components in the brain. The abilities of the approach are proven in excised brain tissue and applied in vivo.

Methods

The basic requirements for proper depiction of uT2 components are short delays between signal excitation and acquisition and rapid spatial encoding. These were met by using RF hardware capable of creating short excitation pulses and switching rapidly from transmit to receive operation, and by employing a high-performance gradient system. These capabilities were exploited in short-T2 sequences tailored to achieve suitable resolution and SNR at the targeted T2s within useful scan times.

Hardware: 3T Philips Achieva; insert gradient providing strength G = 200 mT/m with slew rate 600 mT/m/ms at 100 % duty cycle¹¹; custom-built RF chain¹²; rapid T/R switches¹³; proton-free loop and birdcage coils^14,15.

Imaging: Sequences were used that maximally exploit gradient strength by performing both excitation and acquisition under the gradient employed for radial centre-out encoding¹⁶. In particular, imaging with hybrid filling (HYFI)^17,18 and single-point imaging (SPI)^19,20 were performed. Both techniques reduce the influence of T2* decay on image blurring by providing a plateau of (near-)equal T2* weighting in central k-space at delay TE. See Table 1 for the protocols employed.

Fitting: To prove the existence of uT2 components in the data, a series of SPI images with different TE was acquired. The observed signal evolution was fitted with a complex-valued model^21,22 including components with Super-Lorentzian (SL)⁶ and Lorentzian lineshapes (Figure 2).

Samples and subjects: Porcine brain tissue (thickness 5 mm, diameter 30 mm) at room temperature with and without D2O exchange⁶. One healthy human subject.

Results

Figure 1 shows results obtained in the D2O-exchanged tissue sample. The HYFI image with strongly reduced acquisition range exhibits improved resolution and contrast, indicating dominating uT2 signal. This is confirmed by the rapid signal decay in the SPI series at increasing TE. In the fitted amplitude maps the shortest and strongest component is most characteristic for white matter (WM), indicating a large contribution from myelin.

Figure 2 shows signal evolution and fit results in WM pixels of the SPI series of Figure 1. It demonstrates that a SL component with minimum T2 of 8 μs dominates the signal in the sample. Note that a SL covers a continuum of T2 values over several orders of magnitude.

Figure 3 shows the equivalent results to Figure 1 for the non-D2O-exchanged sample. Here, long-T2 signals dominate the images. Nevertheless, a fitted amplitude map can be generated, exhibiting WM-GM contrast that is comparable to the D2O-exchanged sample.

Figure 4 demonstrates in vivo application of the presented technique capable of imaging large portions of the uT2 signal in the brain. In the raw images, WM signal is clearly reduced at longer TE. The subtraction image showing primarily uT2 signals exhibits a clear WM-GM contrast.

Discussion

It was demonstrated that advanced short-T2 methodology and hardware actually enable imaging of the majority of the ultra-short T2 components in the brain. As these signals are dominated by the myelin membrane⁶, it is expected that the intensities in the presented images have a high specificity for myelin content. Hence, the presented approach may contribute to improved myelin quantification by MRI. In-vivo feasibility was demonstrated but is currently limited by SNR, SAR, and scan time constraints. Improvements are expected from more advanced RF coil design and further sequence optimisation. Moreover, expanding in-vivo scanning to more than two TEs is desirable to enable signal modelling as demonstrated in the in-vitro data.

Acknowledgements

References

[1] C. Laule, I.M. Vavasour, S.H. Kolind, D.K.B. Li, T.L. Traboulsee, G.R.W. Moore, A.L. MacKay, Magnetic Resonance Imaging of Myelin, Neurotherapeutics, 4 (2007) 460-484.

[2] A.C. Seifert, C. Li, M.J. Wilhelm, S.L. Wehrli, F.W. Wehrli, Towards quantification of myelin by solid-state MRI of the lipid matrix protons, NeuroImage, 163 (2017) 358-367.

[3] M. Bloom, K.T. Holmes, C.E. Mountford, P.G. Williams, Complete proton magnetic resonance in whole cells, Journal of Magnetic Resonance (1969), 69 (1986) 73-91.

[4] A. Ramani, A. Aliev, G. Barker, P. Tofts, Another approach to protons with constricted mobility in white matter: pilot studies using wideline and high-resolution NMR spectroscopy, Magn Reson Imag, 21 (2003) 1039-1043.

[5] R.A. Horch, J.C. Gore, M.D. Does, Origins of the ultrashort‐T2 1H NMR signals in myelinated nerve: a direct measure of myelin content?, Magn Reson Med, 66 (2011) 24-31.

[6] M.J. Wilhelm, H.H. Ong, S.L. Wehrli, C. Li, P.H. Tsai, D.B. Hackney, F.W. Wehrli, Direct magnetic resonance detection of myelin and prospects for quantitative imaging of myelin density, Proc Natl Acad Sci, 109 (2012) 9605-9610.

[7] K.S. Nayak, J.M. Pauly, G.E. Gold, D.G. Nishimura, Imaging ultra-short T2 species in the brain, In Proceedings of the 8th Annual Meeting of ISMRM, Denver, USA, 2000, p.509

[8] A. Waldman, J. Rees, C. Brock, M. Robson, P. Gatehouse, G. Bydder, MRI of the brain with ultra-short echo-time pulse sequences, Neuroradiology, 45 (2003) 887-892.

[9] J. Du, G. Ma, S. Li, M. Carl, N.M. Szeverenyi, S. VandenBerg, J. Corey-Bloom, G.M. Bydder, Ultrashort echo time (UTE) magnetic resonance imaging of the short T2 components in white matter of the brain using a clinical 3T scanner, NeuroImage, 87 (2014) 32-41.

[10] S.-J. Fan, Y. Ma, Y. Zhu, A. Searleman, N.M. Szeverenyi, G.M. Bydder, J. Du, 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, (2017).

[11] M. Weiger, J. Overweg, M.B. Rösler, R. Froidevaux, F. Hennel, B.J. Wilm, A. Penn, U. Sturzenegger, W. Schuth, M. Mathlener, M. Borgo, P. Börnert, C. Leussler, R. Luechinger, B.E. Dietrich, J. Reber, D.O. Brunner, T. Schmid, L. Vionnet, K.P. Pruessmann, A high-performance gradient insert for rapid and short-T2 imaging at full duty cycle, Magn Reson Med, 79 (2018) 3256-3266.

[12] M. Weiger, D.O. Brunner, B.E. Dietrich, C.F. Muller, K.P. Pruessmann, ZTE imaging in humans, Magn Reson Med, 70 (2013) 328-332.

[13] D.O. Brunner, L. Furrer, M. Weiger, W. Baumberger, T. Schmid, J. Reber, B.E. Dietrich, B.J. Wilm, R. Froidevaux, K.P. Pruessmann, Symmetrically biased T/R switches for NMR and MRI with microsecond dead time, J Magn Reson, 263 (2016) 147-155.

[14] M.B. Rösler, M. Weiger, T. Schmid, D.O. Brunner, R. Froidevaux, K.P. Pruessmann, Ultrasonic soldering on glass for the construction of MRI coils with minimized background signal in short-T2 images, In Proceedings of the 33rd Annual Scientific Meeting of ESMRMB, Vienna, Austria, 2016, p.87

[15] M.B. Rösler, M. Weiger, D.O. Brunner, T. Schmid, R. Froidevaux, K.P. Pruessmann, An RF birdcage coil designed for an insert gradient coil dedicated to short-T2 MRI, In Proceedings of the 25th Annual Scientific Meeting of ISMRM, Honolulu, Hawaii, USA, 2017, p.2668

[16] M. Weiger, K.P. Pruessmann, MRI with zero echo time, in: eMagRes, John Wiley & Sons, Ltd, Chichester, United Kingdom, 2012, pp. 311-322.

[17] R. Froidevaux, M. Weiger, K.P. Pruessmann, HYFI: Hybrid filling for improved scan efficiency in zero echo time imaging with large dead-time gaps, In Proceedings of the 34th Annual Meeting of ESMRMB, Barcelona, Spain, 2017, p.303

[18] R.N. Froidevaux, M.B. Rösler, D.O. Brunner, M. Weiger, K.P. Pruessmann, HYFI: Hybrid filling of the dead-time gap for faster zero echo time imaging, In Proceedings of the 27th Annual Meeting of ISMRM, Montreal, Canada, 2019

[19] S. Emid, J. Creyghton, High resolution NMR imaging in solids, Physica B+ C, 128 (1985) 81-83.

[20] B.J. Balcom, R.P. MacGregor, S.D. Beyea, D.P. Green, R.L. Armstrong, T.W. Bremner, Single-point ramped imaging with T1 enhancement (SPRITE), J Magn Reson Ser A, 123 (1996) 131-134.

[21] Y. Nam, J. Lee, D. Hwang, D.-H. Kim, Improved estimation of myelin water fraction using complex model fitting, NeuroImage, 116 (2015) 214-221.

[22] T. Boucneau, P. Cao, S. Tang, M. Han, D. Xu, R.G. Henry, P.E.Z. Larson, In vivo characterization of brain ultrashort-T2 components, Magn Reson Med, 80 (2018) 726-735.

Figures

Table 1: Sequence parameters of the five protocols used for isotropic 3D imaging of the uT2 components of the brain. SPI: single point imaging; HYFI: hybrid Cartesian and radial filling of central k-space leading to a T2* plateau (see below); FOV: field of view; Bandwidth: signal bandwidth in FOV; G: gradient strength; TE: echo time corresponding to delay between signal excitation (pulse centre) and acquisition on plateau (including k = 0); Acquisition range: time window engaged by overall data acquisition; Plateau: central fraction of k-space radius acquired with (near-)equal T2* weighting; TR: repetition time.

Figure 1: Short-T2 imaging of porcine brain tissue. By D2O-exchange, long-T2 water was largely removed, leaving primarily the uT2 components. a) The HYFI image at conventional G shows WM-GM contrast but is blurred. Shortening the acquisition range by increasing G and the k-space plateau leads to sharper depiction and stronger contrast, indicating the presence of mainly uT2 components. b) The series of SPI images with increasing TE shows rapid signal decay, confirming dominating signal from uT2 components. c) Fitting the signal evolution (see Figure 2) provides amplitude maps for three components, where the shortest exhibits strong WM-GM contrast (SL=Super-Lorentzian, L=Lorentzian).

Figure 2: Signal evolution of the SPI image series in Figure 1b. Experimental data was averaged over 16 WM pixels. Magnitude and phase show non-exponential and non-linear behaviour, respectively. The data was fitted using a complex-valued model with three signal components, assuming two Super-Lorentzians (SL), characterised by their minimum T2, and one Lorentzian (L). Individual chemical shifts were derived from spectroscopy data (0.43, 0.35, 0 ppm) and T2 of the Lorentzian was set to 50 ms. The fitted values (table) show dominance of the shortest component. To obtain the fitted amplitude maps in Figure 1c, T2 values were held constant.

Figure 3: Short-T2 imaging of excised porcine brain tissue without D2O exchange in analogy to Figure 1. a) Shortening the acquisition range in protocol 2 does not notably change image quality as long-T2 water dominates both images. b) Nevertheless, SPI data at very long TE shows reduced WM signal, indicating decay of the uT2 components. c) The fitted amplitude of the short Super-Lorentzian (SL) component clearly reflects locations of WM and GM in agreement with a). Hence, imaging and mapping performance of uT2 components in untreated and D2O-exchanged tissue is comparable.

Figure 4: In-vivo imaging of the uT2 components in the human brain. Selected views from the 3D isotropic data set are shown. Images at shortest TE of 15 μs and short acquisition range of 29 μs show little WM-GW contrast. Increasing TE leads to reduced signal in WM. Subtraction of the two data sets largely removes long-T2 signal, showing predominantly the uT2 components with a clear WM-GW contrast.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)

1118