Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Mar 11.
Published in final edited form as: NMR Biomed. 2020 Jul 20;33(10):e4326. doi: 10.1002/nbm.4326

Volumetric Imaging of Myelin in vivo Using 3D Inversion-Recovery Ultrashort Echo Time Cones Magnetic Resonance Imaging

Ya-Jun Ma 1,#, Adam C Searleman 1,#, Hyungseok Jang 1, Shu-Juan Fan 1, Jonathan Wong 2,1, Yanping Xue 1, Zhenyu Cai 1, Eric Y Chang 2,1, Jody Corey-Bloom 3, Jiang Du 1
PMCID: PMC7952008  NIHMSID: NIHMS1676461  PMID: 32691472

Abstract

Direct myelin imaging is promising for characterization of multiple sclerosis (MS) brains at diagnosis and in response to therapy. In this study, a 3D inversion recovery prepared ultrashort echo time cones (IR-UTE-Cones) was used for both morphological and quantitative imaging of myelin on a clinical 3T scanner. Myelin powder phantoms with different myelin concentrations were imaged with the 3D UTE-Cones sequence and it showed a strong correlation between concentrations and UTE-Cones signals, demonstrating the ability of the UTE-Cones sequence to directly image myelin in the brain. Quantitative myelin imaging with multi-echo IR-UTE-Cones sequences show similar T2* values for a D2O-exchanged myelin phantom (T2*=0.33±0.04ms), ex vivo brain specimens (T2*=0.20±0.04ms), and in vivo healthy volunteers (T2*=0.254±0.023 ms), further confirming the feasibility of 3D IR-UTE-Cones for direct myelin imaging in vivo. In ex vivo MS brain study, signal loss is observed in MS lesions, which was confirmed with histology. For the in vivo study, the lesions in MS patients also show myelin signal loss using the proposed direct myelin imaging method, demonstrating the clinical potential for MS diagnosis. Furthermore, the measured IR-UTE-Cones signal intensities show a significant difference between normal-appearing white matter in MS patients and normal white matter in volunteers, which cannot be found in clinical used T2-FLAIR sequences. Thus, the proposed 3D IR-UTE-Cones sequence showed clinical potential for MS diagnosis with the capability in direct myelin detection of whole brain.

Keywords: Myelin, Ultrashort Echo Time, Cones

INTRODUCTION

Myelin is a fundamental component of the nervous system, defined by its ultrastructure of multiple lamellae of protein-rich lipid bilayers which insulate axons to facilitate saltatory conduction (1). It is comprised of approximately 40–45% lipids/cholesterol, 10–15% protein, and 40% water (1). The extent of myelination modulates axonal function and health, and is dynamically regulated throughout life in response to a variety of neurologic conditions (2,3). Multiple sclerosis (MS) is defined by the idiopathic formation of demyelinating lesions disseminated in time and space with associated focal neurological deficits, followed by either progression, partial recovery, or full recovery (4). However, MS lesions seen on clinical MRI sequences are nonspecific (i.e. edema or gliosis without demyelination) and may not correlate with the patient’s neurological deficits (5). Additionally, remyelination can occur concurrently with demyelination and is often incomplete, leading to the formation of inactive shadow plaques that are not easily detected with current clinical MR sequences (6). Hence, there is a clinical need for noninvasive myelin imaging for better characterization of MS lesions at diagnosis and in response to therapy.

Ultrashort echo time (UTE) sequences can directly detect signals from myelin protons, which are normally not detectable by conventional MR sequences because of their ultrashort T2 (< 1 ms) (7–10). Multiple advanced MR sequences can indirectly image myelin and have been shown to be sensitive to acute demyelination in MS, including magnetization transfer (MT)-based imaging methods (11–13), T2 relaxometry/myelin water fraction imaging (14–17), and diffusion-based imaging methods (18,19). However, UTE-based methods may be more specific than such indirect methods of myelin quantification in the setting of the heterogeneous pathological changes seen in MS (20–22).

Although myelin is the predominant short T2 component in the brain, over 90% of the UTE signal originates from long T2 water protons, even in myelin-rich white matter (9,23,24). Previous studies have demonstrated that adiabatic inversion recovery (IR) preparation can achieve robust and uniform suppression of the long T2 signals in white matter, allowing for selective imaging of myelin protons on a clinical 3T scanner (9,23,24). This observation has been corroborated by multiple studies showing the similarity of the 2D IR-UTE signal to the 2D UTE signal in white matter after near-complete elimination of the water signal by sequential D2O exchange (24–26). The 2D IR-UTE sequence has been shown to detect MS lesions not previously detected with clinical sequences (27,28), and has demonstrated its ability to obtain high myelin contrast on a clinical 3T scanner (9,27,29). However, 2D UTE sequences are susceptible to eddy current distortion artifacts during slice selection and have inefficient excitation of short T2 components from the relatively long half-sinc excitation pulse. Furthermore, it is difficult to cover the whole brain using 2D UTE sequences, which are subject to stronger eddy currents and gradient distortion for off-center slices.

3D UTE sequences have been developed for volumetric imaging of short T2 tissues (30,31) using short, rectangular excitation pulses for nonselective excitation with greater excitation efficiency and greatly reduced eddy current artifacts. However, the increase in required scan time resulting from center-out sampling of 3D k-space is particularly pronounced with IR preparation (32). To reduce scan time, the 3D spiral sampling trajectory (3D Cones) was recently implemented for more efficient k-space coverage (33), and multiple acquisitions (“spokes”) per IR preparation (34,35) were implemented for time-efficient 3D IR-UTE-Cones imaging of cortical bone and tendons (34–36).

In this study, we performed a comprehensive study to evaluate the use of multi-spoke adiabatic inversion recovery prepared UTE with 3D Cones sampling (3D IR-UTE-Cones) for morphological and quantitative volumetric imaging of myelin on a clinical 3T scanner. First, we evaluated the dynamic sensitivity of the 3D UTE Cones sequence over a physiological range of myelin concentrations in D2O. Second, we imaged cadaveric MS brain specimens with 3D IR-UTE-Cones, and correlated abnormalities with histopathology. Third, we compared 3D IR-UTE-Cones with conventional clinical MR sequences in vivo in healthy volunteers and MS patients. Fourth, the signal intensities in both the normal white matter (NWM) in volunteers and white matters in MS patients (including both MS lesions and normal-appearing white matter (NAWM)) were also measured for quantitative comparison.

METHODS

3D IR-UTE-Cones Pulse Sequence

For efficient volumetric imaging of short T2 components, multiple UTE acquisitions (specified by the number of spokes (Nsp) and time to the start of the next spoke (τ)) are obtained after each IR preparation (Fig. 1A), with inversion times (TI) defined as the average spoke TI. As short T2 components such as myelin are effectively saturated following adiabatic IR pulses (35,38,39), the long T2 suppression achieved is highly dependent on TI. Each UTE acquisition (Fig. 1B) uses a short, rectangular radiofrequency pulse (e.g. from 26 to 52μs) for non-selective excitation, following a minimal nominal TE by efficient sampling from the center of k-space using spiral trajectories. A second echo is also obtained to detect residual long T2 signals after extinction of the myelin signal. The spiral trajectories have conical ordering for more efficient sampling of 3D k-space compared with radial trajectories (33), and can have an anisotropic field-of-view (FOV) for higher in-plane resolution and thicker slices (Fig. 1C). The combination of 3D conical trajectories and multi-spoke acquisition allows for volumetric imaging of short T2 components in a time-efficient manner. By choosing a TI such that the long-T2 signal in white matter is nulled (Fig. 1D), the remaining UTE signal in white-matter tracts will originate from short T2 components, namely myelin. Dual-echo subtraction diminishes residual long T2 signals in gray matter (9,27,32). For the purposes of this study, we imaged myelin phantoms (see preparation details below), cadaveric brains, and in vivo brains.

Figure 1.

Figure 1

Open in a new tab

Pulse sequence for 3D IR-UTE-Cones sequence. (A) Multi-spoke acquisition following adiabatic inversion pulse preparation allows time-efficient sampling at longer average TIs, thus greatly reducing the total scan time. Nsp: number of spokes. Here, TR is defined as the time until the start of the subsequent IR pulse, and tau is the time until the start of the next spoke. (B) A short rectangular pulse allows for efficient non-selective excitation and reduces the influence of eddy currents. Multi-echo acquisition begins after a minimal nominal TE of 32 us. (C) Spiral center-out sampling trajectories are arranged with conical ordering for efficient 3D k-space sampling. (D) The adiabatic IR pulse provides robust inversion of the longitudinal magnetizations of the long T2 components in WML and GM, and saturates myelin because its T2* is much shorter than the IR pulse duration. UTE acquisition starts when WML reaches the null point, leaving signals from myelin and residual GM to be detected. The second echo contains long T2* signals from GM; myelin signals are not detected due to its ultrashort T2*. Subtraction of the second echo from the first provides selective imaging of myelin.

Sample Preparation

Myelin lipid powder (type 1 bovine brain lipid extract, Sigma-Aldrich B1502, St. Louis, MO, USA) was resuspended in 99.9% D2O (Sigma-Aldrich B1502, St. Louis, MO, USA), then lyophilized to remove residual water and solvent contamination. The powder was then resuspended in D2O at concentrations of 6%, 9%, 12%, 18%, and 24% w/v to approximate the physiological range of myelin concentrations of white matter. The samples were loaded into 1.0 mL syringes along with a D2O-only control (0%) prior to imaging.

Cadaveric Brain Study

Four cadaveric heads (two controls without documented neuropathology (82-year-old male and 87-year-old female) and two MS subjects (45-year-old male and 56-year-old female)) were obtained from a nonprofit whole-body donation company (United Tissue Network, Phoenix, AZ, USA) and were stored at −80°C. The specimens were thawed in water at 4°C overnight, then thawed in gently agitated water at room temperature for 18 hours for temperature equilibration. After imaging, the specimen was refrozen at −80°C, then cut into 1-cm axial sections using a band saw (B16, Butcher Boy Machines, Selmer, TN, USA). Regions of interest (ROIs) were identified on review of the images, located on the gross slice of brain, and resected. Samples were fixed in zinc-formalin (Anatech, Battle Creek, MI) for 1 week, paraffin-embedded, and sectioned at 5- and 10-μm thicknesses. Slides were stained overnight in Luxol Fast Blue (LFB) at 60°C and briefly counterstained with neutral red.

In Vivo Study

A total of 12 healthy volunteers (ages 25–69, 6 female and 6 male) and 12 MS patients (ages 39–71, 9 female and 3 male) were recruited for this study, which was reviewed and approved by the UC San Diego Institutional Review Board (IRB). Written informed consent approved by the IRB was obtained prior to each subject’s participation. T2* measurement was only performed on five healthy volunteers due to the prolonged scan time required of this method.

MRI Protocol

The 3D IR-UTE-Cones sequence was implemented on a 3T scanner (MR750, GE Healthcare, Milwaukee, WI, USA) (34). An adiabatic inversion pulse (Silver-Hoult with a duration of 6.048 ms, bandwidth of 1.643 kHz, and maximum B1 amplitude of 17 μT) was used for robust inversion and suppression of the longitudinal magnetization of long T2 components in white matter. The myelin phantoms were imaged simultaneously in a 30-mL birdcage coil using the following parameters: TR = 1000 ms, TE = 0.032 ms, flip angle (FA) = 20°, bandwidth = 41.5 kHz, FOV = 6×6×4 cm3, matrix = 128×128×20. An identical sequence was used to image a 30-mL phantom of 12 mM MnCl2 in 20% H2O/80% D2O for normalization. For T2* measurement, a 15% (w/v) myelin/D2O phantom was imaged using TR = 80 ms; TEs = 0.032, 0.2, 0.4, 0.8, and 2.2 ms; FA = 20°; bandwidth = 41.5 kHz; FOV=4×4×4.8 cm3; matrix = 128×128×24; scan time = 25.6 min.

A 12-channel receive-only head coil was used for all human brains. For the cadaveric study, the following sequences were used: 1) 3D IR-UTE-Cones: TR/TI = 1000/300 ms, TE = 0.032/2.2 ms, Nsp = 9, tau = 7.1 ms, FA = 15°, bandwidth = 250 kHz, FOV = 22×22×20 cm3, matrix = 192×192×50, and scan time = 19.8 min; 2) 3D T1-weighed MP-RAGE sequence: TR = 7 ms, TE = 3 ms, TI = 450 ms, FOV = 22×22×16 cm3, matrix = 256×256×136, scan time = 4.2 min; 3) 2D T2-weighted FSE sequence: TR = 1500 ms, TE = 60 ms, FOV = 22×22 cm2, matrix = 256×160, slice thickness = 3.3 mm, number of slices =38, and scan time = 5 min. For imaging human brains in vivo: 1) 3D IR-UTE-Cones: TR = 1000 ms, TI between 320ms and 330ms, TE = 0.032/2.2 ms, Nsp = 21, tau = 7.1 ms, FA = 20°, bandwidth = 250 kHz, duration of each spoke = 880μs, FOV = 22×22×15.1 cm3, matrix = 192×192×42, scan time = 8.3 min; 2) 3D T1-weighted MP-RAGE sequence: TR = 8.2 ms, TE = 3.2 ms, TI = 450 ms, FOV = 25.6×25.6×17.8 cm3, matrix = 256×256×148, acceleration factor = 4, scan time = 5 min; 3) 3D T2-FLAIR sequence: TR/TI = 7600/2162 ms, TE = 117 ms, FOV = 25.6×25.6×25.6 cm3, matrix = 256×256×256, acceleration factor = 4 and scan time = 6.9 min. For T2* measurement, the same IR-UTE-Cones sequence was repeated with TE = 0.2, 0.4, and 0.8 ms for a total scan time of 80 min for ex vivo and of 33 min for in vivo studies.

TIs of IR-UTE-Cones were experimentally determined. A low-resolution IR-UTE-Cones sequence (FOV = 22×22×15.1 cm3, matrix = 64×64×24, and scan time = 2 min) was used to find the best nulling point of the long T2 components (i.e. no white matter signal in the second echo image). For the ex vivo brain specimen, TI = 300 ms was very close to the best nulling point. For the in vivo study, optimal TIs could vary due to the T1 variation among study participants. However, we found that the best nulling point was typically in the range of 320 ms to 330 ms. Typically, no more than three searches were needed to obtain the best TI.

Data Analysis

T2* values were calculated offline in manually defined ROIs using single exponential fitting in Matlab (The Mathworks Inc., Natick, MA, USA) using the following equation:

S=M0e−TE/T2*+n [1]

in which M0 is the equilibrium magnetization and n is induced to account for background noise and artifacts associated with data acquisition and image reconstruction (25,26,40). The signal equation of the UTE sequence is expressed as follows:

S=Cρsin(θ)1−e−TR/T11−cos(θ)e−TR/T1. [2]

ρ and C are the nominal proton density and the coil sensitivity of the myelin phantom, respectively. Note that ρ is not the absolute proton density because it contaminates signal constant (e.g. receive gains). T2* decay term was not included in Eq. [2] due to the similarity of the T2* values for all the phantoms (see the results section). To calculate the proton density ρ of myelin phantom from Eq. [2], both T1 and C need to be obtained. T1 was measured with the variable-TR UTE technique (41). The coil sensitivity C was measured from UTE imaging of a water phantom. The measured nominal proton densities of the myelin phantoms were correlated with myelin concentrations.

Two experienced radiologists, both with more than 10 years of experience, measured the MS lesion signal intensities in both dual-echo subtracted IR-UTE-Cones and T2-FLAIR images independently. Additionally, the signal intensities in both the NWM in volunteers and NAWM in MS patients (eight regions: left and right of the subcortical white matter, the centrum semiovale and the periventricular area, as well as the genu and splenium of corpus callosum) were also measured for comparison. Intraclass correlation coefficients were calculated for both the UTE and T2-FLAIR measures to test the reproducibility between the two radiologists. The measured signal intensities from the IR-UTE-Cones and T2-FLAIR images were normalized, respectively, using the equation Sn = (S-Smin)/(Smax-Smin) before the statistical analysis. S, Smin, Smax, and Sn are the measured signal intensity, minimal signal intensity, maximal signal intensity, and normalized signal intensity, respectively. Smin and Smax were the minimal and maximal signal intensities in the ROIs from all the subjects (including both patients and healthy controls). The signal correlation between UTE and T2-FLAIR measures were evaluated using the linear regression, and the ANOVA analysis was performed to evaluate the signal difference between lesion and relatively normal white matter (including NAM in volunteers and NAWM in MS patients), as well as the signal difference between NAM in volunteers and NAWM in MS patients.

RESULTS

Fig. 2 shows how proton density-weighted 3D UTE-Cones signal changes with myelin concentration in the myelin lipid D2O phantoms. As expected, there is no signal above background in the D2O-only control. The measured T1 values were 309, 391, 410, 399, 402, and 528 ms for the myelin phantoms with concentrations from 0 to 24%. After normalization to a homogeneous phantom, the nominal proton density and myelin concentration was highly linear, with an R2 of 99% for myelin concentrations ranging from 6% to 24%. A separately prepared 15% myelin-D2O phantom was subjected to T2* analysis (Fig. 3). A single-component relaxation curve has an excellent fit with a T2* of 0.33 ± 0.04 ms. T2* values of all the phantoms are very close ranging from 0.22 to 0.34 ms.

Figure 2.

Figure 2

Open in a new tab

Linearity of the 3D UTE-Cones signal with myelin concentration. Myelin lipid powder resuspended in D2O at concentrations ranging from 0 to 30% was simultaneously imaged using 3D UTE-Cones (A and B). A homogenous phantom of 12 mM MnCl2 in 20% H2O/80% D2O was similarly imaged and used for correction of coil sensitivity (C). The yellow circles represent the phantoms or their corresponding location for normalization (B and C). Linear regression analysis demonstrates the highly linear relationship between myelin concentration and calculated nominal proton density by Eq. [2] (D).

Figure 3.

Figure 3

Open in a new tab

T2* measurement of myelin using 3D UTE-Cones. Myelin lipid powder is resuspended in D2O at 15% w/v and imaged using 3D UTE-Cones with TEs of 0.032, 0.2, 0.4, 0.8, and 2.2 ms (A-D). T2* of 0.33 ± 0.04 ms is calculated by fitting to a single-component model (E).

Comparisons of typical clinical sequences and the proposed IR-UTE-Cones sequences in the ex vivo MS brain study are shown in Fig. 4. The bright signals in IR-UTE-Cones images have T2*s around 0.20 ± 0.04 ms, which is close to the T2* of myelin-D2O phantom shown in Fig. 2, suggesting that myelin protons may be the signal source. MS lesions identified in the clinical images show myelin signal loss in corresponding regions in the IR-UTE-Cones images, with representative lesions subsequently confirmed via histology as demyelinated.

Figure 4.

Figure 4

Open in a new tab

Representative ex vivo MS brain images (45-year-old male donor) with high disease burden using clinical T1-weighted MP-RAGE, T2-weighted FSE, and the proposed 3D IR-UTE-Cones (A). Yellow arrows point to MS lesions in all the images. The 3D IR-UTE-Cones sequence shows signal loss in the MS lesion regions. Tissue blocks used for histology were chosen based on high resolution 3D MP-RAGE images, where MS lesions and NAWM could be accurately located. Representative histology of a sample MS lesion using LFB as a myelin stain, counterstained with neutral red (B). The three histology images from top to bottom were obtained from locations indicated by the red arrowheads from top to bottom in the respective IR-UTE-Cones image. These histology images are regions of NAWM (top), lesion edge (middle), and within lesion (bottom) demonstrating specific loss of myelin staining in the MS lesion.

3D IR-UTE-Cones imaging of the brains of healthy volunteers demonstrates that signal in the white matter decreases to near-background levels at the second echo (Supporting Figure S1). The persistent gray matter signal is due to longer T1 relaxation and is effectively suppressed using dual-echo subtraction, resulting in 3D volumetric images with high myelin contrast. The signal-to-noise ratio (SNR) of the short T2 signal in the first echo image is 22.5 ± 3.8. A numerical phantom with a uniform signal intensity was created with a T2* of 0.3 ms and a matrix size of 192×192. Five images were generated for this phantom with TEs = 0, 0.2, 0.4, 0.8 and 2.2 ms. Gaussian white noise with an identical standard deviation was added to these five images such that the SNR of the first echo image was 22. The measured mean T2* for all the pixels in the phantom was 0.29 ± 0.03 ms, demonstrating robust T2* quantification with an SNR of 22. The T2* measured in the white matter region of a healthy volunteer demonstrates excellent fitting to a single-component model and is 0.254 ± 0.023 ms (Fig. 5). T2* maps from four different slices are also shown in Fig. 5. The mean and standard deviation of T2* value of the myelin in white matter regions of the whole brain is 0.241 ± 0.087 ms (the coefficient of variation = 35.2%), which is very close to the T2* values of the myelin phantom (0.33 ms) and ex vivo brains (0.20 ms), suggesting selective imaging of myelin in vivo.

Figure 5.

Figure 5

Open in a new tab

Myelin T2* measurement using the 3D IR-UTE-Cones for a healthy volunteer (37-year-old male). IR-UTE-Cones with TEs of 0.032, 0.2, 0.4, and 0.8 ms are shown in subfigures A–D, respectively. A T2* of 0.254 ± 0.023 ms was calculated by fitting to a single-component model (E). T2* maps of myelin in white matter regions from four different slices are shown in F–I.

Fig. 6 shows 3D IR-UTE-Cones imaging and clinical T1- and T2-weighted imaging for two representative MS patients. Similar to our observations in the ex vivo study, myelin signal loss in MS lesion regions can be found in IR-UTE-Cones images for the two MS patients. Both the ex vivo and in vivo MS brain studies demonstrate that the proposed IR-UTE-Cones sequence can directly detect myelin loss in MS lesions.

Figure 6.

Figure 6

Open in a new tab

In vivo MS patient brain imaging results of clinical T1-weighted MR-RAGE, clinical T2-weighted FLAIR, and the proposed 3D IR-UTE-Cones of two representative patients (first two rows: 62-year-old female; second two rows: 62-year-old male). MS lesions are highlighted with orange arrows or ovals. The 3D IR-UTE-Cones sequence shows signal loss in all identified MS lesions.

Both the UTE and T2-FLAIR measurement indicate very good reproducibility with intraclass correlation coefficients of 0.965 and 0.947, respectively. The signal intensity correlation between 3D IR-UTE-Cones and T2-FLAIR is very good, with R2 = 0.597 (see Fig. 7A). Both the IR-UTE-Cones and T2-FLAIR measures show significant difference in NWM of healthy volunteers and NAWM of MS patients and in MS lesions (p < 0.001). This demonstrates that both the proposed IR-UIE-Cones and clinical T2-FLAIR are able to detect MS lesions accurately. In addition, there is no signal significant difference in T2-FLAIR images between NWM of healthy volunteers and NAWM of MS patients (p = 0.204, see Fig. 7C). However, the proposed UTE measures show significant difference between the two groups (p < 0.001, see Fig. 7B). Those results suggest that the proposed IR-UTE-Cones technique may provide more useful information in the early detection of demyelination in MS patients compared with the conventional T2-FLAIR sequence.

Figure 7.

Figure 7

Open in a new tab

Statistical analysis for quantitative signal intensity measures of white matter acquired with the proposed 3D IR-UTE-Cones and the clinical T2-FLAIR sequences, respectively. The signal intensity of the 3D IR-UTE-Cones sequence shows a good correlation with the signal intensity of the T2-FLAIR with R2 = 0.597 (A). The measured 3D IR-UTE-Cones signals show significant difference in NWM of healthy volunteers and NAWM of MS patients (P < 0.001) (B), which is not observed in T2-FLAIR measurements (P = 0.204) (C). All the data are used for the statistical analysis. The central mark in boxplots (B and C) indicates the median, and the bottom and top edges of the boxes indicate the 25th and 75th percentiles, respectively. ‘+’ symbol refers to outliers.

DISCUSSION

This study performed a comprehensive study to investigate the potential of the multi-spoke 3D IR-UTE-Cones MR sequence for volumetric, morphological, and quantitative mapping of myelin on a clinical 3T scanner. In the phantom study, the 3D UTE-Cones signal was shown to have a strong linear correlation with myelin at concentrations likely to be encountered under normal and pathological conditions. The 3D IR-UTE-Cones sequence was then shown to generate high-contrast volumetric myelin images that compared favorably to conventional clinical sequences for the detection of lesions in MS brain specimens and MS patients. For the quantitative measures, the proposed IR-UTE-Cones technique show significant difference for the signals between NWM of healthy volunteers and NAWM of MS patients, but no significant difference in T2-FLAIR signals, which demonstrated the potential clinical value of the proposed technique.

The majority of the myelin protons detected by UTE sequences originate from long-chain methylenes of the lipid bilayers, with additional contributions from cholesterol, choline, and proteins (7,8). Most of these myelin protons are not exchangeable with deuterons (7,8,25), which are not detectable in proton MRI (42). Hence, the UTE signal of the myelin-powder phantoms resuspended in D2O should directly correspond to myelin proton density. This study demonstrated that normalized proton density determined using the 3D UTE Cones sequence on phantoms imaged using a clinical 3T scanner was strongly linear in the 6–24% myelin concentration range. The low signal of the phantom with zero-myelin concentration may be generated from background noise (non-zero mean in magnitude images), residual water (the D2O solution in our study had a purity of 99.9%, which meant that there was a very small fraction of H2O in each syringe that might contribute to the UTE signal) and the syringe (the syringe had an extremely short T2*, thus strong spatial blurring which might contribute to the background signal). These results highlight the potential of the 3D UTE-Cones sequence to directly detect and quantify myelin protons.

Similar to 2D IR-UTE sequences, the 3D IR-UTE-Cones sequence was able to generate morphological and quantitative maps of myelin that were able to detect lesions in MS patients. In contrast to 2D IR-UTE, the 3D IR-UTE-Cones sequence had higher excitation efficiency and reduced off-center slice and eddy current artifacts. The total scan time for the in vivo experiment was 8.3 minutes, which could be further reduced by acceleration techniques, such as compressed sensing and deep learning-based image reconstruction.

The T2* value of 0.241 ms measured with 3D IR-UTE-Cones for a healthy volunteer was close to the T2* value of 0.20 ms measured with the same sequence for an ex vivo MS brain and was close to 0.33 ms measured with 3D UTE-Cones of a myelin lipid/D2O phantom at a similar concentration. These values are also similar to the T2* values of myelin phantoms and D2O-exchanged white matter observed in prior studies using 2D UTE sequences and nuclear magnetic resonance at 3T (23–25,27–29). The longer T2* of the myelin phantom is expected from a small amount of residual semi-heavy water (HDO) (26).

Myelin is defined based on its ultrastructural configuration of tightly compacted multilamellar lipid membranes, which requires active maintenance by oligodendrocytes or Schwann cells to maintain. In contrast, the myelin lipids used in the phantoms form a combination of unilamellar and multilamellar vesicles, loosely resembling myelin. In both situations, the detectable myelin protons are in a liquid crystalline state resulting in the primary MRI signal characteristics, as suggested by similar T1 and T2* measurements. More recently, we investigated UTE imaging of white matter isolated from porcine brain which had been mechanically homogenized. Myelin vesicles were purified using discontinuous sucrose gradient ultracentrifugation according to published methods (43). After thoroughly washing out residual sucrose, the myelin was resuspended twice in deuterated tris-Cl buffer to remove residual H2O. 3D UTE images show high signal from the intact myelin vesicles with a short T1 of 367±4 ms and T2* of 225±7 μs on a clinical 3T scanner (44). The T2* of intact myelin vesicles is quite similar to those of myelin lipid powder and myelin-D2O paste (27). It is currently unknown what if any changes in the MRI signal result from the ultrastructure, perhaps through interactions between the lamella. This is a technically challenging study and a topic of active investigation in our group and several other groups.

It should be noted that this T2* relaxometric fit captures only the fraction of myelin signal that is observable with the UTE-Cones sequence employed. There is a substantial fraction of myelin signal that has T2* shorter than 100 μs, which is therefore not observable with the UTE imaging method (7,24). However, the 3D UTE-Cones sequence with a short-hard pulse excitation is more effective in capturing the shorter T2* components than the 2D UTE sequence with a longer soft half pulse excitation. Thus, myelin T2* values measured in this study are shorter than previous results measured with the 2D UTE sequence (9). It is likely that there are multiple T2* components in 3D IR-UTE-Cones imaging of myelin as there are T2* variations which may vary across different brain regions. Another cause of the T2* variation can be the long T2 contamination, since it is very difficult to completely suppress all long T2 components in white matter with a single TI due to T1 variations across the brain. The relatively low in-plane and out-of-plane resolution can also lead to long T2 contamination due to the partial volume effect. It is also well known that the spectrum lineshape of the myelin proton is super-Lorentzian. That the myelin signals were well-fitted by an exponential function suggests that it arises from molecules that are undergoing rapid isotropic reorientation, or perhaps that it is from lipid proton pairs oriented at the magic angle to the magnetic field. Despite the uncertainties, the measured short T2* values still indicated the effective suppression of longer T2 signals in white matter using the adiabatic IR scheme.

So far there are no studies reporting effects of freezing and thawing on MR and tissue properties of the myelin sheath in cadaveric human brain specimens. A previous study has shown no statistically significant difference in mean T2 and T2* values of the Achilles tendon between the fresh specimens and after subsequent cycles of freeze-thaw treatment (46). The freezing/thawing process before MR imaging may damage the myelin ultrastructure and to a lesser extent the myelin components, which may alter the signal characteristics of myelin using the proposed IR-UTE-Cones sequence. However, the measured T1 and T2* of NAWM in the cadaveric specimens were similar to those of myelin paste and NAWM in vivo, suggesting that the freezing/thawing process might have very limited effects on the myelin signals. Clearly, more research is needed to understand potential structural changes in myelin and related MR properties due to freezing and thawing cycles.

The ability to directly detect myelin signal with volumetric imaging using 3D IR-UTE-Cones has several important advantages. First, direct imaging may have improved specificity and interpretability over indirect myelin imaging. Judicious use of both direct and indirect myelin imaging techniques may be important for better lesion characterization. Second, it is possible that IR-UTE signal properties depend on the physicochemical characteristics of the myelin, such that IR-UTE signal analysis (e.g. T2*and T1 measurements) may reflect myelin quality in addition to myelin content, a concept which needs further investigation. Third, partial demyelination may only be visible when using the IR-UTE sequences, as demonstrated by the prior 2D IR-UTE imaging studies of MS brain specimens in which the clinical sequences show no abnormalities in regions with partial demyelination (27). Importantly, the 3D IR-UTE-Cones sequence may potentially allow direct imaging of dynamic demyelination and remyelination post-drug therapy, another possibility which also warrants further investigation.

The presence of the freezing artifacts is an important limitation of the ex vivo study and has to do with the preparation of the specimens by the vendor during the immediate post-mortem period. Proper preparation to prevent ice crystal formation of a large organ like the human brain would require perfusion with a cryoprotectant and rapid but controlled freezing conditions, which would have required pre-mortem arrangements with MS patients that were unavailable to us. In fact, the brain specimens were thawed in endogenous isotonic fluid, the CSF. We thawed whole plastic-wrapped heads in water but did not allow water to penetrate the head and contact the brain. The thawing process was relatively long to allow for even temperature equilibration for MRI and would be expected to result in a radial pattern of degradation, although this was not seen by imaging or histology. The specimens were rapidly refrozen after imaging to allow for later sectioning and histological preparation – the gross morphology of the refrozen brain was unchanged, allowing for correlation between imaging ROIs and histology. Neutral red counterstaining of nuclei is present within figure 4B and manifests as a purplish hue due to the presence of LFB. The neutral red staining was more apparent on routine single dye control slides that were not included in the figure. Note that neither neutral red nor LFB is absolutely specific, and LFB does tend to stain nuclei (47–49). In our study, this tissue tended to have weaker neutral red reactivity, and the nuclei tended to have a greyish, blue to purple hue as shown in the figure. A quantitative analysis of the correlation between histological myelin content and IR-UTE signal was not performed because of these limitations. However, it is reassuring that there are clear qualitative differences in myelin staining by LFB between the MS lesions and adjacent NAWM, which would have been subjected to similar freezing artifacts.

Recent studies have shown nice correlation between T1-based imaging and myelin in microscopic imaging of gray matter and white matter at 7T (50,51). We anticipate that the IR-UTE-based technique will be more specific to myelin, especially in white mater of the brain in vivo at lower field strengths. T1 relaxation is affected by many factors, such as water content, myelin content, iron, etc., leading to reduced specificity when using T1-based imaging to quantify myelin content. On the other hand, the IR-UTE technique is less sensitive to water content as water signal is suppressed through adiabatic inversion and nulling by choosing an appropriate combination of TR and TI. Iron deposition is expected to shorten T1 and T2*. However, typical concentration of iron in the brain ranges from 49 mg/kg in the thalamus to 205 mg/kg (equivalent to 1~4 mM) in the globus pallidus with estimated T2* of 14–40 ms (52). As a result, the adiabatic inversion pulse is expected to invert the longitudinal magnetizations of white matter and/or gray matter in regions with iron deposition or iron loss. The major problem is T1 change due to iron deposition or loss, which may significantly increase the residual signals from long T2 white matter and/or gray matter in regions with high iron deposition or loss. The residual signals are expected to be high in both the first echo and the second echo. Subtraction of the second echo from the first one is expected to greatly reduce long T2 signal contamination. More research is needed to identify iron deposition or loss, and to minimize potential effect on myelin quantification (28,53–55).

There are several other limitations to this study. First, the IR-UTE technique depends on the tissue T1 because a near-optimal TI is needed for suppression of the long T2 signals in white matter. In our experience, good results can be obtained with TI of 325 ms ± 5 ms, with the optimal TI identified using several quick scans. More importantly, the T1 of white matter can be heterogeneous under pathological conditions such as MS. While this has the effect of further highlighting areas of myelin pathology, it is undesirable for quantitative imaging. Second, other challenges of in vivo myelin quantification using IR-UTE sequences include artifacts from short T2 blurring and non-Cartesian sampling. Strategies to reduce or model T1 dependence and improve image reconstruction are under active investigation by our research group. Third, the scan time for T2* measurement in our volunteer study was relatively long at 33 minutes. In comparison, a bi-component T2* fitting study to characterize the ultrashort T2 components in the brain also required a long scan time of 45 min (10). Additional work is required to reduce T2* measurement to clinically feasible scan time using highly undersampled data acquisition together with advanced image reconstruction techniques, such as compressed sensing and/or deep learning-based reconstruction methods (56,57). Fourth, the 3D IR-UTE-Cones sequence was not compared with IR-prepared zero echo time (ZTE) or other short T2 MRI techniques (24), which were beyond the scope of this study. Lastly, more corrections such as B1 inhomogeneity and coil sensitivity correction are needed for more accurate quantitative signal intensity measures. Both B1 and coil sensitivity mapping sequences should be included in future quantitative imaging protocols.

CONCLUSION

The 3D IR-UTE-Cones sequence can be used for volumetric myelin imaging on clinical 3T scanners and demonstrates a strong linear correlation with myelin concentration. This sequence may potentially aid in the detection of demyelinated lesions in MS patients.

Supplementary Material

Supporting figures

Supporting Figure S1 3D IR-UTE-Cones myelin imaging of a healthy volunteer (46-year-old female). The first (TE = 0.032 ms) and second echo (TE = 2.2 ms) IR-UTE-Cones images are shown in the first and second columns, respectively. The corresponding myelin images shown in the third column are obtained by dual echo image subtraction.

ACKNOWLEDGMENTS

The authors acknowledge grant support from the NIH (1R01 NS092650 and T32 EB005970), VA Clinical Science and Rehabilitation Research and Development Services (Merit Awards I01CX001388 and I01RX002604), and GE Healthcare.

ABBREVIATIONS:

MS

multiple sclerosis

UTE

ultrashort echo time

IR

inversion recovery

NWM

normal white matter

NAWM

normal-appearing white matter

3D IR-UTE-Cones

3D adiabatic inversion recovery prepared UTE Cones

LFB

Luxol Fast Blue

Nsp

number of spokes

TI

inversion times

FOV

field of view

ROIs

regions of interest

IRB

Institutional Review Board

FA

flip angle

REFERENCES

  • 1.Siegel GJ. Basic neurochemistry: molecular, cellular and medical aspects.; 1999.
  • 2.Young KM, Psachoulia K, Tripathi RB, et al. Oligodendrocyte Dynamics in the Healthy Adult CNS: Evidence for Myelin Remodeling. Neuron 2013;77:873–885 doi: 10.1016/j.neuron.2013.01.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Duncan ID, Radcliff AB. Inherited and acquired disorders of myelin: The underlying myelin pathology. Exp. Neurol 2016;283, Part B:452–475 doi: 10.1016/j.expneurol.2016.04.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Polman CH, Reingold SC, Banwell B, et al. Diagnostic criteria for multiple sclerosis: 2010 Revisions to the McDonald criteria. Ann. Neurol 2011;69:292–302 doi: 10.1002/ana.22366. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Tillema J-M, Pirko I. Neuroradiological evaluation of demyelinating disease. Ther. Adv. Neurol. Disord 2013;6:249–268 doi: 10.1177/1756285613478870. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Lucchinetti CF, Parisi J, Bruck W. The pathology of multiple sclerosis. Neurol. Clin 2005;23:77–105 doi: 10.1016/j.ncl.2004.09.002. [DOI] [PubMed] [Google Scholar]
  • 7.Horch RA, Gore JC, Does MD. Origins of the Ultrashort-T21H NMR Signals in Myelinated Nerve: A Direct Measure of Myelin Content? Magn. Reson. Med 2011;66:24–31 doi: 10.1002/mrm.22980. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.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 doi: 10.1073/pnas.1115107109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.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 doi: 10.1016/j.neuroimage.2013.10.053. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Boucneau T, Cao P, Tang S, et al. In vivo characterization of brain ultrashort-T2 components. Magn. Reson. Med 2018;80:726–735 doi: 10.1002/mrm.27037. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Filippi M, Campi A, Dousset V, et al. A Magnetization Transfer Imaging Study of Normal-Appearing White Matter in Multiple Sclerosis. Neurology 1995;45:478–482 doi: 10.1212/WNL.45.3.478. [DOI] [PubMed] [Google Scholar]
  • 12.Silver NC, Lai M, Symms MR, Barker GJ, McDonald WI, Miller DH. Serial magnetization transfer imaging to characterize the early evolution of new MS lesions. Neurology 1998;51:758–764. [DOI] [PubMed] [Google Scholar]
  • 13.Khodanovich MY, Sorokina IV, Glazacheva VY, et al. Histological validation of fast macromolecular proton fraction mapping as a quantitative myelin imaging method in the cuprizone demyelination model. Sci. Rep 2017;7:46686 doi: 10.1038/srep46686. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Vavasour IM, Whittall KP, Mackay AL, Li DKB, Vorobeychik G, Paty DW. A comparison between magnetization transfer ratios and myelin water percentages in normals and multiple sclerosis patients. Magn. Reson. Med 1998;40:763–768 doi: 10.1002/mrm.1910400518. [DOI] [PubMed] [Google Scholar]
  • 15.Thiessen JD, Zhang Y, Zhang H, et al. Quantitative MRI and ultrastructural examination of the cuprizone mouse model of demyelination. NMR Biomed. 2013;26:1562–1581 doi: 10.1002/nbm.2992. [DOI] [PubMed] [Google Scholar]
  • 16.Kolind S, Matthews L, Johansen-Berg H, et al. Myelin Water Imaging Reflects Clinical Variability in Multiple Sclerosis. NeuroImage 2012;60:263–270 doi: 10.1016/j.neuroimage.2011.11.070. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Bouhrara M, Spencer RG. Rapid simultaneous high-resolution mapping of myelin water fraction and relaxation times in human brain using BMC-mcDESPOT. NeuroImage 2017;147:800–811 doi: 10.1016/j.neuroimage.2016.09.064. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Song S-K, Yoshino J, Le TQ, et al. Demyelination increases radial diffusivity in corpus callosum of mouse brain. NeuroImage 2005;26:132–140 doi: 10.1016/j.neuroimage.2005.01.028. [DOI] [PubMed] [Google Scholar]
  • 19.Shirani A, Sun P, Schmidt RE, et al. Histopathological correlation of diffusion basis spectrum imaging metrics of a biopsy-proven inflammatory demyelinating brain lesion: A brief report. Mult. Scler. J 2018:1352458518786072 doi: 10.1177/1352458518786072. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Laule C, Vavasour IM, Kolind SH, et al. Magnetic resonance imaging of myelin. Neurotherapeutics 2007;4:460–484 doi: 10.1016/j.nurt.2007.05.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Levesque IR, Giacomini PS, Narayanan S, et al. Quantitative magnetization transfer and myelin water imaging of the evolution of acute multiple sclerosis lesions. Magn. Reson. Med 2010;63:633–640 doi: 10.1002/mrm.22244. [DOI] [PubMed] [Google Scholar]
  • 22.Wood TC, Simmons C, Hurley SA, et al. Whole-brain ex-vivo quantitative MRI of the cuprizone mouse model. PeerJ 2016;4:e2632 doi: 10.7717/peerj.2632. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Du J, Sheth V, He Q, et al. Measurement of T1 of the Ultrashort T2* Components in White Matter of the Brain at 3T. PLOS ONE 2014;9:e103296 doi: 10.1371/journal.pone.0103296. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.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 doi: 10.1016/j.neuroimage.2017.09.054. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Fan S-J, Ma Y, Chang EY, Bydder GM, Du J. Inversion recovery ultrashort echo time imaging of ultrashort T2 tissue components in ovine brain at 3 T: a sequential D2O exchange study. NMR Biomed. 2017;30 doi: 10.1002/nbm.3767. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Fan S-J, 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 doi: 10.1002/mrm.27052. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.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 doi: 10.1016/j.neuroimage.2016.05.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Sheth VR, Fan S, He Q, et al. Inversion Recovery Ultrashort Echo Time Magnetic Resonance Imaging: A Method for Simultaneous Direct Detection of Myelin and High Signal Demonstration of Iron Deposition in the Brain – A Feasibility Study. Magn. Reson. Imaging 2017;38:87–94 doi: 10.1016/j.mri.2016.12.025. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.He Q, Ma Y, Fan S, et al. Direct magnitude and phase imaging of myelin using ultrashort echo time (UTE) pulse sequences: A feasibility study. Magn. Reson. Imaging 2017;39:194–199 doi: 10.1016/j.mri.2017.02.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Glover GH, Pauly JM, Bradshaw KM. Boron-11 imaging with a three-dimensional reconstruction method. J. Magn. Reson. Imaging JMRI 1992;2:47–52. [DOI] [PubMed] [Google Scholar]
  • 31.Du J, Bydder GM. Qualitative and quantitative ultrashort-TE MRI of cortical bone. NMR Biomed. 2013;26:489–506 doi: 10.1002/nbm.2906. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Du J, Bydder M, Takahashi AM, Carl M, Chung CB, Bydder GM. Short T2 contrast with three-dimensional ultrashort echo time imaging. Magn. Reson. Imaging 2011;29:470–482 doi: 10.1016/j.mri.2010.11.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Gurney PT, Hargreaves BA, Nishimura DG. Design and analysis of a practical 3D cones trajectory. Magn. Reson. Med 2006;55:575–582 doi: 10.1002/mrm.20796. [DOI] [PubMed] [Google Scholar]
  • 34.Carl M, Bydder GM, Du J. UTE imaging with simultaneous water and fat signal suppression using a time-efficient multispoke inversion recovery pulse sequence. Magn. Reson. Med 2016;76:577–582 doi: 10.1002/mrm.25823. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Ma Y-J, Zhu Y, Lu X, Carl M, Chang EY, Du J. Short T2 imaging using a 3D double adiabatic inversion recovery prepared ultrashort echo time cones (3D DIR-UTE-Cones) sequence. Magn. Reson. Med 2018;79:2555–2563 doi: 10.1002/mrm.26908. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Chen J, Chang EY, Carl M, et al. Measurement of bound and pore water T1 relaxation times in cortical bone using three-dimensional ultrashort echo time cones sequences. Magn. Reson. Med 2017;77:2136–2145 doi: 10.1002/mrm.26292. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Nazaran A, Carl M, Ma Y, et al. Three-dimensional adiabatic inversion recovery prepared ultrashort echo time cones (3D IR-UTE-Cones) imaging of cortical bone in the hip. Magn. Reson. Imaging 2017;44:60–64 doi: 10.1016/j.mri.2017.07.012. [DOI] [PubMed] [Google Scholar]
  • 38.Larson PEZ, Conolly SM, Pauly JM, Nishimura DG. Using adiabatic inversion pulses for long-T2 suppression in ultrashort echo time (UTE) imaging. Magn. Reson. Med 2007;58:952–961 doi: 10.1002/mrm.21341. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Manhard MK, Horch RA, Harkins KD, Gochberg DF, Nyman JS, Does MD. Validation of quantitative bound- and pore-water imaging in cortical bone. Magn. Reson. Med 2014;71:2166–2171 doi: 10.1002/mrm.24870. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Du J, Diaz E, Carl M, Bae W, Chung CB, Bydder GM. Ultrashort echo time imaging with bicomponent analysis. Magn. Reson. Med 2012;67:645–649 doi: 10.1002/mrm.23047. [DOI] [PubMed] [Google Scholar]
  • 41.Ma Y-J, Lu X, Carl M, et al. Accurate T1 mapping of short T2 tissues using a three-dimensional ultrashort echo time cones actual flip angle imaging-variable repetition time (3D UTE-Cones AFI-VTR) method. Magn. Reson. Med 2018;80:598–608. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Müller S, Seelig J. In vivo NMR imaging of deuterium. J. Magn. Reson. 1969 1987;72:456–466 doi: 10.1016/0022-2364(87)90150-8. [DOI] [Google Scholar]
  • 43.Larocca JN, Norton WT. Isolation of myelin. Curr. Protoc. Cell Biol 2007;Chapter 3:Unit3.25 doi: 10.1002/0471143030.cb0325s33. [DOI] [PubMed] [Google Scholar]
  • 44.Ma Y-J, Searleman AC, Jang H, et al. Whole-Brain Myelin Imaging Using 3D Double-Echo Sliding Inversion Recovery Ultrashort Echo Time (DESIRE UTE) MRI. Radiology 2019:190911 doi: 10.1148/radiol.2019190911. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.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 doi: 10.1016/j.neuroimage.2013.10.053. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Chang EY, Bae WC, Statum S, Du J, Chung CB. Effects of repetitive freeze–thawing cycles on T2 and T2* of the Achilles tendon. Eur. J. Radiol 2014;83:349–353. [DOI] [PubMed] [Google Scholar]
  • 47.Warntjes JBM, Persson A, Berge J, Zech W. Myelin detection using rapid quantitative MR imaging correlated to macroscopically registered luxol fast blue–stained brain specimens. Am. J. Neuroradiol 2017;38:1096–1102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Laule C, Leung E, Li DK, et al. Myelin water imaging in multiple sclerosis: quantitative correlations with histopathology. Mult. Scler. J 2006;12:747–753. [DOI] [PubMed] [Google Scholar]
  • 49.Laule C, Vavasour IM, Leung E, et al. Pathological basis of diffusely abnormal white matter: insights from magnetic resonance imaging and histology. Mult. Scler. J 2011;17:144–150. [DOI] [PubMed] [Google Scholar]
  • 50.Stüber C, Morawski M, Schäfer A, et al. Myelin and iron concentration in the human brain: a quantitative study of MRI contrast. Neuroimage 2014;93:95–106. [DOI] [PubMed] [Google Scholar]
  • 51.Hametner S, Endmayr V, Deistung A, et al. The influence of brain iron and myelin on magnetic susceptibility and effective transverse relaxation-A biochemical and histological validation study. Neuroimage 2018;179:117–133. [DOI] [PubMed] [Google Scholar]
  • 52.Bagnato F, Hametner S, Welch EB. Visualizing iron in multiple sclerosis. Magn. Reson. Imaging 2013;31:376–384. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Hametner S, Wimmer I, Haider L, Pfeifenbring S, Brück W, Lassmann H. Iron and neurodegeneration in the multiple sclerosis brain. Ann. Neurol 2013;74:848–861 doi: 10.1002/ana.23974. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Popescu BF, Frischer JM, Webb SM, et al. Pathogenic implications of distinct patterns of iron and zinc in chronic MS lesions. Acta Neuropathol. (Berl.) 2017;134:45–64 doi: 10.1007/s00401-017-1696-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Hernández-Torres E, Wiggermann V, Hametner S, et al. Orientation Dependent MR Signal Decay Differentiates between People with MS, Their Asymptomatic Siblings and Unrelated Healthy Controls. PloS One 2015;10:e0140956 doi: 10.1371/journal.pone.0140956. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Lustig M, Donoho D, Pauly JM. Sparse MRI: The application of compressed sensing for rapid MR imaging. Magn. Reson. Med 2007;58:1182–1195. [DOI] [PubMed] [Google Scholar]
  • 57.Zhu B, Liu JZ, Cauley SF, Rosen BR, Rosen MS. Image reconstruction by domain-transform manifold learning. Nature 2018;555:487–492 doi: 10.1038/nature25988. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting figures

Supporting Figure S1 3D IR-UTE-Cones myelin imaging of a healthy volunteer (46-year-old female). The first (TE = 0.032 ms) and second echo (TE = 2.2 ms) IR-UTE-Cones images are shown in the first and second columns, respectively. The corresponding myelin images shown in the third column are obtained by dual echo image subtraction.

NIHMS1676461-supplement-Supporting_figures.docx (21.8MB, docx)

RESOURCES