Denoising of diffusion MRI in the cervical spinal cord - effects of denoising strategy and acquisition on intra-cord contrast, signal modeling, and feature conspicuity

Quantitative diffusion MRI (dMRI) is a promising technique for evaluating the spinal cord in health and disease. However, low signal-to-noise ratio (SNR) can impede interpretation and quantification of these images. The purpose of this study is to evaluate several dMRI denoising approaches on their ability to improve the quality, reliability, and accuracy of quantitative diffusion MRI of the spinal cord. We evaluate three denoising approaches (Non-Local Means, Marchenko-Pastur PCA, and a newly proposed Patch2Self algorithm) and conduct five experiments to validate the denoising performance on clinical-quality and commonly-acquired dMRI acquisitions: 1) a phantom experiment to assess denoising error and bias; 2) a multi-vendor, multi-acquisition open experiment for both qualitative and quantitative evaluation of noise residuals; 3) a bootstrapping experiment to estimate uncertainty of parametric maps; 4) an assessment of spinal cord lesion conspicuity in a multiple sclerosis group; and 5) an evaluation of denoising for advanced parametric multi-compartment modeling. We find that all methods improve signal-to-noise ratio and conspicuity of MS lesions in individual diffusion weighted images (DWIs), but MPPCA and Patch2Self excel at improving the quality and intra-cord contrast of diffusion weighted images - removing signal fluctuations due to thermal noise while improving precision of estimation of diffusion parameters even with very few DWIs (i.e., 16-32) typical of clinical acquisitions. These denoising approaches hold promise for facilitating reliable diffusion observations and measurements in the spinal cord to investigate biological and pathological processes.

Sensitivity of portable low-field magnetic resonance imaging for multiple sclerosis lesions

•

Paired, same-day, 3T and 64mT MRI studies were analyzed in 33 MS patients.

•

64mT MRI showed 94% sensitivity for detecting any lesions in 3T confirmed cases.

•

The diameter of the smallest detected lesion was larger at 64mT compared to 3T.

•

Total lesion volume estimates were strongly correlated between 3T and 64mT scans.

•

Portable low-field MRI detects white matter lesions, but smaller lesions may be missed.

Magnetic resonance imaging (MRI) is a fundamental tool in the diagnosis and management of neurological diseases such as multiple sclerosis (MS). New portable, low-field strength, MRI scanners could potentially lower financial and technical barriers to neuroimaging and reach underserved or disabled populations, but the sensitivity of these devices for MS lesions is unknown. We sought to determine if white matter lesions can be detected on a portable 64mT scanner, compare automated lesion segmentations and total lesion volume between paired 3T and 64mT scans, identify features that contribute to lesion detection accuracy, and explore super-resolution imaging at low-field. In this prospective, cross-sectional study, same-day brain MRI (FLAIR, T1w, and T2w) scans were collected from 36 adults (32 women; mean age, 50 ± 14 years) with known or suspected MS using Siemens 3T (FLAIR: 1 mm isotropic, T1w: 1 mm isotropic, and T2w: 0.34–0.5 × 0.34–0.5 × 3–5 mm) and Hyperfine 64mT (FLAIR: 1.6 × 1.6 × 5 mm, T1w: 1.5 × 1.5 × 5 mm, and T2w: 1.5 × 1.5 × 5 mm) scanners at two centers. Images were reviewed by neuroradiologists. MS lesions were measured manually and segmented using an automated algorithm. Statistical analyses assessed accuracy and variability of segmentations across scanners and systematic scanner biases in automated volumetric measurements. Lesions were identified on 64mT scans in 94% (31/33) of patients with confirmed MS. The average smallest lesions manually detected were 5.7 ± 1.3 mm in maximum diameter at 64mT vs 2.1 ± 0.6 mm at 3T, approaching the spatial resolution of the respective scanner sequences (3T: 1 mm, 64mT: 5 mm slice thickness). Automated lesion volume estimates were highly correlated between 3T and 64mT scans (r = 0.89, p < 0.001). Bland-Altman analysis identified bias in 64mT segmentations (mean = 1.6 ml, standard error = 5.2 ml, limits of agreement = –19.0–15.9 ml), which over-estimated low lesion volume and under-estimated high volume (r = 0.74, p < 0.001). Visual inspection revealed over-segmentation was driven venous hyperintensities on 64mT T2-FLAIR. Lesion size drove segmentation accuracy, with 93% of lesions > 1.0 ml and all lesions > 1.5 ml being detected. Using multi-acquisition volume averaging, we were able to generate 1.6 mm isotropic images on the 64mT device. Overall, our results demonstrate that in established MS, a portable 64mT MRI scanner can identify white matter lesions, and that automated estimates of total lesion volume correlate with measurements from 3T scans.

Development of a mobile low-field MRI scanner

Magnetic resonance imaging (MRI) allows important visualization of the brain and central nervous system anatomy and organization. However, unlike electroencephalography (EEG) or functional near infrared spectroscopy, which can be brought to a patient or study participant, MRI remains a hospital or center-based modality. Low magnetic field strength MRI systems, however, offer the potential to extend beyond these traditional hospital and imaging center boundaries. Here we describe the development of a modified cargo van that incorporates a removable low-field permanent magnet MRI system and demonstrate its proof-of-concept. Using phantom scans and in vivo T₂-weighted neuroimaging data, we show no significant differences with respect to geometric distortion, signal-to-noise ratio, or tissue segmentation outcomes in data acquired in the mobile system compared to a similar static system in a laboratory setting. These encouraging results show, for the first time, MRI that can be performed at a participant’s home, community center, school, etc. Breaking traditional barriers of access, this mobile approach may enable imaging of patients and participants who have mobility challenges, live long distances from imaging centers, or are otherwise unable to travel to an imaging center or hospital.

Bedside detection of intracranial midline shift using portable magnetic resonance imaging

Neuroimaging is crucial for assessing mass effect in brain-injured patients. Transport to an imaging suite, however, is challenging for critically ill patients. We evaluated the use of a low magnetic field, portable MRI (pMRI) for assessing midline shift (MLS). In this observational study, 0.064 T pMRI exams were performed on stroke patients admitted to the neuroscience intensive care unit at Yale New Haven Hospital. Dichotomous (present or absent) and continuous MLS measurements were obtained on pMRI exams and locally available and accessible standard-of-care imaging exams (CT or MRI). We evaluated the agreement between pMRI and standard-of-care measurements. Additionally, we assessed the relationship between pMRI-based MLS and functional outcome (modified Rankin Scale). A total of 102 patients were included in the final study (48 ischemic stroke; 54 intracranial hemorrhage). There was significant concordance between pMRI and standard-of-care measurements (dichotomous, κ = 0.87; continuous, ICC = 0.94). Low-field pMRI identified MLS with a sensitivity of 0.93 and specificity of 0.96. Moreover, pMRI MLS assessments predicted poor clinical outcome at discharge (dichotomous: adjusted OR 7.98, 95% CI 2.07–40.04, p = 0.005; continuous: adjusted OR 1.59, 95% CI 1.11–2.49, p = 0.021). Low-field pMRI may serve as a valuable bedside tool for detecting mass effect.

Abstract 33: Qualitative Description of Ischemic Stroke Appearance on Low-Field, Point-Of-Care Magnetic Resonance Imaging

Background and Aims: Advances in low-field MRI have enabled image acquisition at the point-of-care (POC). We aim to characterize ischemic lesions in low-field, POC MRI and assess its relationship with stroke severity in ischemic stroke patients.

Methods: We performed POC MRI exams on ischemic stroke patients. T2-weighted (T2W), fluid-attenuated inversion recovery (FLAIR), and diffusion-weighted imaging (DWI) exams were acquired with a 64mT, portable bedside MRI system. Three raters computed signal intensity ratios (SIR) for each sequence. For every slice showing an infarct, an SIR was generated by dividing the mean signal intensity of the lesion by the mean signal intensity of the contralateral hemisphere. Infarct volumes were obtained by multiplying the lesion area of each slice by the slice thickness (5mm) and summing the cross-sectional areas. Volumes were correlated with National Institutes of Health Stroke Scale (NIHSS) scores at the time of scan.

Results: We studied 18 ischemic stroke patients (50% women; ages 30-95 years). Two patients were studied at two and three serial timepoints, respectively. POC exams were obtained 2.7 ± 2.2 days after symptom onset. A total of 18 T2W, 17 FLAIR, and 18 DWI exams were obtained. Three exams (1 T2W; 1 FLAIR; 1 DWI) were excluded due to motion degradation. High field MRI exams (19 ± 16 hours from POC exams) demonstrated ischemic infarcts in 15 of the 18 patients. All POC T2W and FLAIR exams revealed infarcts in these patients, and 14 of the 17 DWI exams showed infarcts. Ischemic infarcts were seen as hyperintense lesions (SIR: T2W = 1.19 ± 0.10, FLAIR = 1.15 ± 0.08, DWI = 1.36 ± 0.17). Infarct volume significantly correlated with NIHSS scores (T2W: r = 0.71, p < 0.01; FLAIR: r = 0.65, p < 0.05; DWI: r = 0.65, p < 0.05).

Conclusions: These preliminary data suggest that low-field, POC MRI may be useful in the clinical evaluation of ischemic stroke. Further work in larger cohorts is needed to elucidate the appearance of infarction on low-field imaging.

Abstract WP413: Portable, Bedside, Point of Care Magnetic Resonance Imaging in an Intensive Care Setting for Intracranial Hemorrhage

Background: Radiographic diagnosis of intracranial hemorrhage (ICH) is a critical determinant of stroke care pathways requiring patient transport to a neuroimaging suite. Advances in low-field MRI have made it possible to obtain clinically useful imaging at the point of care (POC).

Aim: The aim of this study was to obtain preliminary data regarding the ability of a bedside POC MRI scanner to detect ICH.

Methods: We studied 36 patients with a diagnosis of ICH (n=18) or ischemic stroke (n=18). Five blinded readers independently evaluated T2W and FLAIR exams acquired prospectively on a 64 mT, portable bedside MRI system (Hyperfine Research, Inc). Kappa coefficients (κ) were calculated to determine inter-rater agreement. Ground truth was obtained from the clinical report of the closest conventional imaging study (17.9 ± 10.4 hours) and verified by a core reader. For each exam, majority consensus among raters was used to determine sensitivity.

Results: ICH volume ranged from 4 to 101 cc (median of 13 cc). Exams were acquired within 7 days of symptom onset (51.1 ± 28.8 hours). A pathologic lesion was identified on every exam with 100% sensitivity. Sensitivity for distinguishing any hemorrhage was 89% and specificity was 83%. The mean sensitivity and specificity for individual raters was 79% and 69%, respectively. When limited to supratentorial hemorrhage, consensus sensitivity was 94%. For ICH cases detected by all raters (n=9), there was 100% accuracy for localizing the bleed (lobar vs. non-lobar) with perfect agreement among raters (κ = 1, p <0.0001). There was substantial agreement for identifying intraventricular hemorrhage (IVH) (κ = 0.72, p < 0.0001). Sensitivity for IVH was 100% based on rater consensus. Figure 1 shows a POC exam with an ICH and IVH.

Conclusions: These data suggest that low-field, POC MRI may be used to detect hemorrhagic stroke at the bedside. Further work is needed to evaluate this approach in the hyperacute setting and across a wide range of ICH characteristics.

Diffusion MRI microstructural models in the cervical spinal cord - Application, normative values, and correlations with histological analysis

Multi-compartment tissue modeling using diffusion magnetic resonance imaging has proven valuable in the brain, offering novel indices sensitive to the tissue microstructural environment in vivo on clinical MRI scanners. However, application, characterization, and validation of these models in the spinal cord remain relatively under-studied. In this study, we apply a diffusion "signal" model (diffusion tensor imaging, DTI) and two commonly implemented "microstructural" models (neurite orientation dispersion and density imaging, NODDI; spherical mean technique, SMT) in the human cervical spinal cord of twenty-one healthy controls. We first provide normative values of DTI, SMT, and NODDI indices in a number of white matter ascending and descending pathways, as well as various gray matter regions. We then aim to validate the sensitivity and specificity of these diffusion-derived contrasts by relating these measures to indices of the tissue microenvironment provided by a histological template. We find that DTI indices are sensitive to a number of microstructural features, but lack specificity. The microstructural models also show sensitivity to a number of microstructure features; however, they do not capture the specific microstructural features explicitly modelled. Although often regarded as a simple extension of the brain in the central nervous system, it may be necessary to re-envision, or specifically adapt, diffusion microstructural models for application to the human spinal cord with clinically feasible acquisitions - specifically, adjusting, adapting, and re-validating the modeling as it relates to both theory (i.e. relevant biology, assumptions, and signal regimes) and parameter estimation (for example challenges of acquisition, artifacts, and processing).

Quantification of DTI in the Pediatric Spinal Cord: Application to Clinical Evaluation in a Healthy Patient Population

BACKGROUND AND PURPOSE

The purpose of the study is to characterize diffusion tensor imaging indices in the developing spinal cord, evaluating differences based on age and cord region. Describing the progression of DTI indices in the pediatric cord increases our understanding of spinal cord development.

MATERIALS AND METHODS

A retrospective analysis was performed on DTI acquired in 121 pediatric patients (mean, 8.6 years; range, 0.3-18.0 years) at Monroe Carell Jr. Children's Hospital at Vanderbilt from 2017 to 2018. Diffusion-weighted images (15 directions; b = 750 s/mm²; slice thickness, 5 mm; in-plane resolution, 1.0 × 1.0 mm²) were acquired on a 3T scanner in the cervicothoracic and/or thoracolumbar cord. Manual whole-cord segmentation was performed. Images were masked and further segmented into cervical, upper thoracic, thoracolumbar, and conus regions. Analyses of covariance were performed for each DTI-derived index to investigate how age affects diffusion across cord regions, and 95% confidence intervals were calculated across age for each derived index and region. Post hoc testing was performed to analyze regional differences.

RESULTS

Analyses of covariance revealed significant correlations of age with axial diffusivity, mean diffusivity, and fractional anisotropy (all, P < .001). There were also significant differences among cord regions for axial diffusivity, radial diffusivity, mean diffusivity, and fractional anisotropy (all, P < .001).

CONCLUSIONS

This research demonstrates that diffusion evolves in the pediatric spinal cord during development, dependent on both cord region and the diffusion index of interest. Future research could investigate how diffusion may be affected by common pediatric spinal pathologies.

Multi-compartmental diffusion characterization of the human cervical spinal cord in vivo using the spherical mean technique

The purpose of this work was to evaluate the feasibility and reproducibility of the spherical mean technique (SMT), a multi-compartmental diffusion model, in the spinal cord of healthy controls, and to assess its ability to improve spinal cord characterization in multiple sclerosis (MS) patients at 3 T. SMT was applied in the cervical spinal cord of eight controls and six relapsing-remitting MS patients. SMT provides an elegant framework to model the apparent axonal volume fraction vax , intrinsic diffusivity Dax , and extra-axonal transverse diffusivity Dex_perp (which is estimated as a function of vax and Dax ) without confounds related to complex fiber orientation distribution that reside in diffusion MRI modeling. SMT's reproducibility was assessed with two different scans within a month, and SMT-derived indices in healthy and MS cohorts were compared. The influence of acquisition scheme on SMT was also evaluated. SMT's vax , Dax , and Dex_perp measurements all showed high reproducibility. A decrease in vax was observed at the site of lesions and normal appearing white matter (p < 0.05), and trends towards a decreased Dax and increased Dex_perp were seen. Importantly, a twofold reduction in acquisition yielded similarly high accuracy with SMT. SMT provides a fast, reproducible, and accurate method to improve characterization of the cervical spinal cord, and may have clinical potential for MS patients.

Spinal cord MRI at 7T

Magnetic resonance imaging (MRI) of the human spinal cord at 7T has been demonstrated by a handful of research sites worldwide, and the spinal cord remains one of the areas in which higher fields and resolution could have high impact. The small diameter of the cord (∼1 cm) necessitates high spatial resolution to minimize partial volume effects between gray and white matter, and so MRI of the cord can greatly benefit from increased signal-to-noise ratio and contrasts at ultra-high field (UHF). Herein we review the current state of UHF spinal cord imaging. Technical challenges to successful UHF spinal cord MRI include radiofrequency (B1) nonuniformities and a general lack of optimized radiofrequency coils, amplified physiological noise, and an absence of methods for robust B0 shimming along the cord to mitigate image distortions and signal losses. Numerous solutions to address these challenges have been and are continuing to be explored, and include novel approaches for signal excitation and acquisition, dynamic shimming and specialized shim coils, and acquisitions with increased coverage or optimal slice angulations.

Effects of coplanar shielding for high field MRI

This work investigates the efficacy of "coplanar shielding," in which copper shields are oriented concentric and coplanar to the RF coils rather than implemented as a full ground plane behind them. Following FDTD simulations to determine optimal shielding parameters, two coil geometries were constructed: a circular loop surface coil and a half-volume five-element receive array. Each was evaluated using bench measurements with and without coplanar shielding. Imaging, including accelerated SENSE imaging, was performed with the shielded and unshielded receive arrays on a whole-body 7T scanner. Results from modeled and fabricated coils showed good agreement with improvements in Q factors for all cases. Imaging showed substantial improvements in SNR and g-factors for the coplanar shielded array.

Evaluating single-point quantitative magnetization transfer in the cervical spinal cord: Application to multiple sclerosis

Spinal cord (SC) damage is linked to clinical deficits in patients with multiple sclerosis (MS), however, conventional MRI methods are not specific to the underlying macromolecular tissue changes that may precede overt lesion detection. Single-point quantitative magnetization transfer (qMT) is a method that can provide high-resolution indices sensitive to underlying macromolecular composition in a clinically feasible scan time by reducing the number of MT-weighted acquisitions and utilizing a two-pool model constrained by empirically determined constants. As the single-point qMT method relies on a priori constraints, it has not been employed extensively in patients, where these constraints may vary, and thus, the biases inherent in this model have not been evaluated in a patient cohort. We, therefore, addressed the potential biases in the single point qMT model by acquiring qMT measurements in the cervical SC in patient and control cohorts and evaluated the differences between the control and patient-derived qMT constraints (kmf, T2fR1f, and T2m) for the single point model. We determined that the macromolecular to free pool size ratio (PSR) differences between the control and patient-derived constraints are not significant (p > 0.149 in all cases). Additionally, the derived PSR for each cohort was compared, and we reported that the white matter PSR in healthy volunteers is significantly different from lesions (p < 0.005) and normal appearing white matter (p < 0.02) in all cases. The single point qMT method is thus a valuable method to quantitatively estimate white matter pathology in MS in a clinically feasible scan time.

Application and evaluation of NODDI in the cervical spinal cord of multiple sclerosis patients

INTRODUCTION

There is a need to develop imaging methods sensitive to axonal injury in multiple sclerosis (MS), given the prominent impact of axonal pathology on disability and outcome. Advanced multi-compartmental diffusion models offer novel indices sensitive to white matter microstructure. One such model, neurite orientation dispersion and density imaging (NODDI), is sensitive to neurite morphology, providing indices of apparent volume fractions of axons (vin), isotropic water (viso) and the dispersion of fibers about a central axis (orientation dispersion index, ODI). NODDI has yet to be studied for its sensitivity to spinal cord pathology. Here, we investigate the feasibility and utility of NODDI in the cervical spinal cord of MS patients.

METHODS

NODDI was applied in the cervical spinal cord in a cohort of 8 controls and 6 MS patients. Statistical analyses were performed to test the sensitivity of NODDI-derived indices to pathology in MS (both lesion and normal appearing white matter NAWM). Diffusion kurtosis imaging (DKI) and diffusion tensor imaging (DTI) analysis were also performed to compare with NODDI.

RESULTS

A decrease in NODDI-derived vin was observed at the site of the lesion (p < 0.01), whereas a global increase in ODI was seen throughout white matter (p < 0.001). DKI-derived mean kurtosis (MK) and radial kurtosis (RK) and DTI-derived fractional anisotropy (FA) and radial diffusivity (RD) were all significantly different in MS patients (p < 0.02), however NODDI provided higher contrast between NAWM and lesion in all MS patients.

CONCLUSION

NODDI provides unique contrast that is not available with DKI or DTI, enabling improved characterization of the spinal cord in MS.

Amide proton transfer CEST of the cervical spinal cord in multiple sclerosis patients at 3T

PURPOSE

The ability to evaluate pathological changes in the spinal cord in multiple sclerosis (MS) is limited because T₁ - and T₂ -w MRI imaging are not sensitive to biochemical changes in vivo. Amide proton transfer (APT) chemical exchange saturation transfer (CEST) can indirectly detect amide protons associated with proteins and peptides, potentially providing more pathological specificity. Here, we implement APT CEST in the cervical spinal cord of healthy and MS cohorts at 3T.

METHODS

APT CEST of the cervical spinal cord was obtained in a cohort of 10 controls and 10 MS patients using a novel respiratory correction methodology. APT was quantified using two methods: 1) APT_w , based off the conventional magnetization transfer ratio asymmetry, and 2) ΔAPT, a spatial characterization of APT changes in MS patients relative to the controls.

RESULTS

Respiratory correction yielded highly reproducible z-spectra in white matter (intraclass correlation coefficient = 0.82). APT_w signals in normal-appearing white matter (NAWM) of MS patients were significantly different from healthy controls (P = 0.04), whereas ΔAPT of MS patients highlighted large APT differences in NAWM.

CONCLUSION

Respiration correction in the spinal cord is necessary to accurately quantify APT CEST, which can provide unique biochemical information regarding disease processes within the spinal cord. Magn Reson Med 79:806-814, 2018. © 2017 International Society for Magnetic Resonance in Medicine.

Quantifying the impact of underlying measurement error on cervical spinal cord diffusion tensor imaging at 3T

PURPOSE

To empirically characterize and quantify the impact of gradient weighting schemes on the appearance and fidelity of diffusion tensor imaging of the human spinal cord in vivo in clinically relevant scan time equivalents (STE).

MATERIALS AND METHODS

In five healthy controls at 3T, we evaluated test-retest reproducibility and performed voxelwise analysis of diffusion tensor imaging (DTI)-derived indices (fractional anisotropy [FA], mean [MD], axial [AD], and radial [RD] diffusivity) in the cervical spinal cord to assess spatial dependencies of measurement error and differences across three different sampling schemes (6, 15, and 32 directions) at STE of 4.5, 9, and 18 minutes. A subjective assessment was also performed.

RESULTS

With six directions, column-specific errors are highest (effect size = 2.9%, 4.4%, 7.2% for FA in dorsal column, lateral column, and gray matter) and different than the 15-direction scheme (P < 0.05). STE sequences with 15 and 32 directions exhibited small differences in error (P > 0.05). For FA and AD, measurement errors are prevalent in gray matter, while partial volume effects with cerebrospinal fluid heavily influence RD. Measurement errors decreased with increasing scan time (P < 0.01), albeit with diminishing returns at scan times longer than 9 minutes (P < 0.05).

CONCLUSION

A 15-direction scheme of 9 minutes yields measurements of the cervical spinal cord with low error. J. Magn. Reson. Imaging 2016;44:1608-1618.

A 16-channel receive, forced current excitation dual-transmit coil for breast imaging at 7T

Purpose

To enable high spatial and temporal breast imaging resolution via combined use of high field MRI, array coils, and forced current excitation (FCE) multi channel transmit.

Materials and Methods

A unilateral 16-channel receive array insert was designed for use in a transmit volume coil optimized for quadrature operation with dual-transmit RF shimming at 7T. Signal-to-noise ratio (SNR) maps, g-factor maps, and high spatial and temporal resolution in vivo images were acquired to demonstrate the utility of the coil architecture.

Results

The dual-transmit FCE coil provided homogeneous excitation and the array provided an increase in average SNR of 3.3 times (max 10.8, min 1.5) compared to the volume coil in transmit/receive mode. High resolution accelerated in vivo breast imaging demonstrated the ability to achieve isotropic spatial resolution of 0.5 mm within clinically relevant 90 s scan times, as well as the ability to perform 1.0 mm isotropic resolution imaging, 7 s per dynamics, with the use of bidirectional SENSE acceleration of up to R = 9.

Conclusion

The FCE design of the transmit coil easily accommodates the addition of a sixteen channel array coil. The improved spatial and temporal resolution provided by the high-field array coil with FCE dual-channel transmit will ultimately be beneficial in lesion detection and characterization.