Acceleration of Double Inversion Recovery Sequences in Multiple Sclerosis With Compressed Sensing

Accelerating TOF-MRA: The impact of the combined use of compressed sensitivity encoding and spiral imaging

PURPOSE

To evaluate the image quality of the combined technique of compressed sensitivity encoding (CS) and spiral imaging in time-of-flight magnetic resonance angiography (TOF-MRA), which is approximately 2.5 times faster than conventional methods.

METHODS

Twenty volunteers underwent four TOF-MRA sequences: sensitivity encoding (SENSE) with acceleration factor of 4 (acquisition time: 4:55 min), CS with acceleration factor of 10.9, and spiral and CS-spiral (both 1:55 min). A quantitative image analysis (signal-to-noise ratio [SNR], contrast, and full width at half maximum [FWHM] edge criterion measurements) was performed on four TOF sequences. For qualitative image analysis, two board-certified radiologists evaluated the overall depiction of the proximal, intermediate, and distal branches in CS, spiral, and CS-spiral images using SENSE as a reference.

RESULTS

The SNR of BA in spiral and CS-spiral imaging was significantly lower than that in SENSE (p = 0.009). The contrasts of ACA and BA in CS-spiral were significantly higher and those in spiral were significantly lower than those in SENSE (p < 0.001). The FWHM in the CS image was significantly higher than that of SENSE; however, no significant differences were observed between the spiral or CS-spiral and SENSE. In qualitative analysis, the depiction of proximal vascular branches was significantly impaired in spiral than in others and that of distal vascular branches was significantly impaired in CS than in others (p < 0.001).

CONCLUSIONS

In TOF-MRA, which is approximately 2.5 times faster than conventional methods, the combined use of CS and spiral imaging demonstrated an improvement in image quality compared to either CS or spiral imaging alone.

SUMMARY STATEMENT

The image quality of Compressed SENSE and spiral imaging is particularly poor in the proximal and distal vascular branches, respectively at an extremely high acceleration factor; however, CS-spiral provided stable image quality in all regions as compared with the SENSE technique.

Multiple sclerosis imaging in clinical practice: a European-wide survey of 428 centers and conclusions by the ESNR Working Group

OBJECTIVES

To evaluate compliance with the available recommendations, we assessed the current clinical practice of imaging in the evaluation of multiple sclerosis (MS).

METHODS

An online questionnaire was emailed to all members and affiliates. Information was gathered on applied MR imaging protocols, gadolinium-based contrast agents (GBCA) use and image analysis. We compared the survey results with the Magnetic Resonance Imaging in MS (MAGNIMS) recommendations considered as the reference standard.

RESULTS

A total of 428 entries were received from 44 countries. Of these, 82% of responders were neuroradiologists. 55% performed more than ten scans per week for MS imaging. The systematic use of 3 T is rare (18%). Over 90% follow specific protocol recommendations with 3D FLAIR, T2-weighted and DWI being the most frequently used sequences. Over 50% use SWI at initial diagnosis and 3D gradient-echo T1-weighted imaging is the most used MRI sequence for pre- and post-contrast imaging. Mismatches with recommendations were identified including the use of only one sagittal T2-weighted sequence for spinal cord imaging, the systematic use of GBCA at follow-up (over 30% of institutions), a delay time shorter than 5 min after GBCA administration (25%) and an inadequate follow-up duration in pediatric acute disseminated encephalomyelitis (80%). There is scarce use of automated software to compare images or to assess atrophy (13% and 7%). The proportions do not differ significantly between academic and non-academic institutions.

CONCLUSIONS

While current practice in MS imaging is rather homogeneous across Europe, our survey suggests that recommendations are only partially followed.

CLINICAL RELEVANCE STATEMENT

Hurdles were identified, mainly in the areas of GBCA use, spinal cord imaging, underuse of specific MRI sequences and monitoring strategies. This work will help radiologists to identify the mismatches between their own practices and the recommendations and act upon them.

KEY POINTS

• While current practice in MS imaging is rather homogeneous across Europe, our survey suggests that available recommendations are only partially followed. • Several hurdles have been identified through the survey that mainly lies in the areas of GBCA use, spinal cord imaging, underuse of specific MRI sequences and monitoring strategies.

Uncertainty-Aware and Lesion-Specific Image Synthesis in Multiple Sclerosis Magnetic Resonance Imaging: A Multicentric Validation Study

Generative adversarial networks (GANs) can synthesize high-contrast MRI from lower-contrast input. Targeted translation of parenchymal lesions in multiple sclerosis (MS), as well as visualization of model confidence further augment their utility, provided that the GAN generalizes reliably across different scanners. We here investigate the generalizability of a refined GAN for synthesizing high-contrast double inversion recovery (DIR) images and propose the use of uncertainty maps to further enhance its clinical utility and trustworthiness. A GAN was trained to synthesize DIR from input fluid-attenuated inversion recovery (FLAIR) and T1w of 50 MS patients (training data). In another 50 patients (test data), two blinded readers (R1 and R2) independently quantified lesions in synthetic DIR (synthDIR), acquired DIR (trueDIR) and FLAIR. Of the 50 test patients, 20 were acquired on the same scanner as training data (internal data), while 30 were scanned at different scanners with heterogeneous field strengths and protocols (external data). Lesion-to-Background ratios (LBR) for MS-lesions vs. normal appearing white matter, as well as image quality parameters were calculated. Uncertainty maps were generated to visualize model confidence. Significantly more MS-specific lesions were found in synthDIR compared to FLAIR (R1: 26.7 ± 2.6 vs. 22.5 ± 2.2 p < 0.0001; R2: 22.8 ± 2.2 vs. 19.9 ± 2.0, p = 0.0005). While trueDIR remained superior to synthDIR in R1 [28.6 ± 2.9 vs. 26.7 ± 2.6 (p = 0.0021)], both sequences showed comparable lesion conspicuity in R2 [23.3 ± 2.4 vs. 22.8 ± 2.2 (p = 0.98)]. Importantly, improvements in lesion counts were similar in internal and external data. Measurements of LBR confirmed that lesion-focused GAN training significantly improved lesion conspicuity. The use of uncertainty maps furthermore helped discriminate between MS lesions and artifacts. In conclusion, this multicentric study confirms the external validity of a lesion-focused Deep-Learning tool aimed at MS imaging. When implemented, uncertainty maps are promising to increase the trustworthiness of synthetic MRI.

3-Dimensional Fluid and White Matter Suppression Magnetic Resonance Imaging Sequence Accelerated With Compressed Sensing Improves Multiple Sclerosis Cervical Spinal Cord Lesion Detection Compared With Standard 2-Dimensional Imaging

OBJECTIVES

Fluid and white matter suppression (FLAWS) is a recently proposed magnetic resonance sequence derived from magnetization-prepared 2 rapid acquisition gradient-echo providing 2 coregistered datasets with white matter- and cerebrospinal fluid-suppressed signal, enabling synthetic imaging with amplified contrast. Although these features are high potential for brain multiple sclerosis (MS) imaging, spinal cord has never been evaluated with this sequence to date. The objective of this work was therefore to assess diagnostic performance and self-confidence provided by compressed-sensing (CS) 3-dimensional (3D) FLAWS for cervical MS lesion detection on a head scan that includes the cervical cord without changing standard procedures.

MATERIALS AND METHODS

Prospective 3 T scans (MS first diagnosis or follow-up) acquired between 2019 and 2020 were retrospectively analyzed. All patients underwent 3D CS-FLAWS (duration: 5 minutes 40 seconds), axial T2 turbo spin echo covering cervical spine from cervicomedullary junction to the same inferior level as FLAWS, and sagittal cervical T2/short tau inversion recovery imaging. Two readers performed a 2-stage double-blind reading, followed by consensus reading. Wilcoxon tests were used to compare the number of detected spinal cord lesions and the reader's diagnostic self-confidence when using FLAWS versus the reference 2D T2-weighted imaging.

RESULTS

Fifty-eight patients were included (mean age, 40 ± 13 years, 46 women, 7 ± 6 years mean disease duration). The CS-FLAWS detected significantly more lesions than the reference T2-weighted imaging (197 vs 152 detected lesions, P < 0.001), with a sensitivity of 98% (T2-weighted imaging sensitivity: 90%) after consensual reading. Considering the subgroup of patients who underwent sagittal T2 + short tau inversion recovery imaging (Magnetic Resonance Imaging for Multiple Sclerosis subgroup), +250% lesions were detected with FLAWS (63 vs 25 lesions detected, P < 0.001). Mean reading self-confidence was significantly better with CS-FLAWS (median, 5 [interquartile range, 1] [no doubt for diagnosis] vs 4 [interquartile range, 1] [high confidence]; P < 0.001).

CONCLUSIONS

Imaging with CS-FLAWS provides an improved cervical spinal cord exploration for MS with increased self-confidence compared with conventional T2-weighted imaging, in a clinically acceptable time.

Multi-parameter quantitative mapping of R1, R2*, PD, and MTsat is reproducible when accelerated with Compressed SENSE

Multi-parameter mapping (MPM) magnetic resonance imaging (MRI) provides quantitative estimates of the longitudinal and effective transverse relaxation rates R1 and R2*, proton density (PD), and magnetization transfer saturation (MTsat). Thereby, MPM enables better comparability across sites and time than conventional weighted MRI. However, for MPM, several contrasts must be acquired, resulting in prolonged measurement durations and thus preventing MPM's application in clinical routines. State-of-the-art imaging acceleration techniques such as Compressed SENSE (CS), a combination of compressed sensing and sensitivity encoding, can be used to reduce the scan time of MPM. However, the accuracy and precision of the resulting quantitative parameter maps have not been systematically evaluated. In this study, we therefore investigated the effect of CS acceleration on the fidelity and reproducibility of MPM acquisitions. In five healthy volunteers and in a phantom, we compared MPM metrics acquired without imaging acceleration, with the standard acceleration (SENSE factor 2.5), and with Compressed SENSE with acceleration factors 4 and 6 using a 32-channel head coil. We evaluated the reproducibility and repeatability of accelerated MPM using data from three scan sessions in gray and white matter volumes-of-interest (VOIs). Accelerated MPM provided precise and accurate quantitative parameter maps. For most parameters, the results of the CS-accelerated protocols correlated more strongly with the non-accelerated protocol than the standard SENSE-accelerated protocols. Furthermore, for most VOIs and contrasts, coefficients of variation were lower when calculated from data acquired with different imaging accelerations within a single scan session than from data acquired in different scan sessions. These results suggest that MPM with Compressed SENSE acceleration factors up to at least 6 yields reproducible quantitative parameter maps that are highly comparable to those acquired without imaging acceleration. Compressed SENSE can thus be used to considerably reduce the scan duration of R1, R2*, PD, and MTsat mapping, and is highly promising for clinical applications of MPM.

High-Resolution 3D versus Standard-Resolution 2D T2-Weighted Turbo Spin Echo MRI for the Assessment of Lumbar Nerve Root Compromise

Radiculopathy can be caused by nerve root irritation and nerve root compression at the level of the lateral recess or at the level of the intervertebral foramen. T2-weighted (T2w) MRI is considered essential to evaluate the nerve root and its course, starting at the lateral recess through the intervertebral foramen to the extraforaminal space. With the introduction of novel MRI acceleration techniques such as compressed SENSE, standard-resolution 2D T2w turbo spin echo (TSE) sequences with a slice-thickness of 3–4 mm can be replaced with high-resolution isotropic 3D T2w TSE sequences with sub-millimeter resolution without prolonging scan time. With high-resolution 3D MRI, the course of the nerve root can be visualized more precisely due to a detailed depiction of the anatomical situation and less partial volume effects, potentially allowing for a better detection of nerve root compromise. In this intra-individual comparison study, 55 patients with symptomatic unilateral singular nerve root radiculopathy underwent MRI with both 2D standard- and 3D high-resolution T2w TSE MRI sequences. Two readers graded the degree of lumbar lateral recess stenosis and lumbar foraminal stenosis twice on both image sets using previously validated grading systems in an effort to quantify the inter-readout and inter-sequence agreement of scores. Inter-readout agreement was high for both grading systems and for 2D and 3D imaging (Kappa = 0.823–0.945). Inter-sequence agreement was moderate for both lumbar lateral recess stenosis (Kappa = 0.55–0.577) and lumbar foraminal stenosis (Kappa = 0.543–0.572). The percentage of high degree stenosis with nerve root deformity increased from 16.4%/9.8% to 41.8–43.6%/34.1% from 2D to 3D images for lateral recess stenosis/foraminal stenosis, respectively. Therefore, we show that while inter-readout agreement of grading systems is high for both standard- and high-resolution imaging, the latter outperforms standard-resolution imaging for the visualization of lumbar nerve root compromise.

Spiral 3D time-of-flight MR angiography for rapid non-contrast carotid artery imaging: Clinical feasibility and protocol optimization

PURPOSE

To assess the clinical feasibility of spiral 3D Time-Of-Flight (TOF) MR Angiography (MRA) sequence variants for rapid non-contrast carotid artery imaging.

METHODS

Nine different 3D TOF MRA sequences were acquired in nine healthy volunteers on a standard clinical 1.5 T scanner. Three cartesian sequences (fully sampled (10:15 min), accelerated with SENSE (05:08 min), accelerated with Compressed SENSE (03:32 min)) and six different spiral sequences were acquired (spiral acquisition windows ranging from 10 to 5 ms (01:32 min-03:05 min)). Three readers graded the images qualitatively in terms of overall image quality, vessel sharpness, inhomogeneous intraluminal signal, background noise, visualization of large and small vessels and overall impression of the number of visible vessels. Cross-sectional areas of the vessel lumen were measured and vessel sharpness was quantified.

RESULTS

The SENSE and Compressed SENSE accelerated cartesian sequences and the Spiral 6 ms and 5 ms sequences were deemed comparable to the fully sampled cartesian sequence in most qualitative categories (p > 0.05) based on exact binomial tests. The Spiral 6 ms and 5 ms sequences achieved a scan time reduction of 75.3% and 69.9% respectively compared to the fully sampled cartesian sequence. The spiral sequences (generally) exhibited improved subjective vessel sharpness (p < 0.01-p = 0.13) but increased background noise (p = 0.03-p = 0.25). Cross-sectional area measurements were similar between all sequences (Krippendorff's alpha: 0.955-0.982). Quantitative vessel sharpness was increased for all spiral sequences compared to all cartesian sequences (p = 0.004).

CONCLUSIONS

Spiral 3D TOF MRA sequences with a spiral acquisition window of 5 ms or 6 ms may be used for accurate, rapid, clinical non-contrast carotid artery imaging.

New Prospects for Ultra-High-Field Magnetic Resonance Imaging in Multiple Sclerosis

ABSTRACT

There is growing interest in imaging multiple sclerosis (MS) through the ultra-high-field (UHF) lens, which currently means a static magnetic field strength of 7 T or higher. Because of higher signal-to-noise ratio and enhanced susceptibility effects, UHF magnetic resonance imaging improves conspicuity of MS pathological hallmarks, among them cortical demyelination and the central vein sign. This could, in turn, improve confidence in MS diagnosis and might also facilitate therapeutic monitoring of MS patients. Furthermore, UHF imaging offers unique insight into iron-related pathology, leptomeningeal inflammation, and spinal cord pathologies in neuroinflammation. Yet, limitations such as the longer scanning times to achieve improved resolution and incipient safety data on implanted medical devices need to be considered. In this review, we discuss applications of UHF imaging in MS, its advantages and limitations, and practical aspects of UHF in the clinical setting.

Qualitative and Quantitative Analysis of a Spiral Gradient Echo Sequence for Contrast-Enhanced Fat-Suppressed T1-Weighted Spine Magnetic Resonance Imaging

OBJECTIVES

Pulse sequences with non-Cartesian k-space sampling enable improved imaging in anatomical areas with high degrees of motion artifacts. We analyzed a novel spiral 3-dimensional (3D) gradient echo (GRE) magnetic resonance imaging (MRI) sequence ("spiral," 114.7 ± 11 seconds) and compared it with a radial 3D GRE ("vane," 216.7 ± 2 seconds) and a conventional Cartesian 2D turbo spin echo (TSE) sequence ("TSE," 266.7 ± 82 seconds) for contrast-enhanced fat-suppressed T1-weighted spine imaging.

MATERIALS AND METHODS

Forty consecutive patients referred for contrast-enhanced MRI were prospectively scanned with all 3 sequences. A qualitative analysis was performed by 3 readers using 4- or 5-point Likert scales to independently grade images in terms of overall image quality, occurrence of artifacts, lesion conspicuity, and conspicuity of nerve roots. The numbers of visible nerve roots per sequence and patient were counted in consensus. Coefficient of variation measurements were performed for the paravertebral musculature (CVPM) and the spinal cord (CVSC).

RESULTS

Spiral (median [interquartile range], 5 [4-5]) exhibited improved overall image quality in comparison to TSE (3 [3-4]) and vane (4 [4-5]; both P < 0.001). Vane surpassed TSE in terms of overall image quality (P < 0.001). Spiral (4 [3.75-4]) and vane (3.5 [3-4]) presented with less artifacts than TSE (3 [2.75-3.25]; both P < 0.001). Spiral (4 [4-5]) outperformed vane (4 [3-5]; P = 0.01) and TSE (4 [3-4]; P = 0.04) in terms of lesion conspicuity. Conspicuity of nerve roots was superior on spiral (3 [3-4]) and vane (4 [3-4]) when compared with TSE (1.5 [1-2]; both P < 0.001). Readers discerned significantly more nerve roots on spiral (4 [2.75-8]) and vane (4 [3.75-7.25]) images when compared with TSE (2 [0-4]; both P < 0.001). Interreader agreement ranged from moderate (α = 0.639) to almost perfect (α = 0.967). CVPM and CVSC were significantly lower on spiral as compared with vane and TSE (P < 0.001, P = 0.04). Vane exhibited lower CVPM and CVSC than TSE (P < 0.001, P = 0.01).

CONCLUSIONS

A novel spiral 3D GRE sequence improves contrast-enhanced fat-suppressed T1-weighted spinal imaging qualitatively and quantitatively in comparison with a conventional Cartesian 2D TSE sequence and to a lesser extent with a radial 3D GRE sequence at shorter scan times.

Navigator-Guided Motion and B0 Correction of T2*-Weighted Magnetic Resonance Imaging Improves Multiple Sclerosis Cortical Lesion Detection

BACKGROUND

Cortical lesions are common in multiple sclerosis (MS). T2*-weighted (T2*w) imaging at 7 T is relatively sensitive for cortical lesions, but quality is often compromised by motion and main magnetic field (B0) fluctuations.

PURPOSE

The aim of this study was to determine whether motion and B0 correction with a navigator-guided gradient-recalled echo sequence can improve cortical lesion detection in T2*w magnetic resonance imaging.

MATERIALS AND METHODS

In this prospective study, a gradient-recalled echo sequence incorporating a navigator allowing for motion and B0 field correction was applied to collect T2*w images at 7 T from adults with MS between August 2019 and March 2020. T2*-weighted images were acquired in 1 to 3 partially overlapping scans per individual and were reconstructed using global average B0 correction ("uncorrected") or motion correction and spatially linear B0 correction ("corrected"). Image quality rating and manual segmentation of cortical lesions were performed on uncorrected and corrected images. Lesions seen on a single scan were retrospectively evaluated on the complementary scan. The association of cortical lesions with clinical disability was assessed. Mixed models were used to determine the effect of correction on lesion detection as well as on the relationship between disability and lesion count.

RESULTS

A total of 22 T2*w scans were performed on 11 adults with MS (mean [SD] age, 49 [11] years; 8 women). Quality improved for 20 of 22 scans (91%) after correction. A total of 69 cortical lesions were identified on uncorrected images (median per scan, 2; range, 0-11) versus 148 on corrected images (median per scan, 4.5; range, 0-25; rate ratio [RR], 2.1; P < 0.0001). For low-quality uncorrected scans with moderate to severe motion artifact (18/22, 82%), there was an improvement in cortical lesion detection with correction (RR, 2.5; P < 0.0001), whereas there was no significant change in cortical lesion detection for high-quality scans (RR, 1.3; P = 0.43).

CONCLUSIONS

Navigator-guided motion and B0 correction substantially improves the overall image quality of T2*w magnetic resonance imaging at 7 T and increases its sensitivity for cortical lesions.

Introduction and reproducibility of an updated practical grading system for lumbar foraminal stenosis based on high-resolution MR imaging

In this paper we sought to develop and assess the reproducibility of an updated 6-point grading system for lumbar foraminal stenosis based on the widely used Lee classification that more accurately describes lumbar foraminal stenosis as seen on high-resolution MRI. Grade A indicates absence of foraminal stenosis. Grades B, C, D and E indicate presence of foraminal stenosis with contact of the nerve root with surrounding anatomical structures (on one, two, three or four sides for B, C, D and E respectively) yet without morphological change of the nerve root. To each grade, a number code indicating the location of contact between the nerve root and surrounding anatomical structure(s) is appended. 1, 2, 3 and 4 indicate contact of the nerve root at superior, posterior, inferior and anterior position of the borders of the lumbar foramen. Grade F indicates presence of foraminal stenosis with morphological change of the nerve root. Three readers graded the lumbar foramina of 101 consecutive patients using high-resolution T2w (and T1w) MR images with a spatial resolution of beyond 0.5 mm³. Interreader agreement was excellent (Cohen’s Kappa = 0.866–1). Importantly, 30.6%/31.6%/32.2% (reader 1/reader 2/ reader 3) of foramina were assigned grades that did not appear in the original Lee grading system (grades B and D). The readers found no foramen that could not be described accurately with the updated grading system. Thus, an updated 6-point grading system for lumbar foraminal stenosis is reproducible and comprehensively describes lumbar foraminal stenosis as seen on high-resolution MRI.

Spiral 3-Dimensional T1-Weighted Turbo Field Echo: Increased Speed for Magnetization-Prepared Gradient Echo Brain Magnetic Resonance Imaging

OBJECTIVES

Spiral magnetic resonance imaging acquisition may enable improved image quality and higher scan speeds than Cartesian trajectories. We tested the performance of four 3D T1-weighted (T1w) TFE sequences (magnetization-prepared gradient echo magnetic resonance sequence) with isotropic spatial resolution for brain imaging at 1.5 T in a clinical patient cohort based on qualitative and quantitative image quality metrics. Two prototypical spiral TFE sequences (spiral 1.0 and spiral 0.85) and a Cartesian compressed sensing technology accelerated TFE sequence (CS 2.5; acceleration factor of 2.5) were compared with a conventional (reference standard) Cartesian parallel imaging accelerated TFE sequence (SENSE; acceleration factor of 1.8).

MATERIALS AND METHODS

The SENSE (5:52 minutes), CS 2.5 (3:17 minutes), and spiral 1.0 (2:16 minutes) sequences all had identical spatial resolutions (1.0 mm). The spiral 0.85 (3:47 minutes) had a higher spatial resolution (0.85 mm). The 4 TFE sequences were acquired in 41 patients (20 with and 21 without contrast media). Three readers rated qualitative image quality (12 categories) and selected their preferred sequence for each patient. Two readers performed quantitative analysis whereby 6 metrics were derived: contrast-to-noise ratio for white and gray matter (CNRWM/GM), contrast ratio for gray matter-CSF (CRGM/CSF), and white matter-CSF (CRWM/CSF); and coefficient of variations for gray matter (CVGM), white matter (CVWM), and CSF (CVCSF). Friedman tests with post hoc Nemenyi tests, exact binomial tests, analysis of variance with post hoc Dunnett tests, and Krippendorff alphas were computed.

RESULTS

Concerning qualitative analysis, the CS 2.5 sequence significantly outperformed the SENSE in 4/1 (with/without contrast) categories, whereas the spiral 1.0 and spiral 0.85 showed significantly improved scores in 10/9and 7/7 categories, respectively (P's < 0.001-0.039). The spiral 1.0 was most frequently selected as the preferred sequence (reader 1, 10/15 times; reader 2, 9/12 times; reader 3, 11/13times [with/without contrast]). Interreader agreement ranged from substantial to almost perfect (alpha = 0.615-0.997). Concerning quantitative analysis, compared with the SENSE, the CS 2.5 had significantly better scores in 2 categories (CVWM, CVCSF) and worse scores in 2 categories (CRGM/CSF, CRWM/CSF), the spiral 1.0 had significantly improved scores in 4 categories (CNRWM/GM, CRGM/CSF, CRWM/CSF, CVWM), and the spiral 0.85 had significantly better scores in 2 categories (CRGM/CSF, CRWM/CSF).

CONCLUSIONS

Spiral T1w TFE sequences may deliver high-quality clinical brain imaging, thus matching the performance of conventional parallel imaging accelerated T1w TFEs. Imaging can be performed at scan times as short as 2:16 minutes per sequence (61.4% scan time reduction compared with SENSE). Optionally, spiral imaging enables increased spatial resolution while maintaining the scan time of a Cartesian-based acquisition schema.

Deep Learning on Conventional Magnetic Resonance Imaging Improves the Diagnosis of Multiple Sclerosis Mimics

OBJECTIVES

The aims of this study were to present a deep learning approach for the automated classification of multiple sclerosis and its mimics and compare model performance with that of 2 expert neuroradiologists.

MATERIALS AND METHODS

A total of 268 T2-weighted and T1-weighted brain magnetic resonance imagin scans were retrospectively collected from patients with migraine (n = 56), multiple sclerosis (n = 70), neuromyelitis optica spectrum disorders (n = 91), and central nervous system vasculitis (n = 51). The neural network architecture, trained on 178 scans, was based on a cascade of 4 three-dimensional convolutional layers, followed by a fully dense layer after feature extraction. The ability of the final algorithm to correctly classify the diseases in an independent test set of 90 scans was compared with that of the neuroradiologists.

RESULTS

The interrater agreement was 84.9% (Cohen κ = 0.78, P < 0.001). In the test set, deep learning and expert raters reached the highest diagnostic accuracy in multiple sclerosis (98.8% vs 72.8%, P < 0.001, for rater 1; and 81.8%, P < 0.001, for rater 2) and the lowest in neuromyelitis optica spectrum disorders (88.6% vs 4.4%, P < 0.001, for both raters), whereas they achieved intermediate values for migraine (92.2% vs 53%, P = 0.03, for rater 1; and 64.8%, P = 0.01, for rater 2) and vasculitis (92.1% vs 54.6%, P = 0.3, for rater 1; and 45.5%, P = 0.2, for rater 2). The overall performance of the automated method exceeded that of expert raters, with the worst misdiagnosis when discriminating between neuromyelitis optica spectrum disorders and vasculitis or migraine.

CONCLUSIONS

A neural network performed better than expert raters in terms of accuracy in classifying white matter disorders from magnetic resonance imaging and may help in their diagnostic work-up.

Deep-Learning Generated Synthetic Double Inversion Recovery Images Improve Multiple Sclerosis Lesion Detection

OBJECTIVES

The aim of the study was to implement a deep-learning tool to produce synthetic double inversion recovery (synthDIR) images and compare their diagnostic performance to conventional sequences in patients with multiple sclerosis (MS).

MATERIALS AND METHODS

For this retrospective analysis, 100 MS patients (65 female, 37 [22-68] years) were randomly selected from a prospective observational cohort between 2014 and 2016. In a subset of 50 patients, an artificial neural network (DiamondGAN) was trained to generate a synthetic DIR (synthDIR) from standard acquisitions (T1, T2, and fluid-attenuated inversion recovery [FLAIR]). With the resulting network, synthDIR was generated for the remaining 50 subjects. These images as well as conventionally acquired DIR (trueDIR) and FLAIR images were assessed for MS lesions by 2 independent readers, blinded to the source of the DIR image. Lesion counts in the different modalities were compared using a Wilcoxon signed-rank test, and interrater analysis was performed. Contrast-to-noise ratios were compared for objective image quality.

RESULTS

Utilization of synthDIR allowed to detect significantly more lesions compared with the use of FLAIR images (31.4 ± 20.7 vs 22.8 ± 12.7, P < 0.001). This improvement was mainly attributable to an improved depiction of juxtacortical lesions (12.3 ± 10.8 vs 7.2 ± 5.6, P < 0.001). Interrater reliability was excellent in FLAIR 0.92 (95% confidence interval [CI], 0.85-0.95), synthDIR 0.93 (95% CI, 0.87-0.96), and trueDIR 0.95 (95% CI, 0.85-0.98).Contrast-to-noise ratio in synthDIR exceeded that of FLAIR (22.0 ± 6.4 vs 16.7 ± 3.6, P = 0.009); no significant difference was seen in comparison to trueDIR (22.0 ± 6.4 vs 22.4 ± 7.9, P = 0.87).

CONCLUSIONS

Computationally generated DIR images improve lesion depiction compared with the use of standard modalities. This method demonstrates how artificial intelligence can help improving imaging in specific pathologies.

Age-Related Changes in Relaxation Times, Proton Density, Myelin, and Tissue Volumes in Adult Brain Analyzed by 2-Dimensional Quantitative Synthetic Magnetic Resonance Imaging

OBJECTIVES

Quantitative synthetic magnetic resonance imaging (MRI) enables the determination of fundamental tissue properties, namely, T1 and T2 relaxation times and proton density (PD), in a single scan. Myelin estimation and brain segmentation based on these quantitative values can also be performed automatically. This study aimed to reveal the changes in tissue characteristics and volumes of the brain according to age and provide age-specific reference values obtained by quantitative synthetic MRI.

MATERIALS AND METHODS

This was a prospective study of healthy subjects with no history of brain diseases scanned with a multidynamic multiecho sequence for simultaneous measurement of relaxometry of T1, T2, and PD. We performed myelin estimation and brain volumetry based on these values. We performed volume-of-interest analysis on both gray matter (GM) and white matter (WM) regions for T1, T2, PD, and myelin volume fraction maps. Tissue volumes were calculated in the whole brain, producing brain parenchymal volume, GM volume, WM volume, and myelin volume. These volumes were normalized by intracranial volume to a brain parenchymal fraction, GM fraction, WM fraction, and myelin fraction (MyF). We examined the changes in the mean regional quantitative values and segmented tissue volumes according to age.

RESULTS

We analyzed data of 114 adults (53 men and 61 women; median age, 66.5 years; range, 21-86 years). T1, T2, and PD values showed quadratic changes according to age and stayed stable or decreased until around 60 years of age and increased thereafter. Myelin volume fraction showed a reversed trend. Brain parenchymal fraction and GM fraction decreased throughout all ages. The approximation curves showed that WM fraction and MyF gradually increased until around the 40s to 50s and decreased thereafter. A significant decline in MyF was first noted in the 60s age group (Tukey test, P < 0.001).

CONCLUSIONS

Our study showed changes according to age in tissue characteristic values and brain volumes using quantitative synthetic MRI. The reference values for age demonstrated in this study may be useful to discriminate brain disorders from healthy brains.

Characterization of Leukoencephalopathy and Association With Later Neurocognitive Performance in Pediatric Acute Lymphoblastic Leukemia

OBJECTIVES

The most common form of pediatric cancer is acute lymphoblastic leukemia (ALL). Magnetic resonance (MR) neuroimaging studies have revealed leukoencephalopathy (LE) in pediatric ALL, but the impact of LE on long-term neurocognitive performance remains unknown. This study aims to objectively characterize the prevalence, extent, and intensity of LE, and their association with later neurocognitive performance.

MATERIALS AND METHODS

Pediatric patients (N = 377) treated for ALL without irradiation underwent MR neuroimaging at 4 time points throughout therapy (end of remission induction [MR1], end of consolidation [MR2], and week 31 [MR3] and week 120 [end therapy, MR4] of continuation treatment) and neurocognitive evaluations at the end of therapy and 2 years later. Generalized estimation equation models with logit link were developed to explore the association between LE prevalence and extent with time points throughout therapy, age at diagnosis (≤5 years or >5 years), treatment risk arm (low risk or standard/high risk), and sex. General linear models were also developed to investigate the association between neuroimaging metrics during treatment and neurocognitive performance at 2-year follow-up.

RESULTS

The prevalence of LE was greatest (22.8%, 74/324) after consolidation therapy. The prevalence of LE increased at MR2 relative to MR1 regardless of treatment risk arm (both P's < 0.001), age group (both P's < 0.001), or sex (male, P < 0.001; female, P = 0.013). The extent of white matter affected also increased at MR2 relative to MR1 regardless of treatment risk arm (standard/high risk, P < 0.001; low risk, P = 0.004), age group (both P's < 0.001), or sex (male, P < 0.001; female, P = 0.001). Quantitative relaxation rates were significantly longer in LE compared with that in normal-appearing white matter in the same examination (T1, P < 0.001; T2, P < 0.001). The LE prevalence early in therapy was associated with increased parent ratings of conduct problems (P = 0.039) and learning difficulties (P = 0.036) at 2-year follow-up compared with that at the end of therapy. A greater extent of LE early in therapy was associated with decreasing performance on a measure of processing speed (P = 0.003) from the end of therapy to 2-year follow-up. A larger extent of LE at the end of therapy was associated with decreased performance in reading (P = 0.004), spelling (P = 0.003), and mathematics (P = 0.019) at 2-year follow-up and increasing problems with attention (omissions, P = 0.045; β, P = 0.015) and memory (list A total recall, P = 0.010) at 2-year follow-up compared with that at the end of therapy.

CONCLUSIONS

In this large cohort of pediatric patients treated for ALL without irradiation, asymptomatic LE during therapy can be seen in almost a quarter of patients, involves as much as 10% of the white matter volume, and is associated with decreasing neurocognitive performance, increasing parent reports of conduct problems, and learning difficulties in survivors.

T1-Based Synthetic Magnetic Resonance Contrasts Improve Multiple Sclerosis and Focal Epilepsy Imaging at 7 T

OBJECTIVES

Ultra-high field magnetic resonance imaging (MRI) (≥7 T) is a unique opportunity to improve the clinical diagnosis of brain pathologies, such as multiple sclerosis or focal epilepsy. However, several shortcomings of 7 T MRI, such as radiofrequency field inhomogeneities, could degrade image quality and hinder radiological interpretation. To address these challenges, an original synthetic MRI method based on T1 mapping achieved with the magnetization-prepared 2 rapid acquisition gradient echo (MP2RAGE) sequence was developed. The radiological quality of on-demand T1-based contrasts generated by this technique was evaluated in multiple sclerosis and focal epilepsy imaging at 7 T.

MATERIALS AND METHODS

This retrospective study was carried out from October 2017 to September 2019 and included 21 patients with different phenotypes of multiple sclerosis and 35 patients with focal epilepsy who underwent MRI brain examinations using a whole-body investigative 7 T magnetic resonance system. The quality of 2 proposed synthetic contrast images were assessed and compared with conventional images acquired at 7 T using the MP2RAGE sequence by 4 radiologists, evaluating 3 qualitative criteria: signal homogeneity, contrast intensity, and lesion visualization. Statistical analyses were performed on reported quality scores using Wilcoxon rank tests and further multiple comparisons tests. Intraobserver and interobserver reliabilities were calculated as well.

RESULTS

Radiological quality scores were reported higher for synthetic images when compared with original images, regardless of contrast, pathologies, or raters considered, with significant differences found for all 3 criteria (P < 0.0001, Wilcoxon rank test). None of the 4 radiologists ever rated a synthetic image "markedly worse" than an original image. Synthetic images were rated slightly less satisfying for only 3 epileptic patients, without precluding lesion identification.

CONCLUSION

T1-based synthetic MRI with the MP2RAGE sequence provided on-demand contrasts and high-quality images to the radiologist, facilitating lesion visualization in multiple sclerosis and focal epilepsy, while reducing the magnetic resonance examination total duration by removing an additional sequence.

Lesion load assessment among multiple sclerosis patient using DIR, FLAIR, and T2WI sequences

Background

Magnetic resonance imaging (MRI) is one of the diagnostic imaging modalities employing in lesion detection in neurological disorders such as multiple sclerosis (MS). Advances in MRI techniques such as double inversion recovery (DIR) made it more sensitive to distinguish lesions in the brain. To investigate the lesion load on different anatomical regions of the brain with MS using DIR, fluid attenuated inversion recovery (FLAIR) and T2-weighted imaging (T2WI) sequences. A total of 97 MS patients were included in our retrospective study, confirmed by neurologist. The patients were randomly selected from the major hospital in Saudi Arabia. All images were obtained using 3T Scanner (Siemens Skyra). The images from the DIR, FLAIR, and T2WI sequence were compared on axial planes with identical anatomic position and the number of lesions was assigned to their anatomical region.

Results

Comparing the lesion load measurement at various brain anatomical regions showed a significant difference among those three methods (p < 0.05).

Conclusion

DIR is a valuable MRI sequence for better delineation, greater contrast measurements and the increasing total number of MS lesions in MRI, compared with FLAIR, and T2WI and DIR revealed more intracortical lesions as well; therefore, in MS patients, it is recommended to add DIR sequence in daily routine imaging sequences.

AI in Radiology: Where are we today in Multiple Sclerosis Imaging?

Background MR imaging is an essential component in managing patients with Multiple sclerosis (MS). This holds true for the initial diagnosis as well as for assessing the clinical course of MS. In recent years, a growing number of computer tools were developed to analyze imaging data in MS. This review gives an overview of the most important applications with special emphasis on artificial intelligence (AI).

Methods Relevant studies were identified through a literature search in recognized databases, and through parsing the references in studies found this way. Literature published as of November 2019 was included with a special focus on recent studies from 2018 and 2019.

Results There are a number of studies which focus on optimizing lesion visualization and lesion segmentation. Some of these studies accomplished these tasks with high accuracy, enabling a reproducible quantitative analysis of lesion loads. Some studies took a radiomics approach and aimed at predicting clinical endpoints such as the conversion from a clinically isolated syndrome to definite MS. Moreover, recent studies investigated synthetic imaging, i. e. imaging data that is not measured during an MR scan but generated by a computer algorithm to optimize the contrast between MS lesions and brain parenchyma.

Conclusion Computer-based image analysis and AI are hot topics in imaging MS. Some applications are ready for use in clinical routine. A major challenge for the future is to improve prediction of expected disease courses and thereby helping to find optimal treatment decisions on an individual level. With technical improvements, more questions arise about the integration of new tools into the radiological workflow.

Key Points: 

Citation Format

Technological Advances of Magnetic Resonance Imaging in Today's Health Care Environment

Today's health care environment is shifting rapidly, driven by demographic change and high economic pressures on the system. Furthermore, modern precision medicine requires highly accurate and specific disease diagnostics in a short amount of time. Future imaging technology must adapt to these challenges.Demographic change necessitates scanner technologies tailored to the needs of an aging and increasingly multimorbid patient population. Accordingly, examination times have to be short enough that diagnostic images can be generated even for patients who can only lie in the scanner for a short time because of pain or with low breath-hold capacity.For economic reasons, the rate of nondiagnostic scans due to artifacts should be reduced as far as possible. As imaging plays an increasingly pivotal role in clinical-therapeutic decision making, magnetic resonance (MR) imaging facilities are confronted with an ever-growing number of patients, emphasizing the need for faster acquisitions while maintaining image quality.Lastly, modern precision medicine requires high and standardized image quality as well as quantifiable data in order to develop image-based biomarkers on which subsequent treatment management can rely.In recent decades, a variety of approaches have addressed the challenges of high throughput, demographic change, and precision medicine in MR imaging. These include field strength, gradient, coil and sequence development, as well as an increasing consideration of artificial intelligence. This article reviews state-of-the art MR technology and discusses future implementation from the perspective of what we know today.

Clinical feasibility of ultrafast intracranial vessel imaging with non-Cartesian spiral 3D time-of-flight MR angiography at 1.5T: An intra-individual comparison study

Objectives

Non-Cartesian Spiral readout can be implemented in 3D Time-of-flight (TOF) MR angiography (MRA) with short acquisition times. In this intra-individual comparison study we evaluated the clinical feasibility of Spiral TOF MRA in comparison with compressed sensing accelerated TOF MRA at 1.5T for intracranial vessel imaging as it has yet to be determined.

Materials and methods

Forty-four consecutive patients with suspected intracranial vascular disease were imaged with two Spiral 3D TOFs (Spiral, 0.82x0.82x1.2 mm³, 01:32 min; Spiral 0.8, 0.8x0.8x0.8 mm^3, 02:12 min) and a Compressed SENSE accelerated 3D TOF (CS 3.5, 0.82x0.82x1.2 mm³, 03:06 min) at 1.5T. Two neuroradiologists assessed qualitative (visualization of central and peripheral vessels) and quantitative image quality (Contrast Ratio, CR) and performed lesion and variation assessment for all three TOFs in each patient. After the rating process, the readers were questioned and representative cases were reinspected in a non-blinded fashion. For statistical analysis, the Friedman and Nemenyi post-hoc test, Kendall W tests, repeated measure ANOVA and weighted Cohen's Kappa tests were used.

Results

The Spiral and Spiral 0.8 outperformed the CS 3.5 in terms of peripheral image quality (p<0.001) and performed equally well in terms of central image quality (p>0.05). The readers noted slight differences in the appearance of maximum intensity projection images. A good to high degree of interstudy agreement between the three TOFs was observed for lesion and variation assessment (W = 0.638, p<0.001 –W = 1, p<0.001). CR values did not differ significantly between the three TOFs (p = 0.534). Interreader agreement ranged from good (K = 0.638) to excellent (K = 1).

Conclusions

Compared to the CS 3.5, both the Spiral and Spiral 0.8 exhibited comparable or better image quality and comparable diagnostic performance at much shorter acquisition times.

Compressed SENSE accelerated 3D T1w black blood turbo spin echo versus 2D T1w turbo spin echo sequence in pituitary magnetic resonance imaging

PURPOSE

To compare image quality between a 2D T1w turbo spin echo (TSE) sequence and a Compressed SENSE accelerated 3D T1w black blood TSE sequence (equipped with a black blood prepulse for blood signal suppression) in pre- and postcontrast imaging of the pituitary and to assess scan time reductions.

METHODS AND MATERIALS

For this retrospective study, 56 patients underwent pituitary MR imaging at 3T. 28 patients were scanned with the 2D- and 28 patients with the accelerated 3D sequence. Two board certified neuroradiologists independently evaluated 13 qualitative image features (12 features on postcontrast- and 1 feature on precontrast images).SNR and CNR measurements were obtained. Interreader agreement was assessed with the intraclass correlation coefficient while differences in scores were assessed with exact Wilcoxon rank sum tests.

RESULTS

The interreader agreement ranged from fair (visibility of the ophthalmic nerve, ICC = 0.57) to excellent (presence and severity of pulsation artefacts, ICC = 0.97). The Compressed SENSE accelerated 3D sequence outperformed the 2D sequence in terms of "overall image quality" (median: 4 versus 3, p = 0.04) and "presence and severity of pulsation artefacts" (median: 0 versus 1, p < 0.001). There were no significant differences in any other qualitative and quantitative (SNR, CNR) image quality features. Scan time was reduced by 03:53 min (33.1%) by replacing the 2D with the 3D sequence.

CONCLUSION

The Compressed SENSE accelerated 3D T1w black blood TSE sequence is a reliable alternative for the standard 2D sequence in pituitary imaging. The black blood prepulse may aid in suppression of pulsation artefacts.

Compressed sensing MRI of different organs: ready for clinical daily practice?

OBJECTIVES

The aim was to evaluate the image quality and sensitivity to artifacts of compressed sensing (CS) acceleration technique, applied to 3D or breath-hold sequences in different clinical applications from brain to knee.

METHODS

CS with an acceleration from 30 to 60% and conventional MRI sequences were performed in 10 different applications in 107 patients, leading to 120 comparisons. Readers were blinded to the technique for quantitative (contrast-to-noise ratio or functional measurements for cardiac cine) and qualitative (image quality, artifacts, diagnostic findings, and preference) image analyses.

RESULTS

No statistically significant difference in image quality or artifacts was found for each sequence except for the cardiac cine CS for one of both readers and for the wrist 3D proton density (PD)-weighted CS sequence which showed less motion artifacts due to the reduced acquisition time. The contrast-to-noise ratio was lower for the elbow CS sequence but not statistically different in all other applications. Diagnostic findings were similar between conventional and CS sequence for all the comparisons except for four cases where motion artifacts corrupted either the conventional or the CS sequence.

CONCLUSIONS

The evaluated CS sequences are ready to be used in clinical daily practice except for the elbow application which requires a lower acceleration. The CS factor should be tuned for each organ and sequence to obtain good image quality. It leads to 30% to 60% acceleration in the applications evaluated in this study which has a significant impact on clinical workflow.

KEY POINTS

• Clinical implementation of compressed sensing (CS) reduced scan times of at least 30% with only minor penalty in image quality and no change in diagnostic findings. • The CS acceleration factor has to be tuned separately for each organ and sequence to guarantee similar image quality than conventional acquisition. • At least 30% and up to 60% acceleration is feasible in specific sequences in clinical routine.

Magnetic Resonance Imaging of the Brain Using Compressed Sensing - Quality Assessment in Daily Clinical Routine

PURPOSE

To assess the effect of compressed sensing (CS) on image quality and acquisition speed in routine brain magnetic resonance imaging (MRI).

METHODS

During a 2-month implementation period of CS, two senior neuroradiologists, one MRI physicist and one application specialist optimized the CS acceleration factor to reduce scan time and improve spatial resolution, while maintaining image quality. Afterwards, two neuroradiologists independently scored image quality on a 5-point Likert scale in 3‑dimensional (3D) fluid attenuation inversion recovery (FLAIR), 3D double inversion recovery (DIR), 3D T2, 3D T1, 3D T1 + gadoteric acid, axial T2, axial FLAIR, axial T2*, and 3D arterial time-of-flight MR angiography (art. TOF) sequences acquired during 1 week before (CS-) and after (CS+) the implementation of CS. Time of acquisition was recorded for all sequences.

RESULTS

A total of 51 CS- and 48 CS+ patients were included. The median scan time reduction was 29.3% (range 0.0-58.4%), median voxel size reduction was 10.5% (0.0-33.3%). The CS+ image quality was rated superior for 3D FLAIR (p < 0.001), 3D T2 (p = 0.001), and axial T2* sequences (p = 0.024). For all other sequences, no statistical difference in image quality was observed. Interreader agreement regarding image quality was good for all sequences (weighted Cohen's κ > 0.5).

CONCLUSION

The use of CS saves considerable imaging time while allowing to increase spatial resolution in routine clinical brain MRI without loss in image quality.