Detection Of Multiple Sclerosis Lesions Research Articles

Detection of multiple sclerosis lesions by susceptibility-weighted imaging—A systematic review and meta-analyses

Susceptibility-weighted imaging (SWI) is a valuable method in the diagnosis of multiple sclerosis (MS). SWI can show diagnostic biomarkers and also distinguish MS lesions from similar diseases. This study is a systematic review and meta-analysis of the available literature to investigate the diagnostic accuracy of SWI for the detection of MS lesions in the brain, while, specific aspects of its performance were examined. Using specified keywords, a search of the electronic databases MEDLINE/PubMed, Scopus, Embase/Elsevier-Ovid, and Scholar databases was conducted. The search time filter ranged from the year 2010 to 2023 . Sensitivity and specificity were used as effect sizes, and SEas a precision metric (with 95% confidence interval). The study was performed according to PRISMA 2020 guidelines and the prospective register of systematic reviews (PROSPERO) registration. The number of 19 studies with a total of 2026 patients to examine both automated and nonautomated assessments. The pooled sensitivities for MS lesion (MSL), paramagnetic rim lesions (PRLs), central vein sign (CVS), and PRL with CVS were 62% (53-72), 37% (17-57), and 60% (31-89). Moreover, the pooled specificities were 93% (90-97), 88% (82-94), 100% (100-100), and 78% (50-100), respectively. Adding the SWI method to routine sequences for MS investigation can provide useful diagnostic information.

Ultrafast Brain MRI at 3 T for MS: Evaluation of a 51-Second Deep Learning-Enhanced T2-EPI-FLAIR Sequence.

In neuroimaging, there is no equivalent alternative to magnetic resonance imaging (MRI). However, image acquisitions are generally time-consuming, which may limit utilization in some cases, e.g., in patients who cannot remain motionless for long or suffer from claustrophobia, or in the event of extensive waiting times. For multiple sclerosis (MS) patients, MRI plays a major role in drug therapy decision-making. The purpose of this study was to evaluate whether an ultrafast, T2-weighted (T2w), deep learning-enhanced (DL), echo-planar-imaging-based (EPI) fluid-attenuated inversion recovery (FLAIR) sequence (FLAIRUF) that has targeted neurological emergencies so far might even be an option to detect MS lesions of the brain compared to conventional FLAIR sequences. Therefore, 17 MS patients were enrolled prospectively in this exploratory study. Standard MRI protocols and ultrafast acquisitions were conducted at 3 tesla (T), including three-dimensional (3D)-FLAIR, turbo/fast spin-echo (TSE)-FLAIR, and FLAIRUF. Inflammatory lesions were grouped by size and location. Lesion conspicuity and image quality were rated on an ordinal five-point Likert scale, and lesion detection rates were calculated. Statistical analyses were performed to compare results. Altogether, 568 different lesions were found. Data indicated no significant differences in lesion detection (sensitivity and positive predictive value [PPV]) between FLAIRUF and axially reconstructed 3D-FLAIR (lesion size ≥3 mm × ≥2 mm) and no differences in sensitivity between FLAIRUF and TSE-FLAIR (lesion size ≥3 mm total). Lesion conspicuity in FLAIRUF was similar in all brain regions except for superior conspicuity in the occipital lobe and inferior conspicuity in the central brain regions. Further findings include location-dependent limitations of signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) as well as artifacts such as spatial distortions in FLAIRUF. In conclusion, FLAIRUF could potentially be an expedient alternative to conventional methods for brain imaging in MS patients since the acquisition can be performed in a fraction of time while maintaining good image quality.

Diagnosis of Multiple Sclerosis Lesion using Deep Learning Models

Exploring the identification of Multiple Sclerosis (MS) lesions involves leveraging diverse machine learning classifiers. Using Magnetic Resonance Imaging (MRI) scans, the study aims to detect and characterize these lesions, evaluating their attributes, progression stages, and the pivotal role of Artificial Intelligence (AI) in diagnosis. The focus is on analyzing automated detection algorithms, particularly Deep Learning techniques. Through comprehensive assessment, various classifiers including MLP Classifier, Random Forest (RF), Support Vector Machine (SVM), and DL are evaluated using metrics like precision, recall, F1-score, and accuracy. The DL classifier stands out, achieving remarkable precision (99%), recall (99%), and overall accuracy (99.5%). Comparative analysis confirms its superiority, reinforcing its efficacy over alternative methods. The research underscores the DL model's potential in generalizing to new samples due to its robustness and precision. This study significantly advances automated MS lesion detection, highlighting the promise of AI-based methodologies in medical image analysis.

Resting-state functional magnetic resonance imaging reveals functional connectivity alteration in the experimental autoimmune encephalomyelitis model of multiple sclerosis.

Multiple sclerosis (MS) is an autoimmune degenerative disease targeting white matter in the central nervous system. The most common animal model that mimics MS is experimental autoimmune encephalomyelitis (EAE) and it plays a crucial role in pharmacological research, from the identification of a therapeutic target to the in vivo validation of efficacy. Magnetic resonance imaging (MRI) is largely used to detect MS lesions, and resting-state functional MRI (rsfMRI) to investigate alterations in the brain functional connectivity (FC). MRI was mainly used in EAE studies to detect lesions in the spinal cord and brain. The current longitudinal MRI study aims to validate rsfMRI as a biomarker of the disease progression in the myelin oligodendrocyte glycoprotein 35-55 induced EAE animal model of MS. MR images were acquired 14, 25, and 50 days postimmunization. Seed-based analysis was used to investigate the whole-brain FC with some predefined areas, such as the thalamic regions, cerebellum, motor and somatosensory cortex. When compared with the control group, the EAE group exhibited a slightly altered FC and a decreasing trend in the total number of activated voxels along the disease progression. The most interesting result regards the whole-brain FC with the cerebellum. A hyperconnectivity behavior was found at an early phase and a significant reduced connectivity at a late phase. Moreover, we found a negative correlation between the total number of activated voxels during the late phase and the cumulative disease index. The results obtained provide a clinically relevant experimental platform that may be pivotal for the elucidation of the key mechanisms of accumulation of irreversible disability, as well as the development of innovative therapies for MS. Moreover, the negative correlation between the disease severity and the size of the activated area suggests a possible research pathway to follow for the resolution of the clinico-radiological paradox.

Detection of Spinal Cord Multiple Sclerosis Lesions Using a 3D-PSIR Sequence at 1.5 T.

Multiple sclerosis (MS) is aprevalent autoimmune inflammatory disease. Besides cerebral manifestations, an affection of the spinal cord is typical; however, imaging of the spinal cord is difficult due to its anatomy. The aim of this study was to assess the diagnostic value of a3D PSIR pulse sequencing at a1.5 T magnetic field strength for both the cervical and thoracic spinal cord. Phase sensitive inversion recovery (PSIR), short tau inversion recovery (STIR) and T2-weighted (T2-w)images of the spinal cord of 50patients were separately evaluated by three radiologists concerning the number and location of MS lesions. Furthermore, lesion to cord contrast ratios were determined for the cervical and thoracic spinal cord. Of the lesions 54.81% were located in the cervical spinal cord, 42.26% in the thoracic spinal cord and 2.93% in the conus medullaris. The PSIR images showed ahigher sensitivity for lesion detection in the cervical and thoracic spinal cord (77.10% and 72.61%, respectively) compared to the STIR images (58.63% and 59.10%, respectively) and the T2-w images (59.95% and 59.52%, respectively). The average lesion to cord contrast ratio was significantly higher in the PSIR images compared to the STIR images (p < 0.001) and the T2-w images (p < 0.001). Evaluation of the spinal cord with a3D PSIR sequence at amagnetic field strength of 1.5 T is feasible with ahigh sensitivity for the detection of spinal MS lesions for the cervical as well as the thoracic segments. In combination with other pulse sequences it might become avaluable addition in an advanced imaging protocol.

Segmenting Multi-Sclerosis Lesions: An Enhanced Approach using an Optimized U-Net

Based on MRI images and a U-Net architecture, this study presents an optimized Convolutional Neural Network for segmenting Multiple Sclerosis (MS) lesions. The approach entails a two-step procedure designed to maximize accuracy in lesion detection. First, it identifies the most informative modality from the MRI images at hand. Then, it systematically assesses the critical hyper-parameters of the U-Net architecture to improve the learning process. By selecting the optimal modality and fine-tuning hyper-parameters, significant improvements in MS lesion detection and monitoring accuracy are achieved. This advancement facilitates a deeper understanding of the disease’s progression and its sub-types. The proposed approach holds promise for improving clinical decision-making and patient management in MS diagnosis and treatment.

Thin-slice Two-dimensional T2-weighted Imaging with Deep Learning-based Reconstruction: Improved Lesion Detection in the Brain of Patients with Multiple Sclerosis.

Brain MRI with high spatial resolution allows for a more detailed delineation of multiple sclerosis (MS) lesions. The recently developed deep learning-based reconstruction (DLR) technique enables image denoising with sharp edges and reduced artifacts, which improves the image quality of thin-slice 2D MRI. We, therefore, assessed the diagnostic value of 1 mm-slice-thickness 2D T2-weighted imaging (T2WI) with DLR (1 mm T2WI with DLR) compared with conventional MRI for identifying MS lesions. Conventional MRI (5 mm T2WI, 2D and 3D fluid-attenuated inversion recovery) and 1 mm T2WI with DLR (imaging time: 7 minutes) were performed in 42 MS patients. For lesion detection, two neuroradiologists counted the MS lesions in two reading sessions (conventional MRI interpretation with 5 mm T2WI and MRI interpretations with 1 mm T2WI with DLR). The numbers of lesions per region category (cerebral hemisphere, basal ganglia, brain stem, cerebellar hemisphere) were then compared between the two reading sessions. For the detection of MS lesions by 2 neuroradiologists, the total number of detected MS lesions was significantly higher for MRI interpretation with 1 mm T2WI with DLR than for conventional MRI interpretation with 5 mm T2WI (765 lesions vs. 870 lesions at radiologist A, < 0.05). In particular, of the 33 lesions in the brain stem, radiologist A detected 21 (63.6%) additional lesions by 1 mm T2WI with DLR. Using the DLR technique, whole-brain 1 mm T2WI can be performed in about 7 minutes, which is feasible for routine clinical practice. MRI with 1 mm T2WI with DLR enabled increased MS lesion detection, particularly in the brain stem.

The added value of double inversion recovery (DIR) sequence compared to fluid attenuation inversion recovery (FLAIR) in identifying multiple sclerosis lesions and different cortical subtypes

Background. Diabetic here is a rising concern about multiple sclerosis (MS) lesions that occur in the cortical gray matter (GM), owing to its direct relationship with the disability and cognition dysfunction that occur during the MS course. Two magnetic resonance imaging (MRI) sequences that aid in the precise identification of these lesions are the Double Inversion Recovery (DIR) and Fluid Attenuation Inversion Recovery (FLAIR) sequences. Objective. To study the importance of the DIR for brain lesion detection and the various cortical subtypes distinction in MS-diagnosed patients and its correlations to the level of patient disability by comparing it to FLAIR. Patients and Methods: Thirty MS patients participated in this retrospective cross-sectional study, who had been previously diagnosed with MS and were routinely followed up at the Radiology Department at Ain-Shams University. Results. Comparing the DIR sequence to FLAIR, the overall MS lesion detection enhanced by 8% (p&lt;0.001). Regarding the detection of cortical lesion subtypes (I–IV), it was discovered that the DIR sequence revealed significantly more lesions than FLAIR (p&lt;0.001). The cortical lesions detected by DIR or FLAIR were found to have positive, highly significant correlations with both patients’ disability and the MS duration. Conclusion. For lesion identification, DIR is more effective than FLAIR, particularly for cortical and juxtacortical lesions in MS patients.

Detection of Multiple Sclerosis Lesions in Supra- and Infra-Tentorial Anatomical Regions by Double Inversion Recovery, Flair, and T2 MRI Sequences: A Comparative Study in Iraqi Patients

Background: In young adults, multiple sclerosis is a prevalent chronic inflammatory demyelinating condition. It is characterized by white matter affection, but many individuals also have significant gray matter involvement. A double-inversion recovery pulse (DIR) pattern was recently proposed to improve the visibility of multiple sclerosis lesions. Objective: To find out how well a DIR sequence, FLAIR, and T2-weighted pulse sequences can find MS lesions in the supratentorial and infratentorial regions. Methods: A total of 37 patients with established diagnoses of multiple sclerosis were included in this cross-sectional study. Brain MRI was done using double inversion recovery, T2, and FLAIR sequences. The number of lesions was counted and compared in the three sequences. Results: The DIR sequence detected more infratentorial lesions when compared to the T2 and FLAIR sequences. In the supratentorial region, DIR detected more lesions than T2 and FLAIR. Conclusion: The DIR sequence is highly superior to both the T2 and FLAIR sequences in depicting the lesions, regardless of their anatomical distribution. Moreover, the DIR sequence detected more multiple sclerosis lesions in the infratentorial region than the traditional T2W and FLAIR sequences.

Investigation of an efficient multi-modal convolutional neural network for multiple sclerosis lesion detection

In this study, an automated 2D machine learning approach for fast and precise segmentation of MS lesions from multi-modal magnetic resonance images (mmMRI) is presented. The method is based on an U-Net like convolutional neural network (CNN) for automated 2D slice-based-segmentation of brain MRI volumes. The individual modalities are encoded in separate downsampling branches without weight sharing, to leverage the specific features. Skip connections input feature maps to multi-scale feature fusion (MSFF) blocks at every decoder stage of the network. Those are followed by multi-scale feature upsampling (MSFU) blocks which use the information about lesion shape and location. The CNN is evaluated on two publicly available datasets: The ISBI 2015 longitudinal MS lesion segmentation challenge dataset containing 19 subjects and the MICCAI 2016 MSSEG challenge dataset containing 15 subjects from various scanners. The proposed multi-input 2D architecture is among the top performing approaches in the ISBI challenge, to which open-access papers are available, is able to outperform state-of-the-art 3D approaches without additional post-processing, can be adapted to other scanners quickly, is robust against scanner variability and can be deployed for inference even on a standard laptop without a dedicated GPU.

3D U-Net for automated detection of multiple sclerosis lesions: utility of transfer learning from other pathologies.

Deep learning algorithms for segmentation of multiple sclerosis (MS) plaques generally require training on large datasets. This manuscript evaluates the effect of transfer learning from segmentation of another pathology to facilitate use of smaller MS-specific training datasets. That is, a model trained for detection of one type of pathology was re-trained to identify MS lesions and active demyelination. In this retrospective study using MRI exams from 149 patients spanning 4/18/2014 to 7/8/2021, 3D convolutional neural networks were trained with a variable number of manually-segmented MS studies. Models were trained for FLAIR lesion segmentation at a single timepoint, new FLAIR lesion segmentation comparing two timepoints, and enhancing (actively demyelinating) lesion segmentation on T1 post-contrast imaging. Models were trained either de-novo or fine-tuned with transfer learning applied to a pre-existing model initially trained on non-MS data. Performance was evaluated with lesionwise sensitivity and positive predictive value (PPV). For single timepoint FLAIR lesion segmentation with 10 training studies, a fine-tuned model demonstrated improved performance [lesionwise sensitivity 0.55 ± 0.02 (mean ± standard error), PPV 0.66 ± 0.02] compared to a de-novo model (sensitivity 0.49 ± 0.02, p = 0.001; PPV 0.32 ± 0.02, p < 0.001). For new lesion segmentation with 30 training studies and their prior comparisons, a fine-tuned model demonstrated similar sensitivity (0.49 ± 0.05) and significantly improved PPV (0.60 ± 0.05) compared to a de-novo model (sensitivity 0.51 ± 0.04, p = 0.437; PPV 0.43 ± 0.04, p = 0.002). For enhancement segmentation with 20 training studies, a fine-tuned model demonstrated significantly improved overall performance (sensitivity 0.74 ± 0.06, PPV 0.69 ± 0.05) compared to a de-novo model (sensitivity 0.44 ± 0.09, p = 0.001; PPV 0.37 ± 0.05, p = 0.001). By fine-tuning models trained for other disease pathologies with MS-specific data, competitive models identifying existing MS plaques, new MS plaques, and active demyelination can be built with substantially smaller datasets than would otherwise be required to train new models.

Longitudinal analysis of new multiple sclerosis lesions with magnetization transfer and diffusion tensor imaging

ObjectiveThe potential of magnetization transfer imaging (MTI) and diffusion tensor imaging (DTI) for the detection and evolution of new multiple sclerosis (MS) lesions was analyzed.MethodsNineteen patients with MS obtained conventional MRI, MTI, and DTI examinations bimonthly for 12 months and again after 24 months at 1.5 T MRI. MTI was acquired with balanced steady-state free precession (bSSFP) in 10 min (1.3 mm3 isotropic resolution) yielding both magnetization transfer ratio (MTR) and quantitative magnetization transfer (qMT) parameters (pool size ratio (F), exchange rate (kf), and relaxation times (T1/T2)). DTI provided fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity (RD).ResultsAt the time of their appearance on MRI, the 21 newly detected MS lesions showed significantly reduced MTR/F/kf and prolonged T1/T2 parameters, as well as significantly reduced FA and increased AD/MD/RD. Significant differences were already observed for MTR 4 months and for qMT parameters 2 months prior to lesions’ detection on MRI. DTI did not show any significant pre-lesional differences. Slightly reversed trends were observed for most lesions up to 8 months after their detection for qMT and less pronounced for MTR and three diffusion parameters, while appearing unchanged on MRI.ConclusionsMTI provides more information than DTI in MS lesions and detects tissue changes 2 to 4 months prior to their appearance on MRI. After lesions’ detection, qMT parameter changes promise to be more sensitive than MTR for the lesions’ evolutional assessment. Overall, bSSFP-based MTI adumbrates to be more sensitive than MRI and DTI for the early detection and follow-up assessment of MS lesions.Clinical relevance statementWhen additionally acquired in routine MRI, fast bSSFP-based MTI can complement the MRI/DTI longitudinal lesion assessment by detecting MS lesions 2–4 months earlier than with MRI, which could implicate earlier clinical decisions and better follow-up/treatment assessment in MS patients.Key Points• Magnetization transfer imaging provides more information than DTI in multiple sclerosis lesions and can detect tissue changes 2 to 4 months prior to their appearance on MRI.• After lesions’ detection, quantitative magnetization transfer changes are more pronounced than magnetization transfer ratio changes and therefore promise to be more sensitive for the lesions’ evolutional assessment.• Balanced steady-state free precession–based magnetization transfer imaging is more sensitive than MRI and DTI for the early detection and follow-up assessment of multiple sclerosis lesions.

Use of gadolinium-based contrast agents in multiple sclerosis: a review by the ESMRMB-GREC and ESNR Multiple Sclerosis Working Group.

Magnetic resonance imaging (MRI) is the most sensitive technique for detecting inflammatory demyelinating lesions in multiple sclerosis (MS) and plays a crucial role in diagnosis and monitoring treatment effectiveness, and for predicting the disease course. In clinical practice, detection of MS lesions is mainly based on T2-weighted and contrast-enhanced T1-weighted sequences. Contrast-enhancing lesions (CEL) on T1-weighted sequences are related to (sub)acute inflammation, while new or enlarging T2 lesions reflect the permanent footprint from a previous acute inflammatory demyelinating event. These two types of MRI features provide redundant information, at least in regular monitoring of the disease. Due to the concern of gadolinium deposition after repetitive injections of gadolinium-based contrast agents (GBCAs), scientific organizations and regulatory agencies in Europe and North America have proposed that these contrast agents should be administered only if clinically necessary. In this article, we provide data on the mode of action of GBCAs in MS, the indications of the use of these agents in clinical practice, their value in MS for diagnostic, prognostic, and monitoring purposes, and their use in specific populations (children, pregnant women, and breast-feeders). We discuss imaging strategies that achieve the highest sensitivity for detecting CELs in compliance with the safety regulations established by different regulatory agencies. Finally, we will briefly discuss some alternatives to the use of GBCA for detecting blood-brain barrier disruption in MS lesions. CLINICAL RELEVANCE STATEMENT: Although use of GBCA at diagnostic workup of suspected MS is highly valuable for diagnostic and prognostic purposes, their use in routine monitoring is not mandatory and must be reduced, as detection of disease activity can be based on the identification of new or enlarging lesions on T2-weighted images. KEY POINTS: • Both the EMA and the FDA state that the use of GBCA in medicine should be restricted to clinical scenarios in which the additional information offered by the contrast agent is required. • The use of GBCA is generally recommended in the diagnostic workup in subjects with suspected MS and is generally not necessary for routine monitoring in clinical practice. • Alternative MRI-based approaches for detecting acute focal inflammatory MS lesions are not yet ready to be used in clinical practice.

Imaging immunomodulatory treatment responses in a multiple sclerosis mouse model using hyperpolarized 13C metabolic MRI

BackgroundIn recent years, the ability of conventional magnetic resonance imaging (MRI), including T1 contrast-enhanced (CE) MRI, to monitor high-efficacy therapies and predict long-term disability in multiple sclerosis (MS) has been challenged. Therefore, non-invasive methods to improve MS lesions detection and monitor therapy response are needed.MethodsWe studied the combined cuprizone and experimental autoimmune encephalomyelitis (CPZ-EAE) mouse model of MS, which presents inflammatory-mediated demyelinated lesions in the central nervous system as commonly seen in MS patients. Using hyperpolarized 13C MR spectroscopy (MRS) metabolic imaging, we measured cerebral metabolic fluxes in control, CPZ-EAE and CPZ-EAE mice treated with two clinically-relevant therapies, namely fingolimod and dimethyl fumarate. We also acquired conventional T1 CE MRI to detect active lesions, and performed ex vivo measurements of enzyme activities and immunofluorescence analyses of brain tissue. Last, we evaluated associations between imaging and ex vivo parameters.ResultsWe show that hyperpolarized [1-13C]pyruvate conversion to lactate is increased in the brain of untreated CPZ-EAE mice when compared to the control, reflecting immune cell activation. We further demonstrate that this metabolic conversion is significantly decreased in response to the two treatments. This reduction can be explained by increased pyruvate dehydrogenase activity and a decrease in immune cells. Importantly, we show that hyperpolarized 13C MRS detects dimethyl fumarate therapy, whereas conventional T1 CE MRI cannot.ConclusionsIn conclusion, hyperpolarized MRS metabolic imaging of [1-13C]pyruvate detects immunological responses to disease-modifying therapies in MS. This technique is complementary to conventional MRI and provides unique information on neuroinflammation and its modulation.

Improved detection of juxtacortical lesions using highly accelerated double inversion-recovery MRI in patients with multiple sclerosis

PurposeThe purpose of this study was to compare a highly-accelerated double inversion recovery (fast-DIR) sequence using a recent parallel imaging technique (CAIPIRINHA) with a conventional DIR (conv-DIR) sequence for image quality and the detection of juxtacortical and infratentorial multiple sclerosis (MS) lesions. Materials and methodsA total of 38 patients with MS who underwent brain MRI at 3 T between 2020 and 2021 were included. There were 27 women and 12 men with a mean age of 40 ± 12.8 (standard deviation) years (range: 20–59 years). All patients underwent conv-DIR sequence and fast-DIR sequence. Fast-DIR was obtained with a T2-preparation module to improve contrast and an iterative denoising algorithm to compensate noise enhancement. Two blinded readers reported the number of juxtacortical and infratentorial MS lesions for fast-DIR and conv-DIR, confirmed by further consensus reading that was used as the standard of reference. Image quality and contrast were evaluated for fast-DIR and conv-DIR sequences. Comparisons between fast-DIR and conv-DIR sequences were performed using Wilcoxon test and Lin concordance correlation coefficient. ResultsThirty-eight patients were analyzed. Fast-DIR imaging allowed detection of 289 juxtacortical lesions vs. 238 with conv-DIR, corresponding to a significant improved detection rate with fast-DIR (P < 0.001). Conversely, 117 infratentorial lesions were detected with conv-DIR sequence vs. 80 with fast-DIR sequence (P < 0.001). Inter-observer agreement for lesion detection with fast-DIR and conv-DIR was very high (Lin concordance correlation coefficient ranging between 0.86 and 0.96). ConclusionFast-DIR improves the detection of juxtacortical MS lesions, but is limited for the detection of infratentorial MS lesions.

Multifaceted Evaluation of Double Inversion Recovery Brain MRI in Detection of Cervical Spine Multiple Sclerosis Lesions (P11-3.002)

<h3>Objective:</h3> To quantitatively study the efficacy of Brain Double Inversion Recovery (DIR) MRI for the detection of cervical spinal cord lesions in Multiple Sclerosis (MS) patients using blinded readings as well as contrast to noise ratios (C/N). <h3>Background:</h3> DIR is a well-established MRI pulse sequence for detection of cerebral cortical and juxtacortical MS lesions. Since the superior half of the cervical cord is visible on sagittally acquired volumetric DIR of the brain, we investigated the efficacy of Brain MRI for detecting upper spinal cord of MS patients. <h3>Design/Methods:</h3> We retrospectively examined the MRI of the brain of 64 patients with established MS, who had also undergone cervical spine MRI. Two blinded MS expert readers, who assessed the scans for lesion numbers and rated lesion visibility and overall image quality, reviewed brain 3D DIR sagittal and coronal images. Official cervical spine study reports were used as the “gold standard” in this analysis. Additionally, standardized Mean contrast-to-noise ratios (C/N) and standard deviation (S.D.) were calculated in representative lesions for each patient and compared to those of 3D brain FLAIR images. <h3>Results:</h3> The sensitivity of brain DIR was found to be 87% and specificity was found to 61% when all “definite” lesions were included. When considering both “probable” and “definite” lesions, the sensitivity increased to 91% and specificity decreased to 54%. Contrast to noise ratio of brain DIR lesions was significantly higher compared to brain FLAIR lesions (7.56 +/−1.77 vs. 0.67+/−1.27). <h3>Conclusions:</h3> Sagittally acquired volumetric Brain DIR can provide useful information on cervical spinal cord lesions, with high C/N. In some cases, this should facilitate the attainment of the MRI components of McDonald’s diagnostic criteria without a dedicated cervical spine study. <b>Disclosure:</b> Mr. Patel has nothing to disclose. Marina Creed has nothing to disclose. Dr. Gershon has nothing to disclose. The institution of Dr. Wolansky has received research support from Guerbet. Dr. Imitola has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Novartis . The institution of Dr. Imitola has received research support from biogen. Dr. Imitola has a non-compensated relationship as a Board Member with National MS Society that is relevant to AAN interests or activities. Dr. Imitola has a non-compensated relationship as a Committee Member with International Society for Stem Cell Research that is relevant to AAN interests or activities.

Automatic Intelligent System Using Medical of Things for Multiple Sclerosis Detection.

Malfunctions in the immune system cause multiple sclerosis (MS), which initiates mild to severe nerve damage. MS will disturb the signal communication between the brain and other body parts, and early diagnosis will help reduce the harshness of MS in humankind. Magnetic resonance imaging (MRI) supported MS detection is a standard clinical procedure in which the bio-image recorded with a chosen modality is considered to assess the severity of the disease. The proposed research aims to implement a convolutional neural network (CNN) supported scheme to detect MS lesions in the chosen brain MRI slices. The stages of this framework include (i) image collection and resizing, (ii) deep feature mining, (iii) hand-crafted feature mining, (iii) feature optimization with firefly algorithm, and (iv) serial feature integration and classification. In this work, five-fold cross-validation is executed, and the final result is considered for the assessment. The brain MRI slices with/without the skull section are examined separately, presenting the attained results. The experimental outcome of this study confirms that the VGG16 with random forest (RF) classifier offered a classification accuracy of >98% MRI with skull, and VGG16 with K-nearest neighbor (KNN) provided an accuracy of >98% without the skull.

Single-timepoint low-dimensional characterization and classification of acute versus chronic multiple sclerosis lesions using machine learning

Multiple sclerosis (MS) is a chronic inflammatory and neurodegenerative disease characterized by the appearance of focal lesions across the central nervous system. The discrimination of acute from chronic MS lesions may yield novel biomarkers of inflammatory disease activity which may support patient management in the clinical setting and provide endpoints in clinical trials. On a single timepoint and in the absence of a prior reference scan, existing methods for acute lesion detection rely on the segmentation of hyperintense foci on post-gadolinium T1-weighted magnetic resonance imaging (MRI), which may underestimate recent acute lesion activity. In this paper, we aim to improve the sensitivity of acute MS lesion detection in the single-timepoint setting, by developing a novel machine learning approach for the automatic detection of acute MS lesions, using single-timepoint conventional non-contrast T1- and T2-weighted brain MRI. The MRI input data are supplemented via the use of a convolutional neural network generating “lesion-free” reconstructions from original “lesion-present” scans using image inpainting. A multi-objective statistical ranking module evaluates the relevance of textural radiomic features from the core and periphery of lesion sites, compared within “lesion-free” versus “lesion-present” image pairs. Then, an ensemble classifier is optimized through a recursive loop seeking consensus both in the feature space (via a greedy feature-pruning approach) and in the classifier space (via model selection repeated after each pruning operation). This leads to the identification of a compact textural signature characterizing lesion phenotype. On the patch-level task of acute versus chronic MS lesion classification, our method achieves a balanced accuracy in the range of 74.3–74.6% on fully external validation cohorts.

Accurate Diagnosis of Cortical and Infratentorial Lesions in Multiple Sclerosis Using Accelerated Fluid and White Matter Suppression Imaging.

The precise location of multiple sclerosis (MS) cortical lesions can be very challenging at 3 T, yet distinguishing them from subcortical lesions is essential for the diagnosis and prognosis of the disease. Compressed sensing-accelerated fluid and white matter suppression imaging (CS-FLAWS) is a new magnetic resonance imaging sequence derived from magnetization-prepared 2 rapid acquisition gradient echo with promising features for the detection and classification of MS lesions. The objective of this study was to compare the diagnostic performances of CS-FLAWS (evaluated imaging) and phase sensitive inversion recovery (PSIR; reference imaging) for classification of cortical lesions (primary objective) and infratentorial lesions (secondary objective) in MS, in combination with 3-dimensional (3D) double inversion recovery (DIR). Prospective 3 T scans (MS first diagnosis or follow-up) acquired between March and August 2021 were retrospectively analyzed. All underwent 3D CS-FLAWS, axial 2D PSIR, and 3D DIR. Double-blinded reading sessions exclusively in axial plane and final consensual reading were performed to assess the number of cortical and infratentorial lesions. Wilcoxon test was used to compare the 2 imaging datasets (FLAWS + DIR and PSIR + DIR), and intraobserver and interobserver agreement was assessed using the intraclass correlation coefficient. Forty-two patients were analyzed (38 with relapsing-remitting MS, 29 women, 42.7 ± 12.6 years old). Compressed sensing-accelerated FLAWS allowed the identification of 263 cortical lesions versus 251 with PSIR ( P = 0.74) and 123 infratentorial lesions versus 109 with PSIR ( P = 0.63), corresponding to a nonsignificant difference between the 2 sequences. Compressed sensing-accelerated FLAWS exhibited fewer false-negative findings than PSIR either for cortical lesions (1 vs 13; P < 0.01) or infratentorial lesions (1 vs 15; P < 0.01). No false-positive findings were found with any of the 2 sequences. Diagnostic confidence was high for each contrast. Three-dimensional CS-FLAWS is as accurate as 2D PSIR imaging for classification of cortical and infratentorial MS lesions, with fewer false-negative findings, opening the way to a reliable full brain MS exploration in a clinically acceptable duration (5 minutes 15 seconds).

New multiple sclerosis lesion segmentation and detection using pre-activation U-Net.

Automated segmentation of new multiple sclerosis (MS) lesions in 3D MRI data is an essential prerequisite for monitoring and quantifying MS progression. Manual delineation of such lesions is time-consuming and expensive, especially because raters need to deal with 3D images and several modalities. In this paper, we propose Pre-U-Net, a 3D encoder-decoder architecture with pre-activation residual blocks, for the segmentation and detection of new MS lesions. Due to the limited training set and the class imbalance problem, we apply intensive data augmentation and use deep supervision to train our models effectively. Following the same U-shaped architecture but different blocks, Pre-U-Net outperforms U-Net and Res-U-Net on the MSSEG-2 dataset, achieving a Dice score of 40.3% on new lesion segmentation and an F1 score of 48.1% on new lesion detection. The codes and trained models are publicly available at https://github.com/pashtari/xunet.