National Institute of Neurological Disorders and Stroke
Neuroimmunology Clinic

Author Of 1 Presentation

Imaging Oral Presentation

PS11.05 - Inclusion of small ovoid lesions in central vein sign assessment improves sensitivity for multiple sclerosis

Speakers
Presentation Number
PS11.05
Presentation Topic
Imaging
Lecture Time
10:09 - 10:21

Abstract

Background

The ‘central vein sign’ (CVS) is increasingly recognized as a valuable biomarker with high specificity and sensitivity for multiple sclerosis (MS) MRI lesions. Current consensus North American Imaging in Multiple Sclerosis (NAIMS) guidelines recommend excluding lesions <3mm in diameter in any plane for CVS assessment. However, different lesion-size exclusion cut-offs for CVS have not been systematically evaluated.

Objectives

To evaluate the impact of different lesion size cut-offs and exclusion methodologies on CVS analysis and select3* criteria for MS diagnosis.

Methods

MS patients and non-MS controls were recruited as part of the National Institute of Neurological Disorders and Stroke MS natural history study and underwent 3T MRI on either Siemens Skyra or Philips Achieva scanners. MS lesions were segmented using a deep learning-based method and manually corrected by a single rater. Individual lesions were extracted as clusters of connected voxels, and their principal axes lengths (calculated as the lengths of the major axes of an ellipsoid with the same normalized second central moments) were used to measure lesion size in 3 dimensions. Ground truth CVS assessment was conducted by two raters on all lesions regardless of size. Two paradigms of lesion exclusion were compared: (1) excluding lesions if any dimension was less than threshold (ExcAny), or (2) if all dimensions were less than threshold (ExcAll).

Results

A total of 3920 lesions from 71 subjects (8 healthy controls, 36 RRMS, 12 SPMS, 14 PPMS, 1 CIS) were included in the analysis. CVS+ lesions were more likely to be ovoid and less spherical compared to their CVS- counterparts, as measured by the fractional anisotropy of lesion dimensions (mean difference 0.02, p=0.001). Of the 1679 CVS+ lesions in the cohort, 82% met the ExcAny criteria to be excluded at a 3mm cut-off, which was reduced to 29% when ExcAll criteria were used (McNemar test, p < 0.001). At the subject-level, an increase in the sensitivity of select3* CVS criteria for MS diagnosis was noted at 3mm using the less strict ExcAll (95%) compared to the more conservative ExcAny criteria (61%), without impacting specificity (100% for both methods). There was a reduction in specificity for both ExcAny and ExcAll criteria when size cut-offs less than or equal to 2mm were used (88% for both).

Conclusions

Compared to the current NAIMS guidelines, ExcAll criteria for CVS lesion analysis allow the inclusion of a larger proportion of CVS+ lesions and improve the sensitivity of select3* criteria for MS diagnosis. These findings improve the applicability of the CVS as a diagnostic marker for MS in clinical practice and provide evidence for future modifications of CVS lesion exclusion guidelines.

Collapse

Author Of 3 Presentations

Machine Learning/Network Science Poster Presentation

P0016 - Progressive multifocal leukoencephalopathy lesion and brain parenchymal segmentation from MRI using serial deep convolutional neural networks (ID 1531)

Speakers
Presentation Number
P0016
Presentation Topic
Machine Learning/Network Science

Abstract

Background

Progressive multifocal leukoencephalopathy (PML) is a rare opportunistic brain infection caused by the JC virus associated with significant morbidity and mortality, which can occur in the context of certain MS disease modifying therapies. There are currently no validated automatic methods for quantification of PML lesion burden or brain atrophy on MRI.

Objectives

We assessed whether deep learning techniques can be employed for automated brain parenchymal and lesion segmentation in PML using an approach dubbed “JCnet,” named after the causative viral agent.

Methods

We performed a retrospective analysis of PML patients who were evaluated at the NIH Neuroimmunology Clinic. MRI scans were acquired on either a Siemens Skyra or a Philips 3T MRI scanners. For PML brain and lesion segmentation, we implement a 3D patch-based approach with two consecutive fully convolutional neural networks (CNNs) with a feature pyramid architecture. The first network performs brain extraction as foreground, with meninges and cerebrospinal fluid spaces as background , while the second segments the underlying PML lesion(s). We measured the segmentation accuracy using Dice similarity coefficient (DSC) and absolute volume differences (AVD). We evaluated JCnet against methods designed for normal-appearing brain segmentation, FSL/FMRIB's Automated Segmentation Tool (FAST) and FreeSurfer, as well as MS lesion segmentation, Lesion Segmentation Toolbox (LST) and Lesion-TOpology-preserving Anatomical Segmentation (LTOADS). Comparisons were performed using Wilcoxon matched-pairs signed-ranks test.

Results

A total of 41 PML patients (mean age 55 years, SD 13; 44% female) were included in the analysis. The cohort was empirically divided into 31 training and 10 testing cases sampled at random. The mean time between PML onset and MRI acquisition was 4.5 months (range 0.6 – 44.5 months). JCnet resulted in a 4% and 64cm3 absolute improvement in DSC and AVD compared to FAST (p=0.005 and 0.01), and a 6% and 41cm3 absolute improvement compared to FreeSurfer respectively (p=0.005 and p=0.02). This was driven in part by improved segmentation of brain tissue within T1-hypointense PML lesions. For PML lesion segmentation, there was an absolute improvement of 42% and 14cm3 in DSC and AVD respectively compared to LST, and 53% and 19cm3 absolute improvements compared to LTOADS respectively (p=0.005 for all lesion comparisons). This was driven by improved sensitivity of supra- and infratentorial PML lesion identification and segmentation.

Conclusions

We employ an end-to-end deep learning-based method for automated segmentation of lesion and brain parenchymal volume in PML. By tracking quantitative measurements of PML-related MRI changes, this approach provides a window for clinicians and scientists to accurately monitor PML radiographically and its response to experimental therapies.

Collapse
Diagnostic Criteria and Differential Diagnosis Poster Presentation

P0261 - Paramagnetic rim lesions are specific to multiple sclerosis: an international multicenter 3T MRI study (ID 1025)

Abstract

Background

In multiple sclerosis (MS), a subset of chronic active white matter lesions are identifiable on MRI by their paramagnetic rims, and increasing evidence supports their association with clinical disease severity.

Objectives

To assess the prevalence and MS-specificity of paramagnetic rim lesions (PRL) on 3-tesla susceptibility-based MR brain images in MS vs non-MS cases in a multicenter sample drawn from 5 academic research hospitals at sites in Europe (Brussels, Lausanne, Milan) and the United States (NIH and JHU).

Methods

On submillimetric 3D T2*-segmented EPI brain MRI, the presence of PRL and central vein sign (CVS) were evaluated in the supratentorial brain of adults with MS (n=329) and non-MS neurological conditions (n=83). Non-MS cases were grouped as follows: (1) other-inflammatory neurological diseases (n=41); (2) HTLV-associated myelopathy/tropical spastic paraparesis (HAM/TSP; n=10); (3) HIV-infected (n=10); (4) non-inflammatory neurological diseases (n=22).

ROC curve analysis, with diagnosis as dependent variable (MS vs non-MS), was applied to examine the diagnostic accuracy for each biomarker (PRL and CVS). Youden’s index method was used to obtain the optimal cutoff value for each biomarker.

Results

PRL were detected in 172/329 (52%) of MS cases vs. 6/83 non-MS cases (7%).

In MS, 58% of progressive cases had at least one PRL, compared to 50% of relapsing cases. MS cases with more than 4 PRL were more likely to have higher disability scores (EDSS, MSSS and ARMSS), but not significantly longer disease duration or older age.

In non-MS cases, PRL were seen exclusively in only a few inflammatory/infectious neurological conditions, including Susac syndrome (3 cases), neuromyelitis optica spectrum disorder (1 case), Sjögren disease (1 case) and HAM/TSP (1 case). Unlike in MS, PRL in non-MS cases were not associated with a high frequency of CVS+ lesions.

The identification of at least one PRL (optimal cutoff) was associated with high diagnostic specificity (93%), but relatively low sensitivity (52%) and accuracy (area under ROC curve=0.77), whereas CVS detection alone (optimal cutoff 35.5-38%) could better discriminate MS from non-MS cases with high specificity (96%), sensitivity (99%), and accuracy (area under ROC curve=0.99). The combination of the two biomarkers further improved the specificity (99%), but sensitivity remained low (59%).

Conclusions

PRL yielded high specificity for MS lesions. Future prospective multicenter studies should further validate its role as a diagnostic biomarker.

Collapse
Imaging Poster Presentation

P0623 - Quantifying Cervical and Thoracic Cord Atrophy in Multiple Sclerosis  (ID 1478)

Speakers
Presentation Number
P0623
Presentation Topic
Imaging

Abstract

Background

Spinal cord atrophy contributes to disability in multiple sclerosis (MS), and its quantification along the entire spinal cord may be important to fully characterize the disease.

Objectives

We sought to characterize atrophy of the entire spinal cord in various multiple sclerosis phenotypes and determine its clinical correlates in a cross-sectional study. Further, we sought to evaluate its evolution in a longitudinal study of relapsing remitting MS (RRMS).

Methods

Axial T1-weighted images perpendicular to cord edge were automatically reformatted at each point along the cord. Spinal cord cross‐sectional area (SCCSA) were calculated from C1-T10 vertebral body levels and profile plots were compared across phenotypes. Average values from C2-3, C4-5, and T4-9 regions were compared across phenotypes and correlated with clinical scores, then categorized as atrophic/normal based on z-scores derived from controls, to compare clinical scores between subgroups. In the subset of relapsing-remitting cases with longitudinal scans, cases showing clinical progression (progressive-disability group) were defined as those in whom change in EDSS was ≥ 1 , while all other cases were grouped as having stable-disability. A random coefficient model for longitudinal data was applied to evaluate the change of regional-SCCSA variables over time, including in the model the disability group (progressive vs. stable), age, and the interaction between disability group and age.

Results

The cross-sectional study consisted of 149 adults with RRMS, 49 with secondary-progressive MS, 58 with primary-progressive MS and 48 healthy controls. The longitudinal study included 78 RRMS cases. Compared to controls, all MS groups had smaller average regions except RRMS in T4-9 region. Measures from all regions of the RRMS cohort correlated with clinical measures, whereas the progressive cohorts had fewer clinical correlates. In the RRMS cohort, 23% of cases had at least one atrophic region, whereas in progressive MS the rate was almost 70%. Longitudinal analysis demonstrated a correlation between disability and cervical cord thinning, as the random coefficient model showed a significant interaction between groups (stable- vs. progressive-disability) and age for cervical regional-SCCSA variables, indicating that the rate of decrease in regional-SCCSA with age in the progressive disability group was significantly higher than that in the stable disability group (0.62 mm2/year vs. 0.07 mm2/year for C2–3, p=0.0015; 0.72 mm2/year vs. 0.29 mm2/year for C4–5, p=0.0038).

Conclusions

Spinal cord atrophy was demonstrated in all MS phenotypes, with SCCSA from all regions showing significant correlations with all clinical parameters in RRMS cohort. Longitudinal changes in the cervical regions were significantly higher in RRMS subjects showing clinical progression than those who did not. SCCSA is therefore a potential imaging marker for disease progression.

Collapse