Author Of 1 Presentation
FC03.05 - Reduced thalamic atrophy in patients initiating earlier versus delayed ocrelizumab therapy: results from the OLE of OPERA I/II and ORATORIO
Abstract
Background
In multiple sclerosis (MS), thalamic integrity is affected both directly by demyelination, neuronal loss and increasing iron concentration, and indirectly by remote gray and white matter lesions affecting neural projections into and out of the thalamus. Thalamic atrophy may therefore reflect a large fraction of MS-related brain damage and thus represent a useful marker of overall damage and therapeutic efficacy.
Objectives
To assess the efficacy of ocrelizumab (OCR) in patients switching to or maintaining OCR therapy on thalamic atrophy in patients with relapsing MS (RMS) and primary progressive MS (PPMS), participating in the OPERA I/II (NCT01247324/NCT01412333) and ORATORIO (NCT01194570) Phase III trials, respectively.
Methods
At the end of the double-blind controlled treatment period in OPERA I/II, patients entered the open‑label extension (OLE), and either continued to receive OCR (OCR-OCR) or switched from interferon β-1a (IFN β-1a) to OCR (IFN β-1a-OCR). In ORATORIO, patients entered the OLE ~3–9 months after the double-blind period cut-off and either continued OCR (OCR-OCR) or switched from placebo (PBO) to OCR (PBO-OCR). Changes in thalamic volume from the core trial baseline were computed using Jacobian integration and analyzed using a mixed-effect repeated measurement model, adjusted for baseline volume, age, baseline gadolinium-enhancing lesions (presence/absence), baseline T2 lesion volume, region (US vs rest of the world), Expanded Disability Status Scale category (<4, ≥4), week, treatment, treatment and time interaction, and treatment and baseline volume interaction.
Results
In the OLE of OPERA I/II, changes (%) in thalamic volume from baseline at OLE Week 46, 94, 142, 190, and 238, were: –2.88/–2.12 (p<0.001), –3.31/–2.36 (p<0.001), –3.61/–2.78 (p<0.001), –3.68/–3.03 (p<0.001), and –4.07/–3.41 (p<0.001), for IFN β-1a-OCR/OCR-OCR patients, respectively. During the OLE of ORATORIO, changes in thalamic volume at OLE Day 1, Week 48, 96, and 144, were: –3.46/–2.44 (p<0.001), –3.93/–2.61 (p<0.001), –4.30/–3.25 (p<0.001), and –4.86/–3.62 (p<0.001), for PBO-OCR/OCR-OCR patients, respectively.
Conclusions
In the OLE, patients with RMS and PPMS who were initially randomized to ocrelizumab experienced less thalamic volume loss compared with those initiating ocrelizumab later.
Author Of 6 Presentations
P0006 - Detecting treatment response on T1 gadolinium enhancing lesion burden in clinical trials of multiple sclerosis with deep learning (ID 998)
Abstract
Background
Ocrelizumab (OCR) is a humanized anti-CD20+ monoclonal antibody approved for the treatment of relapsing and primary progressive forms of multiple sclerosis (MS). It suppresses the development of new areas of inflammation as shown by the reduction in the number of T1 gadolinium (Gd) enhancing lesions. Deep learning (DL) based segmentation of lesions has the potential to automate these manual reads; enabling faster and more reproducible quantification.
Objectives
To evaluate a DL model’s ability to detect OCR treatment response on the number of T1 Gd enhancing lesions in clinical trials of relapsing remitting MS (RRMS).
Methods
We used images from Opera I trials (NCT01247324, n=898), to train a Unet model to segment both unenhancing and Gd-enhancing T1 lesions. T1 Gd- enhancing lesions in ground truth (GT) masks were present in ~44% of patients at baseline and ~15% across all time points. We created a dataset with an equal number of imaging volumes with and without enhancing lesions and used 70%-30% data splits for training and validation. The model was tested on images from Opera II trials (NCT01412333, n=905). To detect significant differences between treatment and control arms, we performed a negative binomial regression on the number of T1 Gd-enhancing lesions with baseline imaging and clinical covariates.
Results
The DL model achieved a mean dice coefficient (DC) of 0.72, lesion true positive rate of 0.92 (LTPR), lesion false positive rate (LFPR) of 0.06 and a volume correlation coefficient (CC) of 0.97 for Gd-enhancing lesion segmentation. For unenhancing lesion segmentation, the mean DC was 0.68, LTPR 0.78, LFPR 0.18 and volume CC of 0.97. The model had the highest false positives for lesions smaller than 10 voxels (voxel size: 1x1x3 mm3). For Gd-enhancing lesion segmentation a significant OCR treatment effect (p<0.001) in reducing the mean number of Gd-enhancing lesions at 24, 48 and 96 weeks (92%, 96%, 97% reduction from GT manual reads vs 67%, 71%, 78% from model predictions) was detected.
Conclusions
Our DL model performed Gd-enhancing lesion segmentation comparably to similar DL models in the literature and showed a high correlation to GT manual reads. Our model also had sufficiently high sensitivity to detect an OCR treatment response consistent with neuro-radiologist assessments. To our knowledge, this is the first study to report that a DL model has the sensitivity to detect treatment response on Gd-enhancing lesions.
P0123 - Ocrelizumab reduces thalamic volume loss and clinical progression in PPMS and RMS independent of baseline NfL and other measures of disease severity (ID 1621)
Abstract
Background
Neurofilament light chain (NfL) is a biomarker of neuroaxonal injury in multiple sclerosis (MS). Thalamic atrophy occurs early and may be a sensitive marker of overall brain damage. Ocrelizumab (OCR) reduced brain atrophy and NfL in patients with relapsing MS (RMS) and those with primary progressive MS (PPMS).
Objectives
To examine the independent impact of OCR and baseline (BL) NfL on thalamic volume (TV) and clinical progression in patients with PPMS and RMS, including those with RMS without acute BL activity (i.e. no gadolinium–enhancing [Gd+] lesions or relapse in the last 3 months).
Methods
Patients were from OPERA I/II (RMS, n=1,421) and ORATORIO (PPMS, n=596). Thalamic atrophy was calculated as annualized percentage TV change (PTVC) from Wk 24 to the end of controlled treatment (ORATORIO, Wk 120; OPERA I/II, Wk 96). OCR treatment (vs IFNβ-1a [RMS] or placebo [PPMS]) and log-transformed BL NfL were examined for associations with PTVC (linear regression) and 24-week confirmed disability progression (Cox regression) adjusting for BL demographic and disease characteristics.
Results
In patients with PPMS and RMS, OCR treatment (PTVC: +0.47% and +0.33%, respectively) and lower BL NfL (+0.20% and +0.33% per 2-fold lower NfL) independently associated with a smaller TV reduction (all p<0.005). Adjusting for BL NfL level, Gd+ lesion count, T2 lesion volume and BL disability, OCR still reduced disability progression on Expanded Disability Status Scale (EDSS) (PPMS, hazard ratio [HR]=0.73; RMS, HR=0.65; both p<0.05]), 9-Hole Peg Test (9HPT) (PPMS, HR=0.53, p=0.002; RMS, HR=0.52, p=0.059), Timed 25-Foot Walk (T25FW) (PPMS, HR=0.79, p=0.063), Symbol Digit Modalities Test (RMS, HR=0.54, p=0.002) and time to EDSS 6 (RMS, HR=0.42, p=0.009). In patients with PPMS, higher BL NfL was associated with worsening on 9HPT (HR=1.34 per 2-fold higher NfL), T25FW (HR=1.19) and time to EDSS 7 (HR=1.78) (all p<0.05). In patients with RMS without acute BL activity, higher BL NfL was associated with EDSS worsening (HR=1.49), progression independent of relapse activity (PIRA) (HR=1.61), 9HPT (HR=2.1) and time to EDSS 6 (HR=2.24) (all p<0.05).
Conclusions
Ocrelizumab treatment remained associated with reduced thalamic atrophy and clinical progression after adjusting for baseline NfL and other factors. Higher BL NfL was associated with increased rates of thalamic atrophy and clinical progression in patients with PPMS and those with RMS without acute disease activity.
P0125 - Ocrelizumab treatment induces a sustained blood NfL reduction in patients with PPMS and RMS (ID 1865)
Abstract
Background
Blood neurofilament light chain (NfL) is a biomarker of neuroaxonal injury associated with acute disease activity and may be prognostic for disability progression in patients with multiple sclerosis (MS). Ocrelizumab (OCR) is an anti-CD20 monoclonal antibody indicated for relapsing MS (RMS) and primary progressive MS (PPMS).
Objectives
To assess the impact of OCR on blood NfL distribution in patients with RMS from the OPERA I and II trials and those with PPMS from ORATORIO.
Methods
Pretreatment and posttreatment NfL levels (measured using the SiMOA assay) with OCR vs interferon β-1a (OPERA I and II; n=1,421) or placebo (ORATORIO; n=596) were compared using geometric mean (GM) and GM ratios (GMR). Patients were stratified by presence/absence of acute disease activity at baseline (BL) (T1 gadolinium [Gd]-enhancing lesions and/or relapse in prior 3 months for RMS; T1 Gd-enhancing lesions for PPMS). Age-adjusted NfL distributions (using a linear model for log-NfL and age derived from a healthy donor [HD] cohort) at BL and after OCR were compared with HD using the Kolmogorov-Smirnov test.
Results
Significant reductions in NfL were observed 3 months after OCR initiation (RMS, GMR=0.80; PPMS, GMR=0.89) and sustained through the end of controlled treatment (RMS [96 weeks], GMR=0.56; PPMS [120 weeks], GMR=0.81; all p<0.0001). Age-adjusted BL serum NfL was elevated in patients with RMS disease activity (GM [95% CI]=12.7 [11.9–13.6] pg/mL) vs those without (5.5 [5.3–5.7] pg/mL) and HD (4.1 [3.9–4.4] pg/mL; all p<0.0001). In OCR-treated patients with RMS, GM [95% CI] serum NfL levels after 96 weeks (with activity at BL, 4.4 [4.2–4.6] pg/mL; without activity at BL, 4.1 [4.0–4.3] pg/mL) were comparable to HD (4.1 [3.9–4.4] pg/mL; all p>0.1). Age-adjusted BL plasma NfL was also elevated in PPMS patients with disease activity (GM [95% CI]=8.7 [7.5–10.1] pg/mL) vs those without (4.9 [4.6–5.2] pg/mL) and HD (3.1 [2.9–3.3] pg/mL; all p<0.0001). In OCR-treated patients with PPMS, GM [95% CI] plasma NfL levels after 120 weeks (with activity at BL, 4.6 [4.1–5.1] pg/mL; without activity at BL, 4.2 [4.0–4.4] pg/mL) were reduced from BL (all p<0.005) but remained elevated vs HD (all p<0.001).
Conclusions
NfL is highly elevated in patients with acute MS disease activity, and more subtle elevations are observed in RMS and PPMS patients without detectable disease activity. Ocrelizumab significantly reduces NfL in RMS and PPMS patients with and without detectable disease activity.
P0613 - Ocrelizumab Reduces Regional Brain Atrophy in Relapsing Multiple Sclerosis as Assessed by Longitudinal Deformation-based Morphometry (ID 1228)
Abstract
Background
Ocrelizumab (OCR) is a humanised anti-CD20+ monoclonal antibody approved for the treatment of relapsing and primary progressive forms of multiple sclerosis (MS). It has been reported that OCR reduces brain atrophy in MS patients, however, a voxel-based approach has never been reported.
Objectives
To identify brain regions where atrophy is modified by the OCR treatment in Relapsing MS (RMS), using a fully automatic, voxel-based approach.
Methods
We applied longitudinal deformation-based morphometry (L-DBM) using the Advanced Normalization Tools (ANTs) to T1-weighted 3D MPRAGE brain MRI (voxel=1x1x3 mm3 ) from two clinical trials of OCR for RMS (Opera I: NCT01247324 and Opera II: NCT01412333). For each MRI visit, a Jacobian map was derived based on the deformation generated by registering the scan to a Single-Subject Template (SST) constructed for the same patient. Jacobian maps in patient SSTs were mapped to a population brain template for group analysis. Patients were included if MRI data were available for all visits (baseline, Weeks 24, 48, and 96) and the images were of sufficient quality to be successfully processed through the L-DBM pipeline (683 in the OCR arm, 580 in the active control arm of Interferon beta-1a). Voxelwise ANOVA was iterated over the whole brain to detect an interaction between Treatment and Time (p<0.01 with familywise error correction). Post-hoc analysis was done with mean Jacobian of identified regions, including Linear Mixed Effect (LME) analysis.
Results
Voxelwise ANOVA identified one cluster of voxels in the brain (besides the ventricles) that were contained within parts of thalamus, brainstem, and cerebellum. From the analysis of mean Jacobian of this cluster region, the percent volume reduction from baseline was significantly smaller in the treatment arm than the control arm at every follow-up visit (p<0.0001, t>8.0, relative difference>75%). LME analysis of the mean Jacobian of the cluster region confirmed that there was a strong interaction between Treatment and Time (F=18.35), for which age and sex were controlled.
Conclusions
L-DBM has identified several brain structures where atrophy was modified by the OCR treatment in RMS. The L-DBM analysis was performed at the voxel level and these regional findings are consistent with previous reports. Localization of the treatment-responsive brain structures may have an implication for future assessment of clinical outcomes in MS.
P0856 - Collecting real world MRI in MS: preliminary results from FlywheelMS (ID 1226)
Abstract
Background
Real-world evidence can be used to better characterize the course of multiple sclerosis (MS), care provision and outcomes in clinical practice. Magnetic resonance imaging (MRI) that occurs in the context of usual care is an important source of information that can inform clinical decision-making. Guidelines exist to enhance the clinical impact of routine MRI in MS, but it is unclear whether MRIs acquired as part of routine care in the United States adhere to these guidelines.
Objectives
To describe the clinical routine brain MRIs from patients with MS across different US imaging sites.
Methods
FlywheelMS is a novel patient-centered study that aims to extract and digitize health information not readily available in existing electronic health records of patients with MS. Up to 5,000 consenting adults with a confirmed MS diagnosis will be enrolled. Brain MRI data were retrieved, and summary statistics were computed to describe the sessions, imaging sites, scanner field strengths and slice thickness of T1-weighted and FLAIR (fluid-attenuated inversion recovery) images. Longitudinal acquisition consistency (i.e. MRIs acquired from the same center with the same scanner) was also assessed.
Results
Out of 2,389 patients enrolled, 1555 brain MRI data were retrieved from the first 492 patients (female, 81%; mean age at consent, 49±11 years). The mean number of MRI sessions per patient was 3.2±2.4, and data were captured between 1999 and 2018. Sessions were acquired at 598 different imaging sites, using mainly 1.5T scanners (61.3%), followed by 3T (32.7%) and lower field-strength magnets (3.4%; not available, 2.6%). The mean slice thickness of T1-weighted (3.1±1.7 mm) and FLAIR images (3.1±1.3 mm) was similar. Of the 352 patients (72%) that had more than one MRI session, 85 (24.1%) had consistent acquisition (i.e. same site/scanner), 153 (43.5%) had one site or scanner change, and 114 (32.4%) had more than one site and/or scanner change.
Conclusions
The novel, patient-centered approach of FlywheelMS can successfully extract imaging data from medical records of patients with MS across US imaging sites. These data will help us in describing the clinical routine MRI, determining the compliance to guidelines and understanding which measure (e.g. lesion volume and/or atrophy) could be potentially extracted from MRI data.
P0875 - FlywheelMS: The prevalence of multiple sclerosis subtypes in digitized health records (ID 1882)
Abstract
Background
Data generated from electronic health records (EHRs) offer insight into real-world care of people with multiple sclerosis (MS). Data extracted most readily from EHRs include templated or administrative health information (e.g., MS International Classification of Diseases codes). However, clinical data like disease subtype and characteristics are unlikely to be captured systematically. FlywheelMS is a novel patient-centered study with the aim of digitizing health records of patients with MS and extracting information not readily available in existing EHRs.
Objectives
To evaluate patient characteristics and the prevalence of MS subtypes (i.e., relapsing-remitting MS [RRMS], secondary progressive MS [SPMS], primary progressive MS [PPMS], progressive relapsing MS [PRMS]) in the FlywheelMS cohort and to compare them with existing real-world data sources.
Methods
Adults with MS are recruited across the US via advocacy groups, social media and healthcare professionals. Supervised machine learning with human curation is used to retrieve, digitize and abstract medical records, which are collected as far back as are available and prospectively up to 5 years after enrollment. The most recent non-negated MS subtype from neurology visit records was used as a proxy for the prevalent subtype. Summary statistics were calculated and compared with other MS cohorts.
Results
As of March 1, 2020, 2,389 patients with MS with 24,362 neurology visits across 3,093 neurologists have enrolled in FlywheelMS. Data on MS subtype were available for 973 patients (40.7%); this proportion will increase as abstractions continue. RRMS accounted for 78.9% of patients, followed by SPMS (12%), PPMS (7.3%) and PRMS (1.7%). These findings were comparable to the MSBase Registry (RRMS=76.9%, SPMS=13.0%, PPMS=8.0%, PRMS=2.2%; Kister et al., J Neurol Sci 2012) and NARCOMS Registry (RRMS=65.6%, SPMS=25.1%, PPMS=9.3%; Salter et al., Mult Scler 2018). Mean [SD] age at mention of MS subtype and percent female distribution were as follows: RRMS (46.4 [10.9] years, 80.4%), SPMS (56.4 [9.6] years, 81.2%), PPMS (53.9 [10.7] years, 62.0%), PRMS (mean 53.9 [5.5] years, 82.4%).
Conclusions
The prevalence of MS subtypes in the digitized health records of patients in FlywheelMS was comparable to other real-world data sources. Digitizing and machine-learning guided abstraction of patient healthcare records in MS yields important data about clinical features not readily available in other EHR data sets.