Multicentre assessment of motor and sensory evoked potentials in multiple sclerosis: reliability and implications for clinical trials

Visual evoked potentials in multiple sclerosis: P100 latency and visual pathway damage including the lateral geniculate nucleus

OBJECTIVE

To explore associations of the main component (P100) of visual evoked potentials (VEP) to pre- and postchiasmatic damage in multiple sclerosis (MS).

METHODS

31 patients (median EDSS: 2.5), 13 with previous optic neuritis (ON), and 31 healthy controls had VEP, optical coherence tomography and magnetic resonance imaging. We tested associations of P100-latency to the peripapillary retinal nerve fiber layer (pRNFL), ganglion cell/inner plexiform layers (GCIPL), lateral geniculate nucleus volume (LGN), white matter lesions of the optic radiations (OR-WML), fractional anisotropy of non-lesional optic radiations (NAOR-FA), and to the mean thickness of primary visual cortex (V1). Effect sizes are given as marginal R2 (mR2).

RESULTS

P100-latency, pRNFL, GCIPL and LGN in patients differed from controls. Within patients, P100-latency was significantly associated with GCIPL (mR2 = 0.26), and less strongly with OR-WML (mR2 = 0.17), NAOR-FA (mR2 = 0.13) and pRNFL (mR2 = 0.08). In multivariate analysis, GCIPL and NAOR-FA remained significantly associated with P100-latency (mR2 = 0.41). In ON-patients, P100-latency was significantly associated with LGN volume (mR2 = -0.56).

CONCLUSIONS

P100-latency is affected by anterior and posterior visual pathway damage. In ON-patients, damage at the synapse-level (LGN) may additionally contribute to latency delay.

SIGNIFICANCE

Our findings corroborate post-chiasmatic contributions to the VEP-signal, which may relate to distinct pathophysiological mechanisms in MS.

Motor Evoked Potential-A Pilot Study Looking at Reliability and Clinical Correlations in Multiple Sclerosis

PURPOSE

Multiple sclerosis (MS) is a clinically heterogeneous disease. Biomarkers that can assess pathological processes that are unseen with conventional imaging remain an unmet need in MS disease management. Motor evoked potentials (MEPs) could be such a biomarker. To determine and follow longitudinal MEP reliability and correlations with clinical measures in MS patients.

METHODS

This is a single-center study in alemtuzumab-treated MS patients to evaluate temporal reliability of MEPs, identify MEP minimum detectible differences, and explore correlations with existing clinical scales. Ten MS patients recently treated with alemtuzumab were evaluated every 6 months over 3 years. Clinical evaluations consisted of expanded disability status scale, timed 25-foot walk, 6-minute walk, and nine-hole peg test. MEPs were measured twice, 2 weeks apart, every 6 months.

RESULTS

Eight patients completed all 3 years of study. The intraclass correlation coefficient for MEP parameters ranged from 0.76 to 0.98. TA latency and amplitude with facilitation significantly and strongly correlated with all clinical measures, whereas the MEP duration modestly correlated. Biceps latency with facilitation significantly and moderately correlated with 9-hole peg test. Longitudinal correlations demonstrated good predictive values for either clinical deterioration or improvement.

CONCLUSIONS

MEPs have excellent intrapatient and intrarater reliability, and TA MEPs significantly and strongly correlated with expanded disability status scale, 6-minute walk, and timed 25-foot walk, whereas biceps MEPs significantly and moderately correlated with nine-hole peg test. Further studies using larger cohorts of MS patients are indicated.

CLINICAL TRIAL REGISTRATION

ClinicalTrials.gov, Identifier: NCT02623946.

Validation of Quantitative Scores Derived From Motor Evoked Potentials in the Assessment of Primary Progressive Multiple Sclerosis: A Longitudinal Study

Objective: To evaluate the sensitivity to change of differently calculated quantitative scores from motor evoked potentials (MEP) in patients with primary progressive multiple sclerosis (PPMS). Methods: Twenty patients with PPMS had MEP to upper and lower limbs at baseline, years 1 and 2 measured in addition to clinical assessment [Expanded Disability Status Scale (EDSS), ambulation score]; a subsample (n = 9) had a nine-hole peg test (NHPT) and a timed 25-foot walk (T25FW). Quantitative MEP scores for upper limbs (qMEP-UL), lower limbs (qMEP-LL), and all limbs (qMEP) were calculated in three different ways, based on z-transformed central motor conduction time (CMCT), shortest corticomuscular latency (CxM-sh), and mean CxM (CxM-mn). Changes in clinical measures and qMEP metrics were analyzed by repeated-measures analysis of variance (rANOVA), and a factor analysis was performed on change in qMEP metrics. Results: Expanded Disability Status Scale and ambulation score progressed in the rANOVA model (p < 0.05; post-hoc comparison baseline-year 2, p < 0.1). Lower limb and combined qMEP scores showed significant deterioration of latency (p < 0.01, MEP-LL_CxM-sh: p < 0.05) and in post-hoc comparisons (baseline-year 2, p < 0.05), qMEP_CxM-mn even over 1 year (p < 0.05). Effect sizes were higher for qMEP scores than for clinical measures, and slightly but consistently higher when based on CxM-mn compared to CxM-sh or CMCT. Subgroup analysis yielded no indication of higher sensitivity of timed clinical measures over qMEP scores. Two independent factors were detected, the first mainly associated with qMEP-LL, the second with qMEP-UL, explaining 65 and 29% of total variability, respectively. Conclusions: Deterioration in qMEP scores occurs earlier than EDSS progression in patients with PPMS. Upper and lower limb qMEP scores contribute independently to measuring change, and qMEP scores based on mean CxM are advantageous. The capability to detect subclinical changes longitudinally is a unique property of EP and complementary to clinical assessment. These features underline the role of EP as candidate biomarkers to measure effects of therapeutic interventions in PPMS.

Machine learning analysis of motor evoked potential time series to predict disability progression in multiple sclerosis

Background

Evoked potentials (EPs) are a measure of the conductivity of the central nervous system. They are used to monitor disease progression of multiple sclerosis patients. Previous studies only extracted a few variables from the EPs, which are often further condensed into a single variable: the EP score. We perform a machine learning analysis of motor EP that uses the whole time series, instead of a few variables, to predict disability progression after two years. Obtaining realistic performance estimates of this task has been difficult because of small data set sizes. We recently extracted a dataset of EPs from the Rehabiliation & MS Center in Overpelt, Belgium. Our data set is large enough to obtain, for the first time, a performance estimate on an independent test set containing different patients.

Methods

We extracted a large number of time series features from the motor EPs with the highly comparative time series analysis software package. Mutual information with the target and the Boruta method are used to find features which contain information not included in the features studied in the literature. We use random forests (RF) and logistic regression (LR) classifiers to predict disability progression after two years. Statistical significance of the performance increase when adding extra features is checked.

Results

Including extra time series features in motor EPs leads to a statistically significant improvement compared to using only the known features, although the effect is limited in magnitude (ΔAUC = 0.02 for RF and ΔAUC = 0.05 for LR). RF with extra time series features obtains the best performance (AUC = 0.75±0.07 (mean and standard deviation)), which is good considering the limited number of biomarkers in the model. RF (a nonlinear classifier) outperforms LR (a linear classifier).

Conclusions

Using machine learning methods on EPs shows promising predictive performance. Using additional EP time series features beyond those already in use leads to a modest increase in performance. Larger datasets, preferably multi-center, are needed for further research. Given a large enough dataset, these models may be used to support clinicians in their decision making process regarding future treatment.