Assessing disease activity in multiple sclerosis: biomarkers and clinical measures

Abstract

Background: Multiple sclerosis (MS) is the commonest cause of non-traumatic neurological disability in young adults. Disease-modifying therapies are available which vary in their overall efficacy and safety profiles, with the most effective treatments reserved for those with the most active disease. People with MS (PwMS) who show lack of response to initial treatment may also switch to more effective treatments. Outcome measures are required in MS, for example to inform clinical decisions regarding initiation and switching of disease-modifying therapies. Outcome measures are required both to guide decision-making at an individual level, and to test research questions in large cohorts at a group level. Objectives: i) To evaluate whether relapses or MRI activity after starting the disease-modifying therapy natalizumab predict long-term disability progression. ii) To investigate 18 kDa translocator protein (TSPO) positron emission tomography (PET) with the novel ligand 18F-GE180 as a marker for neuroinflammation, to assess disease activity in MS and response to natalizumab treatment. iii) To investigate whether the machine-learning model ‘Brain Predicted Age’, which estimates chronological age based upon T1-weighted brain MRIs, can assess disease severity in MS. iv) To assess the prognostic value of the patient reported outcome ‘Multiple Sclerosis Impact Scale–29’ (MSIS-29) on hard clinical endpoints including ‘survival time’ in people with MS. Methods: i) Observational cohort of 161 PwMS initiating natalizumab, with up to 7 years of follow-up. ii) Cohort of 16 PwMS starting natalizumab, with 18F-GE180 PET at baseline, 10 weeks after, and 58 weeks after treatment initiation. 11 healthy controls (HCs) had one 18F-GE180 PET scan. In addition, a ‘blocking study’ investigated whether 18F-GE180 PET is labelling TSPO: 6 PwMS had a baseline 18F-GE180 PET followed by repeat PET after administration of the TSPO ligand XBD-173. iii) Observational cohort of 541 subjects with clinically isolated syndrome, MS or HCs, with 1,652 T1-weighted MRI scans and up to 15 years follow-up. iv) Observational cohort of 2,126 PwMS that completed the MSIS-29 questionnaires, with up to 10 years of follow-up. Results: i) Relapses and MRI activity after natalizumab initiation predicted disability progression at years 1 and 2. However, this effect disappeared with longer follow-up. Instead there was a paradoxical trend towards inflammatory activity on treatment predicting a lower risk of disability progression. ii) 18F-GE180 had lower extraction over the blood-brain barrier than expected from preclinical studies, although did show some specific binding. 18F-GE180 signal was increased in MS lesions, and decreased after natalizumab treatment. 18F-GE180 signal was not increased in other regions of interest. iii) Brain Predicted Age was substantially greater than chronological age in all disease subgroups. Brain Predicted Age correlated with disease characteristics and predicted future disability progression. Longitudinal change in Brain Predicted Age increased at a rate ~46% faster in subjects with MS than in HCs. The model compared favourably with conventional univariate measures of brain atrophy. iv) 264 PwMS died over the follow-up period. Higher MSIS-29 physical and psychological scores predicted reduced survival time, independently of other risk factors. Conclusions: i) While relapses and MRI activity remain useful markers of disease activity, there is a need for additional outcome measures which can predict long-term disability progression in PwMS on disease-modifying therapy. ii) 18F-GE180 has low brain extraction (~1%) and poor signal-to-noise ratio. Results should be interpreted with caution, given the unexpected pharmacokinetic characteristics of the tracer. iii) The Brain Predicted Age model could provide conceptually simple and clinically meaningful quantitative data in MS from a standard T1-weighted MRI scan, and will be further evaluated using the MAGNIMS multi-centre cohort. iv) MSIS-29 scores can be prognostic for hard clinical endpoints in MS. This is the first study to associate PROs with survival in any neurological disease. PROs are cheap and easy to administer, and could allow the collection of clinically meaningful outcome data in large MS cohorts.