Prognostic biomarkers of IFNb therapy in multiple sclerosis patients

A systematic review of the application of machine-learning algorithms in multiple sclerosis

INTRODUCTION

The applications of artificial intelligence, and in particular automatic learning or "machine learning" (ML), constitute both a challenge and a great opportunity in numerous scientific, technical, and clinical disciplines. Specific applications in the study of multiple sclerosis (MS) have been no exception, and constitute an area of increasing interest in recent years.

OBJECTIVE

We present a systematic review of the application of ML algorithms in MS.

MATERIALS AND METHODS

We used the PubMed search engine, which allows free access to the MEDLINE medical database, to identify studies including the keywords "machine learning" and "multiple sclerosis." We excluded review articles, studies written in languages other than English or Spanish, and studies that were mainly technical and did not specifically apply to MS. The final selection included 76 articles, and 38 were rejected.

CONCLUSIONS

After the review process, we established 4 main applications of ML in MS: 1) classifying MS subtypes; 2) distinguishing patients with MS from healthy controls and individuals with other diseases; 3) predicting progression and response to therapeutic interventions; and 4) other applications. Results found to date have shown that ML algorithms may offer great support for health professionals both in clinical settings and in research into MS.

Ensemble machine learning identifies genetic loci associated with future worsening of disability in people with multiple sclerosis

Limited studies have been conducted to identify and validate multiple sclerosis (MS) genetic loci associated with disability progression. We aimed to identify MS genetic loci associated with worsening of disability over time, and to develop and validate ensemble genetic learning model(s) to identify people with MS (PwMS) at risk of future worsening. We examined associations of 208 previously established MS genetic loci with the risk of worsening of disability; we learned ensemble genetic decision rules and validated the predictions in an external dataset. We found 7 genetic loci (rs7731626: HR 0.92, P = 2.4 × 10^-5; rs12211604: HR 1.16, P = 3.2 × 10^-7; rs55858457: HR 0.93, P = 3.7 × 10^-7; rs10271373: HR 0.90, P = 1.1 × 10^-7; rs11256593: HR 1.13, P = 5.1 × 10^-57; rs12588969: HR = 1.10, P = 2.1 × 10^-10; rs1465697: HR 1.09, P = 1.7 × 10^-128) associated with risk worsening of disability; most of which were located near or tagged to 13 genomic regions enriched in peptide hormones and steroids biosynthesis pathways by positional and eQTL mapping. The derived ensembles produced a set of genetic decision rules that can be translated to provide additional prognostic values to existing clinical predictions, with the additional benefit of incorporating relevant genetic information into clinical decision making for PwMS. The present study extends our knowledge of MS progression genetics and provides the basis of future studies regarding the functional significance of the identified loci.

The role of machine learning in developing non-magnetic resonance imaging based biomarkers for multiple sclerosis: a systematic review

Background

Multiple sclerosis (MS) is a neurological condition whose symptoms, severity, and progression over time vary enormously among individuals. Ideally, each person living with MS should be provided with an accurate prognosis at the time of diagnosis, precision in initial and subsequent treatment decisions, and improved timeliness in detecting the need to reassess treatment regimens. To manage these three components, discovering an accurate, objective measure of overall disease severity is essential. Machine learning (ML) algorithms can contribute to finding such a clinically useful biomarker of MS through their ability to search and analyze datasets about potential biomarkers at scale. Our aim was to conduct a systematic review to determine how, and in what way, ML has been applied to the study of MS biomarkers on data from sources other than magnetic resonance imaging.

Methods

Systematic searches through eight databases were conducted for literature published in 2014–2020 on MS and specified ML algorithms.

Results

Of the 1, 052 returned papers, 66 met the inclusion criteria. All included papers addressed developing classifiers for MS identification or measuring its progression, typically, using hold-out evaluation on subsets of fewer than 200 participants with MS. These classifiers focused on biomarkers of MS, ranging from those derived from omics and phenotypical data (34.5% clinical, 33.3% biological, 23.0% physiological, and 9.2% drug response). Algorithmic choices were dependent on both the amount of data available for supervised ML (91.5%; 49.2% classification and 42.3% regression) and the requirement to be able to justify the resulting decision-making principles in healthcare settings. Therefore, algorithms based on decision trees and support vector machines were commonly used, and the maximum average performance of 89.9% AUC was found in random forests comparing with other ML algorithms.

Conclusions

ML is applicable to determining how candidate biomarkers perform in the assessment of disease severity. However, applying ML research to develop decision aids to help clinicians optimize treatment strategies and analyze treatment responses in individual patients calls for creating appropriate data resources and shared experimental protocols. They should target proceeding from segregated classification of signals or natural language to both holistic analyses across data modalities and clinically-meaningful differentiation of disease.

Multiple sclerosis diagnosis and phenotype identification by multivariate classification of in vivo frontal cortex metabolite profiles

Multiple sclerosis (MS) is a heterogeneous autoimmune disease for which diagnosis continues to rely on subjective clinical judgment over a battery of tests. Proton magnetic resonance spectroscopy (¹H MRS) enables the noninvasive in vivo detection of multiple small-molecule metabolites and is therefore in principle a promising means of gathering information sufficient for multiple sclerosis diagnosis and subtype classification. Here we show that supervised classification using ¹H-MRS-visible normal-appearing frontal cortex small-molecule metabolites alone can indeed differentiate individuals with progressive MS from control (held-out validation sensitivity 79% and specificity 68%), as well as between relapsing and progressive MS phenotypes (held-out validation sensitivity 84% and specificity 74%). Post hoc assessment demonstrated the disproportionate contributions of glutamate and glutamine to identifying MS status and phenotype, respectively. Our finding establishes ¹H MRS as a viable means of characterizing progressive multiple sclerosis disease status and paves the way for continued refinement of this method as an auxiliary or mainstay of multiple sclerosis diagnostics.

Pharmacodynamic biomarkers of long-term interferon beta-1a therapy in REFLEX and REFLEXION

This post-hoc analysis evaluated candidate biomarkers of long-term efficacy of subcutaneous interferon beta-1a (sc IFN β-1a) in REFLEX/REFLEXION studies of clinically isolated syndrome. Samples from 507 REFLEX and 287 REFLEXION study participants were analyzed. All investigated biomarkers were significantly upregulated 1.5-4-fold in response to sc IFN β-1a treatment versus baseline (p ≤ 0.008). The validity of MX1, 2'5'OAS, and IL-1RA as biomarkers of response to sc IFN β-1a was confirmed in this large patient cohort, with biomarkers consistently upregulated in a dose-dependent manner. Neopterin, TRAIL, and IP-10 were confirmed as biomarkers associated with long-term sc IFN β-1a treatment efficacy over 5 years.

A systematic review of the application of machine-learning algorithms in multiple sclerosis

INTRODUCTION

The applications of artificial intelligence, and in particular automatic learning or "machine learning" (ML), constitute both a challenge and a great opportunity in numerous scientific, technical, and clinical disciplines. Specific applications in the study of multiple sclerosis (MS) have been no exception, and constitute an area of increasing interest in recent years.

OBJECTIVE

We present a systematic review of the application of ML algorithms in MS.

MATERIALS AND METHODS

We used the PubMed search engine, which allows free access to the MEDLINE medical database, to identify studies including the keywords "machine learning" and "multiple sclerosis." We excluded review articles, studies written in languages other than English or Spanish, and studies that were mainly technical and did not specifically apply to MS. The final selection included 76 articles, and 38 were rejected.

CONCLUSIONS

After the review process, we established 4 main applications of ML in MS: 1) classifying MS subtypes; 2) distinguishing patients with MS from healthy controls and individuals with other diseases; 3) predicting progression and response to therapeutic interventions; and 4) other applications. Results found to date have shown that ML algorithms may offer great support for health professionals both in clinical settings and in research into MS.

Validation of Progression-Free Survival Rate at 6 Months and Objective Response for Estimating Overall Survival in Immune Checkpoint Inhibitor Trials: A Systematic Review and Meta-analysis

IMPORTANCE

Progression-free survival (PFS) rate at 6 months has been proposed as a potential surrogate for overall survival (OS) rate at 12 months for immune checkpoint inhibitor (ICI) trials but requires further assessment for validation.

OBJECTIVE

To validate 6-month PFS and objective response rate (ORR) as estimators of 12-month OS in the ICI arms of randomized clinical trials (RCTs).

DATA SOURCES

Electronic databases (Medline, EMBASE, and the Cochrane Central Register of Controlled Trials) were searched for ICI RCTs published between January 2000 and June 2019.

STUDY SELECTION

Eligible studies were phase 2 and phase 3 ICI RCTs in advanced solid cancers that reported ORR, PFS, and OS. A total of 99 articles (from 60 studies) of 2502 articles were selected by consensus.

DATA EXTRACTION AND SYNTHESIS

Data were screened and extracted independently. Estimation models for 12-month OS and to assess correlation coefficient between end points were developed using linear regression. Data were extracted in July 2019, and analyses were conducted in September 2019. This study is reported following the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guideline.

MAIN OUTCOMES AND MEASURES

Validation of previously reported 6-month PFS and ORR estimation models for 12-month OS using contemporary RCTs. Calibration of 6-month PFS and ORR model-estimated vs observed 12-month OS in ICI arms were assessed by correlation coefficient (r) and weighted Brier scores. Secondary analyses were performed for subgroups (ie, ICI-only, ICI-combination, line of therapy, programmed cell death 1 ligand 1 selected, and unselected).

RESULTS

Data from 60 RCTs with 74 experimental ICI arms were used. The development data set included 25 arms from studies published January 2000 to January 2017. The estimation model for 12-month OS using 6-month PFS was: (1.06 × PFS6) + 0.16 + (0.04 × melanoma) - (0.03 × NSCLC) + (0 × other tumors), in which PFS6 indicates 6-month PFS and NSCLC indicates non-small cell lung cancer. The estimation model for 12-month OS using ORR was (0.15 × ORR) + 0.52 + (0 × melanoma) - (0.02 × NSCLC) - (0.01 × other tumors). A total of 49 arms from studies published after January 2017 to June 2019 formed the validation data set. When the models were applied on the validation data set, calibration between the 6-month PFS model estimated vs observed 12-month OS was good (r = 0.89; Brier score, 0.008), but poor for the ORR model (r = 0.47; Brier score, 0.03). Findings were similar across all subgroups.

CONCLUSIONS AND RELEVANCE

The findings of this study suggest that the estimation model using 6-month PFS could reliably estimate 12-month OS in ICI trials. This study could assist in better selection and prioritization of ICI agents for testing in RCTs based on phase 2 single-arm RCT results.

Systematic review of prediction models in relapsing remitting multiple sclerosis

The natural history of relapsing remitting multiple sclerosis (RRMS) is variable and prediction of individual prognosis challenging. The inability to reliably predict prognosis at diagnosis has important implications for informed decision making especially in relation to disease modifying therapies. We conducted a systematic review in order to collate, describe and assess the methodological quality of published prediction models in RRMS. We searched Medline, Embase and Web of Science. Two reviewers independently screened abstracts and full text for eligibility and assessed risk of bias. Studies reporting development or validation of prediction models for RRMS in adults were included. Data collection was guided by the checklist for critical appraisal and data extraction for systematic reviews (CHARMS) and applicability and methodological quality assessment by the prediction model risk of bias assessment tool (PROBAST). 30 studies were included in the review. Applicability was assessed as high risk of concern in 27 studies. Risk of bias was assessed as high for all studies. The single most frequently included predictor was baseline EDSS (n = 11). T2 Lesion volume or number and brain atrophy were each retained in seven studies. Five studies included external validation and none included impact analysis. Although a number of prediction models for RRMS have been reported, most are at high risk of bias and lack external validation and impact analysis, restricting their application to routine clinical practice.

A systematic review of the applications of artificial intelligence and machine learning in autoimmune diseases

Autoimmune diseases are chronic, multifactorial conditions. Through machine learning (ML), a branch of the wider field of artificial intelligence, it is possible to extract patterns within patient data, and exploit these patterns to predict patient outcomes for improved clinical management. Here, we surveyed the use of ML methods to address clinical problems in autoimmune disease. A systematic review was conducted using MEDLINE, embase and computers and applied sciences complete databases. Relevant papers included "machine learning" or "artificial intelligence" and the autoimmune diseases search term(s) in their title, abstract or key words. Exclusion criteria: studies not written in English, no real human patient data included, publication prior to 2001, studies that were not peer reviewed, non-autoimmune disease comorbidity research and review papers. 169 (of 702) studies met the criteria for inclusion. Support vector machines and random forests were the most popular ML methods used. ML models using data on multiple sclerosis, rheumatoid arthritis and inflammatory bowel disease were most common. A small proportion of studies (7.7% or 13/169) combined different data types in the modelling process. Cross-validation, combined with a separate testing set for more robust model evaluation occurred in 8.3% of papers (14/169). The field may benefit from adopting a best practice of validation, cross-validation and independent testing of ML models. Many models achieved good predictive results in simple scenarios (e.g. classification of cases and controls). Progression to more complex predictive models may be achievable in future through integration of multiple data types.

Novel genetic and epigenetic factors of importance for inter-individual differences in drug disposition, response and toxicity

Individuals differ substantially in their response to pharmacological treatment. Personalized medicine aspires to embrace these inter-individual differences and customize therapy by taking a wealth of patient-specific data into account. Pharmacogenomic constitutes a cornerstone of personalized medicine that provides therapeutic guidance based on the genomic profile of a given patient. Pharmacogenomics already has applications in the clinics, particularly in oncology, whereas future development in this area is needed in order to establish pharmacogenomic biomarkers as useful clinical tools. In this review we present an updated overview of current and emerging pharmacogenomic biomarkers in different therapeutic areas and critically discuss their potential to transform clinical care. Furthermore, we discuss opportunities of technological, methodological and institutional advances to improve biomarker discovery. We also summarize recent progress in our understanding of epigenetic effects on drug disposition and response, including a discussion of the only few pharmacogenomic biomarkers implemented into routine care. We anticipate, in part due to exciting rapid developments in Next Generation Sequencing technologies, machine learning methods and national biobanks, that the field will make great advances in the upcoming years towards unlocking the full potential of genomic data.

Common variation near IRF6 is associated with IFN-β-induced liver injury in multiple sclerosis

Multiple sclerosis (MS) is a disease of the central nervous system treated with disease-modifying therapies, including the biologic, interferon-β (IFN-β). Up to 60% of IFN-β-exposed MS patients develop abnormal biochemical liver test results^1,2, and 1 in 50 experiences drug-induced liver injury³. Since genomic variation contributes to other forms of drug-induced liver injury^4,5, we aimed to identify biomarkers of IFN-β-induced liver injury using a two-stage genome-wide association study. The rs2205986 variant, previously linked to differential expression of IRF6, surpassed genome-wide significance in the combined two-stage analysis (P = 2.3 × 10^-8, odds ratio = 8.3, 95% confidence interval = 3.6-19.2). Analysis of an independent cohort of IFN-β-treated MS patients identified via electronic medical records showed that rs2205986 was also associated with increased peak levels of aspartate aminotransferase (P = 7.6 × 10^-5) and alkaline phosphatase (P = 4.9 × 10^-4). We show that these findings may be applicable to predicting IFN-β-induced liver injury, offering insight into its safer use.

Dynamic Response Genes in CD4+ T Cells Reveal a Network of Interactive Proteins that Classifies Disease Activity in Multiple Sclerosis

Multiple sclerosis (MS) is a chronic inflammatory disease of the CNS and has a varying disease course as well as variable response to treatment. Biomarkers may therefore aid personalized treatment. We tested whether in vitro activation of MS patient-derived CD4+ T cells could reveal potential biomarkers. The dynamic gene expression response to activation was dysregulated in patient-derived CD4+ T cells. By integrating our findings with genome-wide association studies, we constructed a highly connected MS gene module, disclosing cell activation and chemotaxis as central components. Changes in several module genes were associated with differences in protein levels, which were measurable in cerebrospinal fluid and were used to classify patients from control individuals. In addition, these measurements could predict disease activity after 2 years and distinguish low and high responders to treatment in two additional, independent cohorts. While further validation is needed in larger cohorts prior to clinical implementation, we have uncovered a set of potentially promising biomarkers.

Myxovirus Resistance Protein A mRNA Expression Kinetics in Multiple Sclerosis Patients Treated with IFNβ

Introduction

Interferon-β (IFNß) is the first-line treatment for relapsing-remitting multiple sclerosis. Myxovirus resistance protein A (MxA) is a marker of IFNß bioactivity, which may be reduced by neutralizing antibodies (NAbs) against IFNß. The aim of the study was to analyze the kinetics of MxA mRNA expression during long-term IFNβ treatment and assess its predictive value.

Methods

A prospective, observational, open-label, non-randomized study was designed in multiple sclerosis patients starting IFNß treatment. MxA mRNA was assessed prior to initiation of IFNß therapy and every three months subsequently. NAbs were assessed every six months. Assessment of relapses was scheduled every three months during 24 months of follow up. The disease activity was correlated to the pretreatment baseline MxA mRNA value. In NAb negative patients, clinical status was correlated to MxA mRNA values.

Results

119 patients were consecutively enrolled and 107 were included in the final analysis. There was no correlation of MxA mRNA expression levels between baseline and month three. Using survival analysis, none of the selected baseline MxA mRNA cut off points allowed prediction of time to first relapse on the treatment. In NAb negative patients, mean MxA mRNA levels did not significantly differ in patients irrespective of relapse status.

Conclusion

Baseline MxA mRNA does not predict the response to IFNß treatment or the clinical status of the disease and the level of MxA mRNA does not correlate with disease activity in NAb negative patients.

Higher expression of IL-12Rβ2 is associated with lower risk of relapse in relapsing-remitting multiple sclerosis patients on interferon-β1b therapy during 3-year follow-up

Cytokines produced by helper T (Th)1 cells, Th17 and regulatory T cells (Treg) are involved in multiple sclerosis (MS) immunopathogenesis. Interferon (IFN)-β alters the numerous genes' expression, but how this alteration affects the treatment response is still elusive. We assessed relative gene expression of nineteen Th1/Th17/Treg-associated mediators in peripheral blood mononuclear cells and plasma levels of GM-CSF, IL-17A and IL-17F, in relapsing-remitting MS (RRMS) patients before IFN-β1b treatment initiation and at 6, 12, 24 and 36 months of therapy. All mRNA levels changed significantly during the IFN-β1b therapy. Higher IL-12Rβ2 mRNA levels were associated with lower risk of relapse. Despite recent reports regarding role of GM-CSF in MS, our study failed to demonstrate its significance as therapy response biomarker, both on the mRNA and protein level.