Accurate classification of secondary progression in multiple sclerosis using a decision tree

Spatiotemporal models of dengue epidemiology in the Philippines: Integrating remote sensing and interpretable machine learning

Previous dengue epidemiological analyses have been limited in spatiotemporal extent or covariate dimensions, the latter neglecting the multifactorial nature of dengue. These constraints, caused by rigid and traditional statistical tools which collapse amidst 'Big Data', prompt interpretable machine-learning (iML) approaches. Predicting dengue incidence and mortality in the Philippines, a data-limited yet high-burden country, the mlr3 universe of R packages was used to build and optimize ML models based on remotely sensed provincial and dekadal 3 NDVI and 9 rainfall features from 2016 to 2020. Between two tasks, models differ across four random forest-based learners and two clustering strategies. Among 16 candidates, rfsrc-year-case and ranger-year-death significantly perform best for predicting dengue incidence and mortality, respectively. Therefore, temporal clustering yields the best models, reflective of dengue seasonality. The two best models were subjected to tripartite global exploratory model analyses, which encompass model-agnostic post-hoc methods such as Permutation Feature Importance (PFI) and Accumulated Local Effects (ALE). PFI reveals that the models differ in their important explanatory aspect, rainfall for rfsrc-year-case and NDVI for ranger-year-death, among which long-term average (lta) features are most relevant. Trend-wise, ALE reveals that average incidence predictions are positively associated with 'Rain.lta', reflective of dengue cases peaking during the wet season. In contrast, those for mortality are negatively associated with 'NDVI.lta', reflective of urban spaces driving dengue-related deaths. By technologically addressing the challenges of the human-animal-ecosystem interface, this study adheres to the One Digital Health paradigm operationalized under Sustainable Development Goals (SDGs). Leveraging data digitization and predictive modeling for epidemiological research paves SDG 3, which prioritizes holistic health and well-being.

The risk of secondary progressive multiple sclerosis is geographically determined but modifiable

Geographical variations in the incidence and prevalence of multiple sclerosis have been reported globally. Latitude as a surrogate for exposure to ultraviolet radiation but also other lifestyle and environmental factors are regarded as drivers of this variation. No previous studies evaluated geographical variation in the risk of secondary progressive multiple sclerosis, an advanced form of multiple sclerosis that is characterized by steady accrual of irreversible disability. We evaluated differences in the risk of secondary progressive multiple sclerosis in relation to latitude and country of residence, modified by high-to-moderate efficacy immunotherapy in a geographically diverse cohort of patients with relapsing-remitting multiple sclerosis. The study included relapsing-remitting multiple sclerosis patients from the global MSBase registry with at least one recorded assessment of disability. Secondary progressive multiple sclerosis was identified as per clinician diagnosis. Sensitivity analyses used the operationalized definition of secondary progressive multiple sclerosis and the Swedish decision tree algorithm. A proportional hazards model was used to estimate the cumulative risk of secondary progressive multiple sclerosis by country of residence (latitude), adjusted for sex, age at disease onset, time from onset to relapsing-remitting phase, disability (Multiple Sclerosis Severity Score) and relapse activity at study inclusion, national multiple sclerosis prevalence, government health expenditure, and proportion of time treated with high-to-moderate efficacy disease-modifying therapy. Geographical variation in time from relapsing-remitting phase to secondary progressive phase of multiple sclerosis was modelled through a proportional hazards model with spatially correlated frailties. We included 51 126 patients (72% female) from 27 countries. The median survival time from relapsing-remitting phase to secondary progressive multiple sclerosis among all patients was 39 (95% confidence interval: 37 to 43) years. Higher latitude [median hazard ratio = 1.21, 95% credible interval (1.16, 1.26)], higher national multiple sclerosis prevalence [1.07 (1.03, 1.11)], male sex [1.30 (1.22, 1.39)], older age at onset [1.35 (1.30, 1.39)], higher disability [2.40 (2.34, 2.47)] and frequent relapses [1.18 (1.15, 1.21)] at inclusion were associated with increased hazard of secondary progressive multiple sclerosis. Higher proportion of time on high-to-moderate efficacy therapy substantially reduced the hazard of secondary progressive multiple sclerosis [0.76 (0.73, 0.79)] and reduced the effect of latitude [interaction: 0.95 (0.92, 0.99)]. At the country-level, patients in Oman, Tunisia, Iran and Canada had higher risks of secondary progressive multiple sclerosis relative to the other studied regions. Higher latitude of residence is associated with a higher probability of developing secondary progressive multiple sclerosis. High-to-moderate efficacy immunotherapy can mitigate some of this geographically co-determined risk.

The impact of healthcare systems on the clinical diagnosis and disease-modifying treatment usage in relapse-onset multiple sclerosis: a real-world perspective in five registries across Europe

Introduction

Prescribing guidance for disease-modifying treatment (DMT) in multiple sclerosis (MS) is centred on a clinical diagnosis of relapsing-remitting MS (RRMS). DMT prescription guidelines and monitoring vary across countries. Standardising the approach to diagnosis of disease course, for example, assigning RRMS or secondary progressive MS (SPMS) diagnoses, allows examination of the impact of health system characteristics on the stated clinical diagnosis and treatment access.

Methods

We analysed registry data from six cohorts in five countries (Czech Republic, Denmark, Germany, Sweden and United Kingdom) on patients with an initial diagnosis of RRMS. We standardised our approach utilising a pre-existing algorithm (DecisionTree, DT) to determine patient diagnoses of RRMS or secondary progressive MS (SPMS). We identified five global drivers of DMT prescribing: Provision, Availability, Funding, Monitoring and Audit, data were analysed against these concepts using meta-analysis and univariate meta-regression.

Results

In 64,235 patients, we found variations in DMT use between countries, with higher usage in RRMS and lower usage in SPMS, with correspondingly lower usage in the UK compared to other registers. Factors such as female gender (p = 0.041), increasing disability via Expanded Disability Status Scale (EDSS) score (p = 0.004), and the presence of monitoring (p = 0.029) in SPMS influenced the likelihood of receiving DMTs. Standardising the diagnosis revealed differences in reclassification rates from clinical RRMS to DT-SPMS, with Sweden having the lowest rate Sweden (Sweden 0.009, range: Denmark 0.103 - UK portal 0.311). Those with higher EDSS at index (p < 0.03) and female gender (p < 0.049) were more likely to be reclassified from RRMS to DT-SPMS. The study also explored the impact of diagnosis on DMT usage in clinical SPMS, finding that the prescribing environment and auditing practices affected access to treatment.

Discussion

This highlights the importance of a healthcare system's approach to verifying the clinical label of MS course in facilitating appropriate prescribing, with some flexibility allowed in uncertain cases to ensure continued access to treatment.

Health related quality of life in the domain of physical activity predicts confirmed disability progression in people with relapsing remitting multiple sclerosis

INTRODUCTION

The diagnosis of the progression phase of Multiple Sclerosis (MS) is still retrospective and based on the objectivation of clinical disability accumulation.

OBJECTIVES

To assess whether the Patient Reported Outcomes Measures (PROMs) scores predict the occurrence of disease progression within three years of follow-up.

METHODS

Observational prospective multicenter study. Stable Relapsing-Remitting MS (RRMS) patients were enrolled. At enrollment, patients completed the following PROMs: Beck Depression Inventory- II, The Treatment Satisfaction Questionnaire for Medications, Medical Outcomes Study Short Form 36- Item (SF36), Fatigue Severity Scale. EDSS was assessed at enrollment and three years later. The outcome measure was defined as the occurrence of confirmed disability progression (CDP) within three years of follow-up. Univariable and multivariable logistic regression models were performed to study the association between the final score of each test and the outcome.

RESULTS

SF36-Physical Functioning (SF36-PF) was the only independent variable associated with the outcome. The ROC curve analysis determined a score of 77.5 at SF36-PF as the cut-off point identifying patients experiencing CDP within three years of follow-up [AUC: 0.66 (95% CI: 0.56-0.75)].

CONCLUSIONS

RRMS patients scoring higher (>77.5) at SF36-PF subscale have a higher likelihood to experience CDP within the next three years.

Proportion and characteristics of secondary progressive multiple sclerosis in five European registries using objective classifiers

Background

To assign a course of secondary progressive multiple sclerosis (MS) (SPMS) may be difficult and the proportion of persons with SPMS varies between reports. An objective method for disease course classification may give a better estimation of the relative proportions of relapsing-remitting MS (RRMS) and SPMS and may identify situations where SPMS is under reported.

Materials and methods

Data were obtained for 61,900 MS patients from MS registries in the Czech Republic, Denmark, Germany, Sweden, and the United Kingdom (UK), including date of birth, sex, SP conversion year, visits with an Expanded Disability Status Scale (EDSS) score, MS onset and diagnosis date, relapses, and disease-modifying treatment (DMT) use. We included RRMS or SPMS patients with at least one visit between January 2017 and December 2019 if ≥ 18 years of age. We applied three objective methods: A set of SPMS clinical trial inclusion criteria ("EXPAND criteria") modified for a real-world evidence setting, a modified version of the MSBase algorithm, and a decision tree-based algorithm recently published.

Results

The clinically assigned proportion of SPMS varied from 8.7% (Czechia) to 34.3% (UK). Objective classifiers estimated the proportion of SPMS from 15.1% (Germany by the EXPAND criteria) to 58.0% (UK by the decision tree method). Due to different requirements of number of EDSS scores, classifiers varied in the proportion they were able to classify; from 18% (UK by the MSBase algorithm) to 100% (the decision tree algorithm for all registries). Objectively classified SPMS patients were older, converted to SPMS later, had higher EDSS at index date and higher EDSS at conversion. More objectively classified SPMS were on DMTs compared to the clinically assigned.

Conclusion

SPMS appears to be systematically underdiagnosed in MS registries. Reclassified patients were more commonly on DMTs.

The uncertainty period preceding the clinical defined SPMS diagnosis and the applicability of objective classifiers - A Danish single center study

BACKGROUND

The clinical transition from relapsing-remitting multiple sclerosis (RRMS) to secondary progressive MS (SPMS) is often related to a period of diagnostic uncertainty delaying diagnosis. With emerging treatment options for SPMS how to identify RRMS patients at risk of SPMS and when to assign a SPMS diagnosis has become a matter of growing clinical concern. This study aimed to determine the period of diagnostic uncertainty among Danish MS patients. Secondly, this study examined the performance of two objective classifiers in a longitudinal setting regarding their ability to shorten the period of diagnostic uncertainty.

METHODS

By using the Danish Multiple Sclerosis Registry, we identified all patients linked to Rigshospitalet with clinically assigned SPMS from 2010 to 2021. We reviewed all patient records and identified the first mentioned sign of progression (FMP). The time between the dates of FMP and clinically assigned SPMS was defined as the period of diagnostic uncertainty. Secondly, we applied two objective classifiers (the Karolinska Decision tree and the MSBase criteria) to generate suggested transition dates and compared them to the ones obtain from the patient records. Detailed descriptions of the population were made at all mentioned timepoints.

RESULTS

In total 138 patients were included. We found a median period of diagnostic uncertainty of 2.12 years. The objective classifiers generated a median suggested transition date 3.44 and 4.48 years earlier than the date of clinically assigned SPMS, but they only provided an earlier SPMS transition date in 50.72% and 55.80% of cases.

CONCLUSIONS

Our findings emphasize the uncertainty related to the transition from RRMS to SPMS illustrating the need of an improved diagnostic approach. Objective classifiers might have the potential to help reduce the period of diagnostic uncertainty in the future, but in their current form they do not perform satisfactorily enough to solve all difficulties related to detecting SPMS-transition.

Bone metastasis risk and prognosis assessment models for kidney cancer based on machine learning

Background

Bone metastasis is a common adverse event in kidney cancer, often resulting in poor survival. However, tools for predicting KCBM and assessing survival after KCBM have not performed well.

Methods

The study uses machine learning to build models for assessing kidney cancer bone metastasis risk, prognosis, and performance evaluation. We selected 71,414 kidney cancer patients from SEER database between 2010 and 2016. Additionally, 963 patients with kidney cancer from an independent medical center were chosen to validate the performance. In the next step, eight different machine learning methods were applied to develop KCBM diagnosis and prognosis models while the risk factors were identified from univariate and multivariate logistic regression and the prognosis factors were analyzed through Kaplan-Meier survival curve and Cox proportional hazards regression. The performance of the models was compared with current models, including the logistic regression model and the AJCC TNM staging model, applying receiver operating characteristics, decision curve analysis, and the calculation of accuracy and sensitivity in both internal and independent external cohorts.

Results

Our prognosis model achieved an AUC of 0.8269 (95%CI: 0.8083-0.8425) in the internal validation cohort and 0.9123 (95%CI: 0.8979-0.9261) in the external validation cohort. In addition, we tested the performance of the extreme gradient boosting model through decision curve analysis curve, Precision-Recall curve, and Brier score and two models exhibited excellent performance.

Conclusion

Our developed models can accurately predict the risk and prognosis of KCBM and contribute to helping improve decision-making.

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.

Prediction of high and low disease activity in early MS patients using multiple kernel learning identifies importance of lateral ventricle intensity

Background

Lack of easy-to-interpret disease activity prediction methods in early MS can lead to worse patient prognosis.

Objectives

Using machine learning (multiple kernel learning – MKL) models, we assessed the prognostic value of various clinical and MRI measures for disease activity.

Methods

Early MS patients ( n = 148) with at least two associated clinical and MRI visits were investigated. T2-weighted MRIs were cropped to contain mainly the lateral ventricles (LV). High disease activity was defined as surpassing NEDA-3 Criteria more than once per year. Clinical demographic, MRI-extracted image-derived phenotypes (IDP), and MRI data were used as inputs for separate kernels to predict future disease activity with MKL. Model performance was compared using bootstrapped effect size analysis of mean differences.

Results

A total of 681 visits were included, where 81 (55%) patients had high disease activity in a combined end point measure using all follow-up visits. MKL model discrimination performance was moderate (AUC ≥ 0.62); however, modelling with combined clinical and cropped LV kernels gave the highest prediction performance (AUC = 0.70).

Conclusions

MRIs contain valuable information on future disease activity, especially in and around the LV. MKL techniques for combining different data types can be used for the prediction of disease activity in a relatively small MS cohort.

Interrogating large multiple sclerosis registries and databases: what information can be gained?

PURPOSE OF REVIEW

Although substantial progress has been made in understanding the natural history of multiple sclerosis (MS) and the development of new therapies, many questions concerning disease behavior and therapeutics remain to be answered. Data generated from real-world observational studies, based on large MS registries and databases and analyzed with advanced statistical methods, are offering the scientific community answers to some of these questions that are otherwise difficult or impossible to address. This review focuses on observational studies published in the last 2 years designed to compare the effectiveness of escalation vs. induction treatment strategies, to assess the effectiveness of treatment in pediatric-onset and late-onset MS, and to identify the clinical phenotype of secondary progressive (SP)MS.

RECENT FINDINGS

The main findings originating from real-world studies suggest that MS patients who will qualify for high-efficacy disease-modifying therapies (DMTs) should be offered these as early as possible to prevent irreversible accumulation of neurological disability. Especially pediatric patients derive substantial benefits from early treatment. In patients with late-onset MS, sustained exposure to DMTs may result in more favorable outcomes. Data-driven definitions are more accurate in defining transition to SPMS than diagnosis based solely on neurologists' judgment.

SUMMARY

Patients, physicians, industry, and policy-makers have all benefited from real-world evidence based on registry data, in answering questions of diagnostics, choice of treatment, and timing of treatment decisions.

Inhibition of colony stimulating factor-1 receptor (CSF-1R) as a potential therapeutic strategy for neurodegenerative diseases: opportunities and challenges

Microglia are specialized dynamic immune cells in the central nervous system (CNS) that plays a crucial role in brain homeostasis and in disease states. Persistent neuroinflammation is considered a hallmark of many neurodegenerative diseases, including Alzheimer’s disease (AD), Parkinson's disease (PD), Huntington’s disease (HD), amyotrophic lateral sclerosis (ALS) and primary progressive multiple sclerosis (MS). Colony stimulating factor 1-receptor (CSF-1R) is predominantly expressed on microglia and its expression is significantly increased in neurodegenerative diseases. Cumulative findings have indicated that CSF-1R inhibitors can have beneficial effects in preclinical neurodegenerative disease models. Research using CSF-1R inhibitors has now been extended into non-human primates and humans. This review article summarizes the most recent advances using CSF-1R inhibitors in different neurodegenerative conditions including AD, PD, HD, ALS and MS. Potential challenges for translating these findings into clinical practice are presented.

Charting a global research strategy for progressive MS-An international progressive MS Alliance proposal

BACKGROUND

Progressive forms of multiple sclerosis (MS) affect more than 1 million individuals globally. Recent approvals of ocrelizumab for primary progressive MS and siponimod for active secondary progressive MS have opened the therapeutic door, though results from early trials of neuroprotective agents have been mixed. The recent introduction of the term 'active' secondary progressive MS into the therapeutic lexicon has introduced potential confusion to disease description and thereby clinical management.

OBJECTIVE

This paper reviews recent progress, highlights continued knowledge and proposes, on behalf of the International Progressive MS Alliance, a global research strategy for progressive MS.

METHODS

Literature searches of PubMed between 2015 and May, 2021 were conducted using the search terms "progressive multiple sclerosis", "primary progressive multiple sclerosis", "secondary progressive MS". Proposed strategies were developed through a series of in-person and virtual meetings of the International Progressive MS Alliance Scientific Steering Committee.

RESULTS

Sustaining and accelerating progress will require greater understanding of underlying mechanisms, identification of potential therapeutic targets, biomarker discovery and validation, and conduct of clinical trials with improved trial design. Encouraging developments in symptomatic and rehabilitative interventions are starting to address ongoing challenges experienced by people with progressive MS.

CONCLUSION

We need to manage these challenges and realise the opportunities in the context of a global research strategy, which will improve quality of life for people with progressive MS.