Abstract
Background: Timely initiation of effective therapy is crucial for preventing disability in multiple sclerosis (MS); however, treatment response varies greatly among patients. Comprehensive predictive models of treatment outcomes are lacking.
Objective: To evaluate a large number of demographic, clinical and paraclinical predictors of individual response to 11 disease modifying therapies and to incorporate these into models informing clinical practice.
Methods: Longitudinal data from MSBase, a large global cohort study, were utilised to identify predictors of on-treatment relapses, disability progression and regression, using univariable survival models. The predictors comprised 51 variables. Multivariable survival models were used to design individual predictive algorithms. Dimensionality of the models was controlled with principal component analysis. Accuracy of the resulting models was tested in a training cohort. External validity was established using a validation cohort.
Results: In the training cohort (n=8513), the most prominent predictors of treatment response comprised age, MS duration, MS course, previous relapse activity, disability, predominant relapse phenotype and previous therapy. Importantly, the magnitude and direction of the associations varied among therapies and disease outcomes. Higher probability of disability progression during treatment with injectable therapies was predominantly associated with a greater disability at treatment start and the previous MS therapy. For fingolimod, natalizumab or mitoxantrone, it was mainly associated with lower pre-treatment relapse activity. The probability of disability regression was predominantly associated with pre-baseline disability, therapy and relapse activity. In contrast, incidence of relapses was associated with pre-treatment
relapse activity, age and relapsing MS course, with the strength of these associations varying among therapies. Accuracy and external validity (n=1196) of the resulting predictive models was high (>80%) for relapse incidence over the initial 2 years and disability outcomes, and moderate (>50%) for relapse incidence in years 3-4.
Conclusion: Demographic, clinical and paraclinical information enables estimation of future individual response to disease modifying therapies. The resulting models constitute a framework that enables incorporation of multiple biomarkers and will be implemented in a web-based tool with the aim of supporting clinical practice.
Objective: To evaluate a large number of demographic, clinical and paraclinical predictors of individual response to 11 disease modifying therapies and to incorporate these into models informing clinical practice.
Methods: Longitudinal data from MSBase, a large global cohort study, were utilised to identify predictors of on-treatment relapses, disability progression and regression, using univariable survival models. The predictors comprised 51 variables. Multivariable survival models were used to design individual predictive algorithms. Dimensionality of the models was controlled with principal component analysis. Accuracy of the resulting models was tested in a training cohort. External validity was established using a validation cohort.
Results: In the training cohort (n=8513), the most prominent predictors of treatment response comprised age, MS duration, MS course, previous relapse activity, disability, predominant relapse phenotype and previous therapy. Importantly, the magnitude and direction of the associations varied among therapies and disease outcomes. Higher probability of disability progression during treatment with injectable therapies was predominantly associated with a greater disability at treatment start and the previous MS therapy. For fingolimod, natalizumab or mitoxantrone, it was mainly associated with lower pre-treatment relapse activity. The probability of disability regression was predominantly associated with pre-baseline disability, therapy and relapse activity. In contrast, incidence of relapses was associated with pre-treatment
relapse activity, age and relapsing MS course, with the strength of these associations varying among therapies. Accuracy and external validity (n=1196) of the resulting predictive models was high (>80%) for relapse incidence over the initial 2 years and disability outcomes, and moderate (>50%) for relapse incidence in years 3-4.
Conclusion: Demographic, clinical and paraclinical information enables estimation of future individual response to disease modifying therapies. The resulting models constitute a framework that enables incorporation of multiple biomarkers and will be implemented in a web-based tool with the aim of supporting clinical practice.
| Original language | English |
|---|---|
| Article number | P729 |
| Pages (from-to) | 358-360 |
| Number of pages | 3 |
| Journal | Multiple Sclerosis |
| Volume | 22 |
| Issue number | Supp: 3 |
| DOIs | |
| Publication status | Published - Sept 2016 |
Keywords
- Multiple Sclerosis
- Therapy modification
- Patient outcomes