Timely intervention, monitoring and education MATTERS in MS (TIME MATTERS in MS): Development of a globally applicable quality improvement tool

[摘要] Background: Lack of easy-to-interpret disease activity prediction methods in early MS can lead to worsepatient prognosis.Objectives: Using machine learning (multiple kernel learning – MKL) models, we assessed the prognosticvalue 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 activitywas defined as surpassing NEDA-3 Criteria more than once per year. Clinical demographic, MRI-extractedimage-derived phenotypes (IDP), and MRI data were used as inputs for separate kernels to predict futuredisease activity with MKL. Model performance was compared using bootstrapped effect size analysis ofmean 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 theLV. MKL techniques for combining different data types can be used for the prediction of disease activity ina relatively small MS cohort.

[关键词] Relapsing/remitting;MRI;disease activity prediction;multiple kernel learning;lateral ventricles;machine learning