Precision medicine in multiple sclerosis with diagnostic uncertainty

Multiple sclerosis (MS) is a debilitating disease mainly affecting young people and women.

The cause of MS is largely unknown, and there are no biomarkers for diagnosing MS early, known risk factors provide no information regarding the severity/progression of MS, and the pathophysiology of MS remains poorly understood.

To overcome these shortcomings, we will develop an AI-assisted diagnostic approach using clinical information from the electronic health record data (EHR) from MS patients in Sweden.

To address significant shortcomings of AI classifiers to include uncertainty in a given prediction, we will produce a methodology that enables a measure of confidence in each prediction.

To enable early disese prediction (forecasting), we will incorporate novel metabolic, lipidomic, and proteomic diagnostic and prognostic biomarkers.

This project capitalizes on well-defined case-control cohorts (>7,600 individuals), including a wealth of clinical, genetic, and environmental factors, two clinical trials in MS, collected in-depth metabolic and lipidomics data, and the developments in machine learning and AI methods and that have made it feasible to manage, process and interpret complex data.This project will be instrumental in realizing precision medicine in MS, with diagnostic uncertainty, thus enabling tailored interventions to the individual patient.