New Open Benchmark for MS Blood-Based Classification.

Diagnosing Multiple Sclerosis (MS) typically involves clinical assessment, MRI, and laboratory evidence. While blood RNA expression data may contain immune signals related to the disease, a blood RNA classifier is not intended to replace a clinical diagnosis. To advance research in this area, a new open benchmark called MS-MLB (Multiple Sclerosis Machine Learning Benchmark) has been introduced. MS-MLB provides a reproducible framework for machine learning-based MS classification using whole blood RNA expression data. It utilizes the publicly available GSE17048 cohort, transforming it into a task of classifying MS versus healthy controls. The benchmark evaluates various algorithms through a shared, leakage-controlled pipeline that researchers can easily rerun. The evaluation process includes nested cross-validation, an untouched stratified holdout set, bootstrap confidence intervals, ROC and precision-recall analysis, and calibration measurement. An exploratory MS Research Score is also provided. In the initial benchmark summary, Gradient Boosting emerged as the top performer on the holdout set, achieving an MS Research Score of 93.83, an AUC-ROC of 0.989, and high sensitivity and specificity. This benchmark is unique in its focus on MS versus healthy control classification from GSE17048 whole blood RNA data and offers a documented external model submission pathway. It's important to note that the score is for research comparison only and has not been clinically validated.