Manual federated simulation for multiple sclerosis integrating XGBoost algorithm with SHAP explanation

Clinical Snapshot

Key Findings

• P Value: Not reported

• Effect Size: Test set accuracy: 81.8%; Federated learning accuracy: 76.3% — representing a 5.5 percentage point degradation with federated simulation

• Primary Outcome: Prediction of MS progression in patients with Clinically Isolated Syndrome using XGBoost classification

• Nnt Or Sensitivity: Sensitivity, specificity, PPV, and NPV not reported. ROC-AUC: 88% on test set; 83.9% in federated simulation. Training AUC of 99% is implausibly high and indicative of overfitting. No NNT or likelihood ratios calculable from available data.

• Confidence Interval: Not reported for any metric — a critical methodological omission

Clinical Application

Clinical deployment is not feasible at this stage. The model requires: (1) external validation on an independent, well-characterised CIS cohort; (2) prospective evaluation against standard clinical decision pathways; (3) real-world federated implementation across actual clinical sites; (4) regulatory review. The computational infrastructure for XGBoost deployment in clinical settings is available, but the evidence base is insufficient. MS affects approximately 33,000 Australians, with incidence rising. Early CIS-to-MS conversion prediction has direct relevance to PBS-listed disease-modifying therapy (DMT) initiation decisions — including interferon beta, glatiramer acetate, and high-efficacy agents such as natalizumab and ocrelizumab. The Australian Privacy Act 1988 and My Health Record framework make federated learning architectures particularly relevant for multi-site data collaboration without centralising sensitive neurological data. However, TGA approval and RACGP/MS Research Australia endorsement would require prospective Australian cohort validation before any clinical integration. The study's Egyptian institutional origin means direct applicability to Australian MS demographics (including Indigenous Australians with distinct MS risk profiles) cannot be assumed. Theoretically applicable to adults with a first demyelinating event (CIS) being evaluated for MS conversion risk. However, the uncharacterised study population, absence of external validation, and in silico-only federated testing preclude any current clinical application.

Abstract

Multiple sclerosis (MS) is a chronic autoimmune disorder of the central nervous system, underscoring the importance of early and accurate diagnosis. In this study investigates the predictive modelling of MS progression in patients with Clinically Isolated Syndrome (CIS), privacy-preserving for a federated and explainable Machine Learning (ML) framework. To address missing data while preserving inter-feature dependencies, Multivariate Imputation by Chained Equations (MICE) with iterative imputers was employed. Classification was performed using the Extreme Gradient Boosting (XGBoost) algorithm. Model interpretability was developed through Explainable Artificial Intelligence (XAI) techniques, specifically Shapley Additive Explanations (SHAP). To ensure data confidentiality and simulate decentralized clinical environments, an in silico federated learning framework was applied. Experimental results demonstrated strong predictive performance, achieving 96.7% accuracy and 99% ROC-AUC during training, 92.5% accuracy in validation, and 81.8% accuracy with an AUC of 88% on the test set. For the Federated Learning (FL) simulation, the model maintained competitive performance, yielding an accuracy of 76.3% and an AUC of 83.9%. The proposed approach supports early diagnosis, enhances clinical trust through interpretability, and promotes secure data collaboration, thereby contributing to more informed and transparent clinical decision-making and improved patient care.

References

1. 1.Ghazy, H. E., Ali, Z. H., & Medhat, T. (2023). Manual federated simulation for multiple sclerosis integrating XGBoost algorithm with SHAP explanation. IEEE Access. https://doi.org/10.1109/ACCESS.2023.3312343