biostudies-literatureUnknown8(1)Pirmani AEarly prediction of disability progression in multiple sclerosis (MS) remains challenging despite its critical importance for therapeutic decision-making. We present the first systematic evaluation of personalized federated learning (PFL) for 2-year MS disability progression prediction, leveraging multi-center real-world data from over 26,000 patients. While conventional federated learning (FL) enables privacy-aware collaborative modeling, it remains vulnerable to institutional data heterogeneity. PFL overcomes this challenge by adapting shared models to local data distributions without compromising privacy. We evaluated two personalization strategies: a novel AdaptiveDualBranchNet architecture with selective parameter sharing, and personalized fine-tuning of global models, benchmarked agaNPJ digital medicine478https://www.ebi.ac.uk/biostudies/studies/S-EPMC12289948biostudies-literaturePersonalized federated learning for predicting disability progression in multiple sclerosis using real-world routine clinical data.PMC12289948Alroughani RSinghal BKalincik TKuhle JJohn NCauchi MGuimaraes JGarber JDe Brouwer EPeeters LMWillekens BRajda CHughes STomassini VMaimone DSas ALugaresi ATaylor BVan Wijmeersch BAl-Asmi ASirbu CAMcCombe PCastillo Trivino THavrdova EKPassemiers AOzakbas Sde Gans KCarroll WMPirmani ABuzzard KHardy TALalive PPerez-Sempere AMoreau YRamanathan SVan Pesch VRozsa CAmato MPSolaro CGrammond PShuey NPrat ABoz CHorakova DOldenhof MCsepany TMason DBlanco YFabis-Pedrini MDecoo DPatti FCartechini EFaes AArany AGouider RGerlach OLechner-Scott JSimu MRoos ILaureys GAltintas ASpitaleri DKermode AGAl-Harbi TPrevost JGray OVan der Walt AButzkueven HKhoury SJSoysal ASanchez-Menoyo JLWeinstock-Guttman BAmpapa RfalsePersonalized federated learning for predicting disability progression in multiple sclerosis using real-world routine clinical data.Early prediction of disability progression in multiple sclerosis (MS) remains challenging despite its critical importance for therapeutic decision-making. We present the first systematic evaluation of personalized federated learning (PFL) for 2-year MS disability progression prediction, leveraging multi-center real-world data from over 26,000 patients. While conventional federated learning (FL) enables privacy-aware collaborative modeling, it remains vulnerable to institutional data heterogeneity. PFL overcomes this challenge by adapting shared models to local data distributions without compromising privacy. We evaluated two personalization strategies: a novel AdaptiveDualBranchNet architecture with selective parameter sharing, and personalized fine-tuning of global models, benchmarked aga2025-01-01T00:00:00Z2025 Jul2026-03-27T17:04:07.93Z2025-08-28T03:08:54.013ZS-EPMC122899484070760110.1038/s41746-025-01788-8