Harmonized magnetic resonance imaging protocols for multiple sclerosis: HARMONY-MS, a pilot project

Introduction: Magnetic Resonance Imaging (MRI) is central to multiple sclerosis (MS) management, yet practice often diverges from guidelines. Large-scale evaluations of harmonized referral and acquisition protocols remain limited. Methods: We conducted a retrospective observational study including all central nervous system (CNS) MRI examinations requested by MS centers within the AUSL Romagna healthcare system (Italy; catchment population ~ 1.3 million) between January 2023 and June 2025. MRI volumes were assessed using Poisson regression and interrupted time-series (ITS) models. Scanner time and gadolinium expenditure were analyzed with Gamma regression. Clinical outcomes included annualized relapse rate (ARR; Poisson regression), MRI inflammatory activity (McNemar’s test, Wilcoxon test), and NEDA-3 (McNemar’s test). Disease-modifying therapy (DMT) distributions were compared using χ2 tests, and time to initiation of high-efficacy therapy (HET) was assessed by Kaplan–Meier analysis, log-rank test, and Cox proportional-hazards regression. Results: Among 143,592 MRI examinations, 2734 were MS-related. Contrast-enhanced MRI decreased by > 90% in Jan–Jun 2025 vs Jan–Jun 2023 (IRR 0.07, 95% CI 0.04–0.12, p < 0.0001). Combined brain + whole spinal gadolinium studies similarly declined by > 90% (IRR 0.08, 95% CI 0.06–0.11, p < 0.0001). Total MRI volume decreased by 86% (IRR 0.14, 95% CI 0.12–0.16, p < 0.0001). Interrupted time-series confirmed significant post-intervention slopes for contrast-enhanced MRI (–4.4 exams/month, p = 0.021) and total MRI examinations (–16 exams/month, p < 0.001). Total time per semester declined from 274 to 29 h, while total gadolinium expenditure per semester declined from €15 730 to €1 678. ARR decreased from 0.048 to 0.018 (IRR 0.37, 95% CI 0.21–0.66, p < 0.001). MRI inflammatory activity remained stable (13.8% vs 13.6%; McNemar χ2 = 0.005, p = 0.90; Wilcoxon V = 14,634, p = 0.50), while NEDA-3 increased from 57.4% to 70.1% (McNemar χ2 = 35, p < 0.0001). Among newly diagnosed patients, first-line high-efficacy DMT use rose from 47.7% to 75.4% (χ2 = 5.5168, p = 0.0188), with shorter time to initiation (HR 1.84, 95% CI 1.09–3.08, p = 0.022). Conclusions: HARMONY-MS is the first large-scale program to show reduction in contrast use, unnecessary imaging, and costs, while improving scanner efficiency. These findings provide a scalable model for guideline-concordant MS imaging, compatible with stable clinical outcomes and supporting a shift toward earlier and more frequent use of high-efficacy DMTs.