Locus coeruleus abnormalities on neuromelanin-sensitive MRI in Parkinson's disease: a systematic review and meta-analysis.

Neuromelanin-sensitive magnetic resonance imaging (NM-MRI) enables in vivo assessment of locus coeruleus (LC) integrity, but findings in Parkinson's disease (PD) vary across cohorts and methods. We searched PubMed/MEDLINE, Embase, Web of Science Core Collection and Scopus from inception to 5 September 2026 and pooled independent PD-healthy control (HC) comparisons using random-effects models. Sixty-five reports were included, of which 17 contributed to at least one main pooled outcome. The primary LC-signal analysis included 14 independent cohorts (1,078 participants with PD and 543 HCs) and showed lower LC signal in PD (Hedges' g = - 0.95, 95% confidence interval (CI) - 1.32 to - 0.57; P = 0.00011). Between-study heterogeneity was substantial (I2 = 88.4%), and the 95% prediction interval crossed zero (- 2.28 to 0.38). The LC-volume analysis included five independent cohorts (215 participants with PD and 134 HCs) and likewise showed lower values in PD (g = - 0.87, 95% CI - 1.13 to - 0.62; P = 0.00071; I2 = 0%). The negative mean LC-signal effect was preserved across sensitivity analyses, but every specification-level prediction interval crossed zero. Of the 65 reports, 64 yielded 243 secondary-outcome records, of which 102 contained directly usable quantitative information; however, no additional outcome cluster simultaneously met the criteria for construct compatibility, cohort independence and numerical poolability. LC NM-MRI shows group-level differences in PD, but LC-signal differences vary across settings. These comparisons do not establish individual diagnostic accuracy. Clinical translation will require harmonized acquisition, analysis and reporting, together with independent, multicentre and longitudinal validation. This review was registered with PROSPERO (CRD420261498031).