首都医科大学

Plasma neurofilament light for early differentiation of multiple system atrophy from Parkinson disease Zhenwei Yu,a,∗∗,m Yuanchu Zheng,b,m Wenyi Kou,c,m Huihui Cai,b Siming Li,b Jiayi Wu,b Chen Yang,b Bingxu Zhu,b Genliang Liu,b Yuke Wang,d,e Hao Tian,f Xinning Zhang,g Ying Zhou,h Xuedong Liu,i Ying Chang,j Zhigang Chen,k,∗∗∗∗ Wei Mao,d,e,∗∗∗ and Tao Fengb,l,∗ aDepartment of Pathophysiology, Beijing Neurosurgical Institute, Capital Medical University, Beijing, China bDepartment of Neurology, Centre for Movement Disorders, Beijing Tiantan Hospital, Capital Medical University, Beijing, China cDepartment of Neurology, Beijing Hospital, National centre of Gerontology; Institute of Geriatric Medicine, Chinese Academy of Medical Sciences, Beijing, China dDepartment of Neurology, Xuanwu Hospital, Capital Medical University, Beijing, China eNational centre for Neurological Disorders, Beijing, China fBeijing University of Chinese Medicine, 100029, Beijing, China gXi’an Hospital of Traditional Chinese Medicine, 710021, Xi’an, China hDepartment of Traditional Chinese Medicine, The Second Affiliated Hospital of Army Medical University, Chongqing, 400030, China iDepartment of Neurology, Xijing Hospital of Air Force Military Medical University, Xi’an, China jDepartment of Neurology, China-Japan Union Hospital of Jilin University, Changchun, China kNeurology Department One, Dongfang Hospital, Beijing University of Chinese Medicine, No. 6, Fangxingyuan Community, Fangzhuang, Fengtai District, Beijing, 100078, China lChina National Clinical Research centre for Neurological Diseases, Beijing, China Summary Background Early differentiation of multiple system atrophy (MSA) from Parkinson’s disease (PD) remains difficult, particularly within two years of symptom onset, when diagnostic uncertainty has major implications for prognosis, referral, and trial enrolment. Because MSA is rare, previous biomarker studies have often been limited by relatively small samples and restricted multicentre validation. We aimed to determine the diagnostic performance of plasma neurofilament light chain (NfL) for differentiating MSA from PD and to assess the incremental value of glial fibrillary acidic protein (GFAP), total tau (t-tau), and phosphorylated tau at threonine 217 (p-tau217). Methods In this multicentre cross-sectional diagnostic study, participants were enrolled from five movement- disorder referral centres in China between Jan 1, 2018, and June 30, 2024, and were divided by enrolment period into discovery and temporally separated validation datasets. Plasma NfL, GFAP, t-tau, and p-tau217 were measured using light-initiated chemiluminescence assays. Group comparisons used age- and sex-adjusted models, and discrimination was assessed using receiver-operating-characteristic analysis with sensitivity, specificity, predictive values, and robustness analyses. Findings The analysis included 2408 participants: 782 (32.5%) with PD, 796 (33.1%) with MSA, and 830 (34.5%) healthy controls. NfL was the best single biomarker for differentiating MSA from PD in the discovery dataset (AUC 0.920, 95% CI 0.903–0.936). The discovery-derived cutoff of 41.3 pg/mL yielded sensitivity of 90.5% and specificity of 83.4% in the discovery dataset, and sensitivity of 86.7% and specificity of 85.5% in the validation dataset (AUC 0.924, 95% CI 0.898–0.948). In the early-stage subgroup, NfL retained strong performance (AUC 0.943, 95% CI 0.915–0.966). Integrated multimarker models provided limited incremental discrimination over NfL alone. Interpretation Plasma NfL may support the differentiation between clinically diagnosed MSA and PD, including early in the disease course. The limited added value of multimarker panels supports a simpler and more imme- diately translatable NfL-first strategy for diagnostically uncertain parkinsonism. *Corresponding author. Department of Neurology, Centre for Movement Disorders, Beijing Tiantan Hospital, Capital Medical University, B