Related Experiment Video
Updated: Jun 26, 2026
Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data
Published on: June 26, 2013
Neuropsychiatric systemic lupus erythematosus subtypes identified by unsupervised clustering: A single-center cohort
Weiting Fang1, Jiuliang Zhao1,2,3,4, Dong Xu1,2,3,4
1Department of Rheumatology and Clinical Immunology, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing, China.
Background And Objectives:
Neuropsychiatric (NP) manifestations are common, heterogeneous and severe in systemic lupus erythematosus (SLE) patients, with attribution to SLE remaining diagnostically challenging. Traditional classification focuses on clinical syndromes, overlooking neuropsychiatric systemic lupus erythematosus (NPSLE) immunological heterogeneity. To address the heterogeneity of NPSLE, this study aimed to delineate distinct disease subgroups by clustering patients based on their antibody profiles. These subgroups were then evaluated for diferences in clinical presentation and prognosis, to better characterize disease subsets and support individualized approaches to diagnosis and management.
Methods:
This retrospective single-center study included hospitalized SLE patients with NP manifestations, collecting demographic, clinical and laboratory data. Patients were classified as NPSLE or non-NPSLE by clinical judgment after excluding alternative causes. Hierarchical cluster analysis explored autoantibody-clinical feature associations.
Results:
Among the 167 patients analyzed, 152 had NP manifestations attributed to SLE. Central nervous system (CNS) involvement was predominant (89.1%), with seizures, cerebrovascular disease, acute confusional state (ACS), and demyelinating syndrome being most prevalent manifestations. Hierarchical clustering of 152 NPSLE patients identified two subgroups: Cluster 1 (23.7%) demonstrated cerebrovascular injury as the predominant manifestation, with higher positivity rates of antiphospholipid antibodies (APLs) (P < 0.01) and a higher incidence of cerebrovascular disease (P < 0.01). Cluster 2 (76.3%) showed immune-mediated inflammatory profile, with higher positivity of anti-SSA (P < 0.01), antidsDNA (P < 0.05), and anti-RiboP antibodies (P < 0.05). Neurological involvement predominantly manifesting as ACS (P < 0.05), accompanied by a higher frequency of fever and joint involvement.
Conclusions:
In this study, NPSLE exhibited distinct serological profiles and segregated into two immunologically defined clusters, reflecting its clinical and biological heterogeneity, and suggesting that immunological profiling may enhance precise classification and personalized management of affected patients.
