Systemic lupus erythematosus phenotypes formed from machine learning with a specific focus on cognitive impairment

Area of Science:

• Rheumatology
• Neuroscience
• Psychiatry

Background:

• Systemic lupus erythematosus (SLE) is a complex autoimmune disease with diverse clinical manifestations.
• Cognitive impairment (CI) is a significant but often under-characterized aspect of SLE.
• Understanding distinct patient phenotypes is crucial for personalized management.

Purpose of the Study:

• To phenotype SLE patients based on symptom burden, including disease damage and patient-reported outcomes.
• To specifically investigate objective and subjective cognitive function in relation to SLE characteristics.
• To identify distinct patient subtypes using a data-driven approach.

Main Methods:

• Adult SLE patients (18-65 years) underwent objective cognitive assessment using the ACR Neuropsychological Battery (ACR-NB).
• Data collected included demographics, clinical variables, disease burden/activity, HRQoL, depression, anxiety, fatigue, and perceived cognitive deficits.
• Similarity Network Fusion (SNF) was employed to identify patient subtypes, with differences analyzed using Kruskal-Wallis and chi-squared tests.

Main Results:

• Two distinct subtypes (A and B) were identified among 238 SLE patients, differing in cognitive function, disease burden, HRQoL, and psychiatric measures.
• Subtype A demonstrated significantly poorer performance on objective cognitive tests compared to Subtype B.
• Subtype A also reported higher levels of subjective cognitive difficulties, greater disease burden, and worse HRQoL and psychiatric outcomes.

Conclusions:

• Cognitive impairment in SLE is complex and presents as multifactorial, individual phenotypes.
• Increased disease burden in SLE is associated with poorer cognitive functioning, both subjectively reported and objectively measured.
• Phenotyping SLE can enhance the understanding of CI trajectories and inform personalized treatment strategies.