New study uses artificial intelligence to uncover Sjögren’s disease subtypes
Findings highlight differences in organ involvement and immune features
Written by |
Primary Sjögren’s disease may be divided into two main clinical subtypes, according to a recent study that used artificial intelligence.
“Our study suggests that primary Sjögren’s syndrome consists of distinct clinical subtypes with different organ involvement patterns and immunological features,” the researchers wrote. “The data-driven classification system complements traditional clinical classifications and offers new perspectives on disease heterogeneity, potentially laying the groundwork for more individualized patient management strategies.”
The study, “Data-driven classification of primary Sjögren’s syndrome: From cluster analysis to clinical immune phenotypes and predictive biomarkers,” was published in the Journal of Translational Autoimmunity.
Understanding how Sjögren’s disease affects the body
Sjögren’s disease is an autoimmune disorder marked by inflammation in the tear and/or salivary glands. However, the disease can affect many parts of the body and cause a wide range of symptoms.
In primary Sjögren’s disease, the condition occurs on its own. This differs from secondary Sjögren’s, which develops alongside another autoimmune disease.
Because primary Sjögren’s can look very different from person to person, researchers have long suspected that the disease may include different clinical subtypes. Until now, however, these subtypes have not been clearly defined.
In this study, researchers in China used machine learning, a form of artificial intelligence that analyzes large datasets, to identify possible Sjögren’s disease subtypes. Their analysis included clinical data from more than 1,000 people with primary Sjögren’s disease who were treated at a hospital in China.
AI analysis reveals two major disease subtypes
The analysis showed that people with primary Sjögren’s disease tend to fall into one of two major clinical subtypes. In the first subtype, all patients had involvement of the musculoskeletal system, meaning muscles and joints, while involvement of other organs was relatively uncommon. By contrast, people in the second subtype showed a wider range of organ involvement, with the tear and salivary glands most often affected.
People in the first subtype also tended to have higher levels of markers linked to body-wide inflammation, particularly rheumatoid factor (RF) and erythrocyte sedimentation rate (ESR). By comparison, patients in the second subtype had higher levels of IgG, a type of antibody.
Based on these patterns, the researchers referred to the first group as a “multisystem inflammatory subtype” and the second as a “glandular-limited high immunoglobulin subtype.”
“The multisystem inflammatory subtype is characterized by widespread musculoskeletal involvement and significantly elevated inflammatory markers,” they wrote. “In contrast, the glandular-limited high immunoglobulin subtype exhibits more “classic” features of Sjögren’s syndrome, with prominent glandular, cardiovascular, and pulmonary involvement, along with elevated IgG levels.”
Disease severity and immune markers differ between subtypes
The researchers noted that patients in the multisystem inflammatory subtype tended to have more severe disease based on traditional clinical measures. The data also suggested that patients who were male were more likely to fall into this more severe subtype, consistent with earlier studies showing that men with Sjögren’s disease often have more severe systemic involvement than women.
In further analyses, the researchers compared laboratory test results to identify differences between the two groups, with the goal of developing a way to distinguish the subtypes using routine lab measures. A prediction model that combined subtype-associated laboratory markers with patients’ biological sex was able to distinguish between the two disease subtypes.
Another key finding was that high levels of the inflammatory marker RF were closely associated with involvement of the musculoskeletal system. This suggests that RF “may become an important marker for future precision subtyping and targeted therapy,” the researchers wrote.
Overall, the researchers noted that this type of classification system could help improve understanding of the underlying biology of Sjögren’s disease and may eventually support more personalized approaches to care. They stressed, however, that further studies are needed to validate these subtypes and determine how useful they may be in real-world clinical settings.
