What question did this study set out to answer?

The study aims to identify clinical and serological signatures associated with anti-centromere antibody positivity in Sjögren’s syndrome using machine learning.

April 27, 2026Frontiers in Immunology0 citationsOpen Access

Identifying the clinical signature of anti-centromere antibody-positive Sjögren’s syndrome: a machine learning-based analysis of a multicenter cohort

WZWenlong ZhuZhejiang Chinese Medical University TFTing FangZhejiang Chinese Medical University XZXinchao ZhuZhejiang Chinese Medical University

Puntos clave

The study aims to identify clinical and serological signatures associated with anti-centromere antibody positivity in Sjögren’s syndrome using machine learning.
Analyzed clinical data from 616 patients with Sjögren’s syndrome, including 81 ACA-positive cases.
Partitioned data into training and validation sets (70:30 ratio) for robust model evaluation.
Employed six machine learning models, evaluated using AUC and SHAP analysis for clinical interpretability.
GBDT model achieved an AUC of 0.812 in training and 0.811 (95% CI: 0.699–0.906) in validation.
Highest sensitivity of 0.750 was observed in the validation test.
SHAP analysis identified key predictors including various serological markers, age, and Raynaud’s phenomenon.

Resumen

Objective We sought to leverage machine learning algorithms to identify the complex clinical and serological signature of anti-centromere antibody (ACA) positivity in Sjögren’s syndrome (SS) patients. Methods This multicenter study analyzed clinical data from a cohort of 616 patients diagnosed SS, comprising 81 ACA-positive and 535 ACA-negative cases. To ensure robust model development, we randomly partitioned the dataset into training and validation subsets in a 7:3 ratio. We implemented and compared six machine learning models after identifying optimal predictors using the LASSO regression. We mainly evaluate the performance of the model through the AUC and a series of comprehensive indicators. To ensure clinical interpretability, we also employed the SHAP analysis method to quantify the influence of each feature on the model’s outcome. Results Among the evaluated models, GBDT exhibited superior predictive efficacy. The model achieved an AUC value of 0.812 in the training set and maintained a robust AUC of 0.811 (95% CI: 0.699–0.906) in the validation cohort. At the same time, the model has the highest sensitivity (0.750 in the validation test). The SHAP analysis revealed that the top predictors influencing the ACA-positive profile included a series of serological markers (anti-SSA/Ro52, anti-SSA/Ro60, anti-AMA-M2, anti-SSB, and IgM), demographic factors (age), and Raynaud’s phenomenon (RP). Furthermore, SHAP interactions captured non-linear synergies, such as the predictive contribution of RP is significantly potentiated by advancing age, and the amplified predictive value of anti-AMA-M2 under lower IgM levels. Conclusion Our machine learning approach effectively structures and quantifies clinical and serological associations, capturing a complex predictive profile for the SS-ACA + subgroup. These findings highlight the value of ML in identifying non-linear patterns within clinical variables, providing a robust quantitative framework for future prospective evaluations.

Preguntar a la IA

Me gusta

Guardar

Ver artículo completo