Predicting thyroid involvement in primary Sjögren's syndrome: development and validation of a predictive nomogram.

Introduction: Patients with Primary Sjögren's syndrome (pSS) are at a higher risk of thyroid disorders than the general population. This retrospective analysis of 202 patients with pSS was conducted to uncover risk factors associated with thyroid involvement and to create a predictive model for this condition. Methods: We analyzed 202 patients with pSS from Guang'anmen Hospital, China Academy of Chinese Medical Sciences, with 105 cases of thyroid involvement and 97 without. The Least Absolute Shrinkage and Selection Operator method was used to identify key variables for our risk model. These variables were then subjected to multivariate logistic regression to develop the model. The accuracy of the model was assessed through the C-index, receiver operating characteristic curves, calibration plots, and decision curve analysis, with internal validation via bootstrapping. Results: High-sensitivity C-reactive protein (HCRP), pulmonary disease, pharyngeal dryness, forgetfulness, night sweats, hyperuricemia, nasal dryness, anxiety, Ro52, and aspartate aminotransferase (AST) were incorporated into the nomogram. The model showed robust discrimination and calibration abilities. Decision curve analysis indicated the clinical utility of our nomogram in intervening on the probability thresholds of thyroid disease. Conclusion: By integrating HCRP, pulmonary disease, pharyngeal dryness, forgetfulness, night sweats, hyperuricemia, nasal dryness, anxiety, Ro52, and AST, our thyroid risk nomogram can predict the risk of thyroid involvement in patients with pSS, aiding in more informed treatment strategies.