SGUS Score and Hyperglobulinemia for an Age-Stratified Prediction Model of Positive Labial Focus Score in Sjögren's Disease

Objective To develop and validate a non-invasive predictive model for labial focus score (FS) ≥1 in patients suspected of having Sjögren's disease (SJD), by integrating salivary gland ultrasound (SGUS) findings with clinical parameters. Methods A novel semi-quantitative SGUS scoring system was applied to 449 patients prospectively recruited. Multivariable logistic regression analysis was conducted to identify independent predictors of FS ≥1, which were subsequently incorporated into a risk stratification matrices model. The model underwent both internal (n=139) and external validation (n=57). Results The derivation cohort included 449 patients (mean age 44.0 years, 94.7% female), with 284 (63.3%) showing labial FS≥1. Patients with FS≥1 were older and had higher total SGUS-scores and hyperglobulinemia. Multivariable logistic regression identified the SGUS-score (odds ratio [OR] 1.44), age (OR 1.51), and hyperglobulinemia (OR 1.91) as independent outcome predictors. This led to an age-stratified prediction model categorizing patients into high-, moderate-, and low-risk groups across three age brackets. This model showed strong predictive value for labial FS≥1 in both high-risk and low-risk populations (AUC: 0.91, 95%CI: 0.87-0.96, P<0.001), with specificity of 81%, sensitivity of 95%, positive predictive value (PPV) of 92%, and negative predictive value (NPV) of 87%. The model performed well in both internal validation (AUC: 0.90) and external validation (AUC: 0.83). Conclusion Our SGUS-based predictive model provides an exploratory, hypothesis-generating tool for guiding LSGB decisions in SJD by integrating imaging with clinical parameters, enabling personalized risk stratification. It effectively identifies high-risk patients needing biopsy and low-risk patients avoiding unnecessary procedures, offering a significant diagnostic advance.