The contribution of salivary gland ultrasound to juvenile Sjögren disease classification: performance of a weighted scoring model.

Abstract

This study evaluated the classification utility of incorporating parotid gland US for juvenile Sjögren's disease (jSjD) by testing its role in a weighted scoring (WS) model and comparing it with the standard 2016 ACR/EULAR classic model (CM). Seventeen patients were retrospectively evaluated. All underwent B-mode parotid US at diagnosis, scored via the OMERACT system. Classification scoring items included positivity of anti-SSA, Schirmer test, labial salivary gland biopsy, US (OMERACT score ≥1). The WS (0-8) was calculated according to positivity of each parameter: (anti-SSA × 3) + (biopsy × 3) + (Schirmer × 1) + (US × 1). ROC analysis assessed the WS for predicting ACR/EULAR classification (≥4) and US alone for identifying patients fulfilling ≥2 conventional items (composite score ≥2). The WS was compared with the CM, which excluded US, based on classification performance parameters. The WS yielded an acceptable AUC of 0.72 for predicting fulfilment of the ACR/EULAR classification criteria. US alone showed strong concordance with composite criteria (AUC: 0.84). Although the WS and CM had comparable overall performance (AUC: 0.85 vs 0.87; P = 0.76), the WS achieved higher specificity (92% vs 77%) and positive predictive value (PPV; 86% vs 73%). Although the inclusion of US in the classification model did not significantly improve overall discrimination, it enhanced specificity and PPV, particularly in borderline cases. Given its concordance with conventional criteria and non-invasive nature, US may be a practical candidate for future jSjD classification systems, warranting prospective validation in larger cohorts.