Ultrasonography as an additional item in the American College of Rheumatology classification of Sjögren's syndrome.

Abstract

In this study we evaluated US as an additional classification item in the ACR classification of SS. Of 581 patients classified as either SS (n = 364) or non-SS (n = 217) based on the minimum requirements of the American-European Consensus Group (AECG) classification, 184 patients (102 SS and 82 non-SS) who had scored two or more positive or two or more negative results according to the ACR criteria were selected. The AECG classification was used as the gold standard. A parotid and/or submandibular gland that was assigned a score ≥G1 was designated as SS positive. We evaluated US alone or with varying combinations of the ACR classification items in the diagnosis of SS. The ACR criteria diagnosed the 184 patients with 91% sensitivity, 90% specificity and 91% accuracy. US alone diagnosed the 184 ACR patients with 79% sensitivity, 90% specificity and 83% accuracy, which was comparable to the results of US diagnosis in the AECG cohort (81%, 86% and 83%, respectively). Incorporating the US criteria as an alternative to one of the three ACR classification items achieved 89-91% sensitivity, 87-96% specificity and 89% or 92% accuracy, which was comparable to that of the original ACR classification. Furthermore, kappa analysis indicated that the results of the original ACR and US-replaced ACR classifications matched completely (κ = 0.960-0.974). These results suggest that US can be used as an alternative to any of the three ACR classification items.