Background/Purpose: Ultrasonography of major salivary glands (sUS) is applied in the diagnostic work-up of primary Sjögren’s syndrome (pSS). This study aims to assess (i) the validity of sUS compared to salivary gland biopsy, sialometry and serology and (ii) the alignment of sUS with classification criteria in patients suspected for pSS.

Methods: 103 consecutive outpatients clinically suspected for pSS underwent sUS of both parotid and submandibular salivary glands between October 2014 and November 2015. Parenchymal echogenicity, homogeneity, hypoechogenic areas, hyperechogenic reflections and clearness of the salivary gland posterior border were scored. Ultrasound total score (UTS) was calculated as the sum of these domains (range 0-48), according to the Hocevar scoring system.(1) All patients underwent a complete diagnostic work-up. For the present analyses, UTS was compared to parotid (n=70) and labial (n=54) gland biopsy outcomes (focus score ≥1), sialometric data (unstimulated (UWS) and stimulated (SWS) whole saliva; n=99), presence of anti-SSA antibodies (n=103) and AECG, ACR and proposed ACR-EULAR classification criteria (n=98/99). ROC analysis with area under the curve (AUC) was performed to define the optimal cut-off point of UTS to predict positive salivary gland biopsy or fulfilling classification criteria.

Results: Of 103 included patients, median age was 51 years (range 18-82), 90% were female and median UTS was 12 (range 3-43). Accuracy of sUS outcomes to predict positive parotid gland biopsy (AUC 0.833) and labial gland biopsy (AUC 0.816) was good. The optimal cut-off point of UTS was 15 for parotid and 14 for labial gland biopsies. Agreement between positive UTS and positive parotid (κ=0.580, sensitivity 70%, specificity 88%) and labial (κ=0.556; sensitivity 70%, specificity 85%) gland biopsy was moderate. UTS showed high negative predictive value for parotid gland biopsy and high positive predictive value for labial gland biopsy. Accuracy of sUS outcomes to predict abnormal UWS (AUC 0.696) was poor and to predict abnormal SWS (AUC 0.731) was fair. There were fair reversed associations of UTS with UWS (ρ=-0.336) and SWS (ρ=-0.371). Accuracy of sUS outcomes to predict presence of anti-SSA antibodies (AUC 0.803) and agreement between positive UTS and anti-SSA antibodies (κ=0.633) were both good. Accuracy of sUS outcomes to predict AECG classification (AUC 0.804) and ACR classification (AUC 0.839) was good and to predict ACR-EULAR classification (AUC 0.782) was fair. The optimal cut-off point of UTS was 14 for AECG and 15 for both ACR and ACR-EULAR. Agreement between positive UTS and AECG (κ=0.573, sensitivity 73%, specificity 85%) and ACR-EULAR classification (κ=0.512, sensitivity 61%, specificity 93%) was moderate and agreement between positive UTS and ACR classification (κ=0.636, sensitivity 71%, specificity 92%) was good.

Conclusion: In our prospective inception cohort derived from daily clinical practice, sUS outcomes showed moderate agreement with salivary gland biopsies, poor to fair agreement with sialometry, good agreement with serology and moderate to good agreement with classification criteria in patients suspected for pSS. References: (1) Hocevar et al. Rheumatology 2005;44:768-72.

« Back to 2016 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/ultrasonography-of-major-salivary-glands-in-patients-suspected-with-primary-sjogrens-syndrome-comparison-with-salivary-gland-biopsy-and-classification-criteria/