RT Journal Article
SR Electronic
T1 Multiobserver Reliability of Ultrasound Assessment of Salivary Glands in Patients with Established Primary Sjögren Syndrome
JF The Journal of Rheumatology
JO J Rheumatol
FD The Journal of Rheumatology
SP jrheum.151220
DO 10.3899/jrheum.151220
A1 Nemanja Damjanov
A1 Vera Milic
A1 Juan Carlos  Nieto-González
A1 Iustina Janta
A1 Lina Martínez-Estupiñan
A1 Belén Serrano
A1 Carmen Mata
A1 María Montoro
A1 Denisa Stanciu
A1 Jelena Marinković-Erić
A1 Francisco Javier  López-Longo
A1 Luis Carreño
A1 Esperanza Naredo
YR 2016
UL http://www.jrheum.org/content/early/2016/08/08/jrheum.151220.abstract
AB Objective To evaluate the multiobserver reliability of salivary gland ultrasonography (SGUS) for scoring greyscale (GS) parenchymal inhomogeneity and parenchymal color Doppler (CD) signal in patients with established primary Sjögren syndrome (pSS). Methods The study comprised 2 multiobserver reliability assessments in patients with pSS in 2 European centers. The first reliability exercise was performed on 24 patients with pSS and 8 controls who were independently evaluated with GS and CD US by 5 observers at the Institute of Rheumatology, Belgrade, Serbia. The second reliability exercise was carried out on 10 patients with pSS who were independently assessed with GS and CD US by 8 observers at the Hospital G.U. Gregorio Marañón, Madrid, Spain. SGUS parenchymal inhomogeneity and parenchymal CD signal were semiquantitatively scored using a 4-grade scoring system. The multiobserver agreement was calculated by the overall agreement and Light’s κ statistics. Results A total of 640 SGUS examinations were performed in the first reliability exercise and a total of 320 examinations in the second reliability exercise. Multiobserver reliability was good (κ = 0.71–0.79) to excellent (κ = 0.81–0.82) for GS parenchymal inhomogeneity in both exercises. There was a moderate (κ = 0.53–0.58) to good (k = 0.70) multiobserver reliability for parenchymal CD signal in the first exercise. However, there was no agreement or only a fair agreement (κ = 0.03–0.29) for parenchymal CD signal in the second exercise. Conclusion US may be a reliable technique in the multiobserver scoring of GS parenchymal inhomogeneity of major SG in patients with established pSS. CD scoring of SG needs further standardization to be used in multicenter studies.