Background:SIJ involvement is a characteristic feature of Spondylarthritis (SpA). Magnetic Resonance imaging (MRI) has been included in the new Assessment of SpA International Society (ASAS) criteria for the classification of Axial SpA. Gray scale US, Color Doppler ultrasound (CDUS), contrast-enhanced CDUS, and spectral Doppler (SD) US has been used in few works to evaluate the inflammatory activity of the SIJ with not conclusive results. Power Doppler ultrasound (PDUS) was not yet applied to the study of SIJ with active SI.Objectives:The aim of this work was to study with PDUS and SD US the SIJ of patients with suspected active SI, to describe inflammatory flows with spectral wave analysis (SWA) in duplex Doppler US, and to correlate US data with clinical characteristics and the presence of bone marrow edema (BME) in MRI.Methods:22 patients (18 females and 4 males, mean age 35 years) with new onset of inflammatory back pain (IBP), were included. Every patient underwent an US examination in prone position. The sonographers were blinded to the clinical data of the patient. A Esaote Twice US machine, equipped with a convex multifrequency 1-8 MHz probe, was used, with standardized parameters: 1-5 MHz for gray scale, 1.9-2.3 MHz frequency for Doppler with Pulse Repetition Frequency (PRF) of 1.0 KHz and a color gain just under the artifact limit. SIJ was located as the hypoechoic triangle delimited between the sacrum and iliac bone, and the posterior SI ligament as the upper margin. The first sacral foramen was always localized to avoid measurement of the normal pre-sacral arteries. The PDUS was applied, and if any signals were detected in the SIJ, they were scored with a 3-points scale: 0= absence of signals, 1= isolate vessels, 2= more than one vessel. The signals were also classified as intra-articular or peri-articular. The same vessels were also evaluated using quantitative SD calculating the Resistive Index (RI=peak of systolic flow- end diastolic flow/peak systolic flow), ranging between 0 and 1. Every patient underwent MRI of SIJ within the same week, before treatment. A statistical analysis was performed, estimating the sensitivity and specificity against the gold standard (presence of BME in the same SIJ according to ASAS criteria). The Spearman rank not-parametric test was applied to correlate the presence and grading of BME with PDUS grading and RI. A regression analysis was applied between PDUS results and clinical characteristics.Results:In 14/22 SIJ MRI revealed BME. In 13 of them, PDUS confirmed abnormal hypervascularisation in the intrarticular portion of SI, and in 3 in the periarticular site too. Two SIJ showed hypervascularisation at PD with no BME in MRI. A significant correlation was demonstrated between positivity and grading of PD and presence of BME in MRI (p=0.0005). SD analysis demonstrated low Resistance Index (RI) values in 14 SIJ (mean 0.57). An inverse correlation was demonstrated between RI and grading of BME in MRI (r= -0,6229, p= 0,044). The diagnostic accuracy of SD for detection of active SI varied on the basis of RI cut-off value. The best values of sensitivity (62,5%) and specificity (61,5%) were obtained with a RI cut-off values of 0.60. A multiple regression model demonstrated a significant relationship between PDUS signals and ASDAS (p=0.0382), but not with inflammatory reactants.Conclusion:PDUS and SD US of SIJ can be useful as first imaging assessment in suspected active SI, demonstrating a good diagnostic accuracy compared with MRI. Intra-articular low RI values (<0.60) on SD indicate active SI with good accuracy. Moreover, PDUS signals into the SIJ correlate with clinical symptoms but not with inflammation reactants.Figure 1.Doppler US in SI.Right SIJ with a Doppler signal along the posterior SIJ ligament, and another Doppler signal into the joint, where SD analysis gave a RI of 0,62.Disclosure of Interests:None declared
Estimating confidence intervals for the difference in diagnostic accuracy with three ordinal diagnostic categories without a gold standard