RMD Open
new set of joints demonstrated a numerically superior, but statistically comparable AUC value (0.83, 95% CI 0.76 to 0.90)with other established scores with a larger number of joints (PsA-Son22 and 28-joint scores), with 95% specificity and 44% sensitivity. Overall, the SONAR-7 offered an optimal balance between diagnostic accuracy and efficiency, suggesting that limited joint assessment may provide comparable discriminatory power to more exten- sive scoring approaches. Our results provide further insights into the asso- ciation between US synovitis and US enthesitis. The SONAR-7 score was positively correlated with the entheseal B-mode (p<0.0001), PD (p<0.0001), erosions (p=0.001) and enthesophytes (p=0.009), in addition to B-mode tenosynovitis and peritendinitis. The associa- tion between enthesitis detected by US and synovitis has been previously explored in other studies. 9 38 Zabotti et al also reported that patients with psoriasis with coex- isting arthralgia presented more tenosynovitis than those without arthralgia. 4 These associations may suggest that the severity of sonographic synovitis may also serve as a marker for sono- graphic inflammatory activity in other musculoskeletal domains. It also supports the interplay between entheseal and synovial tissues (ie, the synovio-entheseal complex), which is a crucial element in the pathogenesis of SpA. 9 39 Regarding the association of US synovitis with the demographic, clinical and patient-reported outcomes, GS synovitis, erosions and NBF all showed significant correlations with age, as reported in other cohorts, 40 while only GS synovitis and NBF maintained this signifi- cant association with age in the multiple regression anal- yses. BMI was also associated with NBF in the univariate analysis but not in the multivariable model. Gender and smoking status showed no associations with any of the US-detected joint lesions. During the evaluation of the convergent construct validity, the SONAR-7 score showed weak-to-moderate correlations with the clinical measures of inflamma- tion at both the joint (ie, TJC-68, SJC-66, TJC-28 and SJC-28) and entheseal (ie, SPARCC and MASES) level. This significant link between US synovitis and SJC has been reported previously in PsA. 38 The SONAR-7 score also showed a significant correla- tion with the quality-of-life measures assessed by the SF-12, both mental and physical components. These results emphasise the impact of synovitis on daily living activities and quality of life. We observed that both GS and PD synovitis correlated with DAS28-CRP and DAPSA scores, similar to other studies. 27 41–43 We also found a significantly lower SONAR-7 score in patients with SpA in remission, as assessed by DAPSA or DAS28-CRP, suggesting a potential ability of this score to differentiate between high and low disease activity states. 44 Furthermore, PD synovitis demonstrated significant associations with the DAPSA, DAS28 and the presence of dactylitis, whereas GS synovitis was associated with a
history of sacroiliitis, SF-12 and FiRST scores in multivari- able regression. Regarding damage lesions, the SONAR-7 NBF mean score was significantly correlated with BASMI, as was the PsA-Son22 score. However, while the SONAR-7 NBF score was also significantly correlated with the HAQ in the multivariable analysis, this was not the case for the PsA-Son22 score. This suggests that NBF is more related to disability and movement indices. Overall, these findings highlight the effectiveness and efficiency of the SONAR-7 score in detecting joint inflam- mation and the degree of structural damage with US in patients with PsA and axSpA. Similar to recent reports, we did not find a significant correlation between US-detected synovitis and certain disease activity scores, such as ASDAS and BASDAI, in patients with axSpA. 41 This suggests that the involvement of the axial skeleton is somewhat independent of periph- eral joint involvement in axSpA. This study has a few limitations worth noting. Its cross- sectional design means that conclusions are derived from associations rather than causality. In addition, the rela- tively long disease duration and the high proportion of patients treated with biologics may have decreased the strength of the associations, since biologic treatment may reduce the prevalence and severity of US-detectable inflammatory lesions, particularly for PD. A previous study concluded that the strength of association between clinical and US outcomes diminishes as patients with RA are in or near remission. 45 Despite this, we describe a large number of highly significant associations. Future longitudinal studies in patients with early-stage, biologic- naïve SpA are needed to evaluate the SONAR-7 score’s ability to distinguish early disease from HC and to explore its sensitivity to change in response to treatment and to evaluate this new score in other patient populations. The evaluation of the small entheses of the fingers was not feasible for this study, but represents an important topic for future research. Using different US machines across centres may represent a potential source of vari- ability, particularly in image resolution and PD sensitivity. However, inclusion of multiple centres and equipment types would enhance external validity and generalisability. This study has numerous strengths, such as the use of a comprehensive US scan that evaluated multiple lesions, including synovitis, enthesitis, tenosynovitis, NBF and erosions, across a large number of sites. We included patients with PsA and axSpA, exploring ultrasonographic differences between the two SpA subtypes, as well as a detailed comparison with a cohort of HC. Additionally, the overall cohort was well-characterised, thus enabling the examination of many variables and the control of various confounders. In summary, the SONAR-7 score demonstrated high efficiency, high specificity and significant correlations with multiple clinical scores, performing as well, if not better, than previous US scores with more joints. More- over, sonographic synovitis at the seven selected joints was
12
Elsehrawy GG, et al . RMD Open 2026; 12 :e006802. doi:10.1136/rmdopen-2026-006802
Powered by FlippingBook