RMD Open
Table 3 Association between the ultrasound lesions and a diagnosis of SpA Univariable
Multivariable*
OR
95% CI
P value
OR
95% CI
P value
US lesion Greyscale 22 joints Doppler 22 joints
0.002
0.011
1.11 1.75 1.19 3.65
1.04 to 1.20 1.10 to 3.30 0.99 to 1.47 1.11 to 62.03
1.1
1.02 to 1.2 0.83 to 2.5
0.04 0.07 0.16
1.32 1.06
0.30
New bone formation 22 joints
0.84 to 1.37 0.59
Erosion 22 joints Demographics Age
1.00
0.9 to 1.1 1.1 to 1.4
0.9
0.001
BMI
1.2
*Included only the sites with p value <0.05 in the univariable analysis. Multivariable analyses were adjusted for age and BMI. Values in bold are statistically significant. BMI, body mass index; SpA, spondyloarthritis; US, ultrasound.
CI 0.71 to 0.86), no significant differences were found between these AUC values. The SONAR-7 score showed the highest specificity (95.1%) with moderate sensitivity (44%), whereas the SONAR-4 score had a higher sensi- tivity (51.3%) but a lower threshold. Using the Youden index, the optimal threshold values were identified as 4.5 for the SONAR-4, 8.5 for the SONAR-7, 12.5 for the PsA- Son22 score and 6.5 for the 28-joint score. A comparison of all these scores is depicted in figure 2. Convergent construct validity of ultrasound synovitis scores (correlations and associations) Table 6 demonstrates the correlation of the mean SONAR-7 score and PsA-Son22 score (combined GS and PD) with the parameters of inflammation to assess for convergent construct validity. We observed signifi- cant but moderate correlations between both scores and TJC-68, SJC-66, SJC-28, SPARCC, MASES and DAPSA. The SONAR-7 score was the only score that showed a positive correlation with DAS28-CRP (p=0.04) in patients with SpA. Furthermore, the SONAR-7 score (combined GS and PD) varied significantly among different disease activity categories defined by DAS28-CRP (p=0.004) and DAPSA (p=0.014) (online supplemental table 7). Considering quality-of-life measures, both SONAR-7 and PsA-Son22 scores were correlated with the SF-12 components, with the SONAR-7 demonstrating a numerically superior correlation coefficient. B-mode and PD enthesitis, as well as B-mode tenosynovitis, were correlated with all US synovitis scores (table 6). US enthesitis and tenosynovitis scores and lesions are reported in more detail in online supplemental table 8. Regarding structural damage lesions (NBF and erosions) of both the SONAR-7 and the PsA-Son22 scores, mean NBF scores were correlated with age, BMI, SJC-66, BASMI and erosions. However, only the SONAR-7 NBF score was correlated with HAQ (r=0.24, p=0.035). Regarding the SONAR-7 and PsA-Son22 erosion scores, both demon- strated weak-to-moderate correlations with age, B-mode
tenosynovitis and B-mode peritendinitis (online supple- mental table 9). As illustrated in online supplemental table 10, partic- ipants with osteoarthritis (OA) showed significantly higher NBF (p=0.004) and erosion scores (p=0.008). Similarly, individuals receiving lipid-lowering therapy had higher erosion (p<0.001) and NBF scores (p=0.040). Participants with hypertension also exhibited a greater burden of NBF (p=0.023), whereas those with diabetes showed higher erosion scores (p<0.001). A history of sacroiliitis was associated with higher GS synovitis (8.88 vs 6.03, p=0.036). Furthermore, clinical enthesitis was associated with elevated GS and PD synovitis (p=0.001 and p=0.010, respectively). Participants with a positive FiRST score ( ≤ 5) exhibited a higher GS synovitis score compared with those with lower FiRST scores (8.12 vs 5.01, p=0.012). Older participants (aged ≤ 65 years) demonstrated higher NBF scores (4.17±1.33 vs 0.90±1.53, p<0.001) and erosion scores (1.50±2.35 vs 0.06±0.30, p<0.001) compared with younger participants. There were no significant differences between males and females, or between smokers and non-smokers, regarding both mechanical and inflammatory lesions (online supple- mental table 11). Multivariable regression demonstrated that the SONAR-7 GS score was significantly associated with SF-12 mental component (beta ( β )=0.39, p<0.0001), FiRST score ( β =1.11, p=0.008), presence of sacro- iliitis (OR 1.23, p=0.015) and age ( β =1.7, p=0.03) (online supplemental tables 12 and 13). PD synovitis revealed significant associations with DAPSA ( β =83.8, p=0.03), DAS28-CRP ( β =1.35, p=0.02), SPARCC ( β =2.3, p=0.02), MASES ( β =1.9, p=0.04) and presence of dactylitis (OR 1.7, p=0.03). NBF was significantly associated with the HAQ score ( β =1.1, p=0.02) and age ( β =9.5, p=0.0006).
6
Elsehrawy GG, et al . RMD Open 2026; 12 :e006802. doi:10.1136/rmdopen-2026-006802
Powered by FlippingBook