Assessment of enthesopathy in conventional pelvis radiography: Another tool for spondyloarthritis diagnosis beyond sacroiliitis

Abstract

Introduction/objectives: Enthesitis is a cardinal feature of spondylarthritis (SpA), and the pelvis is a common site of enthesitis. This study aimed to establish the association between pelvic enthesis involvement on pelvic X-ray and SpA diagnosis through a radiographic enthesis index (REI) and to assess the reliability and accuracy of this REI. Materials and methods: The participants were SpA patients and a control group composed of patients with chronic lumbar pain without SpA. Three blinded observers assessed each pelvic radiography three times. Three zones were used: Zone I (ZI), the iliopubic ramus; Zone II (ZII), the pubic symphysis, and Zone III (ZIII), the ischiopubic ramus. A grading system was created from 0 to 3 [Grade 0, normal; Grade 1, minimal changes (subcortical bone demineralization and/or periosteal wishkering, seen as radiolucency and trabeculation of the cortical bone upon tendon insertion); Grade 2, destructive changes (Grade 1 findings and erosions at the enthesis site); and Grade 3, findings of Grade 2 plus >2 mm whiskering out of the cortical bone) for the REI. The sum of the results of the three zones was called the total REI. For statistical analysis, we used the weighted kappa statistic adjusted for prevalence and bias using Gwet's agreement coefficient. Results: We enrolled 161 patients, 111 of them with SpA (39.6 % with axial SpA and 47.7 % with peripheral SpA) and 50 without SpA. In the SpA group, 36.7 % and 25.7 % had REI Grades 2 and 3 in ZIII, respectively, while only 6 % of the controls had these grades. For ZI, the frequency of Grades 1 to 3 was 42.3 % in the SpA group (8.1 %, 14.4 %, and 19.8 %, respectively), compared to only 2 % in the controls. ZII was unaffected in most of the patients with SpA (82.9 %) and in the controls (98 %). In the control group, Grade 0 was the most common REI grade in all three zones. The agreement was almost perfect for each zone and between the independent readers. The ROC-curve analysis showed that the highest performance areas were the total REI, ZIII, and ZI. Most (75 %) of the SpA patients without sacroiliitis on X-ray were REI-positive. The sensitivity of the REI for SpA diagnosis was 82 %, while the sensitivity of sacroiliitis on X-ray was 38.7 %. Conclusions: The assessment of pelvic enthesis using the REI on pelvic radiography may be useful for SpA diagnosis. Total REI, ZIII, and ZI had the highest accuracy and almost perfect reliability. The REI is especially helpful in patients without sacroiliitis on imaging. © 2024 The Author(s)