gms | German Medical Science

Background: Bone marrow edema (BME), fat metaplasia (FL) and erosions have been identified as relevant for magnetic resonance imaging (MRI) changes in the sacroiliac joints (SIJ) of patients with axial spondyloarthritis (axSpA). However, a high prevalence of MRI changes has recently also been reported in subjects with no evidence of axSpA (Weber U, et al. Arthritis Rheumatol. 2018; De Winter, et al. Arthritis Rheumatol. 2018). In this study, we mapped the MRI lesions suspicious of axSpA in patients diagnosed with axSpA and compare them to those in patients with chronic back pain (cBP, non-SpA).

Methods: Consecutive patients with cBP < 45 years were included if they had at least one pathologic lesion of any type in the SIJ-MRI performed at the time point of cBP symptoms. AxSpA patients diagnosed by 2 experienced rheumatologists in consensus also had to fulfil ASAS classification criteria. Two experienced readers, blinded for diagnosis and patient´s demographics, evaluated all MRIs independently. Lesions were only counted as positive if both readers were in agreement and mean values of both readers were calculated for the final analysis. Both the coronal (assessing the upper and lower sacral and iliac SIJ part) and axial (assessing the ventral, middle and retroauricular as well as the upper and lower sacral and iliac SIJ part) MRI orientations were analyzed for the localization of BME, FL, sclerosis and erosions. In addition, length and width were digitally measured for BME, FL and sclerosis, and signal intensity was measured electronically in each individual patient in comparison to the cerebrospinal fluid signal for BME and the subcutaneous fat for FL (no units).

Comparisons were calculated by Mann-Whitney-U-test for patients classified as positive by both readers for the respective lesions.

Results: A total of 200 consecutive patients (100 axSpA, 100 non-SpA), mean age 36.1±11.3 and 40.3±11.0 years, respectively, were analyzed. BME was found in 85% vs. 80% of patients, while 80% vs. 69% had FL, 54% vs. 40% had sclerosis and 64% vs. 12% had erosions, respectively. The largest surface area covered by BME in axSpA vs. non-SpA was found in the lower and dorsal SIJ: 60±10.1 mm3 in the iliac and 47.3±9.4 mm3 in the sacral part vs. the upper and ventral SIJ: 18.7±3.4 mm3 in the sacral and 5.2±0.1 mm3 in the iliac part.

Patients with axSpA showed a larger surface area covered by FL was found in the upper and anterior sacral SIJ (305.5±56.3 mm3), whereas patients with non-SpA showed larger FL areas in the lower and posterior sacral SIJ (197.9±1.2 mm3).

Regarding sclerosis, the upper and anterior iliac part had more surface involvement in both SpA (139.3±11.6 mm3) and non-SpA (81.8±2.8 mm3) patients.

The mean signal intensity of all lesions and MRI planes differed between axSpA (102.385) and non-SpA (48.995) patients for BME (p<0.001) but not for FL.

Overall, axSpA patients also had significantly more SIJ quadrants with pathologic changes, except for BME and sclerosis in the ventral and fat located in the retroauricular part of the SIJ. The occurrence of erosions in the mid (61 vs. 7) and the ventral (51 vs. 8) part of the SIJ could discriminate best between axSpA and non-SpA (both p<0.001).

Conclusion: These data show that although all types of lesions may be found in both patient groups, the anatomic pattern of SIJ involvement can still distinguish axSpA from non-SpA. The localization and morphological appearance of SIJ-MRI features suggestive of axSpA may serve as an additional feature in the definition of a ‘positive’ MRI both for diagnosis and classification.