Differences in Topographical Location of Sacroiliac Joint MRI Lesions in Patients With Early Axial Spondyloarthritis and Mechanical Back Pain

Rosa Marie Kiil; Clara E. Mistegaard; Anne Gitte Loft; Anna Zejden; Oliver Hendricks; Anne Grethe Jurik


Arthritis Res Ther. 2022;24(75) 

In This Article


In this study, we compared the topographical location and volume of MRI lesions in patients with axSpA and MBP fulfilling or almost fulfilling the ASAS criteria in the early diagnostic state. This is highly clinically relevant, as it is at this time point radiologists and clinicians are challenged, especially in patients with MRI changes suggestive of axSpA. This is underlined by the result of the global assessment of MRI by two experienced musculoskeletal radiologists (Table 2). In this assessment, a not insignificant part of axSpA patients had scores below 0 (deemed not to have axSpA), and some MBP patients had scores above 0 consistent with the radiologist being confident regarding the presence of axSpA based on the MRI appearance.

One-plane MRI assessment with semi-coronal T1 and STIR sequences of the cartilaginous joint compartment is currently most used in the diagnostic assessment in axSpA as well as in axSpA monitoring such as the SPARCC[20] and Berlin[29] methods, with references to the 2009 ASAS definitions.[4] Weber and colleagues demonstrated by adding the semi-axial scan plane to the assessment an overall decrease in the BME prevalence along with fewer lesions observed in the lower ilium,[18] indicating that some of the observed 'BME lesions' could be due to partial volume effects from vessels and anatomical variations. This is important in the initial diagnostic phase to reduce false-positive BME findings, possibly incorrect diagnosis and overtreatment. As the SIJ morphology varies,[30] it is possible that by using solely one-plane assessment, the location of BME in the cartilaginous compartments and the detection in the ligamentous compartments, respectively, is limited as it is difficult to gain an overview of the entire joint topography.

We demonstrated, when using the two-plane scoring system, that patients with MBP had BME predominantly in the middle anterior sacrum (Figure 1B), followed by the upper anterior sacrum. In one previous study, Hoballah et al.[14] reported BME prevalence divided into upper, middle and lower joint portions according to an axial plane, comparable to our study method where we used the semi-axial plane for scoring supported by findings on the semi-coronal sequences. They compared nulliparous with postpartum women and demonstrated a BME prevalence of 14% in nulliparous and 21–33% in postpartum women of which 65–87% was located in the middle joint portion (sacrum and ilium collapsed) in line with our results in the MBP group. As suggested by Weber et al., this may represent a 'strain-related' BME pattern.[7,18]

AxSpA patients had high BME sumscores in the same locations as the MBP patients, but at the same time, they presented with a widespread pattern both anteriorly and posteriorly (Figure 1B) in accordance with the previous findings.[2,6,8,12] Another main finding in axSpA patients was the high BME scores in the ligamentous joint compartments, an infrequent finding in the MBP group, also supported by high OR of having axSpA in case of BME in the ligamentous compartment (Table 4). Another study by Weber et al. did not find any additional diagnostic value by adding observed BME in the ligamentous compartment to that of the cartilaginous compartment. However, in accordance with our findings, they found high BME prevalence ranging between 12 and 80% in non-radiological axSpA and ankylosing spondylitis but only in 2–6% of healthy controls and patients with non-specific back pain, respectively.[31]

As previously reported by Molto et al.,[8] our group of axSpA patients also had widespread FMD and erosions (Figures 2A and 3C), confirmed in the regression analyses where both bilateral FMD and erosions were risk factors of an axSpA diagnosis. Furthermore, significantly more axSpA patients had the presence of BME and FMD depth in most locations, even though also present in the MBP group.

The between-group differences in the distribution of SIJ normal variations and lumbar spine findings did not reveal major differences (Table 3). This is inconsistent with a recent CT study, including four of the same variations analysed in the present study, demonstrating a significantly higher prevalence of atypical SIJ morphologies in patients with mechanical joint disease compared to axSpA patients and healthy controls.[16] With an overall prevalence of five different SIJ variations, including the lumbo-sacral transitions vertebra, of 80% and 90% in the axSpA and MBP groups, respectively, one can argue that we are overcalling them by too wide definitions, and some of the atypical morphologies may simply represent not well-described normal anatomy. On the other hand, recently, we did a CT-MRI comparison study on atypical SIJ morphologies including seven different variations in healthy young adults (mean age of 28 years) using the same definitions as in this study.[19] Here, we detected variations by MRI in 55% of the participants, with a remarkable lower prevalence of accessory SIJ (14%), iliosacral complex (16%), dysmorphic cartilaginous joint facets (14%) and bipartite iliac bony plate (12%) compared to the present results. In the present, axSpA patients' atypical SIJ morphologies were often accompanied by BME and/or structural changes. One theory explaining this, as elaborated by Jacques et al.,[32] is that variations may induce micro-trauma causing inflammation in predisposed persons such as axSpA patients.

Our study has limitations. We had a small sample size of 84 patients. Furthermore, the MDT conference approach of axSpA diagnostics has not been compared to a traditional expert-opinion; however, one of the rheumatologists attending the MDT conferences had examined 60% of the patients. Additionally, it is notable that our MBP group had high morbidity as they fulfilled or almost fulfilled the 2009 ASAS classification criteria being a subgroup of a much larger cohort of patients with low back pain for 2–12 months. This made a comparison to other studies including patients with non-specific low back pain not meaningful. Despite this, we clearly demonstrated the differences in location and volume of MRI lesions. The MRI scoring system is not validated but compared to previous studies demonstrated acceptable interreader agreements.[1,33,34] Further studies are needed to confirm the diagnostic value of two-plane assessment of the SIJ. As a T1 semi-axial sequence was not performed in this research project, it was not possible to perform an as detailed assessment of FMD, sclerosis and erosions as for BME.