How sensitive and specific are Mri features of sacroiliitis for diagnosis of spondyloartHritis in patients witH inflaMMatory back pain ?

OBJECTIVE
To determine the sensitivity and specificity of MRI features of sacroiliitis in spondyloarthritis (SpA).


MATERIALS AND METHODS
A retrospective study reviewed MRI of the sacroiliac (SI) joints in 517 patients with inflammatory back pain. Sensitivity, specificity, positive and negative likelihood ratios of active and structural lesions of sacroiliitis with final clinical diagnosis as golden standard was calculated.


RESULTS
MRI showed active inflammation in 42% of patients (bone marrow oedema (BMO) (41.5%), capsulitis (3.3%), enthesitis (2.5%)) and structural changes in 48.8% of patients (erosion (25%), fat infiltration (31.6%), sclerosis (32%) and ankylosis (7.6%)). BMO was the MRI feature with the highest sensitivity (65.1%) for diagnosis of SpA. Capsulitis (99%), enthesitis (98.4%), ankylosis (97.4%) and erosion (94.8%) had a high specificity for diagnosis of SpA, whereas BMO (74.3%), sclerosis (75.8%) and fat infiltration (84.0%) were less specific. BMO concomitant with enthesitis, capsulitis or erosions increased the specificity. Concomitant presence of BMO and sclerosis or fat infiltration decreased the specificity.


CONCLUSION
BMO is moderately sensitive and specific for diagnosis of SpA in patients with inflammatory back pain. BMO concomitant with enthesitis, capsulitis, ankylosis or erosion increases the specificity. Concomitant fat infiltration or sclerosis decreases the specificity for diagnosis of SpA. Of all lesions, erosion had by far the highest positive likelihood ratio for diagnosis of SpA.

Spondyloarthritis (SpA) is a group of inflammatory joint conditions sharing common clinical, radiological, genetic and even therapeutic characteristics and are often associated with the presence of human leukocyte antigen (HLA)-B27 (1-5).
Early diagnosis of SpA has gained significance for rheumatologists as new medical treatment options have become available (6).MRI of the SI joints shows active inflammatory and structural lesions of sacroiliitis long before radiographic changes become evident (7,8).
Active sacroiliitis on MRI is an important classification criterion.MRI is regarded 'positive' for sacroiliitis if bone marrow oedema (BMO) is clearly present (8).Other MRI features representing active inflammation of the SI joint such as enthesitis or capsulitis alone are not sufficient for a 'positive' MRI for sacroiliitis.Structural lesions in sacroiliitis are sclerosis, fat infiltration, erosion and finally ankylosis (8).
These active and structural changes of the SI joints, in particular the presence of BMO, may also be present in patients presenting with non-rheumatological entities that clinically mimic sacroiliitis such as degenerative disease, lumbosacral the study (185 (35.8%) men, 332 (64.2%) women), with a median age of 33.3 years (range 16.1-44.9).
The aim of this study is to determine the sensitivity and specificity of MRI features of sacroiliitis in SpA.Also, we sought to find if BMO concomitant with other MRI features of sacroiliitis may increase the sensitivity and specificity for diagnosis of SpA.

Materials and methods
The retrospective study was approved by the institutional ethics committee.

Study group
All participants, aged 16

Conclusion: bMo is moderately sensitive and specific for diagnosis of spa in patients with inflammatory back pain. bMo concomitant with enthesitis, capsulitis, ankylosis or erosion increases the specificity. concomitant fat infiltration or sclerosis decreases the specificity for diagnosis of spa. of all lesions, erosion had by far the highest positive likelihood ratio for diagnosis of spa. keywords: spine, Mr -arthritis.
From: 1.Department of 1. Radiology and Medical Imaging, 2. Rheumatology, Ghent University Hospital, Ghent, Belgium, 3. Department of Radiology, University of Alberta Hospital, Edmonton, Alberta, Canada.Address for correspondence: Dr L. Jans, M.D., Ph.D., Dpt of Radiology, Ghent University Hospital, De Pintelaan 185, 9000 Gent, Belgium.E-mail: lennartjans@hotmail.com specificity of 74.6%.Presence of enthesitis and capsulitis had a low sensitivity, but a very high specificity for diagnosis of SpA.Diagnostic utility of active and structural lesions of sacroiliitis for diagnosis of SpA were determined by calculating sensitivity, specificity, positive and negative likelihood ratio for consensus reader data with final clinical diagnosis as golden standard.

Image review
The MR images were reviewed for the presence of active or structural lesions of sacroiliitis by 2 musculoskeletal radiologists with 10 and 14 years of experience (LJ,WH), who were blinded to clinical and other imaging findings.
-Active lesions of sacroiliitis included BMO, enthesitis and capsulitis (8).BMO was defined as 'positive' for sacroiliitis if high T2 FS/STIR signal of the ilium or sacrum typically located periarticularly was present.If there is one signal (lesion) only, this should be present on at least two slices.If there is more than one signal on a single slice, one slice was considered to be enough (8).Representative images are presented in Figs.1-6.

Statistical analysis
Statistical analysis was performed using software package SPSS 20.0 for Windows (SPSS, Chicago, IL, USA).Basic descriptive statistics were performed where appropriate.I).
BMO concomitant with ankylosis, erosion and fat infiltration had a moderate higher positive likelihood ratio (LR+) for diagnosis of SpA, compared to the LR+ of concomitant presence of enthesitis and sclerosis (Table II).

discussion
Early assessment of inflammation in the course of SpA gains importance in the light of new therapeutic strategies.The value of MRI of the SI joints is well established and resulted in a definition of a 'positive MRI' for sacroiliitis (8,11,12).In the current ASAS criteria of axial SpA MRI plays an important role: a 'positive MRI' is a key criterion for disease classification (8,13,14).
However, in daily radiology practice it remains challenging to decide if demonstrated BMO is sufficient to really represent active sacroiliitis as seen in SpA.In our study we found a moderate sensitivity (65%) and -in this context more importantly-only a moderate specificity (74%) of BMO of the SI joints for diagnosis of SpA.These figures are similar to the figures published in the paper by Weber et al., who also reported on the limitations of using BMO as a single criterion in the current ASAS definition of a 'positive' MRI (16).We confirmed their findings, and also showed that concomitant fat infiltration and sclerosis decrease the diagnostic value of MRI.This is not surprising, since these features are commonly seen in degeneration,    with mechanical back pain mimicking inflammation in these patients (15).Our study showed that the concomitant presence of other features of active sacroiliitis such as enthesitis or capsulitis, indicating an ongoing true inflammatory process, increased the specificity.The presence of structural lesions also contributes substantially to the diagnostic utility of MRI for diagnosis of SpA.The concomitant presence of erosions and ankylosis -both hallmarks of the disease -also increased the specificity (8).On the other hand, the concomitant presence of fat infiltration and sclerosis decreased the specificity for diagnosis of SpA, which may be explained by the presence of the same MRI features in degenerative processes (15).
Weber et al stressed the importance of erosions as an MRI feature of SpA (16)(17).In our study we found that of all lesions of the SI joints, the presence of erosion had the highest LR+ (10.4) for diagnosis of SpA.This finding confirms that erosion could play a role as a new or additional criterion in future classification systems, especially as it improves the specificity of MRI of the SI joints for diagnosis of SpA.
In our study we also found a very high specificity of enthesitis (98%) and capsulitis (99%) for diagnosis of SpA.However, as these lesion were not commonly seen (in 4% and 6% respectively), they might not be as useful as classification criteria.Still, the presence of enthesitis or capsulitis may be particularly helpful in equivocal cases or may indicate that active inflammation is present at a certain stage in the disease process.
There are some limitations to our study.Firstly, MRI was the only imaging technique, without correlation with radiography or CT.Secondly, the patient population represented referrals from a single tertiary center; referral patterns for sacroiliitis Enthesitis was defined high T2 FS/STIR signal of an enthesis representing BMO, soft tissue inflammation or joint/bursal fluid.Capsulitis is seen as high signal on STIR images involving the anterior or posterior capsule of the SI joint (7, 8).-Structural lesions of sacroiliitis consisted of sclerosis, fat infiltration, erosion and ankylosis.Sclerosis is depicted as low signal subchondral bands extending at least 5 mm by all sequences.Erosions are low T1 signal bone defects at the joint margin that may occur throughout the cartilaginous joint compartment.Fat deposition depicts as periarticular high T1 signal in the bone.Ankylosis of the SI joint appears as low signal by all sequences bridging the SI joint and may show high T1 signal if bone marrow is present (8).

Fig. 1 .
Fig. 1. -BMO and capsulitis of the SI joint in a 22-year-old male with spondyloarthritis.Semicoronal STIR image shows BMO (arrows) of the left SI joint with concomitant posterior capsulitis (short arrows).

Fig. 2 .
Fig. 2. -BMO and enthesitis of the SI joint in a 24-year-old female with spondylo arthritis.(A-B) Semicoronal STIR images show BMO (arrows) of the right SI joint with concomitant enthesitis of the left retroarticular ligaments (long arrow).

Fig. 3 .
Fig. 3. -BMO and concomitant erosion of the SI joint in a 26-year-old female with spondyloarthritis. A. Semicoronal T1-weighted MR image shows extensive erosions of the SI joints (arrows).B. Semicoronal STIR image shows BMO of both SI joints (arrows).

Fig. 4 .
Fig. 4. -BMO and ankylosis of the SI joint in a 33-year-old male with spondyloarthritis.Semicoronal (A) T1-weighted and (B) STIR MR images show ankylosis of the left SI joint (short arrows) with concomitant BMO of the right SI joint (long arrows).

Fig. 5 .
Fig. 5. -BMO and fat infiltration of the SI joint in a 44-year-old female with degenerative joint changes.Semicoronal (A) T1-weighted and (B) STIR MR images show fat infiltration of the SI joints (short arrows) with concomitant BMO of the left SI joint (long arrow).

Fig. 6 .
Fig. 6. -BMO and sclerosis of the SI joint in a 44-year-old female with degenerative joint changes.Semicoronal (A) T1-weighted and (B) STIR MR images show sclerosis of the right SI joint (short arrows) with concomitant BMO of both SI joints (long arrows).

Table I .
-The sensitivity, specificity, positive and negative likelihood ratios of MRI features of sacroiliitis for the diagnosis of SpA.