Novel imaging modalities in spondyloarthritis


Novel imaging techniques have emerged in the field of spondyloarthritis. This article will cover the role of, and the sensitivity and specificity of magnetic resonance imaging (MRI) and ultrasound in the diagnosis and monitoring of axial and peripheral SpA. It will discuss how the definition of a ‘positive MRI’ of the sacroiliac joints and spine is evolving. Differential diagnoses of inflammatory lesions of both the sacroiliac joints and the spine are addressed due to their importance in image interpretation. The article will also discuss the role of sonography in assessing peripheral entheses, joints, tendon sheaths, nails and soft tissues. The utility for clinical as well as an outcome measure will be discussed. We finally aim to give guidance on when and how to use these new modalities and on how to analyse and interpret the imaging findings in daily practice.

Practice points

  • Imaging is important in early diagnosis of spondyloarthritis

  • MRI plays a key role in diagnosis and classification of patients with axial spondyloarthritis

  • Ultrasound shows arthritis, enthesitis and dactylitis in patients with peripheral spondyloarthritis

Research agenda

  • A more stringent definition of a ‘positive MRI’ may decrease false positive diagnosis of spondyloarthritis

  • Developing validated ultrasound arthritis, enthesitis and dactylitis scoring systems as outcome measures according to the OMERACT filter.

Research agenda

  • A more stringent definition of a ‘positive MRI’ may decrease false positive diagnosis of spondyloarthritis

  • Developing validated ultrasound arthritis, enthesitis and dactylitis scoring systems as outcome measures according to the OMERACT filter.


Seronegative axial spondyloarthritis (SpA) is a chronic inflammatory disease that affects the sacroiliac joints (SIJs) and the spine . Peripheral manifestations of SpA include arthritis of synovial joints in the appendicular skeleton, enthesitis or dactylitis. On average, axial SpA remains undiagnosed for up to 7 years from the onset of clinical symptoms. There is an ongoing debate about the appropriate imaging approach in patients with suspected or confirmed SpA. As magnetic resonance imaging (MRI) of the SIJs enables the early diagnosis of axial SpA, its use in clinical practice has increased dramatically during the last decade . Sonographic evaluation of non-axial manifestations of spondyloarthropathy may aid in diagnosis and may have a role as an outcome measure. In the following article, we will discuss utility of ultrasound in imaging the joints, entheses, soft tissues and digits of patients with spondyloarthropathy.

MR imaging

A thorough knowledge of normal imaging findings, typical features in SpA, and the differential diagnosis are crucial in image analysis and interpretation. This article presents recommendations for the use and interpretation of MRI and ultrasound in SpA based on expert opinion and on literature review.

Why should we image?

Axial SpA can be distinguished on the basis of patient history and typical clinical and laboratory findings. Patients with inflammatory back pain with an insidious onset of symptoms, <45 years of age and a duration of at least 3 months should initially be imaged with radiography, according to the Assessment of SpondyloArthritis Society (ASAS) criteria .

The ASAS classification criteria for axial SpA consists of two arms, an imaging arm and a clinical arm, both considered equally important . The key feature in the imaging arm is sacroiliitis, either diagnosed by MRI in the nonradiographic stage, or by radiographs in the radiographic stage. We should be aware that SpA may remain undiagnosed if only radiographs are obtained, as a negative study does not exclude SpA. If radiographs are negative in patients with a suspicion of SpA, MRI may detect early inflammatory lesions in the nonradiographic stage of the disease ( Fig. 1 ) . MRI enables confirmation of a diagnosis of early SpA, suspected on the basis of clinical aspects, as early as 4 months after symptom onset . MRI is established as the most sensitive imaging modality for the early detection of axial SpA since active and structural lesions in the SIJs and the spine can be detected by MRI long before they become evident on radiography . Early inflammatory changes may lead to later structural changes, with only structural damage visible on radiographs.

Fig. 1

Sacroiliitis in a 17-year-old female patient. (a) Radiograph of the SI joints is normal. (b–c) Paracoronal STIR MR images show subchondral/periarticular bone marrow edema (arrows) of both SI joints in keeping with sacroiliitis.

MRI is the imaging modality of choice for the detection of active inflammatory spinal and sacroiliac lesions since it demonstrates bone marrow edema (BME) on T2-weighted fat-suppressed sequences such as STIR, as a feature of active inflammation in the preradiographic stage. ASAS defines a ‘positive MRI’ in SpA sacroiliitis as the presence of subchondral/periarticular BME that is present on at least two consecutive slices if only one lesion is seen, or, alternatively, at least two lesions on one slice should be present ( Fig. 2 ) . In the absence of BME other signs of inflammation like capsulitis, synovitis, and enthesitis alone are not sufficient for a ‘positive MRI’ in SpA ( Fig. 3 ) .

Fig. 2

Sacroiliitis in a 41-year-old male patient. (a) Paracoronal STIR MR image shows bone marrow edema (arrows) on the iliac and sacral side of the right SI joint fulfilling the ASAS definition of a ‘positive MRI’. (b) T1-weighted MR image shows structural lesions including irregular erosions with ‘string of pearls’ appearance (short arrows) and fat deposition (long arrows).

Fig. 3

Sacroiliitis in a 16-year-old female patient. Paracoronal STIR MRI image shows extensive bone marrow edema (short arrow) of both SI joints. Anterior capsulitis is present bilaterally (long arrows).

The imaging arm of the ASAS criteria has a sensitivity of 66% and a specificity of 97% for classification of axial spondyloarthritis. This low sensitivity indicates that a negative MRI cannot rule out SpA. Also, the lack of high specificity (0.88) of a ‘positive MRI’ of the SIJs in the diagnosis of SpA according to the ASAS definition might be explained by the periarticular BME in the SIJs that also occurs in asymptomatic individuals and in patients with mechanical back pain .

Structural lesions indicating areas of previous inflammation include subchondral periarticular fat infiltration, sclerosis, erosions or ankylosis. These are best seen on T1-weighted images and are common in SpA . Of these, the most specific single feature for a definition of SpA is presence of erosion, defined as a focal cortical irregularity with underlying subcortical signal change . As a relatively late change, this feature may not be as sensitive for SpA as BME.

A ‘positive MRI’ for the spine is defined by the Outcome Measures in Rheumatology Clinical Trials (OMERACT) and ASAS group if at least three anterior and/or posterior corner inflammation lesions (anterior or posterior spondylitis) or fat deposits at the vertebral corners are present .

How should we image?

Technical aspects of MRI

Mandatory MRI sequences for the SIJs and spine are T1 and STIR. When imaging the SI joint, paracoronal T1 and STIR images along the long axis of the sacrum should be obtained. This paracoronal orientation is superior for visualization of the subchondral osseous areas. The para-axial plane is required to evaluate the exact anatomical location of the pathologies and to avoid pitfalls that might occur on paracoronal planes due to partial volume effects. Both hip joints should be included in the scanning range in the para-axial plane .

When imaging the spine, the number of slices in the sagittal plane should be increased to visualize the lateral paravertebral segments. If the same number of slices is used as on standard imaging for mechanical low back pain, inflammatory lesions that are more laterally located, such as at facet joints or costotransverse joints, are easily missed. The total spine should be imaged and not only the lumbar spine because >50% of all active lesions are located in the thoracic spine .

Should contrast medium be included in the MRI protocol?

Currently, it is under debate if the administration of contrast media increases the detection of osteitis, capsulitis, enthesitis and synovitis on MRI of the SIJs . If BME is present, which is the main and mandatory feature in the diagnosis of axial SpA, STIR images only are sufficient. In the European Society of Skeletal Radiology (ESSR) arthritis subcommittee consensus paper, radiologists strongly state that contrast medium is of diagnostic importance for the differentiation of diagnoses other than sacroiliitis and should be applied in doubtful cases . In children, the use of contrast medium is accepted by many for the detection of early subtle inflammatory changes in juvenile SpA .

MR image interpretation

Sacroiliac joints

The SIJs are the most frequently involved site in SpA. The early lesions are tiny discontinuities of the subchondral bone plate along with subcortical BME, typically located at the inferior half of the joints and at the iliac surface, which is lined by fibrocartilage and less resistant to inflammatory damage. Active lesions are typically depicted on the STIR images. BME, one of the most important MRI features in SpA, is considered an active lesion, as well as capsulitis, enthesitis and synovitis. Structural changes are typically seen on the T1-weighted images and include fat deposition, erosions that sometimes result in a widened joint space with “string of pearls” appearance, sclerosis, “backfill” where the eroded joint reossifies and finally ankylosis .


The spine is the second most frequent site of involvement in SpA. When vertebrae are involved, lesions generally present initially as erosions at the anterior aspect of the thoracolumbar vertebral bodies . Later, the erosions are associated with sclerotic changes and syndesmophytes, which, in long-standing disease, tend to fuse. The distribution pattern of active and structural lesions includes corner inflammatory lesions, central inflammatory lesions and BME in the lateral and posterior spinal segments such as the pedicle, costotransverse and costovertebral joints is characteristic of SpA ( Fig. 4 ) . BME in these anatomical locations indicates active inflammation, whereas fatty infiltration or even ankylosis of the small joints indicate structural damage. Syndesmophytes, being thin strands of cortical bone, are generally best seen on radiographs, but, if prominent, they can at times be detected on MRI as well. Noninfectious spondylodiscitis is less common, and when present is characterized by BME and erosion at the vertebral end plates ( Fig. 5 ). BME in the spinous process or at the ligament insertions on the spinous processes is also a feature of SpA . Although spinal lesions typically develop later in the disease course than sacroiliitis, they may be the only active lesions in a symptomatic patient with longstanding disease, and therefore may determine prognosis and need for treatment, motivating routine whole-spine MRI in SpA monitoring at some centres .

Fig. 4

(a–d) Active spinal SpA lesions in a 41-year-old male patient. Sagittal STIR MR images of the spine shows typical anterior and posterior corner inflammatory lesions in the thoracic spine as BME (arrows). Aseptic spondylodiscitis is seen as symmetric bone marrow edema of two adjacent vertebral bodies centred on the intervertebral disc space (arrowheads). BME of the head of rib is also present (long arrow). BME of the synchondrosis of the sternum is also depicted.

Fig. 5

Active spinal inflammation in a 33-year-old male patient. Sagittal STIR MR image of the spine shows aseptic spondylodiscitis (arrows) as bone marrow edema of two adjacent vertebral bodies centred around the intervertebral disc space.

Differential diagnosis


Diagnosing inflammatory sacroiliitis on MRI is not always straightforward and can be challenging. Several alternative diagnoses can be suggested based on a characteristic MRI appearance. Infectious sacroiliitis, osteitis condensans ilii, insufficiency fractures, osteoarthritis, diffuse idiopathic skeletal hyperostosis, anatomical variants and tumour may also result in BME of the SI joints . Infection is most often unilateral and may demonstrate a joint effusion and periarticular soft tissue edema or collections. Osteitis condensans ilii generally presents as bilateral nearly symmetric iliac edema and later sclerosis, often triangular in shape and most often in women who have had one or more children, and is presumably related to mechanical strain during childbirth. The SI joint lines should remain smooth without erosions in this condition. Insufficiency fractures will present with linear edema and sclerosis in one or both sacral ala primarily paralleling the SI joints, rather than being centred within the SI joints themselves, which again should not show erosion. Osteoarthritis may have BME and irregularity of articular margins simulating sacroiliitis, but erosions are not a feature, and bony proliferation occurs at articular margins rather than centrally within the joint. Presence of a transitional lumbosacral junction or history of prior trauma may be associated with osteoarthritis .

Spinal inflammation

There is an overlap between degeneration and inflammation in the spine with regard to clinical symptoms and morphology on MRI. However, knowledge of certain morphological and clinical features facilitates the differentiation between the two entities. The changes seen in noninfectious spondylodiscitis are similar to active osteochondrosis. Analysis of the disc (usually not dehydrated in spondylitis), the location (degeneration typically at L4–S1, spondylodiscitis more often at levels higher in the thoracolumbar spine) and presence of other concomitant features such as corner inflammatory lesions may help differentiate these entities. Other differential diagnoses, such as infection, tumour, and stress reaction/fracture should be carefully excluded .

SpA without active inflammation

Both active and structural lesions of the SIJs and spine can be present in one patient at the same time. Moreover, in late-stage disease or under treatment, inflammatory lesions may turn into structural lesions such as fat deposition . This underlines the importance of careful examination of the T1-weighted images for presence of structural lesions which can be easily missed on inspection of STIR images alone ( Fig. 6 ).

Fig. 6

Structural spinal lesions in a 35-year-old male patient with SpA under treatment. Sagittal T1 MR image of the spine shows multiple sites of fat deposition at the corners of vertebral bodies (arrows) in keeping with structural lesions of SpA. No active lesions were seen on the corresponding STIR image.

Definition of a ‘positive MRI’ for sacroiliitis in SpA

It has been argued that the current ASAS definition of a ‘positive MRI’ of the SIJs and spine lacks specificity and may therefore result in false positive diagnosis of SpA. The definition includes BME as the only feature of active inflammation when it is present on at least two consecutive slices if only one lesion is seen, or, alternatively, at least two lesions on one slice should be present . This definition has been fiercely debated among experts, as illustrated by the ESSR arthritis subcommittee consensus paper on imaging in SpA, stating that, in the absence of additional imaging features of SpA in the SIJs or spine, two tiny lesions <1 cm in diameter are not sufficient for the diagnosis of SpA, particularly if these lesions are located in the proximal or distal margins of the SIJs . The definition does not include other signs of active inflammation like enthesitis, capsulitis or synovitis. However, the presence of these MRI features of SpA may be helpful in doubtful cases . It has been shown that the specificity of MRI of the SIJs in SpA increases if structural lesions and in particular erosions would be included in the definition .

On the other hand, an advantage of the currently applied definition is that BME is relatively easy to demonstrate and depict on the robust STIR sequence whereas erosions are considered more difficult to assess. This leads to a more sensitive, and potentially more reliable, but less specific definition of a ‘positive MRI’ than a definition based on presence of erosions. Ultimately, to be most useful, refinements to this definition may require different MRI criteria depending on whether the clinical intent is to rule in disease already suspected, detect active disease in cases of known SpA, or attempt to rule out SpA. Much further research is needed in this area.

Therapy monitoring and follow-up

Different scoring systems quantify the amount of BME that is seen on MRI of the SIJs or spine . Repeat MRI in patients may show a change in the score thus reflecting changes in the inflammatory status of a patient. This may be useful in clinical practice for therapy monitoring and follow-up. A report on scoring methods on MRI in SpA concluded that the ASspiMRI-a method, the Berlin method and the SPARCC method perform equally well with respect to responsiveness and discrimination. Of the three, the SPARCC method provided consistently higher reliability as measured by intraclass correlation coefficients .

The prognostic role of MRI is currently under debate. For spinal involvement, BME in the vertebral corners is considered a sign of active inflammation, and it may evolve into new syndesmophytes and erosions . Large scale and long term longitudinal studies are necessary to understand the relation of MRI findings to clinical and imaging outcomes and how these can be influenced by therapy. Routine imaging of the SI joints and whole spine using T1 and STIR sequences will allow a more complete understanding of disease activity and progression than imaging the SI joints alone. Diffusion weighted imaging can be rapidly obtained at the SI joints and may also be valuable to routinely include. Use of gadolinium contrast, whether with simple post-contrast imaging or dynamic sequences, is more logistically cumbersome and costly, and may not be of value in routine assessment.

Future imaging techniques

The combined scanning of the whole spine and the SIJs enhances confidence in diagnosing SpA compared with SIJ MRI alone in patients with nonradiographic axial SpA . Diffusion-weighted MRI and whole-body MRI have been used in only a few reports, however, the clinical or research potential of these sequences is not yet well established . Dynamic contrast-enhanced MRI is another advanced MRI sequence which may also allow for quantification of BME, but this and other advanced sequences such as diffusion-weighted imaging and whole-body MRI, are currently only used for research purposes and have not been validated for use in daily practice. As previously mentioned, current consensus opinion is that intravenous contrast should not be routinely administered, but only in cases of diagnostic uncertainty . It is more commonly (but not universally) accepted that in children the use of contrast may be beneficial as the BME is notoriously difficult to assess on STIR images due to the presence of cartilage .


To improve patient outcomes in SpA, early diagnosis and early treatment are crucial prognostic factors. MRI plays a key role in the diagnosis of early axial SpA as it depicts early inflammation. MRI of the SIJs is the cornerstone of imaging in SpA, whereas MRI of the spine may be helpful in troublesome cases, and for evaluation of prognosis in known SpA. Thorough knowledge of the MRI features of active and structural lesions in SpA and differential diagnosis enables appropriate image interpretation.

Ultrasound imaging

Whilst the purview of plain radiography and MRI is to evaluate axial manifestations of Spondyloarthropathy, sonography is an excellent high resolution modality to image peripheral entheses, joints, tendon sheaths, nails and soft tissues. It not only depicts ultrastructural changes, but also is able to demonstrate vascularity of structures utilizing Doppler studies and without use of contrast.

Why should we image peripheral entheses with ultrasound?

Spondyloarthropathies may be associated with peripheral enthesitis in all phases of the diseases. Clinically there are several scoring systems that have evolved to evaluate the entheses . However, a significant proportion of enthesitis may be subclinical . Plain radiography is unable to demonstrate the soft tissue alterations. Although conventional MRI demonstrates bone edema at entheseal insertions, it is not reliable at demonstrating soft tissue alterations at the enthesis. This is due to the lack of water accumulation at the distal enthesis due to tightly packed tendon fibrils or fibrocartilagenous transition . Sonography can not only depict entheseal alterations with increased sensitivity but also may have a role in evaluating peripheral entheses in non-radiographic Spondyloarthritis. It also has a role in evaluating joints for presence of synovitis as well as structural alterations. In a recent study of early Psoriatic arthritis, a significant proportion of patients’ diagnosis was changed from oligoarticular disease to polyarticular disease after ultrasound examination . Dactylitis is a uniform swelling of a digit is pathognomonic of Psoriatic Arthritis but which can also occur in other forms of spondyloarthropathy. There are limited instruments to objectively measure and follow this manifestation. Sonography demonstrates alterations in several digital compartments in dactylitis .

Imaging peripheral entheses

Entheses are defined as tendon or ligamentous attachment to bones. Inflammation at entheses in the past has been loosely termed enthesitis. As we will see below, there is much more to an enthesis and the definition of enthesitis is evolving. Although entheses are widespread in the skeletal system, inflammatory enthesitis only occurs at some of these entheses and predominantly in the lower extremities. Furthermore, entheses can be further divided into fibrous entheses or fibrocartilagenous entheses. The fibres of fibrous entheses such as the deltoid insertion insert directly into the periosteum or bone. Inflammatory enthesitis does not affect these entheses. The other type of enthesis defined by Benjamin et al. is a fibrocartilageous one. Instead of the tendon fibres inserting directly onto the bone, four zones are identified at the tendon attachment – dense connective tissue fibres which is followed by a transition of fibrocartilage prior to the bone . These fibrocartilageous entheses are felt to have formed to adapt to compressive forces . Furthermore, there is increased recognition of further adaption of the adjacent structures to dissipate force. One such adaption is the presence of fibrocartilage on adjacent bone as well as on the tendon . Similarly the role of adjacent fat pads and bursae in distributing forces has led to an alternative concept of enthesis as a synovio entheseal complex . This has reflected a change in consensus definitions of enthesitis .

Structural alterations

A variety of structural changes can be seen on sonographic examination of entheses. These include alteration of tendon structure, thickness, adjacent bursitis and periosteal bone changes including erosions as well as enthesophyte formation and tendon calcification ( Fig. 7 ) . There are only a few studies of histopathologic analysis of enthesitis. The seminal studies by Ball et al. demonstrated multiple foci of inflammation within an enthesis with accompanying erosive changes at the enthesis. Underlying osteitis was described as well as deposition of bone in fibrous connective tissue without prior cartilage formation . McGonagle et al. have demonstrated that erosions at the Achilles enthesitis preferentially occur at the superior pole of the calcaneus possibly due to lower trabecular density of bone in this region, as well as normally occurring minor cortical defects in this region . Alterations in these individual anatomic components have been termed elementary lesions on grey scale scanning. Initial definitions of enthesitis by the Outcome Measures in Clinical Rheumatology (OMERACT) mainly defined enthesitis by the alteration of the tendon or ligament at the bony attachment . More recent OMERACT consensus definitions not only include perientheseal tendon or ligament alterations but also enthesophyte formation, tendon calcification, cortical erosions as well as adjacent Power Doppler signal . Of note bursitis which is largely regarded as part of the synovial entheseal complex did not reach agreement to be included in the definition. The latest OMERACT definition of enthesitis also defines the presence of Power Doppler to be within 2 mm of the bony cortex. This is important to note since neo-angiogenesis may occur as part of a reparative process in the more proximal portion of the tendon. Furthermore, it is important to remember that apart from feeding vessels, Doppler signal is not found at the cortex. In normal individuals, contrast enhanced ultrasonography did not disclose vascularity at the enthesis junction but did show vascularity in the adjacent fat pads and tissues .

Nov 10, 2017 | Posted by in RHEUMATOLOGY | Comments Off on Novel imaging modalities in spondyloarthritis

Full access? Get Clinical Tree

Get Clinical Tree app for offline access