Imaging modalities for the diagnosis of axial spondyloarthritis

Not until recently did the spondyloarthritis (SpA) eventually gain its name rectification after vigorous discussion in the rheumatology community.1 As a member of the inflammatory rheumatic disease family causing arthritis,2-6 spondyloarthritis is not a single type of disease; instead, it indicates a spectrum of rheumatic diseases that share similar clinical features. The most common feature is inflammatory back pain. The affected sites and predominant manifestations differentiate patients into peripheral spondyloarthritis and axial spondyloarthritis (axSpA). The former includes arthritis of the asymmetric lower limbs, enthesitis, and dactylitis, sharing common symptoms with psoriatic arthritis,7 inflammatory bowel disease, and reactive arthritis; the latter primarily affects the spine and sacroiliac joint (SIJ), which leads to non-radiographic axSpA (nr-asSpA) or ankylosing spondylitis (AS). With only a few symptoms, it is not adequate to accurately determine whether the culprit of the back pain originates from inflammation or mechanical causes. If the patient is suspected of having inflammatory back pain, after initial clinical consultation and physical examinations, the appropriate practice is to perform a further imaging test aside from blood tests. As axSpA is a spectrum of chronic progressive disease, image-based staging of the disease is necessary for its diagnosis (Figure 1). Though some traits can be used to diagnose the disease, it is worth noting that the diagnosis of SpA should be made based on the presence of bamboo spine, syndesmophyte, or diffuse idiopathic skeletal hyperostosis, instead of using a single biomarker. The axSpA diagnosis includes two disease stages: For early nr-axSpA to definitive radiographic axSpA, also known as AS, several imaging modalities can be applied to diagnose, monitor, and predict structural damage and treatment response.8 For final stage diseases, the recommended imaging method would be conventional radiography; according to the modified New York criteria,9 it can be used to diagnose AS through the assessment of the SIJ and the lumbar spine (to T10). For radiographic sacroiliitis, classification criteria graded from 0 to 4 according to the severity of bone changes (Table 1).10 However, diagnosis through conventional radiography alone is prone to obstacles in identifying early-stage diseases, or low-grade sacroiliitis given the technical limitations. Although the disease is targeted in the younger age group of patients less than 45 years old, conventional radiographic applications are not sensitive enough to diagnose the early stages of the disease. Previous studies suggested an average delay of up to 5 to 10 years between the onset of the first clinical symptoms and the final diagnosis of AS.11 In this case, magnetic resonance imaging (MRI) is strongly recommended to assess whether the patient has active inflammation that can be mainly observed by the presence of bone marrow edema (BME) or other structural changes, as an alternative diagnostic method. While identifying abnormalities in MRI, osteitis (or BME) is central to the definition of a positive MRI of the SIJ, followed by enthesitis and capsulitis. It was emphasized in the Assessment of SpondyloArthritis International Society (ASAS) handbook 2019 edition that the diagnosis of active sacroiliitis on MRI should meet two conditions: “MRI imaging must be highly suggestive of axSpA”, and the “the presence of BME in a typical anatomical cite of subchondral bone”. With an active inflammatory phase, high signals in fat-suppressed T2-weighted or T1-weighted post-contrast sequences could be found frequently in the bone marrow adjacent to the cartilaginous joint surface. As erosion and sclerosis expand, three manifestations should be highlighted as the main observations of the next phase—repair processes:12 first, the fat lesions show bright T1w-signals that replace the previously inflamed non-eroded bone marrow; second, a new bone formation that appears as a fat signal within areas of erosion, also known as backfill; and third, a progressive new bone formation that crosses the original joint space, developing from bone budding to ankylosis. However, due to the multifactorial etiology and complex definition of the disease and their consequent limitations, the ASAS classification criteria have triggered heated debate. As a result, they are commonly used for research purposes, but not in clinical practice. Further research should focus on finding more accurate diagnostic methods. Nevertheless, there could be some situations in which findings in conventional radiography are negative while MRI cannot be performed either. In these cases, computed tomography (CT) can be a much more accessible option than MRI, given lower costs and clinically applicability. Not only is MRI technically unavailable in some regions, but its reimbursement and authorization may not be as accessible as CT. Recent studies reported that the use of CT might provide additional information on structural abnormalities, including subchondral sclerosis, ankylosis, and erosions of SIJ. Given that there are no reliable diagnostic criteria for axSpA based on X-rays or MRI, the development and application of CT will inevitably become a focus in diagnosing axSpA. The advantages of CT include the fact that the physical principles of which are the same as radiography, which is sensitive to the detection of calcification and ossification, and even capable of further evaluation of the thoracic spine. As a result, subtle changes in bone erosion of an articular surface can be detected, which may be easily ignored through MRI-based diagnosis. On the other hand, the time limitation is another factor that needs to be considered when acquiring the detection images from MRI. While the image quality of MRI and its spatial resolution depend heavily on the time required to take a scan, clear information within only a few seconds can be obtained in CT examinations. However, one challenging issue for CT would be the exposure to radiation. Fortunately, CT has been developed as an appropriate modality with similar or even lower radiation exposure (lower than 1 mSv)13 than radiography. At the same time, it has been proved to be more sensitive than conventional radiographs (X-rays) in monitoring disease progression in patients with AS, according to reports in the Annual European Congress of Rheumatology (EULAR) in 2017. Especially in identifying evidence of radiographic change of nr-axSpA imaging stage, low-dose CT scanning has an outstanding performance on both specificity and accuracy.14 Computing power and artificial intelligence have significantly reduced radiation dose while maintaining high detection efficacy and imaging quality. Low-dose CT certainly has tremendous research capacity and great value because subtle abnormalities can be demonstrated clearly, which may aid the resolution of equivocal findings obtained with X-ray or MRI. Furthermore, it can be applied in clinical practice as a technology supporting other imaging tools that can guide the most suitable treatment for each individual; importantly, unnecessary treatments may bring about higher risks of adverse drug reactions,15, 16 which warrants education.17 This is not to say that low-dose CT could replace other methods and monopolize the imaging detection market, but the methods mentioned above can support each other with their respective advantages and weaken the disadvantages (Table 2). Of course, in line with the holy grail of precision medicine, patients suspected of having AS can also undergo further genetic testing when there is a lack of radiological evidence. According to previous studies,18, 19 nearly 90% of people with AS carry a human leukocyte antigen (HLA) that is a specific gene for B27 (HLA-B27), for which there may be a correlation between different clinical manifestations and HLA-B27 status in patients with AS.11, 12 Nevertheless, carrying this gene is not directly linked to the development of AS, but can only predict the risk of disease in advance to allow preventive measures. The final goal is to receive a timely diagnosis that leads to effective management of axSpA. As such, clinicians may avoid misdiagnosing ordinary back pain and nr-axSpA, and ultimately achieve improved symptoms, recovered physical function, and better quality of life. CC drafted the manuscript, and KSM and JCW revised the manuscript. The authors declare no conflicts of interest.