Introduction

“Mirror, mirror on the wall, who’s the fairest one of all?” Imaging has come a long way from simple surface reflections, but is yet to reach the prognostic accuracy of the mirror interrogated by the Wicked Queen. All the same, the analogy is an interesting one as all forms of imaging “mirror” aspects of reality, which in the rheumatology context usually means articular (or periarticular) inflammation and damage. Many ingenious methods have been developed to make such mirrors function, from the application of ultrasonic waves, to altering the moment of spinning protons using magnetic fields, to bombarding tissues with X-rays or forcing them to emit gamma rays after injection with radioactive isotopes. There have been huge developments for rheumatologists in the world of imaging over the last three decades. Although plain radiography (XR) remains the most common form to be used in every day clinical practice, there has been a steady increase in interest in other advanced modalities for diagnosis, prognostication, monitoring of disease activity and damage and assessment of the response to therapy. In some situations, diagnostic criteria are being reframed to include advanced imaging features , and increasingly, these modalities are being used to monitor outcome clinically and in the setting of clinical trials ( Fig. 1 ).

A conceptual framework for the impact of advanced imaging in the diagnosis and management of rheumatic disease.

With the current emphasis on early intervention in the inflammatory arthritides, the detection of sub-clinical inflammation and pre-radiographic joint damage has become a priority, and advanced imaging tools are ideal to provide this information. In addition, imaging can reveal aspects of pathology and is providing novel insights into the anatomical, pathological and molecular basis of disease. However, with the proliferation of these new techniques there are challenges for the radiologist and non-radiologist alike as the information provided is often complex, prone to different interpretation by different observers and frequently requires specialist expertise and equipment. Exposure to ionising radiation remains an issue for some techniques and access to many modalities remains limited for reasons including high cost. Nevertheless, a working knowledge of the strengths and weaknesses of advanced imaging is becoming mandatory for the modern rheumatologist, especially given the great leap forward provided by modern therapeutics, including the biological disease-modifying anti-rheumatic drugs (bDMARDs), which have revolutionised our specialty. The following issue of Best Practice and Research Clinical Rheumatology entitled ‘Investigation of Musculoskeletal Conditions: Imaging in Arthritis’ seeks to provide detailed information from experts in the field of imaging, concentrating on areas of special clinical interest where important advances have recently occurred.

Imaging, pathology and prognosis

Imaging has triggered a re-examination of pathogenetic pathways in many forms of arthritis as it provides a (mostly) non-invasive way to study the tissues involved and importantly their responses to therapy. In the pre-biologic era (1998), a study of rheumatoid arthritis (RA) patients enrolled within 6 months of symptom-onset revealed that 15% had baseline XR erosions and 45% baseline magnetic resonance imaging (MRI) erosions at the wrist . Results from that cohort showed MRI bone marrow oedema (BME) to be present in 64% and in long-term follow-up studies, to be associated with significantly adverse radiographic and functional outcomes . Multiple subsequent studies in larger cohorts have confirmed this negative prognostic influence . Comparative MRI/histopathological studies have revealed BME in RA to represent osteitis, comprising a vascular, inflammatory infiltrate of macrophages, B cells, plasma cells and T cells in periarticular bone . This lesion was largely unknown prior to the application of MRI to musculoskeletal imaging in the 1980s, and appears to be of major importance in the pathology of RA , supporting a two-compartment model (involving both synovium and bone) initially proposed by Watson et al. . Osteitis has been shown to respond to biological therapies and as such is assuming prominence as an imaging biomarker . Its response to B-cell-depleting therapy (BCDT) and anti-tumour necrosis factor (anti-TNF) agents sheds light on pathways underpinning RA pathogenesis . Moreover, the lymphocytes, plasma cells and macrophages within the osteitic lesion sit adjacent to osteoclasts that are closely apposed to bone trabeculae, suggesting a mechanism whereby bone oedema could lead to erosion, based on imaging evidence. Recently, Harre et al. have shown that osteoclasts are activated by antibodies to citrullinated peptide antigens (ACPAs), providing evidence for a link between the adaptive immune system and bone.

The pathological features of psoriatic arthritis (PsA) are even more diverse than those of RA, and advanced imaging has propelled these into the collective rheumatological consciousness. The pre-eminence of XR in the 1970s contributed to the description by Moll and Wright of the typical PsA sub-types , which emphasise radiographic features and distribution of disease. The application of MRI and ultrasound (US) to the investigation of PsA has led to new emphases on soft-tissue inflammation and the previously invisible enthesis . From these observations, McGonagle et al formulated the hypothesis that enthesitis is the primary pathological lesion in PsA and that synovitis is secondary to this . Laura Coates and others discuss the characterisation of enthesopathy by MRI and US in this condition, as well as the features of dactylitis and nail disease revealed by imaging. Grading systems designed to capture the entire spectrum of disease have been developed for both MRI (the psoriatic arthritis MRI scoring system (PsaMRIS)) and US . These combine scores for articular and periarticular inflammation plus bone erosion to provide an overall indicator encompassing the entire spectrum of disease. Whole-body MRI is another new technique that may be particularly relevant in PsA and spondyloarthropathies (SpAs) to define the total burden of disease .

New insights into the pathological basis of joint damage in gout have recently been gleaned from imaging studies, and these are reviewed by Nicola Dalbeth. Here, high-resolution computed tomography (HRCT) scanning has been pivotal in showing that tophi are closely associated with erosions and most likely a major contributor to their development . CT, and its offspring dual-energy computed tomography (DECT), have a special role in gout in view of their ability to image tophi (and adjacent bone) with exceptional three-dimensional clarity. DECT has also recently revealed a wider spectrum of gout pathology than previously suspected, with extensive tophaceous deposits observed within tendons and bursae as well as adjacent to joints . DECT and CT should be ideal modalities to use longitudinally in monitoring tophus resolution after urate lowering therapy (ULT), as already described in feasibility studies of US and MRI .

Nuclear imaging is an extremely sensitive technique that provides quite a different way to view the molecular pathology of rheumatic disease, by imaging cellular function and responses to inflammation. Quantitative positron emission tomography (PET) techniques are discussed by Conny van der Laaken and include the targeting of peripheral benzodiazepine receptors (PBRs), expressed on activated macrophages to image clinically inflamed synovium. Diagnosing pre-clinical RA has become a possibility using this technique and in a recent pilot study, ACPA-positive patients were shown to progress to clinical arthritis at scan-positive finger joints . The tracer [18F] fluoro-deoxy-glucose ([18F] FDG) accumulates in metabolically active tissue and this imaging tool has been used to demonstrate disease activity in clinically inactive joints . The evolving hybrid techniques PET–CT and PET–MRI are also discussed, and these have the potential to allow more precise anatomical localisation of the PET signal.

Imaging, pathology and prognosis

Imaging has triggered a re-examination of pathogenetic pathways in many forms of arthritis as it provides a (mostly) non-invasive way to study the tissues involved and importantly their responses to therapy. In the pre-biologic era (1998), a study of rheumatoid arthritis (RA) patients enrolled within 6 months of symptom-onset revealed that 15% had baseline XR erosions and 45% baseline magnetic resonance imaging (MRI) erosions at the wrist . Results from that cohort showed MRI bone marrow oedema (BME) to be present in 64% and in long-term follow-up studies, to be associated with significantly adverse radiographic and functional outcomes . Multiple subsequent studies in larger cohorts have confirmed this negative prognostic influence . Comparative MRI/histopathological studies have revealed BME in RA to represent osteitis, comprising a vascular, inflammatory infiltrate of macrophages, B cells, plasma cells and T cells in periarticular bone . This lesion was largely unknown prior to the application of MRI to musculoskeletal imaging in the 1980s, and appears to be of major importance in the pathology of RA , supporting a two-compartment model (involving both synovium and bone) initially proposed by Watson et al. . Osteitis has been shown to respond to biological therapies and as such is assuming prominence as an imaging biomarker . Its response to B-cell-depleting therapy (BCDT) and anti-tumour necrosis factor (anti-TNF) agents sheds light on pathways underpinning RA pathogenesis . Moreover, the lymphocytes, plasma cells and macrophages within the osteitic lesion sit adjacent to osteoclasts that are closely apposed to bone trabeculae, suggesting a mechanism whereby bone oedema could lead to erosion, based on imaging evidence. Recently, Harre et al. have shown that osteoclasts are activated by antibodies to citrullinated peptide antigens (ACPAs), providing evidence for a link between the adaptive immune system and bone.

The pathological features of psoriatic arthritis (PsA) are even more diverse than those of RA, and advanced imaging has propelled these into the collective rheumatological consciousness. The pre-eminence of XR in the 1970s contributed to the description by Moll and Wright of the typical PsA sub-types , which emphasise radiographic features and distribution of disease. The application of MRI and ultrasound (US) to the investigation of PsA has led to new emphases on soft-tissue inflammation and the previously invisible enthesis . From these observations, McGonagle et al formulated the hypothesis that enthesitis is the primary pathological lesion in PsA and that synovitis is secondary to this . Laura Coates and others discuss the characterisation of enthesopathy by MRI and US in this condition, as well as the features of dactylitis and nail disease revealed by imaging. Grading systems designed to capture the entire spectrum of disease have been developed for both MRI (the psoriatic arthritis MRI scoring system (PsaMRIS)) and US . These combine scores for articular and periarticular inflammation plus bone erosion to provide an overall indicator encompassing the entire spectrum of disease. Whole-body MRI is another new technique that may be particularly relevant in PsA and spondyloarthropathies (SpAs) to define the total burden of disease .

New insights into the pathological basis of joint damage in gout have recently been gleaned from imaging studies, and these are reviewed by Nicola Dalbeth. Here, high-resolution computed tomography (HRCT) scanning has been pivotal in showing that tophi are closely associated with erosions and most likely a major contributor to their development . CT, and its offspring dual-energy computed tomography (DECT), have a special role in gout in view of their ability to image tophi (and adjacent bone) with exceptional three-dimensional clarity. DECT has also recently revealed a wider spectrum of gout pathology than previously suspected, with extensive tophaceous deposits observed within tendons and bursae as well as adjacent to joints . DECT and CT should be ideal modalities to use longitudinally in monitoring tophus resolution after urate lowering therapy (ULT), as already described in feasibility studies of US and MRI .

Nuclear imaging is an extremely sensitive technique that provides quite a different way to view the molecular pathology of rheumatic disease, by imaging cellular function and responses to inflammation. Quantitative positron emission tomography (PET) techniques are discussed by Conny van der Laaken and include the targeting of peripheral benzodiazepine receptors (PBRs), expressed on activated macrophages to image clinically inflamed synovium. Diagnosing pre-clinical RA has become a possibility using this technique and in a recent pilot study, ACPA-positive patients were shown to progress to clinical arthritis at scan-positive finger joints . The tracer [18F] fluoro-deoxy-glucose ([18F] FDG) accumulates in metabolically active tissue and this imaging tool has been used to demonstrate disease activity in clinically inactive joints . The evolving hybrid techniques PET–CT and PET–MRI are also discussed, and these have the potential to allow more precise anatomical localisation of the PET signal.

Imaging in the diagnosis of rheumatic disease

Imaging is having an increasing impact on the diagnosis of rheumatic conditions as diverse as RA, ankylosing spondylitis (AS), PsA, gout and large vessel vasculitis (LVV). The 1987 American College of Rheumatology (ACR) classification criteria for RA included an imaging parameter (XR erosions) as one of seven items, four of which were required for classification as RA. The revised 2010 American College of Rheumatology/European League Against Rheumatism (ACR/EULAR) criteria were devised in the context of an overall change in rheumatology practice that now aims at much earlier diagnosis, when patients may still be in the pre-erosive phase . A baseline snapshot of joint inflammation and damage can be provided by advanced imaging at presentation, ideally before the patient is started on DMARD therapy or biologics. The new criteria include a requirement for definite clinical synovitis in one or more joints, but this can be difficult to confirm where there are other causes for joint swelling (such as osteoarthritis) or joint tenderness (such as fibromyalgia). In these individuals, confirmation of synovitis by either US or MRI scanning could be helpful, and increasing evidence suggests that imaging synovitis may exist well below the level of clinical detection . Therefore what might have been apparent clinically as monoarthritis, may be revealed as polyarthritis using imaging, with telltale involvement of typical RA joints.

Imaging is making even more of a contribution to the diagnosis of AS and the other spondyloarthropathies (SpAs), again using MRI scanning. Susanne Pedersen and others describe the recently developed Assessment of Spondyloarthritis international Society (ASAS) classification criteria for axial and peripheral SpA, which include the MRI features of sacroiliitis. The time interval between the disease trigger and the development of radiographic change indicating sacroiliitis remains unknown in AS, but may be up to 10 years during which time, diagnostic criteria which previously relied on this XR evidence could not be fulfilled . MRI offers a major advance, especially with its ability to reveal inflammatory bone oedema in the sacroiliac region , which characteristically occurs much earlier. However, new ways to image the spine bring new difficulties related to diagnostic certainty, a commodity that is hard to find in this field at the best of times. How often, for example, can the finding of moderate bone oedema at one or both sacroiliac joints indicate degenerative disease or even occur in healthy controls? Dedicated work, particularly by the ASAS Outcome Measures in Rheumatology (OMERACT) group, has been put in to try to allow differentiation of the various causes of back pain on the basis of MRI features, bearing in mind that these must always be interpreted in conjunction with other clinical and serological evidence. It should be recalled that MRI becomes unnecessary in those patients who present late and already have radiographic sacroiliitis, so XR is still a required assessment tool.

Another quite separate area where imaging is making a diagnostic difference for the rheumatologist is in forms of LVV, particularly giant cell arteritis (GCA). Although not yet incorporated into classification or diagnostic criteria, both US and PET scanning hold promise for assisting in diagnosis of the GCA patient. Fred Joshua discusses the use of US in this situation and the “halo sign” which indicates the presence of inflammatory tissue around the lumen of the vessel . While the well-known US bugbear of operator variability may make clinicians wary of adopting this application immediately, features of accompanying polymylagia rheumatica (PMR) including for example subdeltoid bursitis, have now been included as part of the classification criteria for that condition . Whole-body [18F]-FDG PET has recently also been shown to have value in patients with a high clinical suspicion of GCA but a negative temporal artery biopsy , a definite clinical advance if these data can be replicated in other cohorts. The challenge for workers in this area is to provide clinicians with a technique that is accurate, informative and accessible in terms of resource and cost, while also being repeatable and useful in a practical sense to monitor therapy.

US is also making an impact as a potential tool to assist in the diagnosis of gout . The ‘double contour sign’ has been described and is regarded as indicating monosodium urate (MSU) crystals covering hyaline cartilage in a fine crystalline layer. Proponents have also described US abnormalities in asymptomatic hyperuricaemia , blurring the line between this condition and the fully fledged clinical syndrome. The additional advantage of US in gout is that it can be used to assist joint aspiration and therefore help reach the diagnostic gold standard by demonstration of MSU crystals. However, very few studies have compared US with other imaging modalities for the diagnosis of gout, and those cited above originate from a few centres of excellence; hence, the level of proof required for widespread clinical acceptance is yet to be reached. DECT is another new form of imaging that has exciting potential for gout diagnostics, having very high specificity for identification of tophaceous deposits ; but this also requires validation against other forms of imaging including high field MRI .