European Journal of Nuclear Medicine and Molecular Imaging

, 38:1939

Evolving role of FDG PET imaging in assessing joint disorders: a systematic review

Authors

  • Kathleen Carey
    • Department of Radiology, School of MedicineUniversity of Pennsylvania
  • Babak Saboury
    • Department of Radiology, School of MedicineUniversity of Pennsylvania
  • Sandip Basu
    • Department of Radiology, School of MedicineUniversity of Pennsylvania
  • Alex Brothers
    • Department of Radiology, School of MedicineUniversity of Pennsylvania
  • Alexis Ogdie
    • Department of Radiology, School of MedicineUniversity of Pennsylvania
  • Tom Werner
    • Department of Radiology, School of MedicineUniversity of Pennsylvania
  • Drew A. Torigian
    • Department of Radiology, School of MedicineUniversity of Pennsylvania
    • Department of Radiology, School of MedicineUniversity of Pennsylvania
    • Department of RadiologyHospital of the University of Pennsylvania
Review Article

DOI: 10.1007/s00259-011-1863-4

Cite this article as:
Carey, K., Saboury, B., Basu, S. et al. Eur J Nucl Med Mol Imaging (2011) 38: 1939. doi:10.1007/s00259-011-1863-4

Abstract

Assessing joint disorders has been a relatively recent and evolving application of 18F-2-fluoro-2-deoxy-D-glucose (FDG) positron emission tomography (PET) imaging. FDG is taken up by inflammatory cells, particularly when they are active as part of an ongoing inflammatory process. Hence FDG PET has been employed to assess a wide array of arthritic disorders. FDG PET imaging has been investigated in various joint diseases for diagnostic purposes, treatment monitoring, and as a prognostic indicator as in other disorders. In some of the diseases the ancillary findings in FDG PET have provided important clues about the underlying pathophysiology and pathogenesis processes. While substantial promise has been demonstrated in a number of studies, it is clear that the potential utility of PET in this clinical realm far outweighs that which has been established to date.

Keywords

18F-FluorodeoxyglucosePositron emission tomographyArthritisOsteoarthritisInflammation

Introduction

Inflammation and related inflammatory factors play a prominent role in the pathogenesis and progression of arthritis, in a range of joint diseases ranging from osteoarthritis (OA) to rheumatoid arthritis (RA). It is well known that 18F-2-fluoro-2-deoxy-D-glucose (FDG) is taken up by inflammatory cells, particularly when they are activated as part of the ongoing process [1]. As such, FDG is an ideal biologic marker along with positron emission tomography (PET) imaging for assessing arthritic disorders [2]. The application of FDG PET to evaluate joint diseases is rapidly increasing, and numerous studies are emerging which demonstrate increased uptake of FDG in inflamed joints compared to normal joints. FDG uptake in inflammatory cells is due to overexpression of glucose transporter type 1 (GLUT1) and type 3 (GLUT3), particularly in stimulated leukocytes including neutrophils and macrophages [3, 4]. This glucose analogue thus is taken up in the regions of inflammation proportional to increases in the metabolic activity of the inflammatory cells at those sites. FDG PET has several advantages over conventional radiologic and scintigraphic techniques, making it an ideal imaging modality for the assessment of joint disease in many settings. The strengths are further enhanced by hybrid PET/CT and PET/MRI [58].

In recent years, other imaging modalities have been proposed as an important modality for examining patients with joint disease. For instance, it has been demonstrated that gadolinium (Gd) contrast-enhanced MRI is capable of depicting knee synovitis and provides a potential quantitative marker of synovial inflammation. Thus, MRI has been proposed by some as the new gold standard for the assessment of synovitis. [In particular, the rate of early enhancement (REE) on dynamic Gd-enhanced MRI has been highly correlated with microscopic evidence of active inflammation including vessel proliferation and mononuclear leukocyte infiltration.] However, MRI, unlike FDG PET, is limited to the evaluation of large joints, making global disease assessment throughout the body a difficult task [9, 10]. Also, concerns have been raised about the safety of MRI contrast agents, and their ability to quantitatively assess disease activity. One of the most important characteristics of FDG PET imaging is its ability to provide quantitative measurements of activity in metabolism at the molecular level. In practice, the standardized uptake value (SUV) is measured within a 2-D or 3-D region of interest (ROI) that has been placed over the disease site or any tissue of interest in the body [11]. One of the concerns when generating SUVs in the past has been the impact of patient blood glucose. Of note however, unlike tracer uptake in malignant lesions which may be falsely reduced due to serum glucose levels, inflammatory disease assessment has not proved to be affected by serum glucose levels. In fact, high concentration levels of serum glucose up to 250 mg/dl do not appear to affect the uptake of FDG in inflammatory lesions [11, 12].Early detection is essential for the optimal management of joint inflammation and will lead to earlier therapeutic intervention and improved patient outcomes. Overall, FDG PET provides complementary information to structural imaging techniques such as plain film radiography, ultrasonography (US), CT, and MRI. In addition, the latter modalities allow for accurate quantification of the data generated by PET.

The utility of FDG PET for the assessment of arthritis has been substantiated by many reports in the literature. Therefore, the purpose of this article is to provide a systematic review of the current literature available on the application of FDG PET for the assessment of OA, RA, other arthropathies, and juxta-articular soft tissue inflammation, which have been described in three major categories in this scientific communication.

Search strategy and selection criteria

For this review, a PubMed database was generated for papers published after 1993 in German, French, or English. The search strategy used the free text (“Fluorodeoxyglucose F18” [Mesh] OR FDG) AND (“Positron-Emission Tomography” [Mesh] OR PET OR “PETCT” OR “PET-CT” OR “PET/CT” OR “PET_CT”) AND (Arthritis [Mesh] OR osteoarthritis [Mesh] OR osteoar* OR “rheumatoid arthritis” OR (joint AND inflammation)). In addition, specific searches were performed for the following subjects paired with “FDG,” “PET,” “positron emission”: tendonitis, bursitis, tendinopathy, enthesitis, synovitis, and degenerative joint disease. Bibliographies of identified articles were checked for other additional studies as well. Inclusion criteria in this review were: original articles assessing the value of PET in articular or juxta-articular disorders (whether in relation to prognosis, diagnosis, or response assessment) with provision of an adequate description of the materials and methods and results. There was no restriction based on the number of study subjects. Both reports of preliminary results and case report studies were not included in our review.

Arthritis

Rheumatoid arthritis

Diagnosis

RA is an autoimmune disease of unknown etiology which leads to chronic progressive systemic inflammation and synovial changes [13]. The synovium in RA is characterized by massive leukocyte infiltration, proliferation, and hypervascularization, leading to pannus formation [14]. For a long time, RA has been a challenging disease for diagnosis, because there is no gold standard for the diagnosis and many times the physicians rely on a “gestalt” of disease pattern derived from anecdotal clinical experiences [15]. However, classification systems and diagnostic criteria for the diagnosis of RA have been available since 1956. For much of that time, the criteria used in systematic-based approaches have lacked sensitivity and specificity for diagnosis at an early stage of disease; however, this has been largely rectified by the 2010 American College of Rheumatology (ACR) classification criteria, which have shown significantly higher sensitivity and specificity [16]. Current RA biomarkers do not seem to rectify this problem since many patients (up to 30%) may have no elevation in usual biomarkers such as rheumatoid factor (RF), anti-cyclic citrullinated peptide (anti-CCP), erythrocyte sedimentation rate (ESR), or C-reactive protein (CRP) [15, 17].

The paradigm of RA treatment currently rests upon early detection and initiation of aggressive therapy which has shown to improve clinical outcomes and disease-associated morbidity [1820].

Like MRI and diagnostic ultrasound (US) which are capable of detecting early synovitis prior to exam, FDG PET is able to detect changes in synovium at the molecular level. This observation was made when RA patients with cancer, undergoing FDG PET scans for assessing disease activity, were noted to demonstrate additional FDG uptake in the areas localized to their joint (see Fig. 1) [1]. As a result, studies evaluating FDG PET for the diagnosis of RA have been performed.
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-011-1863-4/MediaObjects/259_2011_1863_Fig1_HTML.gif
Fig. 1

FDG PET whole-body scan acquired without attenuation correction showing intense uptake in elbow joints in a patient with RA. (Reproduced with permission from [1])

A study by Lin and Sicuro recently demonstrated the diagnostic possibilities of FDG PET at a molecular level in RA. There are numerous proinflammatory cytokines, such as tumor necrosis factor (TNF)-alpha, interleukin (IL)-1, and IL-8, which are thought to be involved in RA, although they are not classically employed in the workup of this joint disease since they are technically difficult to measure. By using two rabbit models of acute inflammatory arthritis, these investigators found that FDG uptake correlates well with synovial fluid TNF-alpha concentration and provides an accurate means of detecting early disease in animal models of RA [21].

Likewise Matsui et al. demonstrated the ability of FDG PET to detect early synovial inflammation using a collagen-induced arthritis model in rats. On day 14 when clinical signs of RA developed, the swollen joints were detected by FDG PET as well. When histologic findings were compared with those of FDG PET, increased FDG uptake was seen to correspond with regions of bone destruction and pannus formation [22].

Disease burden assessment

The quantitative evaluation of RA has always been very challenging. Detection based on changes seen on plain film radiography is not always sensitive and structural alterations such as osseous erosions in general lag behind clinical improvements. Use of semiquantitative measures such as ESR and CRP are not always directly correlated with disease activity and are at best weak predictors of the clinical course of RA in any individual patient [23]. FDG PET, however, may be very useful in the individual quantitative assessment of RA disease activity and its course. FDG PET also allows the use of soft tissue inflammation throughout the body as a direct evidence for disease activity. Numerous studies have highlighted the possibility of employing this technique for this purpose (see Table 1).
Table 1

Studies investigating the role of FDG PET in assessing disease activity in RA

Author

Year

n

Age (years)

Joints examined

Study design

Imaging procedure

Interpretation of PET

Additional clinical parameters

Beckers et al. [24]

2004

21 (RA)

34–69 (mean)

Knee (all)

Prospective

PET

SUVlbm, visual identification, and cumulative SUV

DAS28

Wrist/MCP (n=13)

SDAI

Ankle/MTP-1 (n=8)

CRP

US (conventional and power Doppler)

ESR

HAQ

PGA

Beckers et al. [9]

2006

16 (RA)

48 (mean)

Knees

Prospective

PET

SUV, visual identification

Serum CRP

US

MMP-3 (n=3)

Dynamic MRI

 

Roivainen et al. [25]

2003

10

36 ± 13

Knee (n=9)

Prospective

PET

SUV

Methyl-[11C]choline

2 (RA)

Ankle (n=1)

MRI

7 (unspecified)

1 (AS)

Elzinga et al. [23]

2007

25

57 ± 7 (RA)

Hand, wrist

Prospective

PET

Visual identification

Physical exam

14 (RA)

55 ± 11 (OA)

6 (OA)

32 ± 11 (FM)

5 (FM)

 

Palmer et al. [27]

1995

12

35–75

Wrist

Prospective

PET

Visual identification, SUV, radiography

Paulus clinical index

9 (RA)

MRI

3 (PA)

Goerres et al. [33]

2006

7 (RA)

24–63

Multiple

Prospective

PET radiography

Visual assessment

CRP

4-point uptake scale

ESR

DAS28

Kubota et al. [17]

2009

18 (RA)

67 ± 11

Multiple

Prospective

PET/CT

SUV

WBC

Visual uptake scale

CRP

DAS28

Sato et al. [34]

2001

6 (RA)

61

Knee

Prospective

PET

SUV

ESR

CRP

MHAQ

VAS

Rosen et al. [44]

2006

150

60.4 ± 14.5

Spine

Retrospective

PET

Visual assessment scale

NA

CT

CT

Stumpe et al. [10]

2002

30

54

Spine

Prospective

PET

FDG accumulation

Bone biopsy

MRI

Culture

lbm lean body mass, DAS28 disease activity score that includes 28-joint counts, MCP metacarpophalangeal, SDAI simplified disease activity index, HAQ Health Assessment Questionnaire, PGA patient global assessment, MMP-3 metalloproteinase 3, FM fibromyalgia, PA psoriatic arthritis, MHAQ Modified Health Assessment Questionnaire, VAS visual analogue scale, NA not available

Beckers et al. compared FDG uptake in patients with RA to healthy control subjects and showed that SUV of the synovium in arthritic patients is correlated with US measurements of synovial thickness (see Fig. 2) [24]. Twenty-one patients with RA as defined by the American Rheumatism Association (ARA) criteria, three healthy subjects and ten patients, who were being staged for melanoma, underwent FDG PET imaging and were included in this report. The maximum SUV (SUVmax) of joints of both the upper and lower extremities were measured. A significant linear correlation between the SUVs and synovial thickness was found for all affected joints investigated except for those of the metatarsophalangeal (MTP) joints. The cumulative SUV (defined as the sum of SUVs) of all PET-positive joints, as well as the number of positive joints in each patient, correlated significantly with clinical parameters as well (including the number of tender and swollen joints, the patient’s and physician’s global assessment scores, biologic measures such as ESR and CRP, US findings such as number of US-positive and Doppler-positive joints, and composite indices such as 28-joint counts and the simplified disease activity index) [24].
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-011-1863-4/MediaObjects/259_2011_1863_Fig2_HTML.gif
Fig. 2

18F-FDG PET images of healthy control subject (a and b) and RA patient with active disease (c and d). a 3-D projection image of normal tracer distribution in knee. b Normal distribution in hand and wrist. c Rheumatoid knee. d Rheumatoid hand and wrist. (Reproduced with permission from [24])

Likewise, Roivainen et al. conducted a study to examine the usefulness of PET imaging in the assessment of synovial inflammation and compared MRI measurements of synovial thickness with the degree of FDG uptake. Ten patients with established inflammatory joint disease and with clinical signs of joint inflammation (nine patients with knee joint involvement and one patient with ankle joint involvement) were studied. The SUVs in the inflamed synovia were measured and then compared to synovial volumes measured on T1-weighted MR images. All patients showed a high accumulation of FDG at sites of clinically apparent arthritic change, and SUV correlated highly with synovial volume [25]. This is an important point since MRI assesses vascularity and capillary permeability, whereas FDG PET measures glucose utilization and metabolic activity in the inflammatory cells [26, 27].

RA is a disease which predominately affects small joints, and therefore Elzinga et al. examined FDG uptake in the hand and wrist joints in RA patients. A total of 25 patients were included (14 with RA, 6 with OA, and 5 with fibromyalgia, which served as the control group since there are no associated anatomic abnormalities in this disorder). All RA patients had active synovial swelling at the time of scanning and all OA patients had at least one joint involved. They found that 29% of RA joints with clinical signs of inflammation demonstrated increased FDG uptake compared to 6% of the OA patients and 0% of the fibromyalgia controls [23].

In a more recent report, the inflammatory activity of the synovium in RA was assessed by FDG PET/CT, which was found to be superior in delineating inflammatory changes when compared to conventional radiographic techniques [28]. This is especially important in RA when the FDG uptake by ligaments and tendinous attachments to bone must be discriminated from synovitis [29].

Lastly, FDG PET/CT may allow for highly accurate assessment of atlantoaxial instability in patients with RA. This is an important application since the cervical spine is a common site of synovitis in RA leading to nerve root compression and instability [30].

Response assessment

An important point has been raised by Brenner who stated if the cost-effectiveness of PET was high it would allow this modality to be translated to everyday clinical practice in rheumatology. PET, which is currently a relatively expensive modality, must provide additional critical information that is not attainable by the clinical or other laboratory-based assessments for its routine clinical application and potential impact. The critical approach would require demonstrating that PET can serve as a tool to assess changes in disease activity in response to therapy [13].

One of the first studies to conclude that contrast-enhanced MRI and FDG PET allow for comparison of efficacies of treatment was reported by Palmer et al. in 1995 [27]. Through performance of FDG PET of the wrist joints in 12 patients with inflammatory arthritis (9 of whom had RA), the authors were able to correlate clinical measures such as pain, tenderness, and swelling with the degree of inflammatory joint activity based on SUV measurement and the volume of enhancing pannus via MRI. Clinical and imaging parameters were acquired before and after 2 weeks of no treatment, before and after 2 weeks of treatment with nonsteroidal anti-inflammatory drugs (NSAIDS) or steroidal drugs, or before and after 12–14 weeks of treatment with methotrexate (MTX). FDG uptake and pannus volume decreased with therapy and were linearly correlated. In addition, both PET and MRI data were significantly correlated with improvement of clinical parameters in the wrist (p < 0.002). However, neither modality was associated with treatment success or failure as measured by the Paulus index [31]. (The Paulus index is a composite index for estimating improvement in RA in response to disease-modifying antirheumatic drugs (DMARDs). An improvement by 20% in each of 4 of 6 possible measures is required to demonstrate a Paulus 20 response. The measures are: improvement in tender and swollen joint counts, morning stiffness, patients’ disease assessment, physicians’ disease assessment, and ESR [31].)

Similarly, Polisson et al. hypothesized that measurements of synovial volume by MRI and synovial metabolism by FDG PET should mirror findings by standard clinical evaluation before and after the initiation of treatment with low-dose prednisone and MTX. They examined two patients with RA and active synovitis involving the carpus, and found that clinical parameters (including joint count, morning stiffness, and ESR) did parallel the improvement in synovial volume after treatment with low-dose prednisone or MTX (60 and 76%, respectively) after 14 weeks of treatment. In parallel with these results, the uptake of FDG was also reduced after therapy (66 and 69%, respectively). They therefore concluded that a volumetric measure of inflamed synovium (using MRI) and quantification of synovial glucose metabolism (using FDG PET) correlate well with clinical parameters [32].

Subsequently, Beckers et al. compared FDG PET (qualitatively and semiquantitatively), dynamic contrast-enhanced MRI, and US in the assessment of RA synovitis before and after treatment with anti-TNF-alpha. In addition, CRP and matrix metalloproteinase (MMP) 3 were evaluated as markers of inflammation. They evaluated 16 knees in 16 patients at baseline and 4 weeks after initiation of anti-TNF-alpha treatment. FDG PET was positive in 69%, MRI in 9%, and US in 75% of knees. As expected, PET-positive knees had significantly higher Gd contrast enhancement and greater synovial thickness than PET-negatives knees. Likewise, SUVs were significantly correlated with the MRI parameters, synovial thickness, and serum CRP and MMP-3 levels. Importantly, changes in SUVs after 4 weeks were correlated with changes in MRI parameters and in serum CRP and MMP-3 levels, but not with changes in synovial thickness. Synovial tissues, in turn, took 6 weeks to decrease in thickness [9]. This demonstrates the predictive capacity of molecular imaging with FDG PET which occurs earlier than morphologic changes and predicts the outcome.

In addition, Goerres et al. showed matched clinical and FDG PET results in 78% of joints of RA patients responding to infliximab therapy. Unlike other studies which used semiquantitative methods of FDG assessment, this study developed a qualitative visual assessment scale between 0 and 4 and involved the assessment of multiple joints in each patient. In 12% of joints, PET was able to document a decrease of disease activity, whereas clinical evaluation did not, and in 16% the clinical evaluation revealed a decrease of disease activity, which was not detected by visual scoring of PET images. This discrepancy may underscore the importance of semiquantitative measures [33].

Finally, assessment of treatments for RA using FDG PET has not been limited to medications but has even extended to alternative treatments such as acupuncture. Sato et al. recently compared joint SUVs, serum inflammatory markers, and patient subjective reports of improvement before and after ten treatments of acupuncture therapy. After treatment, physical examination revealed less swelling and tenderness to palpitation of the knee joint as well as a decrease in subjective scales of pain, although there was no significant change noted in serum ESR, serum CRP, synovial SUVmax, or synovial volume [34].

Treatment outcomes as a whole are difficult to assess in RA and rely heavily on the definition of success. This difficulty, paired with a limited number of studies available, makes it challenging to truly assess the utility of PET for treatment assessment. Measuring one joint vs overall joint activity in systemic diseases such as RA may provide greater insight into the impact of therapy [35]. Likewise, RA and other inflammatory processes have long suffered from a lack of objective criteria for response assessment. As a result, improvement in joint inflammation due to a therapeutic intervention may be masked by patients’ reports of pain or stiffness, which may be responsible for a disconnect between scales that involve patient reports of symptoms and objective PET findings.

Prognosis assessment

FDG PET has the possibility to serve as a prognostic indicator as well. In cancer, the pre-therapeutic SUV of tumor has been shown to predict both the outcome of chemotherapy and the disease-free survival and overall survival for many tumor types [6, 11]. It is hypothesized that more aggressive tumors are metabolically more active and thus show greater levels of FDG uptake. This type of correlation may be applicable to RA, and as such it will dramatically change the individual risk assessment, given the highly unpredictable course of RA. Thus, RA can serve as an example of how FDG PET may allow tailoring of therapy for the individual patient.

The first evidence for this concept came from Beckers et al. who showed that SUV in joints was correlated with positive Doppler signal. A positive Doppler signal, in turn, indicates neovascularization, a characteristic of aggressive synovitis in RA [24].

Recently, Kubota et al. performed a study using FDG PET/CT in patients with RA and noted that this modality is highly sensitive and accurately detected large joint inflammation (see Fig. 3). The authors concluded that FDG PET/CT may be helpful for early disease assessment of global RA activity including assessment for future high risk atlantoaxial instability. They found that patients with higher FDG uptake in their large joints were more likely to have positive FDG uptake in the atlantoaxial joint (see Fig. 4) [17, 33].
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-011-1863-4/MediaObjects/259_2011_1863_Fig3_HTML.gif
Fig. 3

A 71-year-old woman with a 7-year history of RA who experienced a recurrence. a Anterior and right lateral views of maximum intensity projection (MIP) images obtained using FDG PET/CT. b Axial images at the atlantoaxial joint (PET, CT, and fused image: top to bottom). Strong FDG uptake was seen in the atlantoaxial joint (score 3), the right and left axillary lymph nodes (3, 3), the right and left knees (4, 4), the right and left hips (1, 1), the right and left carpals (2, 2), the right and left wrists (3, 3), the right and left elbows (4, 3), and the right and left shoulders (4, 4). (Reproduced with permission from [17])

https://static-content.springer.com/image/art%3A10.1007%2Fs00259-011-1863-4/MediaObjects/259_2011_1863_Fig4_HTML.gif
Fig. 4

Coronal and transverse PET images through the level of the atlantoaxial joints of the cervical spine of a 57-year-old male patient with a 4-year history of RA and known involvement of the cervical spine. Symmetrically increased FDG uptake is seen at the atlantoaxial joints indicating synovial inflammation (long arrows). Increased uptake is also found in the acromioclavicular joint due to inflammatory disease (short arrow). (Reproduced with permission from [33])

Linn-Rasker et al. examined the predictive value of the distribution of inflamed joints at first presentation for the severity and disease course of RA using physical examination and various laboratory parameters. They concluded that the presence of arthritis in large joints, particularly arthritis in the knee joint, was predictive of a destructive disease course and severe RA. This finding was correlated with the serum CRP level, which was found to be a stronger predictor for destructive disease. However, as laboratory findings are not always specific and do not provide regional information, PET may indeed play an important role in assessment of prognosis [36].

Osteoarthritis

Diagnosis

While RA is a well characterized systemic inflammatory disorder, the pathogenesis of OA is poorly understood. This disorder is thought to result from both mechanical and biochemical effects of aging and degenerative changes [29, 37]. More recently, OA has been recognized as having an important inflammatory component. The disease process affects the entire joint structure, including the synovial membrane, subchondral bone, ligaments, and periarticular muscles [38]. It is thought, however, that the initiation of the process occurs at the chondrocyte level, and a number of current studies are designed to better understand the underlying molecular pathophysiology of this disorder [39, 40]. Much like in RA, the synovium in OA is infiltrated with a mixed population of inflammatory cells and this leads to synovial hyperplasia and hypertrophy. This is reflected in certain clinical symptoms of OA, namely swelling and tenderness of the joints [4, 41].

Currently, plain film radiography and clinical correlation serve as the standard practice for diagnosis and evaluation of the severity of OA, which is generally assessed by cartilage destruction [29]. MRI is also able to accurately assess cartilage volume. However, as in RA, PET may provide a means of detecting early metabolic changes in OA as a result of the high concentrations of cartilage glycosaminoglycans that are maintained by glucose metabolism in chondrocytes. The theory that PET can detect metabolic changes in glucose metabolism at the chondrocyte level was first recognized when the PET portion of FDG PET/CT scans of subjects referred for cancer imaging showed uptake at sites of OA that had been diagnosed on the CT portion of the examination as joints demonstrating OA [42, 43].

In addition, FDG PET has demonstrated increased radiotracer uptake in a wide variety of joints affected by OA including those of the spine. Rosen et al. assessed the relationship between the severity of degenerative joint disease (DJD) and anatomic variations of the spine by comparing FDG PET and CT images. They noted that incidental findings suggesting DJD were common on FDG PET primarily in the lumbosacral area, and that the severity of disease as assessed by FDG PET corresponded with that of CT findings. More importantly, they were able to differentiate metastatic spinal lesions from degenerative changes using FDG PET/CT. Of note, the correlation of the degree of FDG uptake to the severity of CT findings although significant was weak, and therefore the authors hypothesized that this observation is related to the occurrence of metabolic changes preceding structural changes, which are delayed for an extended period of time [44].

One of the first studies to show that the FDG PET findings precede symptomatic findings in OA was that of von Schulthess et al. (see Table 2). They retrospectively evaluated FDG accumulation in major joints in a large group of patients who underwent FDG PET imaging. FDG uptake was graded using a qualitative visual assessment scale. The authors found a strong correlation of FDG uptake with age, and attributed this to subclinical chronic inflammation in the early stages of OA. Of note, FDG uptake did not correlate with joint symptoms, and the authors felt that this observation demonstrated the high sensitivity of PET in the detection of joint disease [42].
Table 2

Studies investigating the role of FDG PET in assessing disease activity in OA

Author

Year

n

Age (years)

Joints examined

Cohort assembly

Imaging procedure

Interpretation of PET

Additional clinical parameters

Nakamura et al. [41]

2007

15 (RA)

71.5

Knee

Prospective

PET

SUV

NA

3 (healthy)

MRI

Wandler et al. [45]

2005

24 (shoulder disease)

58.5 ± 17.5

Shoulder

Prospective

PET

SUV

Questionnaire regarding joint pain

von Schulthess et al. [42]

2001

354 (no joint disease)

7–84

Multiple

Retrospective

PET

Visual

Questionnaire regarding joint pain

Semiquantitative

NA not available

Wandler et al. took this one step further and identified FDG uptake patterns on PET for specific types of joint disease. They observed FDG uptake in 21 patients after a clinical diagnosis of shoulder disease had been established by clinical examination. Of these 21 patients, 14 had clinical findings consistent with a specific diagnosis in the abnormal shoulder. It was primarily shown that diffuse shoulder uptake on FDG PET was associated with OA, and in addition numerous other findings were made. For instance, two of four patients with focal greater tuberosity uptake of FDG had findings of rotator cuff injury, and two of four patients with focal glenoid uptake had findings of a frozen shoulder. This study thus raised the possibility that FDG uptake patterns may elucidate the type of joint disease [45].

One of the potential problems with using PET in the assessment of arthritis is that there are few studies that have scientifically analyzed age-matched controls. Thus, it is not clear how much the normal aging process may contribute to the appearance of OA-related uptake. This point is brought up by von Schulthess et al. who recognized that without histologic evidence, there was no way to confirm that the strong correlation of FDG uptake and age could be attributed to osteoarthritic changes alone [42].

Another problem of using FDG PET in the diagnosis of OA-related joint damage is a possible lack of specificity. The reason for this is that although the etiology of arthritis may be multifactorial, the degree of FDG uptake may be similar for many processes. This was reported in a study by Elzinga et al. By assessing smaller joints, they found that the number of PET-positive joints in RA patients was significantly higher than that in OA patients (29 vs 6%). These numbers reflect the incidence of pathologic uptake among all joints studied, including those with no clinical evidence of inflammation. In RA and OA patients, respectively, 76 and 100% of joints with clinical evidence of inflammation showed increased FDG uptake. Conversely, no increased uptake was found in the joints of control patients with fibromyalgia. The authors concluded that FDG uptake alone is difficult to use for differentiation between RA and OA since secondary OA or OA-induced synovitis may falsely contribute to joint uptake in RA [23].

Additionally, it is difficult to differentiate septic from aseptic inflammatory uptake [6]. Some studies have even used FDG uptake as an indicator of infection and to exclude DJD in a given setting. For instance, Stumpe et al. concluded that FDG uptake was 100% sensitive and 100% specific for infection of a joint, and used this approach to differentiate between infectious and inflammatory processes [10].

Disease burden assessment

Although there are only a few studies which have assessed FDG PET as a means to measure the disease burden in OA, FDG PET has the potential to measure both disease burden and the extent of OA-related changes throughout joint structures in the body.

Nakamura et al. were one of the first groups to study FDG PET in the evaluation of OA (see Fig. 5). They established that FDG uptake is higher in knees involved by OA than in normal knees. More importantly, they also used fusion of PET and MR images to specify anatomic regions of FDG accumulation, and found that the medial femoral condyles had higher SUVs than the lateral femoral condyles. The posterior cruciate ligament (PCL), intercondylar notch, and osteophytes were also found to demonstrate increased FDG uptake. Thus, they concluded that PET provides specific in vivo regional information about the inflammation associated with OA [46].
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-011-1863-4/MediaObjects/259_2011_1863_Fig5_HTML.gif
Fig. 5

18F-FDG accumulation was found around the medial osteophyte in a patient with OA (arrow). a Fusion image. b MRI. c PET axial view. d PET coronal view. (Reproduced with permission from [46])

Despite the correlation of SUV with age and presumed osteoarthritic changes, there are few studies which have been able to show a correlation of SUV with clinical parameters. Parsons et al. [47], from our group, examined patients with OA and painful knee joints. Of the 18 knees reported to be painful, 78% had knee joint space SUVmax which exceeded the average SUVmax of control knees. Similarly, 83% of painful knees had synovial SUVmax that exceeded that of control knees. The difference between the two groups was found to be significant for both joint space and synovial uptake of FDG. The authors concluded that pain in the joint due to inflammation is associated with increased metabolic activity as seen on FDG PET. Further studies are needed to define the nature of the findings by FDG PET in patients with OA.

Ankylosing spondylitis

Ankylosing spondylitis (AS) is a chronic inflammatory joint disease which has a predilection for the spine, sacroiliac joints, and pelvic joints, leading eventually to spinal fusion. Aseptic spondylodiscitis is a classic complication of AS [48]. This entity has been shown to correlate with FDG uptake. Wendling et al. studied three patients with MRI documented lumbar aseptic spondylodiscitis requiring anti-TNF-alpha treatment. They were studied before and after 6 weeks of therapy using FDG PET in parallel with MRI and clinical evaluation. FDG uptake was evident at all sites of discitis in the three patients. Overall, PET was less capable of correctly identifying clinical response when compared to MRI, which was attributed to global FDG uptake in the red marrow of the lumbar spine [49].

Juvenile idiopathic arthritis

Juvenile idiopathic arthritis (JIA) is a childhood rheumatic disease defined as a clinically heterogeneous group of arthritides involving one or more joints with swelling, pain, or limited range of movement for at least 6 weeks, with age of onset younger than 16 years. It is the most common form of arthritis that affects children and differs significantly from the types of arthritis seen in adults. JIA can be divided into three categories: systemic, oligoarticular, and polyarticular. As implied by the name, the cause is unknown. However, with proper treatment, individuals can recover and lead normal lives [50]. As with other arthritic assessments described above, therapy in JIA is currently guided by changes detected via structural imaging, even though radiography is insensitive to detect acute erosive changes in the cartilage. This suggests that FDG PET may be a promising modality for optimal patient management through earlier detection of JIA.

Tateishi et al. conducted a retrospective study to examine the clinical validity of FDG PET in JIA. They found that joint tenderness and swelling had a positive association with abnormal FDG uptake in the joint [odds ratio (OR) 5.37 and 7.12, respectively]. The SUVmax correlated with the neutrophil count, serum CRP, ESR, and MMP-3. Joint erosion (OR 6.17), soft tissue swelling (OR 3.77), major joint involvement (OR 3.50), tenderness (OR 5.22), and serum CRP concentration (OR 1.81) were also associated with the FDG uptake as measured by SUVmax [51] (see Fig. 6).
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-011-1863-4/MediaObjects/259_2011_1863_Fig6_HTML.gif
Fig. 6

A 9-year-old girl with polyarticular JIA. Whole-body 18F-FDG PET performed at presentation shows abnormal uptake of six major joints (arrows). The SUVmax of the involved joints ranged from 0.9 (left elbow) to 2.3 (left hip and right ankle). Physical examination revealed tenderness and swelling in elbows, wrists, knees, and ankles as well as limited range of motion in elbows. (Reproduced with permission from [51])

Polymyalgia rheumatica

Polymyalgia rheumatica (PMR) is a clinical syndrome characterized by stiffness and proximal muscle pain, usually afflicting patients over 50 years of age. PMR is often diagnosed by exclusion of other disorders that can cause similar complaints and by its rapid response to low-dose corticosteroids. The exact nature of PMR is unknown, but it is thought to either be due to a vasculitis limited to the subclavian or axillary arteries or a synovitis of the shoulder and/or hip joints. While many studies have documented increased FDG uptake in the major vessels, several studies and case reports have recently demonstrated more consistent uptake in the large joints in these patients [52, 53].

In a prospective study by Blockmans et al., 35 patients with isolated PMR underwent FDG PET imaging before, 3 months after, and 6 months after treatment with corticosteroids. Total vascular score (TVS) was calculated as well as joint FDG uptake in the shoulders, hips, and spinous processes of the vertebrae. Overall, it was found that while mean TVS was low, FDG uptake in the shoulders, hips, and spinous processes was high (94, 89, and 51%, respectively) [52].

In another study by Blockmans et al., a similar finding was noted, although 32% of the control patients, who were matched for age and inflammatory parameters with 25 giant cell arteritis (GCA)/PMR patients, demonstrated increased FDG uptake [54]. In retrospect, five of these control patients were found to actually suffer from RA, reactive arthritis, or psoriatic arthritis, and therefore may not be considered normal controls. It was concluded that FDG uptake in the shoulders or hips has a low specificity for PMR, but a high sensitivity.

In both of these studies, FDG uptake was not useful for monitoring response to corticosteroid therapy. Instead, after 3 and 6 months, laboratory tests such serum CRP and ESR decreased and provided the same amount of information as FDG uptake levels which also decreased. Thus, the authors concluded that repetitive PET scans in PMR patients offered no advantage for assessing response. However, the power of both studies was severely limited. In the latter study, follow-up data were limited, as only one third of subjects enrolled into the study completed the planned three serial FDG PET scans, and there was higher FDG uptake noted 6 months after baseline in the shoulders, hips, and spinous processes of relapsing patients compared to those without relapse, although differences did not reach statistical significance [52, 54]. Thus, the value of FDG PET in PMR still needs to be explored by well-designed prospective studies in a large population of patients with this disorder.

Other etiologies of joint inflammation

In addition to the inflammatory disease entities described above, FDG PET has shown utility in a wide variety of additional arthritic disorders. Case reports cite the use of FDG PET imaging as a means for the diagnosis of synovitis, acne, pustulosis, hyperostosis, osteitis (SAPHO) syndrome [55], psoriatic arthritis [56], gouty tophus of the patella and thoracic spine [34, 57], and the assessment of herbal remedies for inflammatory joint disease [58].

Other arthropathies

Diabetic neuropathic arthropathy

FDG PET appears to be of significant value in patients with complicated diabetic joint disease. A study by Basu et al. demonstrated that FDG PET may help distinguish neuropathic osteoarthropathy in the setting of the complicated diabetic foot, which is diagnostically challenging, by using clinical assessment and structural imaging techniques. This prospective study was designed to investigate the usefulness of FDG PET in the complicated diabetic foot and specifically to determine if PET can show a difference between the uptake patterns in osteomyelitis vs Charcot’s neuropathy. In this study, 63 patients (17 patients with clinical diagnosis of Charcot’s neuropathy, 21 patients with uncomplicated diabetic foot, 5 patients with osteomyelitis secondary to a complicated diabetic foot, and 20 nondiabetic patients with normal lower extremities) were examined with FDG PET and MRI consecutively. Abnormal uptake patterns were identified visually by comparison to the contralateral foot and by detecting areas of focal abnormality in the foot. Quantitative assessment was performed by measuring SUVmax of the affected sites. Imaging findings were compared with histopathologic results and clinical outcome when possible [19] (see Fig. 7).
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-011-1863-4/MediaObjects/259_2011_1863_Fig7_HTML.gif
Fig. 7

FDG PET in one patient with Charcot’s neuroarthropathy with foot ulcer. Note the focal uptake in the ulcer and the relatively low-grade diffuse uptake in the neuropathic osteoarthropathy (arrows). (Reproduced with permission from [19])

A low degree of FDG uptake was noted in Charcot joints which was greater than that in normal joints. SUVmax in osteomyelitis of the foot as a complication of diabetes mellitus was significantly higher than in Charcot joints. The SUVmax in Charcot joints ranged from 0.7 to 2.4 (mean 1.3 ± 0.4), whereas the SUVmax of normal midfoot joints and of the uncomplicated diabetic foot ranged from 0.2 to 0.7 (mean 0.42 ± 0.12) and from 0.2 to 0.8 (mean 0.5 ± 0.16), respectively. Importantly, in the setting of a diabetic foot ulcer, PET was able to exclude the presence of osteomyelitis and had an overall accuracy and sensitivity for the diagnosis of Charcot foot greater than those of MRI. Overall sensitivity and accuracy in the diagnosis of Charcot foot were 100 and 93.8%, respectively, for FDG PET without coregistered CT and for MRI were 76.9 and 75%, respectively. FDG PET also showed focal abnormalities which suggested soft tissue inflammation (n=7), which were proven histopathologically to be secondary to infection [19].

Juxta-articular disease

While the studies described above have focused on PET as a tool for the assessment of joint pathology, PET also provides a means to study the juxta-articular components of the joint including tendons, ligaments, and bursae. Case reports demonstrate PET detection of large ischiogluteal bursitis [59], inflammation of the tarsal joints [60], and meniscal tears [61] (see Fig. 8).
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-011-1863-4/MediaObjects/259_2011_1863_Fig8_HTML.gif
Fig. 8

A 59-year-old woman with a history of thyroid cancer, thyroidectomy, and radioiodine therapy who underwent an MRI of the right knee for pain. MRI demonstrated a benign lesion in the medial femoral condyle, most likely a nonossifying fibroma or enchondroma (a, dashed arrow). Also, there was a tear of the anterior horn of the medial meniscus of the right knee, best appreciated on the sagittal fat-saturated T2-weighted images (a, solid arrow). The patient underwent a whole-body FDG PET scan for restaging. Additional images of the lower extremities were obtained in order to assess the lesion in the distal right femur seen on MRI. There was no abnormal FDG uptake corresponding to the benign lesion of the right femur. However, there was an increase in FDG uptake in the medial region of the right knee in an inverted C-shaped pattern seen on the axial images (b). After fusing the PET and MRI pictures, the area of increased FDG uptake corresponded to the peripheral border of the medial meniscus, suggestive of meniscal tear associated with synovitis (c). (Reproduced with permission from [61])

Goerres et al. found that during the PET assessment of joint disease increased FDG uptake was also found in the tendon sheaths and bursae [33]. FDG accumulation on PET has even been noted in multiple extra-articular cysts complicating RA [62].

Fusion images such as those provided via PET/CT and PET/MRI may provide improved means to differentiate juxta-articular from articular processes in complicated cases, thus allowing for better individualized treatments. For instance, Fischer et al. recently conducted a prospective study on 28 patients with nonspecific foot pain using FDG PET/CT. In 46% of patients, therapeutic decisions were altered after PET findings revealed important results, with diagnoses ranging from os trigonum syndrome to OA of several joints to insertional tendinopathy [63]. Similarly, case reports have demonstrated the use of FDG PET/MRI for the diagnosis of Achilles tendonitis (see Fig. 9) [64]. The use of fusion images has also helped to evaluate enthesitis in patients with spondyloarthritis and is speculated to lead to earlier diagnosis [60].
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-011-1863-4/MediaObjects/259_2011_1863_Fig9_HTML.gif
Fig. 9

a FDG PET images of a 67-year-old man with a history of worsening left ankle pain who underwent a PET study of the distal lower extremities for the evaluation of a suspected left ankle osteomyelitis. The images did not demonstrate any deep tissue or bone infection in the scanned region. However, abnormally increased FDG uptake in the posterior left ankle in a region of the Achilles tendon 2–3 cm superior to the left talus was noted (arrow). The findings are not consistent with osteomyelitis. Instead, Achilles tendonitis is suggested. b Short τ inversion recovery (STIR)-weighted MRI images of the left ankle demonstrated thickening and an abnormal signal (large arrows) in the Achilles tendon with associated edema in the pre-Achilles fat and a partial tear of the Achilles tendon (small arrow), representing tendonitis. The patient was treated conservatively for tendonitis with rest and a nonsteroidal anti-inflammatory drug. His symptoms improved significantly over the 2-year follow-up period. (Reproduced with permission from [64])

Finally PET may allow for better pain management for chronic conditions such as back pain where the source of the pain may be difficult to characterize. Various structures have been incriminated as possible sources of chronic lower back pain, including the posterior longitudinal ligament, dorsal root ganglia, dura, annular fibers, muscles of the lumbar spine, and the facet joints, and delineating the source may lead to more targeted and effective management. This was recently exemplified in a case report by Houseni et al. They describe a 56-year-old woman who had a history of a lung nodule and lower back pain who underwent an FDG PET scan for assessing the etiology of the lung nodule. However, the FDG PET scan was able to determine the source of the patient’s back pain as originating from the zygpophyseal joint (see Fig. 10) [65].
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-011-1863-4/MediaObjects/259_2011_1863_Fig10_HTML.gif
Fig. 10

Approximately 60 min after intravenous administration of 560 MBq (15 mCi) of FDG, images were acquired with a combined PET/CT camera. The images did not reveal any abnormal activity in the lungs. However, on PET images (upper row), significantly increased FDG activity is noted in the region of the posterior aspect of the spine at the level of L4-L5 (black arrow). On CT images (middle row, bone window), in the region of the facet joints at the level of L4-L5 vertebrae, there is a narrowing of joint spaces and facet hypertrophy with disruption of the normal bony architecture (white arrow). The combined PET/CT images (lower row) demonstrated that the two foci of abnormal FDG activity were located at the facet joints. The facet joint, which is also known as the zygapophyseal joint, has been implicated as a cause of chronic pain in 15–45% of patients with chronic lower back pain. Identifying the zygapophyseal joints as a source of a pain syndrome is a clinical challenge. Foci of hypermetabolism revealed by FDG PET have been reported in numerous inflammatory and infectious processes. However, intense FDG activity in facet joint arthropathy has not been reported in the English literature. FDG PET/CT might be able to play a role in the evaluation of lower back pain in such clinical settings. (Reproduced with permission from [65])

Future directions

The future outlook for FDG PET in the assessment and evaluation of joint disease and inflammation appears greatly promising. Detection of the joint disease at the molecular level via PET allows for early noninvasive in vivo visualization and quantification of disease activity. It enables improved understanding of key pathophysiologic processes in vivo and has the potential to provide prognostic information and improved response assessment to novel therapeutic agents which are urgently needed for many joint diseases, potentially leading to optimal individualized patient care. Despite this broad set of potential utilities, FDG PET is currently not used in the initial diagnosis of joint disease. Coregistered PET/CT or PET/MRI hybrid acquisition may be necessary in the early evaluation of joint disease, in order to compare the molecular information provided by PET with the morphologic data generated from CT and MR imaging.

One area of active research involves the development of novel radiotracers and molecular imaging agents other than FDG. Recent strategies to target specific processes in the arthritic synovium include antibodies and receptor antagonists for common inflammatory cytokines which are specifically upregulated in RA. For instance, radiolabeled leukocytes [66], IL-1 receptor antagonists [67], and anti-TNF-alpha monoclonal antibodies [68] have been used to visualize and quantify synovial inflammation[6668]. Likewise, radiolabeled targeting of inflammatory cells such as macrophages has shown some promise. For example, Kropholler et al. used PET to monitor macrophage migration into inflamed areas in RA using the radiotracer (R)-[11C]PK11195 which binds to benzodiazepine receptors on macrophages [69]. This is similar to the earlier usage of 99mTc-labeled nanocolloids to monitor inflammation [70].

Methyl-[11C]choline is also undergoing assessment as a possible molecular marker of inflammation. This is based on the knowledge that [11C]choline is a precursor for phospholipids in cell membranes and that the greatly enhanced number of cells in the arthritic synovium makes this radiotracer a marker for cellular proliferation. Roivainen et al. compared FDG and [11C]choline uptake on PET in the inflamed synovium of ten patients with inflammatory joint disease, and found that both the uptake of [11C]choline and the uptake of FDG were highly correlated with synovial volume; the highest correlation was observed with the Ki of [11C]choline (r = 0.954, p < 0.0001) [25].

Finally, there are numerous other PET radiotracers that are under development for assessing cancer that eventually may have applications in joint disorders. One of these includes 3′-deoxy-3′-[18F]fluorothymidine (FLT), which is currently used as a marker of cellular proliferation [71].

Conclusion

The purpose of this review article was to survey the existing literature available on the use of FDG PET in joint diseases. In reviewing the literature, it is clear that there is an enormous potential for PET in these disorders. Prospective longitudinal studies are necessary in order to better characterize the role of FDG PET in characterizing articular or juxta-articular structures at different stages of these disorders. Many of the previously conducted studies did not have histopathologic assessments to confirm the PET findings of inflammation, and more studies need to utilize quantitative measurements from FDG PET such as SUV or global metabolic assessment as opposed to qualitative visual assessment scales alone. This would enhance the role and reliability of this powerful imaging tool for measuring therapeutic response early in the disease course. One of the limitations of FDG PET and FDG PET/CT in assessing joint disorders is the inability to differentiate joint disorders due to RA from OA. Because of the nonspecificity of FDG, the potential use of FDG PET/CT is not useful for pinpointing the specific etiologic diagnosis of joint disorders but perhaps only for the monitoring of therapy. In addition, the reproducibility of PET-based measurements should be assessed in future studies. Overall, FDG PET is a powerful and robust diagnostic modality that holds great promise for the comprehensive assessment of joint disease for clinical and research purposes.

Conflicts of interest

None.

Copyright information

© Springer-Verlag 2011