Conventional radiography is the standard reference method for assessing destructive skeletal changes in rheumatoid arthritis (RA) and is part of the American College of Rheumatology (formerly, the American Rheumatism Association) 1987 revised criteria for the classification of RA (1). Magnetic resonance imaging (MRI) is more sensitive for the detection of synovitis and bone erosions in RA than are conventional radiography and clinical assessment (2–4). However, the specificity of the MRI findings in this context remains to be established.
A review of the literature describing erosions, synovitis, and bone marrow edema in wrist and metacarpophalangeal (MCP) joints on MRI revealed that very few studies have assessed the prevalence of these findings in healthy persons (Table 1). To our knowledge, only one reported study had the primary aim of describing signs of arthritis in healthy persons. Investigators in that study only examined signs of synovitis in wrist joints (5).
Table 1. Studies of MRI signs of rheumatoid arthritis (synovitis, bone erosions, and bone marrow edema) that included healthy controls*
|Author, year (ref.)||Joints||No. of healthy controls||MRI parameters||Ero||Syn||BME||T-syn||Cysts||dMRI||Radiograph||TJC||SJC||CRP||ESR||IgM-RF|
|Tan et al, 2003 (15)||MCP||28||Cor + ax T1-SE + T2-SE, 20 subjects without Gd, 8 subjects ± Gd||+||+||−||−||−||+||−||NA||NA||−||−||−|
|Partik et al, 2002 (5)||Wrist||18||Ax STIR, cor + ax T1-SE-fs ± Gd||−||+||−||−||−||−||−||NA||NA||−||−||−|
|Lindegaard et al, 2001 (28)||Wrist + MCP||3||Cor STIR, cor + ax T1-SE ± Gd||+||+||−||−||+||−||−||NA||NA||−||−||−|
|Klarlund et al, 1999 (29)||MCP||5||Cor + ax T1-SE ± Gd||+||+||−||−||−||−||+||NA||NA||+||+||+|
|McGonagle et al, 1999 (6)||MCP||31||Cor T1-SE, cor T2-SE-fs, no contrast||+||+||+||−||+||−||−||+||+||−||−||−|
|Offidani et al, 1998 (7)||MCP + PIP||12||Cor + sag T1-SE and T2*-GE, no contrast||+||+||+||+||+||−||+||+||+||−||−||−|
|Pierre-Jerome et al, 1997 (8)||Wrist||42||Ax T2-fast-SE, no contrast||+||+||+||+||+||−||−||+||+||−||−||−|
|Nakahara et al, 1996 (11)||Wrist||10||Cor T1 + T2-SE ± fs ± Gd||−||+||+||+||−||−||−||NA||NA||NA||NA||NA|
|Tonolli-Serabian et al, 1996 (12)||Wrist||10||Cor T1-SE ± Gd||+||+||−||−||+||−||+||NA||NA||−||−||−|
|Ostergaard et al, 1995 (30)||Wrist||3||Cor + ax T1-SE ± Gd||+||+||−||−||−||−||+||NA||NA||−||−||−|
|Jorgensen et al, 1993 (14)||Wrist||4||Cor T2*-GE + T1-SE ± Gd||+||+||−||+||+||−||−||NA||NA||−||−||−|
|Yanagawa et al, 1993 (13)||Wrist||10||Cor T1-SE ± Gd||−||+||−||−||−||−||+||NA||NA||−||−||−|
|Corvetta et al, 1992 (9)||Wrist + MCP||10||Cor + ax T1-SE + T2-SE, no contrast||−||+||NA||NA||NA||−||NA||NA||NA||NA||NA||NA|
|Beltran et al, 1987 (10)||Wrist||6||Cor T1-SE + T2-SE, no contrast||+||+||−||+||−||−||−||NA||NA||−||−||−|
Studies involving healthy controls are often performed without the use of contrast media (6–10), thus reducing the ability to detect arthritic changes, especially synovitis. Furthermore, in these studies, the MR images are most often obtained and evaluated in 1 plane only (6, 8, 10–14), increasing the risk of misinterpretations, particularly due to partial volume artifacts being wrongly interpreted as erosive changes or synovitis. Bone marrow edema was assessed in only a few studies (6–8, 11), and only 1 report describes dynamic contrast-enhanced MRI of MCP joints in the peripheral joints of healthy persons (15).
The objectives of the present study were 1) to explore the presence of bone changes resembling erosions, bone marrow edema, and soft tissue changes resembling synovitis in MCP and wrist joints of healthy persons on MR images obtained and evaluated according to the latest recommendations of the international initiative to harmonize outcome measures in rheumatology (Outcome Measures in Rheumatology Clinical Trials [OMERACT]) (16); and 2) to compare the MRI findings with conventional radiographic, clinical, and biochemical findings.
PATIENTS AND METHODS
- Top of page
- PATIENTS AND METHODS
Twenty-eight healthy persons (20 women and 8 men with a mean age of 47 years [range 24–67 years]) without present or prior arthralgias, joint swelling, or joint tenderness were included. All subjects were examined on the same day by MRI, conventional radiography, clinical examination for tender and swollen joints and morning joint stiffness, administration of the Health Assessment Questionnaire, and laboratory testing for IgM rheumatoid factor (IgM-RF) and serum C-reactive protein (CRP) levels.
Local ethics committee approval was obtained prior to beginning the study. All subjects were recruited from among staff members (and their families) of the Departments of Rheumatology and Radiology at Copenhagen University Hospital at Hvidovre, Denmark.
MRI of the wrist and second through fifth MCP joints of the dominant hand was performed using a 1.0T Siemens Impact whole body MRI unit (Siemens, Erlangen, Germany) equipped with a circular polarized transmit–receive coil. The subjects were positioned supine with the hand fixed above the head (superman position) in the center of the coil. T1-weighted spin-echo (T1-SE) sequences were obtained in the coronal and axial planes before and after intravenous administration of 0.2 ml/kg of body weight of contrast medium (Omniscan; Amersham Health, Copenhagen, Denmark). In addition, a STIR sequence and a T2-SE fat-suppressed sequence were obtained in the coronal plane before the contrast medium was administered.
Dynamic MRI measurements were acquired as a series of 20 T1-weighted fast low-angle shot MR images, each of 10 seconds' duration, and obtained in the same preselected coronal slice covering both the wrist and the second through fifth MCP joints. After the fourth acquisition, the contrast agent was injected intravenously as a bolus, and the following 16 acquisitions covered the initial 160 postcontrast seconds. The postcontrast T1-SE images were obtained subsequent to the dynamic MR images. The imaging parameters for the T1-SE images were as follows: repetition time (TR) 600 msec, echo time (TE) 15 msec, slice thickness 3 mm, field of view (FOV) 109 × 145 mm, matrix 192 × 256, number of acquisitions 2, and number of repetitions 1. For the STIR sequence, the following imaging parameters were used: TR 4,500 msec, TE 30 msec, inversion time 150 msec, slice thickness 3 mm, FOV 109 × 145 mm, matrix 182 × 256, number of acquisitions 3, and number of repetitions 1. The imaging parameters for the T2-SE fat-suppressed sequence were as follows: TR 4,500 msec, TE 96 msec, slice thickness 3 mm, FOV 109 × 145 mm, matrix 182 × 256, number of acquisitions 3, and number of repetitions 1. For the dynamic MRI, the imaging parameters were as follows: TR 40 msec, TE 12 msec, flip angle 70°, slice thickness 3 mm, FOV 105 × 140 mm, matrix 192 × 256, number of acquisitions 1, and number of repetitions 20.
Bone erosions, synovitis, and bone marrow edema identified on MRI were defined according to the latest OMERACT consensus (16). MRI bone erosions were evaluated separately in the metacarpal head and the phalangeal base of each MCP joint and in each wrist bone. An MRI bone erosion was defined as a sharply marginated lesion, with correct juxtaarticular localization, visible in 2 planes, with a cortical break seen in at least 1 plane. On T1-weighted images, this type of lesion will show loss of normal low signal intensity of cortical bone and loss of high signal intensity of the trabecular component of the bone.
Synovitis was evaluated on T1-weighted images and defined as an area in the synovial compartment with above-normal enhancement after intravenous administration of contrast agent and a thickness greater than the width of the normal synovium. Synovitis in MCP joints was evaluated on a “whole-joint basis” compared with the wrist, which was divided into 3 regions: the radioulnar area, the radiocarpal area, and the intracarpal–carpometacarpal area.
Early enhancement was defined as an increase in the signal intensity in the synovial compartment during the first 60 seconds after injection of contrast medium. Dynamic MRI was performed to assess signs of early enhancement in the synovial compartment. Every region of interest (ROI) was assessed by manual delineation of the synovial compartment in the MCP and wrist joints. All calculations of ROIs were done using a program (RIP) developed in-house that runs under the MATLAB environment (The MathWorks, Natick, MA).
MRI bone marrow edema was defined as a lesion within the trabecular bone with ill-defined margins that shows high signal intensity on STIR or T2-weighted fat-suppressed images and low signal intensity on T1-weighted images. MR images were evaluated by a person experienced in evaluating MR images in RA patients and who was unaware of the radiographic findings.
Plain films of both wrists and hands were obtained in the posterior–anterior, lateral, and Nørgaard projections (17). All radiographs were evaluated by the same person, who was experienced in musculoskeletal radiology and was unaware of the MRI findings. Radiographic bone erosions were evaluated separately in each wrist and MCP joint bone.
Clinical examination and biochemical parameters.
All clinical examinations were performed by the same rheumatologist. The examinations included assessment of joint swelling and joint tenderness as recommended by the European League Against Rheumatism (18). Standard laboratory tests, including serum CRP (Vitros crp-slide; Johnson & Johnson, Copenhagen, Denmark) and IgM-RF levels, were performed and analyzed at the central laboratory unit of Copenhagen University Hospital at Hvidovre.
- Top of page
- PATIENTS AND METHODS
Conventional radiography showed changes classified as bone erosions in 2 subjects (in the fifth metacarpal base and the radial aspect of the phalangeal base in the third MCP joint) (Table 2). In MCP joints, precontrast MRI revealed erosion-like changes in 5 of 224 bones (metacarpal head/phalangeal base) (Figure 1). None of these showed enhancement on postcontrast images, but occasionally, slightly increased water content was observed on STIR and T2-weighted fat-suppressed images.
Table 2. Erosion-like changes on MRI and conventional radiography, by anatomic location*
| ||No. of bones evaluated (all subjects/subjects with normal CRP and IgM-RF levels)||Erosion-like changes on MRI||Erosion-like changes on radiography†|
|All subjects||Subjects with normal CRP and IgM-RF levels||All subjects||Subjects with normal CRP and IgM-RF levels|
|Second MCP joint||56/54||2 (3.6)||1 (1.9)||0 (0.0)||0 (0.0)|
|Third MCP joint||56/54||1 (1.8)||0 (0.0)||1 (1.8)||1 (1.9)|
|Fifth MCP joint||56/54||2 (3.6)||1 (1.9)||0 (0.0)||0 (0.0)|
|Capitate||28/27||3 (10.7)||2 (7.4)||0 (0.0)||0 (0.0)|
|Lunate||28/27||4 (14.3)||3 (11.1)||0 (0.0)||0 (0.0)|
|Fifth metacarpal base||28/28||0 (0.0)||0 (0.0)||1 (3.6)||1 (3.6)|
|Fourth MCP joint, trapezium, trapezoid, hamate, pisiform, scaphoid, triquetrum, first through fourth metacarpal bases, radius, and ulna||392/350||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)|
|Total||644/594||12 (1.9)||7 (1.2)||2 (0.3)||2 (0.3)|
Figure 1. Erosion-like changes in the wrist and second metacarpophalangeal (MCP) joint. a + b and c + d, Axial + coronal T1-weighted spin-echo (T1-SE) images of the wrist before and after injection of contrast medium, respectively. Arrows depict a lesion classified as an erosion-like change in the capitate bone. On postcontrast images, this lesion shows enhancement. This was the only bone lesion showing enhancement after injection of contrast medium in the study. e + f and g + h, Axial + coronal T1-SE images of the second and third MCP joints before and after injection of contrast medium, respectively. Arrows depict a lesion classified as an erosion-like change at the second metacarpal head. In this example, the lesion is not enhanced after injection of contrast medium.
Download figure to PowerPoint
In wrist joints, erosion-like changes were found in 7 of 420 bones (3 in the capitate and 4 in the lunate) (Figure 1). Only 1 of these showed enhancement on postcontrast images and high signal on T2-weighted fat-suppressed and STIR images. Findings of erosion-like changes on MRI are shown in Table 2. All erosion-like changes were small and equivalent to grade 1 (0–10 scale) according to the 2002 OMERACT RA MRI scoring system (RAMRIS) (16). Bone marrow edema–like changes were not found in any of the joints examined.
Synovitis-like changes were visualized on T1-weighted contrast-enhanced spin-echo MRI in 10 of 112 MCP joints (Figure 2) (detected in 7 subjects) and in 8 of 84 wrist areas (Figure 3) (detected in 6 subjects, with an overlap of 4 subjects between the two groups). The observed areas with enhancement were low grade (OMERACT grade 1 [0–3 scale]) and only found in small areas, except in 1 wrist, where an OMERACT grade 2 enhancement was observed in the radiocarpal area (Figure 3). STIR and T2-weighted fat-suppressed images occasionally showed signs of slightly increased water content in the areas with postcontrast enhancement. Table 3 illustrates the distribution of the synovitis-like changes. Dynamic MRI revealed only minimal, but measurable, synovial enhancement in the examined areas. The results of dynamic MRI are shown in Figure 4, accompanied by examples from 2 RA patients.
Figure 2. Synovitis-like changes in the second and third MCP joints of a healthy subject (a–f), as well as an example from a patient with rheumatoid arthritis (RA) (g and h). a + b and c + d, Coronal + axial T1-SE images of the second through fifth MCP joints before and after injection of contrast medium, respectively. Arrows in c and d depict areas with enhancement in the synovial compartment of the second and third MCP joints. e, Coronal color-coded image showing enhancement within the region of interest (delineated with a red line) during the first 60 seconds after injection of contrast medium. Only minimal enhancement is seen when measured voxel by voxel (see Figure 3f for color coding scale). f, Computerized fusion of image in e and the coronal T1-weighted fast low-angle shot magnetic resonance image (not shown) where voxels with a signal intensity increase of more than 50% are color coded. Arrows depict the areas where increased signal was detected on the corresponding coronal T1-SE image (c). In these areas, no early enhancement is seen within the first 60 seconds after injection of contrast medium. g and h, Example (analogous to e and f, respectively) from an RA patient with synovitis (arrows in h) in the second through fifth MCP joints. The color scale is the same as that in images e and f and calibrated as shown in Figure 3. See Figure 1 for other definitions.
Download figure to PowerPoint
Figure 3. Synovitis-like changes in the wrist of a healthy subject (a–g), as well as an example from a patient with rheumatoid arthritis (RA) (h and i). a + b and c + d, Coronal + axial T1-SE images of the wrist and second through fifth MCP joints before and after injection of contrast medium, respectively. Arrows in c and d depict areas with enhancement in the synovial compartment of the wrist joint. e, Coronal T1-weighted fast low-angle shot magnetic resonance image, before injection of contrast medium, of the wrist and second through fifth MCP joints. The region of interest (ROI) is delineated with a red line. f, Coronal color-coded image analogous to e showing enhancement within the ROI during the first 60 seconds after injection of contrast medium. The scale indicates percentage enhancement. Only minimal enhancement is seen when measured voxel by voxel. g, Computerized fusion of images in e and f where voxels with a signal intensity increase of more than 50% are color coded. Arrows depict the areas where increased signal was detected on the corresponding coronal T1-SE image (c). In these areas, no early enhancement is seen within the first 60 seconds after injection of contrast medium. h and i, Example (analogous to f and g, respectively) from an RA patient with synovitis (arrows in i) in the radioulnar, radiocarpal, and intracarpal–carpometacarpal compartment of the wrist joint. Identical color scales were used for f and h. See Figure 1 for other definitions.
Download figure to PowerPoint
Table 3. Synovitis-like changes on MRI, by anatomic location*
| ||No. of joint areas evaluated (all subjects/subjects with normal CRP and IgM-RF levels)||No. of areas with synovitis-like changes on MRI|
|All subjects||Subjects with normal CRP and IgM-RF levels|
|Second MCP joint||28/25||6 (21.4)||3 (12.0)|
|Third MCP joint||28/25||4 (14.3)||1 (4.0)|
|Radioulnar area of wrist||28/28||2 (7.1)||2 (7.1)|
|Radiocarpal area of wrist||28/26||6 (21.4)||4 (15.4)|
|Fourth and fifth MCP joints, Intracarpal–carpometacarpal area of wrist||84/75||0 (0.0)||0 (0.0)|
|Total||196/179||18 (9.2)||10 (5.6)|
Figure 4. Graphic presentation of early (first 60 seconds after injection of contrast medium) enhancement in the wrist and second through fifth metacarpophalangeal joints in healthy subjects as well as examples from the 2 patients with rheumatoid arthritis (RA) illustrated in Figures 2 and 3. Broken line represents the mean enhancement (3.69) for the healthy subjects; dotted lines represent the mean ± 2SD (SD = 3.48). dMRI = dynamic contrast-enhanced magnetic resonance imaging.
Download figure to PowerPoint
The results of clinical examinations and biochemical tests were normal in all subjects, except for 1 subject with slightly elevated levels of IgM-RF (31 kIU/liter) and serum CRP (13 mg/liter) and 2 other subjects with slightly elevated levels of serum CRP (16 mg/liter and 13 mg/liter). The subject with elevated levels of both IgM-RF and serum CRP had synovitis-like changes in the second and third MCP joints (both OMERACT grade 1). Furthermore, the only erosion-like lesion with enhancement on postcontrast MR images was found in this subject. The 2 other subjects with slightly elevated serum CRP levels both had synovitis-like changes in the second and third MCP joints and in the radiocarpal region of the wrist joint (all OMERACT grade 1). Thus, 6 of the 10 MCP joints and 2 of the 8 wrist areas with synovitis-like changes were found in the 3 subjects with elevated serum CRP levels. Regarding the erosion-like changes, the 2 subjects with elevated serum CRP levels and the 1 subject with elevated levels of both serum CRP and IgM-RF accounted for 60% of the erosion-like changes in MCP bones and 28.6% of those in wrist bones.
- Top of page
- PATIENTS AND METHODS
The present study showed that changes resembling mild synovitis and small bone erosions are occasionally found in the MCP and wrist joints of healthy control subjects when standard MRI sequences are used. The synovitis-like changes were all at the lowest grade assessed on pre- and postcontrast T1-weighted spin-echo MR images, except in 1 subject, in whom an OMERACT grade 2 enhancement was observed in the wrist (Figure 3).
STIR and T2-weighted fat-suppressed images showed, although not as frequently as T1-weighted postcontrast spin-echo images, an area with minimally increased signal intensity in the synovium. This may suggest that STIR/T2-weighted fat-suppressed images have a higher specificity for detection of synovitis than do contrast-enhanced T1-weighted images. However, preliminary data suggest that STIR/T2-weighted fat-suppressed images may have a lower sensitivity for synovitis compared with postcontrast T1-weighted MRI (19).
As a consequence of the established close correlation between early synovial enhancement and the histopathologic degree of inflammation (20–23), the healthy subjects in this study underwent dynamic MRI. We found that healthy subjects had a measurable, but only minimal, early enhancement in the synovial compartment as assessed by dynamic MRI. This is consistent with data reported recently by Tan et al, showing some low-grade early enhancement in normal joints (15).
The early enhancement on dynamic MRI is almost exclusively dependent on capillary permeability and perfusion, while the late static enhancement is the result of a poorly defined mixture of perfusion and diffusion. It has been shown that the postgadolinium enhancement in less-inflamed joints is slower, but still present, compared with that in more intensely inflamed joints (24, 25). Consequently, marked enhancement could be seen in less-inflamed joints on late static postgadolinium images, but not on early dynamic images (24, 25). The same mechanism can explain the findings in healthy controls in our study. The available data indicate that dynamic MRI may provide a more specific distinction than conventional MRI between synovitis-like changes in healthy persons and synovitis in RA patients.
An explanation for finding synovitis-like changes in healthy persons could be related to employment. We expect that the mild synovitis-like changes observed could be induced by physical exertion, but further studies should be performed to examine this possibility. The fact that 1 subject had elevated levels of IgM-RF and serum CRP and 2 other subjects had elevated levels of serum CRP (despite no joint symptoms), coupled with the fact that these 3 subjects (11% of the study group) accounted for 44.4% of the synovitis-like changes, may suggest the presence of subclinical arthritis in some of the subjects.
Bone erosion–like changes were found in 1.9% of the bone areas assessed. All bone erosion–like changes were small, and the majority did not show increased signal intensity on postcontrast images. Only 1 erosion-like lesion showed increased signal on postcontrast images and high signal on T2-weighted fat-suppressed and STIR images, and this lesion was found in the subject with elevated levels of serum CRP and IgM-RF.
The finding of a few changes classified as radiographic erosions in healthy subjects, which has also been reported by other investigators (12), illustrates that even well-established methods are not 100% specific. There is an inherent, usually divergent, balance between the sensitivity and specificity of a test. In clinical trials, in which the diagnosis is established, a high sensitivity is often of fundamental importance. However, in a diagnostic setting, a high specificity may have the highest priority, since the diagnosis has important implications for classification and treatment.
MRI is known to be considerably more sensitive than conventional methods for the detection of RA joint pathologies, while its specificity has not been established. Thus, it is highly relevant to consider approaches to maximizing the specificity without seriously compromising the sensitivity. It could be a requirement that the erosion-like changes show enhanced signal on postcontrast MR images. Of course, this aspect is most important when only diminutive and few (one or two) bone changes are present. Furthermore, RA bone erosions will not always show enhancement (e.g., when inactive and fibrotic). Another approach to increasing the specificity with respect to bone changes would be to require the presence of more than 1 erosion (if no enhancement is seen within the erosion) or to exclude the capitate and lunate bones from evaluation, since ∼60% of all bone erosion–like changes were found in these bones, and these bones represent only 8.7% of all assessed bones.
The latter “solution” of excluding certain bones is, of course, subject to underestimation of possible erosions, because it is not known whether the prevalence of lesions in these sites is especially high in early RA. The exclusion of “difficult” bone areas is well known from radiographic methods (e.g., the Sharp/van der Heijde method ). Another method for further augmenting the specificity of bone lesions could be to require a lesion size above a certain threshold (e.g., greater than OMERACT grade 1), but this would probably reduce very significantly the sensitivity of MRI in early disease and therefore cannot be recommended. It should be mentioned that the RAMRIS already includes 1 requirement that increases the specificity:erosions shall be visible in 2 imaging planes.
In RA, bone marrow edema is a frequent finding and is associated with an increased (up to 6.5-fold) risk of developing erosive disease in patients with early disease (27). We did not find any signs of increased water content in the bone marrow of healthy subjects in the present study. This indicates that the presence of bone marrow edema is an indicator of a “real pathologic event.” However, bone marrow edema is not disease- specific, since it is a well-known finding in osteoarthritis, posttraumatic conditions, and RA.
In conclusion, we found that changes resembling mild synovitis or small bone erosions are occasionally found in the MCP and wrist joints of healthy controls. It cannot be ruled out that some of the findings represent subclinical arthritis. Signs of synovitis on dynamic MRI, signal enhancement within bone erosion–like changes, and signs of bone marrow edema appear rarely or are absent in healthy controls. These signs may prove to be very specific for distinguishing arthritic joints from normal joints.