Ultrasonography of the metatarsophalangeal joints in rheumatoid arthritis: Comparison with magnetic resonance imaging, conventional radiography, and clinical examination

Authors


Abstract

Objective

To compare ultrasonography (US) with magnetic resonance imaging (MRI), conventional radiography, and clinical examination in the evaluation of bone destruction and signs of inflammation in the metatarsophalangeal (MTP) joints of patients with rheumatoid arthritis (RA).

Methods

Two hundred MTP joints of 40 patients with RA and 100 MTP joints of 20 healthy control subjects were assessed with B-mode US, contrast-enhanced MRI, conventional radiography, and clinical examination for signs of bone destruction and joint inflammation.

Results

With MRI considered the reference method, the sensitivity, specificity, and accuracy of US for the detection of bone erosions were 0.79, 0.97, and 0.96, respectively, while the corresponding values for radiography were 0.32, 0.98, and 0.93. The sensitivity, specificity, and accuracy of US for the detection of synovitis were 0.87, 0.74, and 0.79, while for clinical examination, the corresponding values were 0.43, 0.89, and 0.71. Erosive disease was identified in 26 patients by US, compared with 20 patients by MRI and 11 patients by radiography. Evaluation by US indicated signs of inflammation in 36 patients, while MRI and clinical examination revealed signs of inflammation in 31 patients and 20 patients, respectively. US and MRI volume-based gradings of synovitis showed intraclass correlation coefficients of 0.56–0.72 (P < 0.0001). The MRI and radiographic visualizations of US-detected bone changes were closely related to their size-based gradings on US.

Conclusion

US enables detection and grading of destructive and inflammatory changes in the MTP joints of patients with RA. By comparison with MRI, US was found to be markedly more sensitive and accurate than clinical examination and conventional radiography. Considering the early and frequent involvement of the MTP joints, evaluation of these joints by US may be of major clinical importance in RA.

Rheumatoid arthritis (RA) is a chronic, systemic autoimmune disease characterized by symmetric joint inflammation and destruction that often involves the small joints of the hands and feet, with progressive destruction, deformity, and disability of the joints. Criteria for diagnosing RA include the presence of bone destruction and signs of inflammation in the joints of the hands and feet (1). With the new, aggressive treatment strategies in early RA, there is a growing need for sensitive and specific methods for earlier detection of the signs of disease and monitoring of disease activity.

Visualization of bone erosions with radiography is essential in the diagnosis and monitoring of RA. However, radiography is not a sensitive imaging method in RA, especially in the early stages of the disease (2, 3). Clinical examination of the joints, which is one of the important tools in diagnosing and monitoring RA, is not an optimal outcome measure due to poor reproducibility and accuracy (4). There is increasing evidence that assessment of inflammatory and destructive joint changes in RA with magnetic resonance imaging (MRI) has high accuracy and strong diagnostic and prognostic value (5–9). Ultrasonography (US) is a new method of assessment of the small joints in RA (10). With the US equipment now available, including high-frequency linear-array transducers, it is possible to visualize and grade signs of inflammation and destruction in the small joints of the hands and feet (11).

The metatarsophalangeal (MTP) joints are among the joints affected early in the course of RA (12). Very few reports on the role of US in assessment of the MTP joints in RA exist (13), and none have included comparisons with other imaging techniques or with clinical examination of the joints. The aim of the present study was to compare US with MRI, conventional radiography, and clinical examination for the evaluation of bone destruction and signs of inflammation in the MTP joints in RA.

PATIENTS AND METHODS

Patients.

Two hundred MTP joints of 40 patients with RA and 100 MTP joints of 20 healthy control subjects were examined. The healthy control subjects had no history of joint symptoms or inflammatory activity in the joints. The RA patients were recruited from 2 outpatient hospital–based arthritis clinics. The diagnosis of RA in all patients fulfilled the 1987 American College of Rheumatology (formerly, the American Rheumatism Association) classification criteria (1). The inclusion criterion for the study was the presence of swelling or tenderness in at least 3 joints in the following regions: the metacarpophalangeal (MCP) joints, proximal interphalangeal joints of the fingers, or MTP joints. The exclusion criteria were severe deformity of the MTP joints and contraindications for MRI. Corticosteroid injections, either intramuscular or intraarticular, were not allowed within 1 month prior to study enrollment.

The median age of the RA patients and control subjects was 56 years (range 23–78 years) and 52 years (range 25–78 years), respectively. The ratio of women to men was 3:1 in the RA group and 4:1 in the control group. The median disease duration in patients with RA was 2 years (range 0–20 years). The group of patients with RA comprised 20 patients with early disease (disease duration <2 years) and 20 patients with established disease (disease duration >2 years). Twenty-five of the 40 patients with RA (14 in the group with established disease and 11 in those with early RA) were positive for IgM rheumatoid factor. Thirty-five patients with RA were being treated with disease-modifying antirheumatic drugs (26 receiving methotrexate, 7 receiving sulfasalazine, 1 receiving aurothiomalate, and 1 receiving D-penicillamine).

The study was conducted in accordance with the Declaration of Helsinki and approved by the local ethics committee. Signed informed consent was obtained from each participant.

The US, MRI, and clinical and laboratory examinations of the individual subjects were performed on the same day. Radiography was performed within 4 weeks of the other examinations.

Clinical examination.

Prior to undergoing evaluation by US, all patients were assessed for clinical disease activity (presence or absence of swelling and/or tenderness) in the MTP joints. The number and localization of swollen and/or tender joints were determined by a consultant rheumatologist.

Ultrasonography.

US was performed with a General Electric LOGIQ-500 unit using a 7–13-MHz linear-array transducer. B-mode US was done in the accessible aspects of the first through fifth MTP joints of the right foot: in the first MTP joint, the dorsal and medial aspects were evaluated (due to the presence of the sesamoid bones under the plantar surface of the joint), in the second to fourth MTP joints, the dorsal and plantar aspects were evaluated, and in the fifth MTP joint, the dorsal, lateral, and plantar aspects were evaluated. The assessed changes in the MTP joints were defined as follows: joint effusion was defined as the compressible anechoic intracapsular area, synovitis as a hypoechoic synovial thickening (noncompressible hypoechoic intracapsular area), and bone erosions as pathologic changes in the bone surface of the area adjacent to the joint, visualized in 2 planes.

The US examinations were primarily carried out by quadrant for the presence or absence of bone erosions (Figure 1), and by joint for the presence or absence of signs of inflammation (joint effusion and synovial thickening) (Figure 2). Furthermore, the findings on B-mode US were scored according to a semiquantitative scale (ranging 0–3) for bone destruction and for joint effusion. The scoring system was introduced in an earlier study (11) and also included a 0–3 scale for synovial thickening. For the present study, grade 3 synovial thickening was further divided into grades 3 and 4 to encompass the more advanced stages of synovial thickening. As in the previous study, grades 0–1 were considered to be physiologic and grades 2–3 (or 4) were considered to be pathologic.

Figure 1.

Detection of an erosion in the right foot of a patient with early rheumatoid arthritis, which was visualized by magnetic resonance imaging (MRI) (A and B) and ultrasonography (US) (C and D) but not by conventional radiography. In coronal (A) and axial (B) T1-weighted spin-echo MR images, in the proximal lateral quadrant of the fifth metatarsophalangeal (MTP) joint, a bone erosion (arrows) is evident. The first MTP joint is covered by a grid (A) (black perpendicular lines) to illustrate division of the joint into 4 quadrants: medial proximal, lateral proximal, medial distal, and lateral distal. B-mode US in the longitudinal and transverse planes of the lateral aspect of the fifth metatarsal head shows an erosive (grade 3) change (C and D) (arrows).

Figure 2.

Signs of synovitis on MRI and US of the right foot of a patient with established rheumatoid arthritis. Arrows indicate an intracapsular area with synovitis. The patient had no clinical signs of synovitis. Axial T1-weighted MR images were obtained before (A) and after (B) intravenous contrast injection. In the second MTP joint, a distinct postinjection contrast enhancement (grade 3 synovitis) is evident. B-mode US in the longitudinal plane of the dorsal aspect of the second MTP joint (C) shows signs of synovitis (grade 3). See Figure 1 for definitions.

All US examinations were performed by the same rheumatologist (MS), who was trained in the examination of the small joints of the hands and feet (11). The visualized changes were documented on hard-copy films and scored during the US examinations on score sheets printed beforehand. The US examiner was not informed of the results of the clinical and MRI examinations. The examination time was ∼15 minutes.

Magnetic resonance imaging.

On the same day as US, MRI examinations were subsequently performed by acquiring continuous axial and coronal pre- and post–Gd-DTPA T1-weighted, spin-echo MR sequences of the first through fifth MTP joints of the right foot, with a 1.0T Siemens Impact MR unit equipped with a transmit/receive knee coil. The Gd-DTPA was administered intravenously at a concentration of 0.1 mmoles/kg body weight. The examined subjects were placed in a supine position with the foot positioned in the coil and maintained in neutral position. The parameters of the applied coronal sequences were a repetition time (TR) of 500 ms, echo time (TE) of 15 ms, slice thickness of 3 mm, field of view (FOV) of 140 mm, and matrix of 192 × 256; for the axial sequences, the parameters were a TR of 450 ms, TE of 14 ms, slice thickness of 3 mm, FOV of 160 mm, and matrix of 320 × 512.

RA joint pathologies, as determined on MRI, were defined in accordance with the Outcome Measures in Rheumatoid Arthritis Clinical Trials (OMERACT) international recommendations (14). Bone erosions were defined as a sharply marginated bone lesion, with correct juxtaarticular localization and typical signal characteristics, which is visible in 2 planes, with a cortical break seen in at least 1 plane. Synovitis was defined as an area in the synovial compartment that shows above-normal post-gadolinium enhancement of a thickness greater than the width of the normal synovium. The following definition of joint effusion (not defined by the OMERACT consensus) was also used: an intraarticular area, with correct signal characteristics, that is not enhanced on post-gadolinium sequences and is surrounded by the synovial membrane.

The MRI examinations were carried out by quadrant for the presence or absence of bone erosions (Figure 1), and by joint for the presence or absence of signs of inflammation (joint effusion and synovial thickening) (Figure 2). Furthermore, synovitis was scored according to a semiquantitative scale (ranging 0–4) as suggested by Klarlund et al (15). All MRI examinations were evaluated on hard-copy films by the same observer (MØ), who was trained in evaluation of MR images in RA and who was blinded to the clinical and ultrasonographic data.

Conventional radiography.

Radiography of the right foot was performed within 4 weeks of the other examinations, using standard posteroanterior views. On radiography, bone erosions, defined as irregular punched-out areas of subchondral bone and irregular articular surfaces (16), were evaluated. The films were assessed for erosions by quadrant in the first through fifth MTP joints by an experienced musculoskeletal radiologist (EN), who was unaware of the findings of the other examinations.

Biochemical assessment.

A blood sample was obtained from each patient, by withdrawing the blood through a cannula inserted in the cubital vein, before the MRI examination. The blood samples were assessed to determine the serum C-reactive protein level. Information on the IgM rheumatoid factor status was obtained from the medical records of the patients.

Statistical analysis.

The agreement between imaging methods and in comparison with clinical examination findings was reported as overall agreement, which was defined as the proportion of exact agreements to the overall number of trials, with results expressed as a percentage. With MRI considered as the reference method, the accuracy, sensitivity, and specificity of B-mode US for the evaluation of bone destruction and signs of inflammation were calculated. Moreover, the agreement between US and MRI in terms of assessment of synovitis was evaluated with an intraclass correlation coefficient (ICC), by means of the statistical software package SPSS, release 9.0.1 (Chicago, IL).

RESULTS

Signs of bone destruction.

Comparison of US, MRI, and radiography.

Among the 40 patients with RA and 20 control subjects examined by US, MRI, and radiography, 1,200 quadrants in the 300 MTP (first through fifth) joints were assessed (Table 1). The overall agreement between US and MRI was 96% (1,153 of 1,200 quadrants), while the overall agreement between US and radiography was 93% (1,118 of 1,200 quadrants) (data not shown). Eleven percent of the examined MTP joints (134 of 1,200 quadrants) were scored as having erosions by at least one of the methods. Agreement on the presence of erosions between all methods was 21% (28 of 134 quadrants), while agreement between US and MRI on the detection of erosions was 54% (73 of 134 quadrants). By US only, 16% (22 quadrants) showed bone erosions, while by MRI only, 13% (18 quadrants) showed bone erosions. Most of the erosions on US that were not detected with MRI were located in the first and fifth MTP joints (16 of 22 quadrants), while the MRI-detected erosions that were not visualized with US were located predominantly in the second through fourth MTP joints (14 of 18 quadrants).

Table 1. Detection of bone erosions by US, MRI, and radiography*
Joint (total quadrants), groupQuadrants with erosionsQuadrants with no erosions on US, MRI, or RADAgreementSensitivitySpecificity
US + MRI + RADUS + MRIMRI + RADUS + RADUS onlyMRI onlyRAD onlyUS vs. MRI, no. (%)RAD vs. MRI, no. (%)USRADUSRAD
  • *

    Except where indicated otherwise, values are the number of quadrants. Accuracy, sensitivity, and specificity were determined with magnetic resonance imaging (MRI), considered to be the reference method. US = ultrasonography; RAD = plain radiography; MTP = metatarsophalangeal; RA = rheumatoid arthritis.

MTP1 (240)              
 Total11300931213228 (95)223 (93)0.820.060.960.99
 Established RA110005206273 (91)68 (85)    
 Early RA02003107476 (95)77 (96)    
 Healthy controls01001017779 (99)78 (97)    
MTP2 (240)              
 Total5903254212230 (96)219 (91)0.740.260.980.97
 Established RA59031545371 (89)59 (74)    
 Early RA00001007979 (99)80 (100)    
 Healthy controls00000008080 (100)80 (100)    
MTP3 (240)              
 Total9900143214235 (98)224 (93)0.820.410.990.99
 Established RA99001435475 (94)64 (80)    
 Early RA00000008080 (100)80 (100)    
 Healthy controls00000008080 (100)80 (100)    
MTP4 (240)              
 Total1400352225232 (97)229 (95)0.500.110.990.99
 Established RA14003426673 (91)70 (88)    
 Early RA00000107979 (99)79 (99)    
 Healthy controls00000008080 (100)80 (100)    
MTP5 (240)              
 Total121013714202228 (95)222 (92)0.920.540.950.97
 Established RA1210133044773 (91)63 (79)    
 Early RA00003007777 (96)80 (100)    
 Healthy controls00001107878 (97)79 (99)    
Total (1,200)              
 Total2845162218141,0661,153 (96)1,117 (93)0.790.320.970.98
 Established RA284216131513282365 (91)324 (81)    
 Early RA0200720389391 (98)396 (99)    
 Healthy controls0100211395397 (99)397 (99)    

All but 1 of the patients with RA had erosions detected by US in the first or fifth MTP joints. The bone changes registered as erosions by US in the healthy control subjects were also located in the first or fifth MTP joints.

Specificity and sensitivity of US and radiography, versus MRI.

With MRI as the reference method, the sensitivity of US for the detection of bone erosions in the toe joints was 0.79, while it was 0.32 for radiography (Table 1). The specificity of US as compared with MRI was 0.97, and that of radiography as compared with MRI was 0.98. The accuracy of US, i.e., the overall agreement between US and MRI for the detection of bone erosions in the toe joints, was 0.96, while the accuracy of radiography was 0.93.

Identification of subjects with erosive disease in the MTP joints using the different methods.

When erosive disease was defined as the presence of at least 1 bone erosion in the examined MTP joints, 11 patients were considered to have erosive disease on radiography, 17 on MRI, and 18 on US, among the group of patients with established RA. All of the patients who were classified as having erosive disease with radiography were also considered to have erosive disease on US and MRI. Among the group of patients with early RA, none had erosive disease on radiography, 3 had erosive disease on MRI, and 8 had erosive disease on US. One healthy person was considered to have erosive disease by all methods, while 1 healthy person had indications of erosive disease on MRI, and 1 healthy person showed erosive disease on US.

Comparison between grading of bone changes on US and detection of bone erosions with MRI and conventional radiography.

Grade 1 bone changes on US were not detected either by MRI or by radiography, whereas 47 of 75 grade 2 bone changes on US were visualized with MRI. All 26 grade 3 bone changes on US were visualized with MRI and almost half of them (14 of 26) with radiography (Table 2).

Table 2. MRI and radiographic detection of the bone changes visualized with US, as compared with the size-based scoring of the US findings*
US bone changesBone erosions on MRIBone erosions on RADNo bone erosions on MRI or RAD
  • *

    Values are the number of joints. US grading: grade 1 = irregularity of the bone surface without formation of a defect seen in 2 planes; grade 2 = formation of a defect in the surface of the bone seen in 2 planes; grade 3 = bone defect creating extensive bone destruction. See Table 1 for definitions.

Grade 1 (n = 21)0021
Grade 2 (n = 75)472022
Grade 3 (n = 26)26140

Signs of inflammation.

Overall agreement between US and MRI.

The 300 MTP (first through fifth) joints of the 40 patients with RA and 20 control subjects were evaluated by US, MRI, and clinical assessment for signs of inflammation. The overall agreement between US and MRI on the detection of synovitis in the MTP joints was 79% (237 of 300 MTP joints). Of the examined MTP joints with normal MRI findings, US revealed synovitis in 16% (47 of 300); MRI revealed synovitis in 5% (16 of 300) of the MTP joints with normal US findings (Table 3). US and MRI simultaneously showed effusion in only 1 of the 300 examined MTP joints. US alone revealed joint effusion in 27 MTP joints, while MRI alone showed joint effusion in 2 MTP joints.

Table 3. Detection of signs of synovitis by US, MRI, and clinical examination*
Joint (total joints), groupJoints with synovitisNo synovitis by any of the methodsAgreementSensitivitySpecificity
US + MRI + clinical assessmentUS + MRIMRI + clinical assessmentUS + clinical assessmentUS onlyMRI onlyClinical assessment onlyUS vs. MRI, no. (%)Clinical assessment vs. MRI, no. (%)USClinical assessmentUSClinical assessment
  • *

    Except where indicated otherwise, values are the number of joints. Accuracy (agreement), sensitivity, and specificity were determined with MRI, considered to be the reference method. See Table 1 for definitions.

MTP1 (60)              
 Total1220142511548 (80)30 (50)0.840.340.730.77
 Established RA9702000218 (90)11 (55)    
 Early RA3712121314 (70)8 (40)    
 Healthy controls06001301016 (80)11 (55)    
MTP2 (60)              
 Total9101112022546 (77)47 (78)0.950.500.680.93
 Established RA8510201317 (85)14 (70)    
 Early RA1201801711 (55)16 (80)    
 Healthy controls03002001518 (90)17 (85)    
MTP3 (60)              
 Total711166102846 (77)42 (70)0.900.400.700.83
 Established RA4812000517 (85)10 (50)    
 Early RA3104410711 (55)14 (70)    
 Healthy controls02002001618 (90)18 (90)    
MTP4 (60)              
 Total1110025302950 (83)45 (75)0.880.460.810.94
 Established RA8301300516 (80)16 (80)    
 Early RA3501130715 (75)11 (55)    
 Healthy controls02001001719 (95)18 (90)    
MTP5 (60)              
 Total95018423147 (78)48 (80)0.780.500.790.93
 Established RA7400401416 (80)15 (75)    
 Early RA20012311114 (70)15 (75)    
 Healthy controls01002101617 (85)18 (90)    
Total (300)              
 Total485631433135128237 (79)212 (71)0.870.430.740.89
 Established RA3627259021984 (84)66 (66)    
 Early RA12151916933565 (65)64 (64)    
 Healthy controls014008407488 (88)82 (82)    

Sensitivity and specificity of US and clinical examination, versus MRI.

The sensitivity of US for the detection of signs of synovitis, with T1-weighted MRI as the reference method, was 0.87, while the sensitivity of the clinical examination was 0.43 (Table 3). The specificity of US was 0.74 and that of the clinical examination was 0.89, as compared with MRI. The accuracy of US, i.e., the overall agreement between US and MRI, for the detection of signs of synovitis in the toe joints was 0.79, while the accuracy of clinical examination was 0.71.

Overall agreement between US and clinical examination.

The overall agreement on the presence or absence of signs of inflammation between US and clinical assessment was 63% (190 of 300 examined toe joints) (Table 4). Joint effusion and synovitis were visualized with B-mode US in 102 of 300 joints, whereas these signs of inflammation were detected in only 8 joints on the clinical assessment (as joint swelling and/or tenderness).

Table 4. Signs of inflammation (synovitis and/or joint effusion) by US compared with clinical joint assessment (swelling and/or tenderness) in the MTP joints of all study participants*
JointJoints with signs of inflammationJoints with no signs of inflammation on US or clinical assessmentNo. of joints examinedAgreement, US vs. clinical assessment, no. (%)
US + clinical assessmentUS onlyClinical assessment only
  • *

    Values are the number of joints. See Table 1 for definitions.

MTP118282126030 (50)
MTP214223216035 (58)
MTP315181266041 (68)
MTP414170296043 (72)
MTP59172326041 (68)
Total701028120300190 (63)

Correlation between US and MRI gradings of synovitis.

The ICCs for grading of synovitis on US and grading of synovitis on MRI in the respective joints were as follows: for the first MTP, the ICC was 0.56, for the second MTP, 0.67, for the third MTP, 0.69, for the fourth MTP, 0.72, and for the fifth MTP, 0.56 (P < 0.0001). The detailed results concerning agreement between scoring on US and scoring on MRI are presented in Table 5.

Table 5. Scoring of synovitis in all study participants by US and by MRI, using their respective volume-based scales (0–4)*
Grade on MRIGrade on US
01234
  • *

    Values are the number of joints. The numbers in boldface represent exact agreements between grades on US and grades on MRI. See Table 1 for definitions.

08892370
1324890
211214250
32113292
4000138

Identification of subjects with signs of joint inflammation by the different methods.

Signs of inflammation in 1 or more of the MTP joints were found in 31 patients on MRI, in 36 on US, and in 20 on clinical assessment. Among the healthy control subjects, who by definition (in accordance with the inclusion criteria) had no joint symptoms or clinical signs of inflammatory activity in the joints, 9 showed signs of inflammation on MRI and 10 on US, albeit at the lowest grade of synovial abnormality considered to be pathologic (grade 2).

DISCUSSION

The increasing use of US in rheumatology and lack of data on the US assessment of the toe joints in RA prompted us to examine the performance of US in imaging the MTP joints of a group of RA patients and a control group of healthy subjects. Radiographic studies have shown that the MTP joints are among the earliest affected in the course of RA (17). Based on experiences with assessment of signs of bone destruction and inflammation in RA finger joints, we have investigated the role of US in assessment of the MTP joints in RA, with MRI as the reference method.

Our study showed high overall agreement between US and MRI on assessment of bone erosions in the MTP joints. Moreover, US had higher sensitivity and equal specificity, compared with radiography, when MRI was considered the reference method. There was a clear site dependency of visualization of the destructive bone changes, with best results for US in the first and fifth MTP joints and poorer results in the second through fourth MTP joints. These results are consistent with those presented by Wakefield et al (18) for the MCP joints. In their study, on the radial, dorsal, and palmar aspects of the second MCP joint, where the imaging options for US are optimal, the results of assessment of bone destruction on US and MRI were equal.

The assessment of potential target joints, e.g., the first and fifth MTP joints, which was suggested earlier by Alarcón et al (19) and Schmidt (20), did not provide unequivocal results. Although all but 1 of the examined RA patients with erosive disease on US had evidence of bone erosions in the first or fifth MTP joints, all of the observed bone changes in the healthy persons also had the same localization.

The advantage of US compared with radiography and MRI was most striking in the group of patients with early disease (<2 years' disease duration). All of the patients with early RA had nonerosive disease on radiography, while almost half of them (8 of 20) already had bone erosions on US and only 3 on MRI. Thus, US showed more erosions than did MRI. However, this finding may have been influenced by the fact that our group used 3-mm–thick MR slices to assess the examined joints. Backhaus et al (21), who examined the finger joints of a heterogeneous group of patients with joint symptoms and used 1-mm–thick MR slices, showed no benefit of US over MRI when assessing bone destruction. At the same time, Backhaus et al used a 7.5-MHz transducer with a distance pad, an approach that may be considered inferior to the use of modern high-frequency transducers.

The detection of erosions on MRI and radiography was dependent on how advanced the bone changes were considered to be on US (Table 2). In particular, none of the mildest bone changes on US were visualized by the other methods, suggesting that US may be capable of showing bone changes earlier than both MRI and radiography, due to its higher-detail resolution.

Similar to the visualization of bone destruction, the assessment of signs of synovitis with B-mode US showed good correlation with T1-weighted, contrast-enhanced MRI. A considerable benefit of US as compared with clinical examination was shown, since US was found to have higher sensitivity with a certain loss of specificity, using MRI as the reference method.

B-mode US and T1-weighted, contrast-enhanced MRI correlated well, with regard to both detection (as expressed by overall agreement) and semiquantitative volume-based grading (as expressed by the ICC) of synovitis, suggesting that the visualized inflammatory changes were similar or identical. A proportion of the discrepancy between US and MRI may be explained by US visualization of mild osteoarthritic changes, which were not scored as synovitis on MRI due to minimal postcontrast enhancement. Osteoarthritis in the MTP joints may also be the reason that many of the control subjects showed signs of synovitis on both US and MRI. Furthermore, the higher frequency of detection of synovitis by US suggests that B-mode US shows both “active” and fibrotic synovial tissue in the examined joints. Our results are in accordance with those of Backhaus et al (21), who showed a higher frequency of visualizing synovitis in the finger joints with US than with MRI. The improved accuracy of US, compared with conventional methods, may be particularly clinically important in the toe joints, because these joints are very frequently involved and are considered among the most difficult to assess clinically (22).

US of the small joints of the hands and feet has not been studied with a histopathologic reference method. Contrast-enhanced, T1-weighted MRI has been shown to correlate well with the histopathologic and miniarthroscopic signs of synovitis in the finger and knee joints (7, 23, 24). On this basis, we found it justified to use MRI as an indirect reference method.

MRI revealed joint effusion in only a few cases; a contributing factor may have been that the analysis of the MR images on films was carried out without the possibility of adjusting the image contrast or magnifying the joint structures. Moreover, the MRI contrast agent may have diffused into the joint cavity, rendering the assessment of joint effusion with MRI more difficult. In contrast, US in our study enabled considerable magnification of the examined joints and the findings were analyzed in real time, both in motion and under compression, suggesting that the detection of small joint effusions is more effective with US compared with MRI. For joints in which accessibility is problematic in US or in which the amount of synovitis is minimal, joint effusion may be used as indirect proof of the presence of inflamed synovial tissue.

This report presents the results of evaluation of the RA MTP joints by B-mode US. Further analyses and quantification of the US-detected signs of inflammation by Doppler techniques, as suggested by several groups (25–28), are warranted. Our study has shown that the use of US allows detection and grading of destructive and inflammatory changes in the MTP joints of patients with RA, and the results are in high overall agreement with those of MRI. After comparison with MRI, our results revealed that US is markedly more sensitive and accurate than clinical examination and conventional radiography.

Acknowledgements

We thank Ms Susanne Østergaard for assistance with the medical images.

Ancillary