Conventional radiography remains the “gold standard” for the evaluation of joint damage and its progression in rheumatoid arthritis (RA), but changes may not be evident radiographically until late in the disease course (1). Magnetic resonance imaging (MRI) has been used to study patients with early RA, but high cost and lack of standardization limit its use (2). Several studies have now shown that high-resolution ultrasound (US) of the metacarpophalangeal (MCP) joints can be very informative in the evaluation of RA of various durations (3–5). Erosions have been visualized in selected hand joints (second MCP) and foot joints (fifth metatarsophalageal [MTP]) (5). In fact, detection of more erosions per joint and per patient with US as compared with radiography was demonstrated by Wakefield et al in a recent study reported in Arthritis & Rheumatism (3).
The present pilot study was undertaken to assess the feasibility of performing US in selected (target) joints of patients with RA, and to determine if this technique offers additional information beyond that obtained with conventional radiography. MRI was chosen as a comparator method because of its high sensitivity. Our Institutional Review Board approved this study.
Ten unselected patients with RA according to the criteria of the American College of Rheumatology (formerly, the American Rheumatism Association) (6) and 5 healthy volunteers who, as a group, were of comparable age (mean 45 years), sex (80% women), and ethnicity (80% white) with the patients were studied. The duration of RA was <2 years in 2 patients, 3–5 years in 5, and >5 years in 3. Six of the 10 patients were positive for IgM rheumatoid factor. Five of the 10 were receiving a disease-modifying antirheumatic drug (methotrexate in 4 of the 5); 3 other patients were receiving anti–tumor necrosis factor therapy.
Joints selected for US imaging included the second and fifth MCP joints and the fifth MTP joints. These joints were selected on the basis of their likelihood of early involvement in RA as well as their easy accessibility with the US probe. The first MTP joints, not considered targets for involvement early in the course of RA, were also examined because of their easy accessibility with the US probe.
Conventional radiographs of the hands/wrists and feet were obtained. These included 3 views (anteroposterior, lateral, and oblique) obtained using standard techniques (Kilovolt peak 54, milliampere-second 2.5, fine detail film-screen combination). Coronal T1-weighted MR images of the hands and feet were obtained using a 1.5T body scanner (Signa; GE, Milwaukee, WI), dedicated extremity coils, and the following parameters: repetition time 500 msec, echo time 15 msec, 256 × 192 matrix, 3 mm thickness with 0.5 mm gap, field of view 20 cm). Finally, US of the joints as noted above was performed by one of us (RL-B), using 10–15 MHz high-resolution linear array transducers with small footprints. Individual scanning parameters were optimized for musculoskeletal detail for the US machines utilized (Sequoia; Accuson, Mountain View, CA and HDI 3000; Advanced Technologies Laboratories, Bothell, WA). In each subject, the 3 studies were performed between a few hours and 10 days apart.
Radiography was considered the gold standard and used to calculate the sensitivity, specificity, and predictive value of US. Radiographic films from patients and controls were coded and blinded and were read independently by 2 experienced rheumatologists (GSA, LWM) for erosions in the same joints imaged by US; the final erosion count recorded per joint and per patient was the average from the 2 readers. US and MR images from patients and controls were mixed and then read for erosions by a fellowship-trained musculoskeletal radiologist (RL-B) with experience in sonography. An erosion by sonography was defined as a cortical defect >2 mm in width with an irregular floor seen in longitudinal and transverse planes. An erosion by MRI was defined as a focus of decreased T1 marrow signal occurring in a juxtaarticular or subchondral distribution.
Table 1 shows the comparison of the findings with the 3 imaging modalities as well as the sensitivity, specificity, and overall accuracy of US and MRI compared with radiography. As in the study by Wakefield et al (3), all erosions found by radiography were also identified by US and MRI. US was superior to conventional radiography in the identification of erosions; moreover, it was superior to MRI in some instances, perhaps due to volume averaging of pixels given the technique used. Erosions detected by US, but not by MRI, were all small (2–3 mm). US, with in-plane axial resolution >1 mm at such superficial depth for the transducers used, would be better able to resolve such small cortical defects.
|Method||No. of erosions in joints examined||No. of patients with erosions in joints examined||Sensitivity, %†||Specificity, %†||Overall accuracy, %†|
We have shown that erosions can be identified by US in patients with RA of various durations and that the study can be performed in a busy outpatient clinic. The time to complete the study (and its costs) can be reduced substantially if only joints that are typically affected early in the course of the disease and/or whose surface is easily accessible with the US probe are examined. These joints include the second and fifth MCP joints and the fifth MTP joints. Of interest, no additional information was derived from the assessment of the first MTP joints, since no erosions were detected in these joints. US of target joints in RA patients meets the Outcome Measures in Rheumatology Clinical Trials filter requirements (truth discrimination and feasibility) (7). Demonstrating the presence of erosions early in the course of the disease, before radiographic findings are apparent, may allow the clinician to initiate more aggressive therapy, which in the long term may result in better patient outcomes. This is quite important considering the new therapies now available for RA (8–10).