Objectives To study the reliability and construct validity of ultrasound in interphalangeal finger joints affected by erosive osteoarthritis (EOA) and non-EOA with MRI as the reference method.
Methods 252 joints were examined by ultrasound, conventional radiography and clinical examination. Ultrasound was performed using a high-frequency linear transducer (12×18 MHz). On the same day, magnetic resonance images of 112 joints were obtained on a 3.0 T magnetic resonance unit. The ultrasound and MRI images were re-read independently by other readers unaware of the diagnosis, clinical and other imaging findings. Interobserver reliability was calculated by the percentage of exact agreement obtained and κ statistics. With MRI as the reference method, the sensitivity and specificity of ultrasound in detecting structural (bone erosions and osteophytes) and soft tissue (effusion and grey-scale synovitis) changes in EOA were calculated.
Results Ultrasound and MRI were found to be more sensitive in detecting erosions than conventional radiography in EOA. A high agreement between ultrasound and MRI in the assessment of bone erosions (77.7%), osteophytes (75.9%) and synovitis (86.5%) was present. A high percentage of inflammatory changes was found in EOA, and in smaller amount in non-EOA, both confirmed by MRI. Good interobserver reliability of ultrasound was obtained for all variables (all median κ >0.8).
Conclusion Grey-scale ultrasound proved to be a reliable and valid imaging technique to assess erosions and soft tissue changes, compared with MRI as a reference method in EOA.
Statistics from Altmetric.com
Erosive osteoarthritis (EOA) of the interphalangeal finger joints is considered an inflammatory subset of osteoarthritis of the fingers.1 2 Its destructive nature and inflammatory clinical presentation are unmistakable.3 Nevertheless, details on the pathogenesis of the disease remain unknown. Diagnosis of EOA is supported by the presence of subchondral erosions on conventional radiography. Radiography is capable of detecting progression over time,4 5 but the appearance of erosions on radiographs is delayed with respect to the clinical presentation. Ultrasound has been proved to be more sensitive in detecting bone erosions than conventional radiography in osteoarthritis6 and other rheumatic disorders.7 8 Therefore, ultrasound appears to be a promising technique for the diagnosis of EOA before conventional radiography changes become evident.
In addition, ultrasound is capable of visualising the soft tissue structures and changes (eg, synovial hyperproliferation, synovitis, tendon lesion). The high occurrence of these soft tissue changes in EOA, that is, grey-scale synovitis, power Doppler signal and effusion, has recently been reported, emphasising the inflammatory nature of the disease.6 However, no control group was assessed, and questions remained on the specificity.
Despite the appearance of reports of ultrasonographic assessment in EOA in the literature, the more basic issues of reliability and construct validity remain to be addressed. We have therefore undertaken a study in order to assess the reliability and construct validity of ultrasound in interphalangeal finger joints affected by EOA with MRI as the reference method. To address specificity issues, we have additionally examined a small control group of patients with non-EOA.
Patients and methods
Nine patients with EOA and five with non-erosive finger osteoarthritis were examined with ultrasound and MRI. All patients were recruited from the outpatient clinic and fulfilled the American College of Rheumatology criteria for hand osteoarthritis.9 Diagnosis of EOA was based upon the presence of at least one radiographically erosive interphalangeal finger joint. Six women and three men were included, with a median age of 60.9 years (range 48.5–72.5) and a median disease duration of 9.3 years (range 1.5–25.3). They had a median of six of 18 interphalangeal joints affected by erosions on conventional radiography. The non-EOA group comprised five women with a median age of 63.2 years (range 57.1–68.1) and a median disease duration of 11.3 years (range 3.3–14.3). None of them showed radiographic signs of erosions at the interphalangeal joints. Patients taking any immunomodulating agent (disease-modifying antirheumatic drugs, corticosteroids, biological agents) or chondroitin sulphate or glucosamine were excluded from the study. The intake of analgesics or anti-inflammatory drugs was allowed.
The study was conducted in accordance with the Declaration of Helsinki and was approved by the local ethics committee. Signed informed consent was obtained from each participant.
Ultrasound was performed with a MyLab 25 unit (Esaote, Genova, Italy) using a high-frequency (12–18 MHz) linear array transducer set at 18 MHz. All persons were examined by the same sonographer (RW, US1) and examination was repeated in 90 joints (from five patients) by another trained sonographer (PC, US2). Both US1 and US2 have a rheumatological background and were blinded to diagnosis and other imaging findings. The dorsal and volar aspects of each joint were examined in a longitudinal and transverse view. All views were obtained with the hands in a neutral position, but extended and flexed as required to visualise pathology. All proximal interphalangeal (PIP) and distal interphalangeal (DIP) joints of both hands (including interphalangeal 1) were assessed for soft tissue changes (grey-scale synovitis or synovial hyperproliferation and joint effusion), and structural changes (bone proliferations or osteophytes, and bone erosions). Ultrasound definitions were employed according to the Outcome Measures in Rheumatology Clinical Trials (OMERACT) definitions.10 Ultrasound findings are illustrated in figure 1.
Magnetic resonance imaging
On the same day, magnetic resonance images of one hand with erosive features was taken. In case both hands were showing erosions, the most painful hand was selected (level of pain registered on a 100-mm visual analogue scale for both hands individually). The second to fifth PIP and DIP joints were imaged in a receive only, four-channel loop coil (diameter 70 mm; Siemens Medical, Erlangen, Germany), firmly fixated on the dorsal side of the fingers. The patients were in the supine position with the hand in the coil along the femur. The sequences obtained during MRI examination included: coronal T1-weighted turbo spin echo (slice thickness: 1 mm; repetition time/echo time (TR/TE): 573/11 ms); coronal fat saturated T2-weighted images (slice thickness: 1 mm; TR/TE: 3770/62 ms); T2-weighted double-echo steady-state three-dimensional sequences (slice thickness: 0.4 mm; TR/TE: 14.28/5.03 ms); transverse T1-weighted turbo spin echo (slice thickness: 1 mm; TR/TE: 2130/11 ms). Immediately after the Gd-DTPA administration (0.1 mmol/kg body weight intravenously; Magnevist; Bayer Healthcare Pharmaceuticals, Berlin, Germany) coronal dynamic turboflash (slice thickness: 3.5 mm; TR/TE: 1100/3.24 ms) and fat saturated T1-weighted (slice thickness: 1 mm; TE/TR: 865/10 ms) and transverse T1-weighted spin echo magnetic resonance sequences (slice thickness: 1 mm; TR/TE: 2940/10 ms) were taken. The MRI assessments were obtained from a 3.0 T MRI unit (Magnetom Trio; Siemens Medical).
The images were assessed for the presence of bone erosions, osteophytes, synovitis, effusion and bone marrow oedema by a radiologist experienced in musculoskeletal MRI (LJ, MRI1), who was blinded to diagnosis, clinical and other imaging findings (figures 1 and 2). OMERACT definitions for rheumatoid arthritis (RA) joint pathologies were applied.11 The acquired images were re-interpreted by another experienced radiologist, also blinded to diagnosis, clinical data and other imaging findings (VL, MRI2).
The number of finger joints assessed by MRI for the presence of synovitis and structural changes differ because one patient (from the non-EOA group) was excluded from the post-gadolinium MRI examination due to contrast allergy.
Posteroanterior radiographs were taken of both hands and were assessed for the presence of osteophytes and bone erosions by an experienced rheumatologist (GV) blinded to clinical and other imaging findings.
All patients underwent clinical examination before ultrasound to determine tenderness, palpable effusion and nodules of all interphalangeal joints of both hands.
The low dependency between interphalangeal joints in the same patient (derived from the correlation matrix calculated) allowed the analysis to be performed on a joint level instead of a patient level. Interobserver reliability was calculated by the percentage of exact agreement (PEA) and unweighted κ values (with 95% CI) between both US1 and US2, and MRI1 and MRI2. κ Values below 0.20 were considered poor, 0.21–0.40 fair, 0.41–0.60 moderate, 0.61–0.80 good and 0.81–1.00 very good.12 MRI was used as the standard reference method for the calculation of the sensitivity and specificity of ultrasound, conventional radiography and clinical examination. The statistical software package SPSS 15.0 for Windows was used.
A total of 252 joints (162 EOA; 90 non-EOA) were examined by ultrasound, conventional radiography and clinical examination.
The observations by ultrasound, conventional radiography and clinical examination in EOA and non-EOA are shown in the supplementary file (p. 1), available online only.
Observations by ultrasound, MRI and conventional radiography in the MRI-examined hand
A total of 112 joints (72 EOA; 40 non-EOA) were examined by MRI. The observations by ultrasound, MRI and conventional radiography in the MRI examined hand are shown in table 1.
More erosions were seen by ultrasound and MRI compared with conventional radiography in all DIP and PIP joints, and in the subgroups EOA and non-EOA. Osteophytes were detected less frequently by MRI and conventional radiography than ultrasound, except for DIP joints in the EOA group, but very high percentages were reported there by all three methods (both >80%). Synovitis was seen slightly more frequently in PIP joints by ultrasound compared with MRI, both in EOA and non-EOA. In DIP joints, the opposite was true for EOA patients. In non-erosive DIP joints, no synovitis was detected by MRI and in only one joint by ultrasound. Joint effusion was seen in relatively high numbers in PIP joints in both EOA and non-EOA by MRI and ultrasound (all >70%). Bone marrow oedema was present in high percentages, especially in DIP joints of EOA patients (81%).
Interobserver reliability of ultrasound and MRI
PEA and κ values for ultrasound and MRI readings are listed in table 2. Excellent agreement was achieved overall for the sonographic detection of erosions (PEA 96%, κ=0.897), osteophytes (PEA 92%, κ=0.833), synovitis (PEA 98%; κ=0.931) and effusion (PEA 92.4%, κ=0.841). The reproducibility of MRI was very good for erosions (PEA 90%, κ=0.758), good for the detection of synovitis (PEA 85%, κ=0.581), but only moderate for effusion (PEA 75%; κ=0.500) and even poor for osteophytes (PEA 61%; κ=0.152).
Agreement between ultrasound and MRI
The agreements between ultrasound and MRI for bony and inflammatory changes are listed in table 2. The exact agreement between MRI and ultrasound for erosions and osteophytes was good (78% and 76%, respectively). The κ values were moderate (0.55 and 0.51). Erosions were detected in 31 joints in EOA patients by both MRI and ultrasound. In 16 joints, MRI was the only modality that detected the erosion(s). Six erosive joints were detected exclusively by ultrasound. In the non-EOA group, nine erosions were detected by MRI. Four out of seven erosions found by ultrasound were confirmed by MRI.
The exact agreement between ultrasound and MRI for inflammatory changes was good for synovitis (87%) and slightly lower for effusion (73%) with moderate κ values (κ=0.552 for synovitis and κ=0.410 for effusion).
The agreement between ultrasound and MRI and other imaging constructs, conventional radiography and clinical examination are given in the supplementary file (p. 2), available online only.
Sensitivities and specificities of ultrasonography with MRI as the standard reference method
The sensitivity and specificity of ultrasound with MRI as the standard reference method are listed in table 3. The specificity of ultrasound was high for bone erosions and for synovitis. For detecting synovitis, ultrasound showed high sensitivity for PIP joints but poor sensitivity in DIP joints. Specificity was high for both groups. It was noted that ultrasound was more sensitive in PIP compared with DIP joints, except for erosions. The sensitivities and specificities of conventional radiography and clinical examination with MRI as the reference method are shown in the supplementary file (p. 4), available online only.
Several paper have recently reported ultrasonographic findings in EOA but, in fact, no comparative study with other imaging constructs has been performed.6 13 14 In this study, we have investigated the construct validity of ultrasound in EOA and non-EOA, with MRI as the reference method. We found a high agreement between ultrasound and MRI in the assessment of bone erosions, osteophytes and grey-scale synovitis or synovial hyperproliferation (median agreement, all >70%).
Our study showed that ultrasound detected true erosions when compared with MRI and only a small number of false positives were seen in PIP joints. A previous study demonstrated an increase in sensitivity compared with conventional radiography when using MRI, especially for marginal erosions.15 This current study and other studies also showed an increase in the sensitivity of ultrasound compared with conventional radiography.6 13 This can be explained by the fact that conventional radiography is only able to show erosions contacted tangentially by the beam, whereas ultrasound and MRI might visualise hidden areas. In our study both central and marginal erosions were considered.
This study confirmed the evidence of grey-scale synovitis being present in EOA joints. However, ultrasound seemed less sensitive for DIP joints than MRI. This is probably explained by the more difficult visualisation of DIP joints by ultrasound. Our findings are in agreement with the study performed by Vlychou et al6 who reported frequencies of thickened synovium in 24.1% of EOA joints.
High percentages of effusion were detected in both EOA and non-EOA joints by MRI and ultrasound. The percentages of effusion found in our study are a multitude of previously reported numbers.6 The lowest agreement between the imaging tools was observed in assessing the presence of effusion (ultrasound vs MRI: 73%; κ=0.41). This might be caused because no clear cut-off levels were used to interpret the ultrasound images; instead, this feature was rather assessed subjectively for presence or absence.
There is only little evidence about the validity of ultrasound in detecting osteophytes. Although not the aim of this study, it showed that ultrasound detected more osteophytes than conventional radiography. This is in agreement with previous studies.16 17 This observation is most likely a reflection of the multiplanar imaging capability of ultrasound and, in particular, its ability to image joints in the dorsal longitudinal plane, which is not routinely done with conventional radiography. We could only obtain moderate agreement between ultrasound and conventional radiography, which is in line with a previous report.17 MRI is less optimal to visualise osteophytes due to the signal void of densely packed calcium as is present in osteophytes. As such, it is debatable whether MRI may be considered an accurate reference method for this finding. This probably explains the low sensitivity and specificity of ultrasound in this study. CT, although not better validated than MRI, could be potentially superior as a reference method for these often small and marginal bone changes.
Ultrasound is often criticised for being operator dependent. Our analysis of interobserver reliability showed excellent agreement for both destructive and inflammatory changes (all median κ>0.8). The interobserver reliability of MRI was high for erosions. Very good agreement has been reported by others.15
In this study we assessed a small control group of radiographic non-EOA patients. Erosions and inflammatory changes were also detected by ultrasound in this group, although in smaller percentages than in the EOA group. The sonographic erosions were confirmed to be true erosions by MRI in four of seven joints. Similar observations were made for synovitis. Previous studies did not include a control or non-erosive patient group.
There are some limitations to this study. Both the patient and control groups assessed were rather small; however, all interphalangeal joints of both hands of each patient were examined by ultrasound and all interphalangeal joints of one hand by MRI, instead of selecting just one target joint per patient.
Another potential limitation was that, in common with other clinical studies of hand osteoarthritis hitherto, no explicit diagnostic criteria are available for EOA. Patients were considered to have EOA whenever at least one erosive interphalangeal joint was present on radiographs. However, the minimal number of erosive joints that needs to be present is still undecided. Moreover, there are no specific criteria for the diagnosis of erosions in osteoarthritis by MRI and as a result we had to rely on the definition of erosions produced by the OMERACT group for RA.11 There are no arguments that these features could not be applied to osteoarthritis joints.
The presence of both erosions and soft tissue changes in radiographic non-EOA illustrates the high sensitivity of ultrasound and MRI. Previous reports in RA have demonstrated that sonographic erosions became true cortical radiographic defects in time.18 However, other studies in RA reported on the regression of erosions by conventional radiography and MRI.19 20 Whether the erosions found by MRI and ultrasound will become radiographically visible in osteoarthritis should thus be explored in longitudinal studies.
The relationship between conventional osteoarthritis and EOA has been the subject of considerable debate.3 21 22 It has been stated by others that the debate about EOA can be simplified by stating that dividing osteoarthritis into erosive and non-erosive on the basis of conventional radiography is somewhat artificial because many more erosions were seen on MRI in both the erosive and non-erosive radiographic form of osteoarthritis.15 Likewise, we showed that soft tissue changes are present both in EOA and in non-EOA; however, in a minor degree, but confirmed by MRI. The question about the nature and specificity of EOA will, however, not be resolved by the nature of tissue changes as detected by different imaging techniques. What drives the ‘inflammatory clinic’ of EOA remains unknown and should be the subject of future research.
In conclusion, we found that ultrasound is a valid imaging technique to assess erosions and soft tissue changes, such as effusion and synovial hyperproliferation (or grey-scale synovitis), compared with MRI as a reference method in EOA. Ultrasound should be used for diagnostic purposes and could select a patient population for clinical trials in EOA.
The authors would like to thank all patients who volunteered to participate in the study and all the radiology staff for their assistance during the MRI examinations.
Funding RW is a research fellow supported by a Ghent University Coordinated Research Initiative (GOA) grant (BOF07/GOA/002).
Competing interests None.
Patient consent Obtained.
Ethics approval This study was conducted with the approval of the University Hospital Ghent, De Pintelaan 185, 9000 Gent, Belgium.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.