clinical examination, ultrasound and mri imaging of … · clinical examination, ultrasound and mri...

14
ORIGINAL RESEARCH Clinical Examination, Ultrasound and MRI Imaging of The Painful Elbow in Psoriatic Arthritis and Rheumatoid Arthritis: Which is Better, Ultrasound or MR, for Imaging Enthesitis? Clare Groves . Muthusamy Chandramohan . Ne Siang Chew . Tariq Aslam . Philip S. Helliwell Received: October 14, 2016 / Published online: February 8, 2017 Ó The Author(s) 2017. This article is published with open access at Springerlink.com ABSTRACT Introduction: The purpose of the current study was to examine the painful elbow, and in par- ticular enthesitis, in psoriatic arthritis (PsA) and rheumatoid arthritis (RA) using clinical exami- nation, ultrasonography (US) and magnetic resonance imaging (MRI). Methods: Patients with elbow pain (11 with PsA and 9 with RA) were recruited. Clinical exami- nation, US and MRI studies were performed on the same day. For enthesitis, the common extensor and flexor insertions and the triceps insertion were imaged (20 patients, giving a total of 60 sites with comparative data). Imaging was performed with the radiologists blinded to the diagnosis and clinical findings. US was used to assess ‘inflammatory activity’ (Power Doppler signal, oedema, tendon thickening and bursal swelling) and ‘damage’ (erosions, cortical roughening and enthesophytes). MRI was used to assess ‘inflammation’ (fluid in paratenon, peri-entheseal soft-tissue oedema, entheseal enhancement with gadolinium, entheseal oedema and bone oedema) and ‘damage’ (ero- sion, cortical roughening and enthesophyte). Results: Complete scan data were not available for all patients as one patient could not tolerate the MRI examination. No significant differences in imaging scores were found between PsA and RA. Analysis of damage scores revealed com- plete agreement between US and MRI data in 43/55 (78%) comparisons; in 10/55 (18%) cases the US data were abnormal but the MRI data normal; in 2/55 (4%) cases, the MRI data were abnormal and the US data normal. Analysis of the inflammation scores revealed complete agreement between US and MRI data in 33/55 (60%) comparisons; in 3/55 (5%) cases US data were abnormal but MRI data normal; in 19/55 (35%) cases the MRI data were abnormal and the US data normal. There was a poor relation- ship between assessments based on clinical examination and imaging studies. Readers could not accurately identify the disease from imaging findings. Conclusion: Based on our results, at the elbow, US and MR have different roles in assessing enthesitis, with US apparently the better diag- nostic tool for assessing damage and MR the better tool for assessing inflammation. In this study enthesitis and synovitis in the painful elbow were found equally in cases of established RA and PsA. Enhanced content To view enhanced content for this article go to http://www.medengine.com/Redeem/CD7 7F06057687EB8. C. Groves Á M. Chandramohan Á N. S. Chew Á T. Aslam Á P. S. Helliwell St. Luke’s Hospital, Bradford Teaching NHS Foundation Trust, Bradford, UK P. S. Helliwell (&) Leeds Institute of Rheumatic and Musculoskeletal Medicine, University of Leeds, Leeds, UK e-mail: [email protected] Rheumatol Ther (2017) 4:71–84 DOI 10.1007/s40744-017-0053-7

Upload: hoangdang

Post on 07-Sep-2018

227 views

Category:

Documents


0 download

TRANSCRIPT

ORIGINAL RESEARCH

Clinical Examination, Ultrasound and MRI Imagingof The Painful Elbow in Psoriatic Arthritisand Rheumatoid Arthritis: Which is Better, Ultrasoundor MR, for Imaging Enthesitis?

Clare Groves . Muthusamy Chandramohan . Ne Siang Chew .

Tariq Aslam . Philip S. Helliwell

Received: October 14, 2016 / Published online: February 8, 2017� The Author(s) 2017. This article is published with open access at Springerlink.com

ABSTRACT

Introduction: The purpose of the current studywas to examine the painful elbow, and in par-ticular enthesitis, in psoriatic arthritis (PsA) andrheumatoid arthritis (RA) using clinical exami-nation, ultrasonography (US) and magneticresonance imaging (MRI).Methods: Patients with elbow pain (11 with PsAand 9 with RA) were recruited. Clinical exami-nation, US and MRI studies were performed onthe same day. For enthesitis, the commonextensor and flexor insertions and the tricepsinsertionwere imaged (20 patients, giving a totalof 60 sites with comparative data). Imaging wasperformed with the radiologists blinded to thediagnosis and clinical findings. US was used toassess ‘inflammatory activity’ (Power Dopplersignal, oedema, tendon thickening and bursalswelling) and ‘damage’ (erosions, corticalroughening and enthesophytes). MRI was used

to assess ‘inflammation’ (fluid in paratenon,peri-entheseal soft-tissue oedema, enthesealenhancement with gadolinium, enthesealoedema and bone oedema) and ‘damage’ (ero-sion, cortical roughening and enthesophyte).Results: Complete scan data were not availablefor all patients as one patient could not toleratethe MRI examination. No significant differencesin imaging scores were found between PsA andRA. Analysis of damage scores revealed com-plete agreement between US and MRI data in43/55 (78%) comparisons; in 10/55 (18%) casesthe US data were abnormal but the MRI datanormal; in 2/55 (4%) cases, the MRI data wereabnormal and the US data normal. Analysis ofthe inflammation scores revealed completeagreement between US and MRI data in 33/55(60%) comparisons; in 3/55 (5%) cases US datawere abnormal but MRI data normal; in 19/55(35%) cases the MRI data were abnormal andthe US data normal. There was a poor relation-ship between assessments based on clinicalexamination and imaging studies. Readerscould not accurately identify the disease fromimaging findings.Conclusion: Based on our results, at the elbow,US and MR have different roles in assessingenthesitis, with US apparently the better diag-nostic tool for assessing damage and MR thebetter tool for assessing inflammation. In thisstudy enthesitis and synovitis in the painfulelbow were found equally in cases of establishedRA and PsA.

Enhanced content To view enhanced content for thisarticle go to http://www.medengine.com/Redeem/CD77F06057687EB8.

C. Groves � M. Chandramohan � N. S. Chew �T. Aslam � P. S. HelliwellSt. Luke’s Hospital, Bradford Teaching NHSFoundation Trust, Bradford, UK

P. S. Helliwell (&)Leeds Institute of Rheumatic and MusculoskeletalMedicine, University of Leeds, Leeds, UKe-mail: [email protected]

Rheumatol Ther (2017) 4:71–84

DOI 10.1007/s40744-017-0053-7

Keywords: Enthesitis; Disease activity;Magnetic resonance imaging; Psoriaticarthritis; Rheumatoid arthritis; Ultrasound

INTRODUCTION

Enthesitis is an important aspect of spondy-loarthropathy. It plays a pivotal part in thespinal changes which occur in ankylosingspondylitis (AS) and may have a similar role inthe peripheral arthritis of psoriatic arthritis(PsA) and other spondyloarthropathies. Enthe-sitis is assessed separately in clinical studies ofAS and PsA and was placed in the ‘outer circle’(recommended for clinical trials) of domainsdetermined for PsA at the Outcomes inRheumatology Clinical Trials (OMERACT)meeting in Malta 2006 [1]. Further, enthesitisassessment is included in the domains in threecomposite measures of disease activity in PsA[2–4].

RA is a different disease to PsA in terms ofgenetics, pathogenesis, presentation and out-come. It has been hypothesised that the pri-mary pathological lesion occurs in thesynovium in RA and at the enthesis in PsA [5].According to this hypothesis, synovitis in PsA isconsidered to be secondary to the extra-capsulardisease. On the other hand, a number ofauthors have recognised the occurrence ofenthesitis as part of the RA disease spectrum[6–8]. It is therefore clear that, by the time thepatient presents, both synovitis and enthesitismay occur in both conditions, occasionallymaking it difficult to distinguish between themon radiological grounds [9]. However, in estab-lished disease, differences have been descri-bed—symmetrical involvement, periarticularosteopenia and marginal erosions in RA; rela-tively preserved bone stock, paramarginal ero-sions, osteolysis, perisostitis and periarticularnew bone in PsA [10].

A clinical enthesis index specific to PsA hasrecently been developed [11]. This indexexamines tenderness at six sites: lateral epi-condyles of the humerus, medial condyles ofthe femur and the insertion of the Achillestendon. The proximity of the lateral epicondyleof the humerus to the elbow joint (EJ)

exemplifies the difficulties of separating enthe-seal from articular tenderness, particularlywhen both occur together. The aim of the studyreported here was threefold: firstly, to examinethe relationship between enthesitis and syn-ovitis by comparing data from clinical andimaging examinations of this anatomical areain PsA and RA; secondly, to examine the rela-tionship between clinical enthesitis and enthe-sitis found on imaging; thirdly, to compareultrasound (US) and magnetic resonance imag-ing (MRI) in terms of identifying features ofenthesitis. We hypothesised (1) that synovitiswould be more prominent in RA, and enthesitisin PsA; (2) that there would be a poor relation-ship between clinical enthesitis and enthesitisidentified by US and MRI; (3) that US wouldprovide more information than MRI on enthe-sitis at the elbow.

METHODS

Full ethical committee approval was given forthis study, and all patients provided theirsigned, informed consent to participate (Brad-ford REC approval 09/H1302/113). Subjectswere seen in rheumatology out-patient clinicsand, after consent procedures had been com-pleted, examined using a standard clinical pro-tocol. Patients with a physician diagnosis of PsAand RA were eligible if they complained of painin the elbow region. The protocol consisted ofgathering sufficient clinical information toassess the Classification of Psoriatic Arthritis(CASPAR) criteria [12], an acute phase markerand a swollen joint count. It included anassessment of the entheses at the elbow,namely, the common extensor origin (CEO) atthe lateral epicondyle, the common flexor ori-gin (CFO) at the medial epicondyle and thetriceps insertion. Pressure was exerted at theenthesis sufficient to blanch the thumb nail ofthe examiner (approximately 4 kg). In addition,the examiner assessed the presence of soft-tissueswelling at the enthesis and performed ‘stress’testing at each site (for medial and lateral epi-condyles, resisted extension and flexion of thewrist with the arm fully extended; for triceps,resisted extension of the elbow).

72 Rheumatol Ther (2017) 4:71–84

Active inflammatory arthritis in the EJ wasassessed by palpation. For the radio-humeralcomponent (RHJ) the examiner faced thepatient and grasped the hand of the patient in agentle handshake (with the right hand for theright RHJ and the left hand for the left RHJ).With the thumb of the other hand the examinerthen palpated over the RHJ while supinating/pronating the forearm. Tenderness and pain onmovement indicated a positive test. Swelling atthis location could not reliably be identified.For the EJ the examiner palpated the joint lineanteriorly and the groove between the lateralepicondyle and the olecranon process, notingsoft-tissue swelling and tenderness at this point.Articular damage was scored if the patient hadmore than 50% loss of movement and/or hadbony swelling of the joint margins.

The US and MRI examinations were carriedout on the same day as the clinical evaluation,with the US images assessed by one radiologistand the MRI images assessed by a second radi-ologist; each radiologist was ‘blinded’ to thediagnosis and to the findings of the ‘other’imaging modality. After each radiologicalexamination, the radiologists made a decisionon the diagnosis based on their respectiveinterpretation of the imaging data. Each radi-ologist had over 15 years of experience in mus-culoskeletal radiology. A third radiologistperformed a second reading of both the MRIand (static) US scans; this radiologist was alsoblinded to the diagnosis and to the findings ineach imaging modality.

Ultrasound Protocol

The scans were performed on a Siemens S2000machine using a 10-15 MHz linear probe (Sie-mens AG, Munich, Germany). The sono-graphic assessments were made at the CEO, theCFO and the attachment of the triceps tendonat the olecranon. The CEO (anatomically at thelateral epicondyle of the elbow) was examinedwith the patient seated and the hand restingon the knee with the elbow slightly flexed andthe wrist in gentle internal rotation. The CFO(anatomically at the medial epicondyle of theelbow) was examined with the arm fully

extended and the arm resting on the knee withthe palm facing anteriorly. Grey scale imagingin the longitudinal and transverse planes wasused to assess the enthesis for the presence oferosions, enthesophytes, entheseal thickeningand perientheseal soft-tissue oedema. Lesionswere scored as present (score 1) or absent (score0). An erosion was defined as a step-downcortical contour defect seen in two planes andmeasuring C2 mm in depth. An enthesophytewas defined as a step-up bony prominence atthe end of a normal bone contour and forminga bony spur seen within the tendinous portionof the enthesis. Entheseal thickening wasscored as present if there was a discrepancybetween the thickness at the contralateralenthesis or if the normally smooth enthesealcontours appeared to be bulky. Peri-enthesealsoft-tissue oedema was scored as present ifthere was compressible fluid within the softtissues on the outer margin of the enthesis.Assessment of entheseal vascularisation wasperformed using Power Doppler mode, withthe pulse repetition frequency set at 600.Entheseal vascularity was scored as present(score 1) or absent (score 0).

The following scores were calculated as anapproximation of ‘inflammation’ and ‘damage’at the enthesis:(1) Inflammation: vascularisation, fluid in

paratenon, perientheseal soft-tissueoedema, altered echogenicity and thicken-ing of tendon (score range 0–5).

(2) Damage: erosion, cortical roughening andenthesophyte (score range 0–3).

For the joint, both grey scale of effusion/syn-ovitis and Power Doppler mode were used inassessment (articular score, range 0–2).

This protocol is in keeping with the recom-mendations of the OMERACT US in Spondy-loarthropathy Study Group [13].

MRI Protocol

Scans were performed on a Siemens Avanto 1.5Tscanner (Siemens AG). Patients were imagedprone in the ‘superman position’ with the armextended into the centre of the bore using awrap-around coil.

Rheumatol Ther (2017) 4:71–84 73

For the purposes of this study, the followingsequences were performed:(1) Axial T1-weighted. Repetition time(TR)/

echo time (TE), 300/24 ms; echo-trainlength (ETL), 1; matrix, 144 9 192; fieldof view (FOV), 11 cm; number of excita-tions, 1; slice thickness 3 mm with 0.3-mminterslice gap.

(2) Coronal T1-weighted. TR/TE, 333–361/24ms; ETL, 1; matrix, 218 9 257; FOV, 13 cm;number of excitations, 1; slice thickness3 mm with 0.3-mm interslice gap.

(3) Axial T2 fat-saturated. TR/TE, 2290–4800/80– ms; ETL, 9–16; matrix, 192 9 256; FOV,12 cm, number of excitations, 1; slice thick-ness, 3 mm with 0.3-mm interslice gap.

(4) Coronal short-tau T1 inversion recovery(STIR). TR/TE, 4630–5130/34 ms; ETL, 7;matrix, 205 9 256; FOV, 13 cm, number ofexcitations, 1; slice thickness, 3 mm with0.3-mm interslice gap.

A standard dose of 0.2 mL/kg body weight ofgadoterate dimeglumine (Dotarem; Guerbet,Aulney-sous-Bois, France) was administeredintravenously by hand injection, followed by aflush of 20 mL saline solution (0.9%). T1fat-saturated sequences in three orthogonalplanes after gadolinium administration (TR/TE,580–710/24 ms; ETL, 1; matrix, 218 9 256; FOV,10 cm; number of excitations, 1; slice thickness,3 mm with 0.3-mm interslice gap) were subse-quently obtained.

This protocol is in keeping with the recom-mendations of the OMERACT MRI Study Groupfor PsA [14].

Images were scored in a similar fashion tothe US images as described above. Scores werecalculated as an approximation of ‘inflamma-tion’ and ‘damage’ at the enthesis, as follows:(1) Inflammation: fluid in paratenon, peri-en-

theseal soft-tissue oedema, enthesealenhancement with gadolinium, enthesealoedema, andboneoedema (score range 0–5).

(2) Damage: erosion, cortical roughening andenthesophyte (score range 0–3).

Further, images were scored for synovitis in theelbow and radio-humeral joints, as defined bythe presence of a joint effusion and synovialenhancement following gadolinium adminis-tration (Articular score, range 0–2).

Statistics

All statistics were carried out using SPSS soft-ware v 21.0 (IBM Corp., Armonk, NY). Contin-uous data were compared using non-parametricstatistics, dichotomous data using Chi-squaredtests and where insufficient cell numbers werefound, we performed a Fisher’s exact test.Agreement between readers was examined withthe kappa (j) statistic. By general consensus,kappa values of \0 indicate no agreement,0–0.20 slight agreement, 0.21–0.40 fair agree-ment, 0.41–0.60 moderate agreement,0.61–0.80 substantial agreement and 0.81–1almost perfect agreement. Intra-class correla-tion was used to compare the scores of the twoobservers.

RESULTS

The study included 20 patients. Of these, 11 hadwith PsA [5 males, 6 females; mean age48.9 years; mean duration of disease 6.8 years;mean C-reactive protein (CRP) 2.3 mg/dL; meanswollen joint count 3.4], and nine had RA (4males, 5 females; mean age 61.9 years; meanduration of disease 12.4 years; mean CRP10.5 mg/dL; mean swollen joint count 3.5). AllRA patients fulfilled the 1987 criteria for RA [15]and eight of the 11 patients with PsA fulfilledthe CASPAR criteria.

Clinical Examination

All patients had a painful elbow. Generallymore patients in the PsA group had clinicalenthesitis (PsA: N = 10 and 8 at the CEO andCFO, respectively; RA: N = 6 and 4 for CEOand CFO, respectively), and more patientswith RA had articular involvement (PsA: N = 4and 3 for EJ and RHJ, respectively; RA: N = 8and 3 for EJ and RHJ, respectively). The onlystatistically significant difference between thediagnostic groups was for EJ tenderness andswelling (Chi-squared test 5.7; p = 0.28). Veryfew cases of clinical articular damage werefound [2 cases of PsA and 3 cases of rheuma-toid arthritis (RA)].

74 Rheumatol Ther (2017) 4:71–84

Table1

Imagingscores

from

ultrasound

andmagneticresonanceim

agingstudiesat

each

enthesisandjointaccordingto

disease

Imagingmod

ality

Psoriatic

arthritis

Rheum

atoidarthritis

Inflammation

Dam

age

Inflammation

Dam

age

Vasc

Fluid

STO

Echo

Thick

Total

Erosion

CR

Enth

Total

Vasc

Fluid

STO

Echo

Thick

Total

Erosion

CR

Enth

Total

Ultrasound

a

TRIC

EPS

00

00

00

00

0.30

0.30

00

0.22

00.22

0.44

0.22

00.11

0.33

CEO

0.45

0.27

0.18

0.45

0.45

1.82

00

0.55

0.55

0.38

0.13

00.56

0.44

1.57

0.33

0.63

0.67

1.50

CFO

0.09

00

00

0.09

00

00

00

00

00

00

00

Effusion

Power

Doppler

?ve

Effusion

Power

Doppler

?ve

Articular

0.36

0.27

0.64

0.22

0.13

0.38

Imagingmod

ality

Psoriatic

arthritis

Rheum

atoidarthritis

Inflammation

Dam

age

Inflammation

Dam

age

Fluid

STO

Enh

anEnthO

Bon

eOTotal

Erosion

CR

Enth

Total

Fluid

STO

Enh

anEnthO

Bon

eOTotal

Erosion

Rou

ghEnth

Total

MRIb TRIC

EPS

0.18

0.27

0.18

0.18

00.82

00

00

00.25

0.25

0.25

00.75

00

00

CEO

0.36

0.64

0.73

0.64

0.18

2.55

0.18

0.27

0.09

0.55

0.13

0.13

0.75

0.75

0.13

1.88

0.13

0.25

0.13

0.50

CFO

0.18

0.36

0.27

0.27

01.09

00.09

00.09

00.38

0.38

0.13

00.88

00

00

Effusion

Enh

ancementwithcontrast

Effusion

Enh

ancementwithcontrast

Articular

1.18

0.63

2.00

0.63

0.82

1.25

Allscores

reported

inthetablearemeans.N

oneof

thecomparisons

(scores)betweenPsAandRAin

term

sof

inflammation,

damage,andarticularweresignificantlydifferent

TRICEPS

Triceps

insertion,

CEO

common

extensor

origin,C

FOcommon

flexororigin,A

rticu

larjointscores

aUltrasound

abbreviations:VascVascularisation

atenthesis,Fluidfluidin

paratenon,ST

Operienthesealsoft-tissue

oedema,Echoalteredechogenicityof

tend

onatenthesitis,T

hick

thickening

oftend

onat

enthesis

Erosionerosionat

enthesis,C

Rcorticalroughening,E

nthenthesophyte

bMRI(m

agneticresonanceim

aging)abbreviations:FluidFluidin

paratenon,ST

Operienthesealsoft-tissue

oedema,Enhan

enthesealenh

ancementwithgadolin

ium,E

nthO

enthesealoedem

a,BoneO

bone

oedema,Erosionerosionat

enthesis,C

Rcorticalroughening,E

nthenthesophyte

Rheumatol Ther (2017) 4:71–84 75

Imaging Scores by Disease Group

Complete imaging data were not available for

all patients due to one patient not tolerating the

MRI procedure; overall, data on 55/60

comparisons at the three entheseal sites wereavailable. Scores for both enthesitis and articu-lar involvement by diagnosis are given inTable 1. For both diseases the highest scoreswere found for inflammation and damage at the

Fig. 1 a Coronal magnetic resonance (MR) short-tau T1inversion recovery (STIR) imaging. Normal commonextensor origin (CEO) with homogenous low signalthroughout (arrowhead). Note the signal at the CEOcannot be identified separately from the lateral collateralligament (arrow). There is a small effusion in the joint.

b Ultrasound image. Longitudinal section through thenormal CEO. Fibrils are apparent within the ligamentsubstance. There is a bright, regular, echogenic paratenon(arrow), and the slice also shows the edge of the radiocarpaljoint with normal fat pad (star)

76 Rheumatol Ther (2017) 4:71–84

CEO by both imaging modalities (US: PsA meaninflammation and damage scores of 1.82 and0.55, respectively; RA mean inflammation and

damage scores of 1.57 and 1.50, respectively;MRI: PsA mean inflammation and damagescores of 2.55 and 0.55, respectively; RA mean

Fig. 2 Ultrasound image. Longitudinal section through the CEO. Enthesopathy is present with swollen, hypoechoic CEO,marked neovascularity in the tendon substance (arrows) and an enthesophyte (arrowhead) close to the tendon insertion

Fig. 3 Coronal MR T1 fat-saturated sequence plus con-trast. Severe lateral epicondylitis with enhancing, inflam-matory tissue between the oedematous fibres of thecommon extensor origin (arrowheads) and multiple small

erosions in the lateral epicondyle (arrow) are visualized. Thelateral collateral ligament is shown separately from theCEO and is oedematous with a small erosion at the origin

Rheumatol Ther (2017) 4:71–84 77

inflammation and damage scores of 1.88 and0.50 respectively). None of the differences werestatistically significant.

Relationship Between US and MRIfindings at the Enthesis

The scores for damage differed from those forinflammation. For damage scores, completeagreement between US and MRI data was foundin 43/55 (78%) comparisons; in 10/55 (18%)cases the US data were abnormal but the MRIdata were normal; in 2/55 (4%) cases the MRIdata were abnormal and the US data werenormal.

For inflammation scores complete agree-ment between US and MRI data was found in33/55 (60%) comparisons; in 3/55 (5%) casesthe US data were abnormal but the MRI werenormal; in 19/55 (35%) cases the MRI data wereabnormal and US data were normal. Thus, USappeared to be the more sensitive imagingmodality for evaluating damage and MRI themore sensitive imaging modality for evaluatinginflammation. Overall unadjusted kappa scoresbetween US and MRI data were 0.34 for damagescores and 0.23 for inflammation scores.

Examples of US and MR images at the elboware shown in Figs. 1–6.

Relationship Between the Resultsof Clinical Examinations and ImagingFindings

Overall there was a poor agreement betweenthe clinical findings and the imaging results.Further, no significant relationship was foundbetween clinical enthesitis and the presence ofinflammation in the components of the EJ(data not shown). Kappa scores for agreementbetween the clinical and imaging results werecalculated, aggregated for all sites. For the USinflammation score (any positive) and clinicaltests, the kappa scores for agreement were 0.1,0.16 and 0.12 for stressing, swelling and ten-derness, respectively. For the MRI inflamma-tion scores and clinical tests, the kappa scoresfor agreement were 0.28, 0.04 and 0.11 forstressing, swelling and tenderness, respec-tively. Levels of agreement at individual loca-tions were also generally poor, except for theCEO (stress) and MR inflammation (j = 0.41),and for CEO tenderness and MR inflammation(j = 0.28).

Fig. 4 Coronal and axial MR T1 fat-saturated sequence plus contrast. These images show severe ‘osteitis’ with bone oedemaat the origin of the common extensor origin (arrowheads). The CEO is swollen and enhancing (arrows)

78 Rheumatol Ther (2017) 4:71–84

Agreement Between Observers on the ImagingAgreement between observers was slight tomoderate when aggregated across all enthesealsites, and for each modality (see Table 2). Kappascores were highest for vascularisation (PowerDoppler signal) detected by US (j = 0.51) andentheseal enhancement detected by MRI (j =0.32). Interclass correlation (ICC) scores werealso modest for enthesitis, with the largestscores found for the inflammation scores in theMRI studies (Triceps 0.59; CEO 0.42; CFO 0.39).However, ICC scores were good for agreementon MR inflammation in the EJ (j = 0.86) and USinflammation in the EJ (j = 0.52).

Observers’ Diagnosis Based on Imaging

There was a poor relationship between thediagnosis of the primary radiologist, based onimaging findings, and the clinical diagnosis.The radiologists were unable to distinguishbetween RA and PsA on the basis of imagingfindings. The radiologist interpreting the USscans ‘misclassified’ 10/17 cases, and the radi-ologist interpreting the MRI scans ‘misclassified’7/16 cases. The kappa scores were 0.29 and 0.08for the level of agreement between US andclinical diagnoses and between MRI and clinicaldiagnoses, respectively.

Fig. 5 a Ultrasound. Longitudinal section through thecommon extensor origin. Normal cortical margins and thefibrils of the CEO are well shown, but there is a partialthickness tear of the ligament with intra-substance exten-sion (arrowheads). b Coronal MR T1 fat-saturated

sequence plus contrast. There is a partial thickness tear ofthe common extensor origin with distal retraction of thetorn fibres (arrowheads) and normal lateral collateralligament (arrows)

Rheumatol Ther (2017) 4:71–84 79

DISCUSSION

The aim of this study was to focus on theentheses around the EJ in terms of clinical andimaging findings. US and MR imaging, whilegiving contrasting information on inflamma-tion and damage, remain complementarymodalities. The relationship between clinicaland imaging enthesitis and synovitis was poor.Further, it was impossible to accurately identifya disease group from the imaging results alone,either with US or MRI.

To place some of these findings in context itis worth noting that abnormalities at theentheses around the elbow have been demon-strated in asymptomatic individuals [16] and inpatients with RA in whom clinical inflamma-tion has subsided [7]. Both these factors arelikely to confound studies looking at the rela-tionship between clinical and imaging findingsat the enthesis, and in comparing imaging andclinical findings between diseases. Eshed et al.recognised that while MRI is sensitive forenthesopathy, the findings lack specificity anddifferentiating between the different aetiologiesof enthesitis on the basis of their MRI charac-teristics is not always possible [17]. Our study

would appear to confirm this conclusion andmay suggest that the process of enthesealinflammation is similar in both RA and PsA. USappeared to be more sensitive for detectingdamage and MRI more sensitive for detectinginflammation, although a ‘gold standard’, suchas histology, was not available. Nevertheless,this result is to be expected as the resolution ofMRI is much lower than that of US and, con-sequently, the former is relatively insensitive toareas of bony abnormality, such as entheso-phytes or small erosions.

The discrepancy between clinical and imag-ing enthesitis in spondyloarthropathy is not anovel finding [18–20]. This may in part be dueto the fact that clinical examination and imag-ing studies are measuring different things. Forexample, US cannot visualise osteitis, yet ostei-tis may cause tenderness at the enthesis. Ostei-tis, or bone oedema, can be visualised at theenthesis in spondyloarthropathy using MRI,and this has been interpreted as indicatingenthesitis [21]. However, it was clear that therelationship between the MRI and clinicalfindings in the current study were poor. In asimilar study by Aydin and colleagues in theknee found an equally poor relationship

Fig. 6 Axial MR STIR sequence. Oedema is evident in the paratenon of the common extensor origin (arrows)

80 Rheumatol Ther (2017) 4:71–84

Table2

Agreementbetweenobserversandultrasound

andmagneticresonanceim

aging

Variables

Inflammationcriteria

Dam

agecriteria

Vascularisation

Fluidin

parateno

nPerientheseal

soft-tissueoedema

Altered

echo

genicity

Thickeningof

tend

onErosion

Corticalroughening

Enthesoph

yte

Ultrasound

Prevalence

(%)a

169

423

2110

1129

Agreement(%

)b87

9491

7176

9070

69

Kappa

c0.51

n/a

0.04

0.22

0.15

0.03

0.14

0.27

Variables

Inflammationcriteria

Dam

agecriteria

Bon

eoedema

Fluidin

parateno

nEntheseal

oedema

Entheseal

enhancem

ent

Peri-enthesealsoft-tissueoedema

Erosion

Corticalroughening

Enthesoph

yte

MRI

Prevalence

(%)a

516

3742

355

117

Agreement(%

)b93

7068

6853

9382

88

Kappa

c0.30

0.07

0.23

0.32

0.08

0.03

0.07

0.16

Figuresareaggregated

forallenthesealsites(N

=60)

n/aDatanotavailable

aPrevalence

isthepercentage

ofsiteswiththeabnorm

alitybasedon

thefirstassessor

bAgreementisthepercentage

ofscores

withcompleteagreem

ent

cKappa

statisticisameasure

ofagreem

entandranges

from

0to

1,with1indicating

completeagreem

ent

Rheumatol Ther (2017) 4:71–84 81

between imaging and clinical scores [20]. In thecurrent study one reason for the poor associa-tion could have been the low prevalence ofabnormality, particularly in the triceps inser-tion and CFO, and this may disadvantage thekappa statistic. That a ‘good’ relationshipbetween MR inflammation and CEO stressingwas found reflects the value of this test and itsrelationship to osteitis at the enthesealinsertion.

It had been thought that this study wouldshow a clear difference between the clinical andimaging findings of RA and PsA. However, itmust be noted that all the patients included inthis study had established disease, which mayhave ‘blurred’ the differences between them.Further, the mean age of the patients meantthat degenerative enthesopathy could be aconfounding factor. It is possible that thegreater age of the RA patients enrolled in ourstudy could have influenced the frequency ofdegenerative changes at the enthesis but, nota-bly, inflammatory changes were also foundequally in these diseases. The commonestabnormalities at each of the three enthesealsites were found at the CEO, and ‘lateral epi-condylalgia’ is a common clinical presentationin primary care in the UK [22].

How do these results inform the use of theLeeds Enthesitis Index, where the lateral epi-condyle of the elbow is one of the three sitesexamined? The poor relationship betweenclinical examination and imaging findings hasbeen confirmed but, on the positive side, therewas no relationship between inflammation inthe joint and inflammation in the adjacententhesis on imaging, and no relationshipbetween pain at the enthesis and underlyingarticular inflammation. The anatomical rela-tionship of the radio-humeral joint, its capsu-lar insertion and the CEO suggest thatinflammation in the EJ cannot extend to theCEO without disruption of the articular cap-sule at this point, and which in itself mayrepresent a process of entheseal inflammationat the capsular insertion. Discriminatingbetween synovitis and enthesitis at the elbow,and at other joints, may therefore, and partic-ularly in established disease, be a futile exercise

because of the inter-relationship between thetwo processes.

CONCLUSIONS

In summary, a poor relationship was foundbetween clinical and imaging assessments, andbetween imaging modalities. US seemed thebetter modality for assessing damage at theenthesis, and MRI the better modality forassessing inflammation. However, MRI is gen-erally the more sensitive of these two modalitiesand should be considered as the next line ofinvestigation following a normal US result. Nodifferences were found between PsA and RA interms of enthesitis.

ACKNOWLEDGEMENTS

The study was part funded by the BradfordRadiology Discretionary Fund. All namedauthors meet the International Committee ofMedical Journal Editors (ICMJE) criteria forauthorship for this manuscript, take responsi-bility for the integrity of the work as a wholeand have given final approval for the version tobe published. Dr. Helliwell conceived the study,collected data and undertook the analysis; Dr.Aslam collected clinical study data. All authorshave reviewed the paper.

Author contributions. Dr. Helliwell is theguarantor of integrity of the entire study,established the study concepts and designed thestudy; literature research is not applicable tothis study; Dr. Helliwell and Dr. Aslam per-formed the clinical studies; Dr. Groves per-formed the ultrasound examinations; Dr.Chandramohan interpreted the MRI images andDr. Chew read both the ultrasound and MRIimages; Dr. Helliwell performed the experi-mental studies/data analysis and the statisticalanalysis. All authors contributed to manuscriptpreparation and editing.

Disclosures. Clare Groves, MuthusamyChandramohan, Ne Siang Chew, Tariq Aslam

82 Rheumatol Ther (2017) 4:71–84

and Philip S Helliwell declare that they havenothing to disclose.

Compliance with Ethics Guidelines. Fullethical committee approval was given for thisstudy, and all patients provided signed,informed consent to take part (Bradford RECapproval 09/H1302/113).

Open Access. This article is distributedunder the terms of the Creative CommonsAttribution-NonCommercial 4.0 InternationalLicense (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommer-cial use, distribution, and reproduction in anymedium, provided you give appropriate creditto the original author(s) and the source, providea link to the Creative Commons license, andindicate if changes were made.

REFERENCES

1. Gladman DD, Mease PJ, Strand V, Healy P, HelliwellPS, Fitzgerald O, et al. Consensus on a core set ofdomains for psoriatic arthritis. J Rheumatol.2007;34(5):1167–70.

2. Helliwell PS, FitzGerald O, Fransen J, Gladman DD,Kreuger GG, Callis-Duffin K, et al. The developmentof candidate composite disease activity andresponder indices for psoriatic arthritis (GRACEproject). Ann Rheum Dis. 2013;72(6):986–91.

3. Mumtaz A, Gallagher P, Kirby B, Waxman R, CoatesLC, Veale JD, et al. Development of a preliminarycomposite disease activity index in psoriaticarthritis. 2011;70(2):272–7.

4. Coates LC, Fransen J, Helliwell PS. Defining mini-mal disease activity in psoriatic arthritis: a proposedobjective target for treatment. 2010;69(1):48–53.

5. McGonagle D, Gibbon W, Emery P. Classification ofinflammatory arthritis by enthesitis. Lancet.1998;352(9134):1137–40 (see comment).

6. Helliwell PS, Porter G, group Cs. Sensitivity andspecificity of plain radiographic features of periph-eral enthesopathy at major sites in psoriatic arthri-tis. Skelet Radiol. 2007;36(11):1061–6.

7. Ibrahim GH, Groves C, Chandramohan M, Valle R,Beltran A, Reyes B, et al. Clinical and ultrasoundvalidation of the Leeds Enthesitis Index in psoriatic

arthritis and rheumatoid arthritis. ISRN Rheumatol.2011;731917:5. doi:10.5402/2011/731917.

8. Frediani B, Falsetti P, Storri L, Allegri A, Bisogno S,Baldi F, et al. Ultrasound and clinical evaluation ofquadricipital tendon enthesitis in patients withpsoriatic arthritis and rheumatoid arthritis. ClinRheumatol. 2002;21(4):294–8.

9. Marzo-Ortega H, Tanner SF, Rhodes LA, Tan AL,Conaghan PG, Hensor EMA, et al. Magnetic reso-nance imaging in the assessment of metacar-pophalangeal joint disease in early psoriatic andrheumatoid arthritis. Scan J Rheumatol.2009;38(2):79–83.

10. McGonagle D. Imaging the joint and enthesis:insights into pathogenesis of psoriatic arthritis.Ann Rheum Dis. 2005;64[Suppl II]:ii58–60. doi:10.1136/ard.2004.034264.

11. Healy P, Helliwell PS. Measuring clinical enthesitisin psoriatic arthritis:assessment of existing mea-sures and development of an instrument specific forpsoriatic arthritis. Arthritis Care Res.2008;59(5):686–91.

12. Taylor WJ, Gladman DD, Helliwell PS, MarchesoniA, Mease PJ, Mielants H. Classification criteria forpsoriatic arthritis. Arthritis Rheum.2006;54(8):2665–73.

13. Terslev L, Naredo E, Iagnocco A, Balint PV, Wake-field RJ, Aegerter P, et al. Defining enthesitis inspondyloarthritis by ultrasound: results of a Delphiprocess and of a reliability reading exercise.Arthritis Care Res. 2014;66(5):741–8.

14. McQueen F, Lassere M, Bird P, Haavardsholm EA,Peterfy C, Conaghan PG, et al. Developing a mag-netic resonance imaging scoring system forperipheral psoriatic arthritis. J Rheumatol.2007;34(4):859–61.

15. Arnett FC, Edworthy SM, Bloch DA. The AmericanRheumatism Association 1987 revised criteria forthe classification of rheumatoid arthritis. ArthritisRheum. 1988;31(3):315–24.

16. Levin D, Nazarian L, Miller T, O’Kane P, Feld R,Parker L, et al. Lateral epicondylitis of the elbow: USfindings. Radiology. 2005;237(1):230–4.

17. Eshed I, Bollow M, McGonagle DG, Tan AL, AlthoffCE, Asbach P, et al. MRI of enthesitis of theappendicular skeleton in spondyloarthritis. AnnRheum Dis. 2007;66(12):1553–9.

18. D’Agostino MA, Said-Nahal R, Hacquard-Bouder C,Brasseur JL, Dougados M, Breban M. Assessment ofperipheral enthesitis in the spondylarthropathies

Rheumatol Ther (2017) 4:71–84 83

by ultrasonography combined with power Doppler:a cross-sectional study. 2003;48(2):523–33.

19. Balint P, Kane D, Wilson H, McInnes I, Sturrock R.Ultrasonography of entheseal insertions in thelower limb in spondyloarthropathy. Ann RheumDis. 2002;61(10):905–10.

20. Aydin S, Tan AL, Hodgson R, Grainger A, Emery P,Wakefield RJ, et al. Comparison of ultrasonographyandmagnetic resonance imaging for the assessmentof clinically defined knee enthesitis in spondy-loarthritis. Clin Exp Rheumatol. 2013;31:933–6.

21. McGonagle D, Gibbon W, O’Connor P, Green M,Pease C, Emery P. Characteristic magnetic reso-nance imaging entheseal changes of knee synovitisin spondylarthropathy. Arthritis Rheum.1998;41(4):694–700.

22. Hay EM, Paterson SM, Lewis M, Hosie G, Croft P.Pragmatic randomised controlled trial of local cor-ticosteroid injection and naproxen for treatment oflateral epicondylitis of elbow in primary care. BMJ.1999;319(7215):964–8.

84 Rheumatol Ther (2017) 4:71–84