Assessment of anterior subcutaneous hypersignal on proton-density-weighted MR imaging of the knee and relationship with anterior knee pain

ORIGINAL

ARTICLE

/Musculoskeletal

imaging

Assessment

of

anterior

subcutaneous

hypersignal

on

proton-density-weighted

MR

imaging

of

the

knee

and

relationship

with

anterior

knee

pain

E.N.

Unlu

,

Y.

Turhan

,

D.M.

Kos

,

A.A.

Safak

a_{DüzceUniversity,FacultyofMedicine,DepartmentofRadiology,Düzce,Turkey} b_{DüzceUniversity,FacultyofMedicine,DepartmentofOrthopaedicSurgeryand}

Traumatology,Düzce,Turkey

c_{AnkaraOccupationalDiseasesHospital,DepartmentofInternalMedicine,Ankara,Turkey}

KEYWORDS Bursitis; Boneedema; Knee; Magneticresonance imaging(MRI) Abstract

Purpose:Thepurposeofthisstudywastoevaluatetheprevalenceofanteriorsubcutaneous hypersignalindicatingedemaonproton-density(PD)-weightedMRIofthekneeandtodetermine whetherreportinganterioredemaisclinicallyrelevant.

Materials and methods:One hundred and ninety-one knee MRIs from 162 patients were reviewed for anterior subcutaneous edema. There were 92 men and 70 women with a mean age of 41.72years±13.92 (SD) (range, 15—80years) years and a mean body weight of 75.94kg±12.54 (SD) (range, 50—130kg). The MRI findings were compared with patient age, gender, body weight, historyof repetitive microtrauma and clinical findings. Patellar andtrochlearchondropathy,medialplica,jointeffusion,synovitis, infrapatellarfat-pad sig-nalintensity,suprapatellarfat-padsignalintensitywithmasseffect,quadricepsandpatellar tendonabnormalitieswerealsoreviewed.

Results:AnanteriorhypersignalonPD-weightedMRIwasdetectedin158/191MRexaminations (82.7%)and 104 (84.6%)ofthese cases had histories ofanterior knee pain. Nocorrelation between anterior pain and anterior edema was found (P=0.42). Age (P<0.0001), weight (P<0.0001), and repetitive microtrauma (P=0.001) were identified as significant variables associatedwithanterioredema.

∗_{Correspondingauthor.DuzceUniversity,FacultyofMedicine,DepartmentofRadiology,KonuralpTR-81000,Düzce,Turkey.}

E-mailaddresses:nisaunlu@yahoo.com,enisaunlu@gmail.com(E.N.Unlu),yturhan2000@yahoo.com(Y.Turhan),

mehmetkos@gmail.com(D.M.Kos),alpalpersafak@yahoo.com(A.A.Safak).

http://dx.doi.org/10.1016/j.diii.2016.08.008

Pleasecitethisarticleinpressas:Unlu EN,etal.Assessmentofanteriorsubcutaneoushypersignal on

proton-density-Conclusion:Anterioredemamaybeaphysiologicalphenomenonordegenerativechangerelated topatient age, weight, andknee movement or mechanics.It shouldnot be reportedas a pathologicalfindingonMRIunlessclinicalfindingssupportregionalinfectionorinflammation. ©2016Editionsfranc¸aisesderadiologie.PublishedbyElsevierMassonSAS.Allrightsreserved.

Anterior knee pain is the most common knee complaint andhasvariousunderlyingcauses,includingpatellar chon-dromalacia,anterior kneebursitis, quadriceps or patellar tendon abnormalities, plica syndromes, Hoffa’s disease, bone abnormalities, and traction apophysitis syndromes

[1—3].Bursitismaymimicseveralperipheraljointand

mus-cle abnormalities clinically. Thus, it is important for the radiologisttoidentifybursalpathology andprovidedirect managementgearedtowardsbursitis[4].Anterior subcuta-neousadiposetissue,includingprepatellar andsuperficial infrapatellarbursae,mayoftenshowedematoushighsignal intensitiesonfat-suppressedproton-density(PD)-weighted images of the routine magnetic resonance imaging (MRI) ofthe knee, which maybe misdiagnosed asanterior bur-sitisor inflammation radiologically. Some previous studies haveshownthatthisfindingiscommononkneeMRI[1,4]. It has been also reported that the peripatellar lesions involving prepatellar and superficial infrapatellar bursae demonstratedonMRIarecommoninolderindividualswith equal prevalence in both symptomatic and asymptomatic patientswithosteoarthritis[5].

Thepurposeofthisstudywastoevaluatetheprevalence ofanterior subcutaneous hypersignal indicating edema on PD-weighted MRI of the knee and to determine whether reportinganterioredemaisclinicallyrelevant.

Materials

and

methods

Patients

WeretrospectivelyevaluatedconsecutiveroutinekneeMRI examinationsperformedbetween June2014 andFebruary 2015. MRI examinations were selected from the picture archivingandcommunicationsystem(PACS)ofourhospital. Patientswithahistoryofregionalradiotherapy,generalized edema,acutesportsinjury,historyof anoperation within 6months, prosthesis and metallic artefacts and regional softtissueorboneinfectionwereexcludedfromthestudy. AlsoforpatientsundergoingseveralMRIexaminations dur-ingthisperiod,onlythefirstexamination wasincluded in ourstudy.Therefore,94examinationsof80patientswere excluded. A total of 191 knee MRIsof 162 patients were included inthe study.In total,92 (56.8%) ofthe patients weremales,and70(43.2%)werefemales.Theirmeanage was41.72years±13.92(SD)(range,15—80)andthemean weightwas75.94kg±12.54 (SD)(range,50—130) kg. The examined kneewasthe right knee in 68(42%), leftin 65 (40.1%),andbilateralin29(17.9%)patients.

The patients were classified according to age, gender, bodyweight, and the presence of repetitive microtrauma

orchronicirritation(e.g.,occupationalkneeling,activities requiring a significant amount of kneeling, such ashouse cleaningorprayinginIslamicculture).Theclinical presen-tationsofallpatientswereclearlydocumented.

Local ethics committee approval was obtained for this retrospective study and the requirement for patient informedconsentwaswaived.

MRI

protocol

MRIwasperformedusinga1.5-Tscanner(Echelon®_,Hitachi,

Tokyo, Japan) with dedicated extremity coil. Body coil was used for overweight patients. MRI protocol included T1-weighted spin-echo-sequences in the transverse and sagittalplanes(TR/TE:400-450/11-12),fat-suppressedfast spin-echo PD-weighted sequences in the transverse and sagittal planes(TR/TE: 1550—2000/12—24), and gradient-echosequence in the sagittaland coronalplanes(TR/TE: 500—550/20,flip angle 20◦C).Thefieldofview(FOV)was 20cm,theslicethicknesswas4—4.5mm,andthe intersec-tion gap was from4 to5mm. The number of acquisition was either one or two and the matrix ranged 256×320, 224×320,256×384and320×320.

Intravenousadministrationofagadoliniumchelatewas performed in4 patientsat adose of0.1mmol/kg ofbody weight because of a suspicion of synovitis or arthritis. T1-weightedspin-echoorfat-suppressedT1-weighted spin-echo imaging(TR/TE: 450—600/12—20) in the transverse, sagittaland/orcoronalplaneswereaddedtotheprotocol forthesepatients.

Imaging

analysis

MR images were analyzed in consensus by two radiolo-gistswith8and10yearsofexperienceinmusculoskeletal imaging.MRIimageswereevaluatedintermsofthesignal ofthe anteriorsubcutaneous adiposetissue,cysticlesions related to bursitis, patellar and/or trochlear chondropa-thy, medial plica, effusion and synovitis in the joint, a signalfromtheinfrapatellar(Hoffa)fat-pad,thesignaland masseffectofthesuprapatellar(quadriceps)fat-pad,and patellarandquadricepstendinopathies.Thecomplaintsof patientswereclassifiedasanteriorpain,regional inflamma-tion(pain,increasedtemperature,swelling,andredness), andapalpablemass.

Reticularfluidintensitiesordiffuseill-definededematous signalchangesonthefat-suppressedPD-weightedsequence at the anterior subcutaneous adipose tissue, affecting the prepatellar andsuperficial infrapatellarregions, were noted as‘‘anterior subcutaneous PD-weighted high signal

intensity’’.Encapsulatedfluidcollectionattheprepatellar and/orsuperficialinfrapatellarregionsisdefinedasbursitis. Patellarandtrochlearchondropathywasnoted(grade0, normal,grade1,signalabnormalitywithoutdefect,grade2, lessthan50%cartilagedefect,grade3,morethan50% car-tilagedefect,grade4,full-thicknesscartilagedefectwith subchondralbonychanges).Thepresenceofmedialpatellar plicawasnoted.

Lessthan5mmwideningatthesuprapatellarrecesson sagittal PD-weightedimage was accepted asnormal joint fluid,whereasmorethan5mmwideningwasdeemed effu-sion.

Thesuprapatellarorquadricepsfat-padliesonthe patel-lar base and fills the gap between superior aspect of the patellar cartilage andposterior aspect of the insertionof thequadricepstendon.Thepre-femoralfat-pad islocated immediately anterior tothe femur, and the suprapatellar joint recess separates the suprapatellar fat-pad from the pre-femoral fat-pad. Truncation or scalloping of the pre-femoral fat-pad, defects or displacement or irregularity oftheinfrapatellarfat-pad(Hoffa),andthickeningand/or enhancementofsynovialsurfaceswererecordedas synovi-tis.Hoffaintensityequaltofatwasconsideredasnormal. Focal, linearor diffuse increased intensity,heterogeneity and fluid intensities were noted as pathological findings. Suprapatellarfat-padintensityequaltothatoffatwas con-sideredasnormal.Highersignalthan fatorfluidintensity withmasseffect(evidentbyconvexposteriorcontour)were notedaspathologicalfindings.

Thequadricepsandpatellartendonswerecharacterized as having partial or full-thickness tears, thickening, and increasedsignal intensities that indicated focalor diffuse tendinitis.Hyperintensestriationsrelatedtoadiposetissue ofthequadricepstendonswereevaluatedasnormal.Focal insertional increasedsignalsin a regular andhomogenous tendonwithoutthickeningorirregularityofthetendonand withoutadjacentbonemarrowedemawerenotconsidered atendonabnormality.

Statistical

analysis

Kolmogorov—Smirnov test was used to evaluate the normality of the distribution of quantitative variables. Mann—Whitney U-test was used to compare quanti-tative variables. Continuous variables were given as mean±standard deviation (SD) and range (min—max). Categoricalvariableswereexpressedasrawnumbers, pro-portions andpercentages, andanalyzed withFisher exact test. Statistical analyses were performed using SPSS for Windows software(ver 15.0, SPSS Inc., Chicago, IL, USA) packageandPvalues<0.05wereconsideredtoindicate sta-tisticalsignificance.

Results

Reticularorill-definedhyperintensesignalsweredetected intheanteriorsubcutaneousadiposetissueof 158(82.7%) knees on fat-suppressed PD-weighted sagittal and trans-verseMRIimages.Thesignal wasextended fromthelevel of thepatella tothetuberosity ofthe tibialongitudinally

(Figs. 1and 2).Adipose tissueintensity wasnormal in 33

Figure1. Forty-nine-year-old woman weighing 105kg without anteriorkneepainorregionalinflammationonphysical examina-tion.Fat-suppressed PD-weightedMR imageinthesagittalplane showsprepatellarandsuperficialinfrapatellarreticularfluid inten-sities(arrowheads).

Figure2. A56-year-oldwomanweighing90kgwithoutanterior kneepain and clinical evidence of bursitis. Fat-suppressed PD-weighted MR image in the sagittal planeshows anterior edema prominentintheprepatellarregion(arrowheads).

Pleasecitethisarticleinpressas:Unlu EN,etal.Assessmentofanteriorsubcutaneoushypersignal on

proton-density-Figure3. A71-year-oldandoverweighted(130kg)womanwithoutanteriorkneepain.A.Fat-suppressedPD-weightedMRimageinthe sagittalplaneshowsmoderateanterioredema(arrowheads).B.Noenhancementonpost-contrastfat-suppressedT1-weightedimageis visibleinthisregion.Notethesubchondralchangesofpatellarchondropathy(smallarrows,AandB),andenhancementofsynovialsurfaces consistentwithsynovitis(largearrows,B).

(17.9%)knees.AnanteriorPD-weightedhighsignalintensity withoutenhancementwasdetected inallof the contrast-enhancedstudies(n=4)(Fig.3).

The relationship between mean age and the anterior PD-weighted high signal intensity was statistically signifi-cant (P<0.0001). Therewas nosignificant gender or side (right-left)differencebetweenthegroupswithandwithout anteriorPD-weightedhighsignalintensity(P=0.24and0.99, respectively). The relationship between the mean weight andtheanteriorPD-weightedhighsignalintensitywas sta-tisticallysignificant(P<0.0001).

Within the subgroup of patients for whom both knees wereexamined, bilateralanteriorPD-weightedhighsignal intensitywasdetectedin23/29(79%)patients(Fig.4).Four ofthem(14%)hadnoPD-weightedhighsignalintensity bilat-erally,and2(7%)hadhyperintensityinonlyoneknee.

A history of repetitive microtrauma was noted in 92 (48.2%) knees. The association between anterior PD-weightedhigh signal intensity andrepetitivemicrotrauma wasstatisticallysignificant(P=0.001).

Anterior knee pain was present in 123 (64.4%) knees. AnteriorPD-weightedhighsignalintensitywasdetectedin 104(84.6%)ofkneeswithpain,andin54(79.4%)ofknees withoutpain;thedifferencewasnotstatisticallysignificant (P=0.42).

Only2(1%)patientshadsignsofregionalanterior subcu-taneousinflammationclinically.Thesignalwassimilartothe othercases(Fig.5).Nofindingrelatedtoregional inflamma-tionwasfoundinotherpatientsonphysicalexamination.

Whilepatellarand/ortrochlearchondropathywereseen in 92 (74.8%) of the patients withanterior knee pain,31 (45.6%)of theknees without painhad chondropathy; the differencewasstatisticallysignificant(P<0.001).The asso-ciationbetweenpatellofemoralchondropathyandanterior PD-weightedhigh signalintensitywasalsostatistically sig-nificant(P=0.046).

Amedialplicawasidentifiedin43(35.0%)ofkneeswith anteriorpainandin16(23.5%)of kneeswithoutpain;the differencewasnotstatisticallysignificant(P=0.14).There wasalsonosignificantassociationbetweenanterioredema andmedialplica(P=0.83).

Effusion wasdetected in 65 (52.8%) of the knees with anteriorpainandin24(35.3%)of thekneeswithoutpain. The relationship between anterior pain and effusion was statistically significant (P=0.02). However, there was no significant association between anterior edema and joint effusion(P=0.12).

Whilesynovitiswasseenin84(44%)ofallknees,ofwhom anterior painwas detected in 64 (76.2%), 59/107 (55.1%) of the knees without synovitishad anteriorpain; the dif-ference was statistically significant (P=0.003). But there wasnosignificantassociationbetweenanteriorPD-weighted highsignalintensityandsynovitis(P=0.33).

Hoffapathologieswereseenin59(30.9%)oftheknees, and 43 (72.9%) had anterior pain. Pain was detected in 80/132 (60.6%) knees with normal Hoffa. The pain rates were similar (P=0.102). The relationship between patho-logicalHoffaintensityandanteriorPD-weightedhighsignal intensitywasnotstatisticallysignificant(P=0.082).

Aquadriceps fat-padmasseffectwasseen in16(8.4%) of all knees, of whom pain was detected in 10 (62.5%). Also,113/175(64.6%)ofthekneeswithoutmasseffectshad anteriorpain.The relationship betweenanteriorpainand quadricepsfat-pad masseffectwasnotstatistically signif-icant(P=0.868).Therewasalsonostatisticallysignificant associationbetweenanteriorPD-weightedhighsignal inten-sityandquadricepsmasseffect(P=0.314).

Patellar tendon abnormalities were seen in 3 (2.4%) knees.Twokneeswithproximalpatellartendinitisand adja-centbonemarrowedemahadpaininthisregion.Theywere alsoaccompaniedbychronicOsgood-Schlatterdisease.One patient hada postoperative defect and tendinosisdue to

Figure4. Fourty-five-year-old,87kgweightedmanwithbilateralanteriorkneepainandclinicalfindingsofpatellofemoralosteoarthritis. Fat-suppressedPD-weightedMRimagesinthesagittalplaneoftheright(A)andtheleft(B)kneeshowanterioredema(arrowheads)and end-stagepatellofemoralchondropathy(arrows)prominentontheleftside.Notethejointeffusionbilaterally(*_).

Figure 5. Thirty-three-year-old, 55kg weighted woman pre-sentingwithanterior kneepain,swelling,rednessandincreased temperature(regionalinflammation).Fat-suppressedPD-weighted MR imageinthesagittalplaneshowsprepatellaredema (arrow-heads)withill-definedborders.Nofluidcollectionconsistentwith anteriorbursitisisvisible.

aprevious(∼1yearago)anteriorcruciateligament recon-structionsurgery. This patient hadnoanterior kneepain. AnteriorPD-weightedhighsignalintensitywasdetectedin allofthesepatients.Noquadricepstendontearortendinitis wasseeninanypatient.

Fluidcollectionconsistent withanteriorbursitis,acute Osgood-Schlatter disease, or traction apophysitis was not seeninanycase.

The association between anteriorknee pain and ante-rioredema,andthesignificantandnon-significantvariables associatedwithananteriorPD-weightedhighsignalintensity areshowninTable1.

Discussion

Wefoundahighprevalenceofanteriorsubcutaneousedema (82.7%)inourstudy,andtherewasnosignificantassociation betweenthissignal andanteriorkneepain.Wealsofound nosignificantassociationbetweensubcutaneousedemaand otherpossiblecausesofanteriorpain,suchasmedialplica, effusion, synovitis, Hoffa fat-pad and quadriceps fat-pad abnormalities,and patellar tendon pathologies.We found significantassociationsbetweenpatientage,bodyweight, andhistoryofrepetitivemicrotraumaorchronicirritation, andpatellofemoralchondropathy.

Anterior subcutaneous edema is a common finding on routine knee MRI studies, and we noticed that this find-ing may confused with bursitis or local inflammation by radiologists,especiallywhentheyareinexperiencedin mus-culoskeletalimagingornotsufficientlyinformedaboutthe clinicalfindings.Inaddition,clinicianswhowantto evalu-atetheMRimagesoftheirpatientsmayconsiderthisfinding asan inflammatorychange in patients withanteriorknee pain.Thus,anincorrectassessmentofthisfindingmaylead tounnecessaryanti-inflammatorytreatmentofpatientsby clinicians.

Theprepatellarbursaislocatedbetweenthepatellaand overlyingsubcutaneoustissue.Thesuperficialinfrapatellar bursaislocatedinthesubcutaneousfatbetweenthedistal third of the patellar tendon, the tibial tubercle,and the overlyingskin,andthisbursawasfoundin55%ofcasesin acadaveric investigation [4,6]. Mostof the bursaein the bodyarenotnormallyvisualizedonimaging,andtheterm ‘‘bursitis’’referstopathologicalenlargementofthebursa

[4,7]. Bursitis can be caused by excessive local friction,

infection, arthritides, or direct trauma. Inflammation of prepatellar and superficial infrapatellar bursae usually occursfromchronicrepetitivetraumaduetooccupational kneeling or activities requiring a significant amount of kneeling, e.g., housemaids, clergy (clergyman’s knee), carpet-layers,and wrestlers [4]. The effusion and edema ofprepatellar bursitis mayextendand communicate with the superficial infrapatellar bursa [8]. Bursitis should be

Pleasecitethisarticleinpressas:Unlu EN,etal.Assessmentofanteriorsubcutaneoushypersignal on proton-density-Table1 Correlationbetweenanteriorpainandanterioredema,andthesignificantandnon-significantvariables asso-ciatedwithananteriorPDhighsignalintensityonMRimagingoftheknee.

AnteriorPD

highsignalintensity[+]

n=158(82.7%)

AnteriorPD

highsignalintensity[—]

n=33(17.3%) P Anteriorpain Yes,n(%) 104(84.6) 19(15.4) 0.42 No,n(%) 54(79.4) 14(20.6) Age

mean,SD[min—max],years

44.76±13.35[15—80] 28.48±6.82[16—45] <0.0001 Gender Female,n(%) 73(86.9) 11(13.1) 0.24 Male,n(%) 85(79.4) 22(20.6) Side Right,n(%) 82(82.8) 17(17.2) 0.99 Left,n(%) 76(82.6) 16(17.4) Weight mean,SD[min—max],kg 77.25±12.39[50—130] 67.82±9.89[50—86] <0.0001 Repetitivemicrotrauma 0.001 Yes,n(%) 85(92.4) 7(7.6) No,n(%) 73(73.7) 26(26.3) Patellofemoralchondropathy Yes,n(%) 107(87.0) 16(13.0) 0.046 No,n(%) 51(75.0) 17(25.0) Medialplica Yes,n(%) 48(81.4) 11(18.6) No,n(%) 110(83.3) 22(16.7) 0.83 Effusion Yes,n(%) 78(87.6) 11(12.4) 0.12 No,n(%) 80(78.4) 22(21.6) Synovitis Yes,n(%) 69(86.2) 15(13.8) No,n(%) 89(80.2) 18(19.8) 0.33 Hoffapathology Yes,n(%) 53(89.8) 6(10.2) 0.082 No,n(%) 105(79.5) 27(20.5)

Quadricepsfat-padmass effect

Yes,n(%) 15(93.8) 1(6.2) 0.314

No,n(%) 143(81.7) 32(18.3)

Numbersinbracketsareranges;SDindicatesstandarddeviation.Boldindicatessignificantdifferences(P<0.05).

differentiated from subcutaneous edema on imaging. In contrast to edema, bursitis appears as a localized fluid collectionwithwell-definedborders[7].

InthestudybyRothetal.,whichevaluatedthe quadri-ceps fat-pad signal intensity and enlargement on MRI, prepatellar edema was present in 90/92 (98%) patients

[1]. They reported that, although the cause of this find-ingwasunclear,thehighprevalenceandlackofsignificant correlation withanterior kneepain suggested that it is a physiological phenomenon related to knee movement or mechanics. Our results were similar, and we agree with Rothetal.Wealsothinkthatthisfindingmaybea degen-erativechange, whichbecomesprominent withagingand overuseoftheknee.AlthoughRothetal.[1]andShabshin etal.[9]reportedthatquadricepsfat-padedemawithmass effectis significantly associated with anterior knee pain,

TsavalasandKarantanasconcludedthatthisfindingis com-mononMRIandrarely associatedwithanteriorpain[10]. Wealsodidnotfindsignificantcorrelationbetween quadri-cepsfat-padedemawithmasseffectandanteriorkneepain in our study. We found a significant correlation between subcutaneousedemaandpatellofemoralchondropathy.We suggest that this correlation may depend on the similar-ity of predisposing factors of both processes, including age(forosteoarthritis),weightandrepetitivemicrotrauma

[11,12].

Therearesomestudiesaboutedematoussignalchanges inotherpartsofthebody,whichmayberelated.Haliloglu etal.evaluatedtheprevalenceofperitrochanterichighT2 signal (peritendinitis, peritrochanteric edema) on routine hipMRI[13].Theyreportedthatbilateralperitrochanteric edemawasacommonfindingonMRIinthoseover40years

ofageandwasnotalwaysrelatedtotheclinical findings. Theysuggestedthatthisfindingmaybeapartof degener-ative process and may not be reportednecessarily if the clinical findings do not also support greater trochanteric painsyndrome.Genuetal.evaluatedtheT2-STIRhigh sig-nalintensityinthelumbosacraladiposetissueandfounda significant relationshipbetween non-inflammatory infiltra-tionandoverweight,age,andhospitalizedstatus[14].They suggested that this phenomenon seemed to be an inter-stitial edema, related to subcutaneous stasis, and should not be confused with local inflammation, unless there is no localunderlying pathology, such asinterspinous infec-tiousdiseases,myositis,orarecentsurgery.Shietal.also reportedthatposteriorlumbarsoft-tissueedemaorfluidis frequentlynotedandmorecommonlyseeninobesepatients

[15].Kunindefinedbridgingseptaintheperirenaladipose tissue,whicharefibrouslamellaethatcan actasbarriers tolimit distribution of diseaseprocesses [16]. Thickening oftheperinephric bridgingseptais anon-specificimaging finding,andseptamaybethickenedduetoedemaor fibro-sis, as a result of large parapelvic cysts causing chronic, tense capsular congestion [17]. Thickened bridging septa shouldalso notbe considered pathologicalif thereis any underlying cause, such as inflammation, hemorrhage, or a tumor. In light of these studies, anterior subcutaneous edema seems to be a physiological and/or degenerative phenomenon, whichmay berelated toage, weight,knee movementsandrepetitivemicrotrauma. Itisimportantto distinguishthisimagingfindingfrompathologicalconditions to protect patients from unnecessary anti-inflammatory treatments.

Ourstudyhassomelimitations:itsretrospectivenature, the use of contrast material for a few patients, absence of anterior bursitis cases, and unknown past histories of anti-inflammatory therapy before the MRI examination in all patients. Also, our MRI protocol did not include fat-suppressed T2-weighted and/or STIR sequences, and we couldnotcomparethedifferenceofsequences.Inaddition, wedidnothavehistologicalorpathologicaldata;however, biopsy or surgery is not necessary for this very common andasymptomaticfinding.Furtherinvestigationswithlarger serieswouldbeuseful.

In conclusion,anedematoussignal inthe anterior sub-cutaneous adipose tissue of the knee is a very common findingonPD-weightedMRI andseemstobea physiologi-calphenomenonordegenerativechangerelatedtopatient age, weight,andknee movement or mechanics.It should notbereportedasapathologicalfindingonMRIunlessthe clinicalfindingsalsosupportregionalinfectionor inflamma-tion.

Financial

support

Nofinancialsupportwasreceivedforthissubmission.

Disclosure

of

interest

Theauthorsdeclarethattheyhavenocompetinginterest.

References

[1]Roth C,JacobsonJ, JamadarD, CaoiliE, MoragY, Housner

J. Quadriceps fat pad signal intensity and enlargement on

MRI:prevalenceandassociatedfindings.AJRAmJRoentgenol

2004;182:1383—7.

[2]Samim M, Smitaman E,Lawrence D,Moukaddam H. MRIof

anteriorkneepain.SkeletalRadiol2014;43:875—93.

[3]FaruchBilfeldM,LapègueF,BrunC,BakoucheS,CambonZ,

BrucherN,etal.Boneabnormalitiesoftheknee:MRIfeatures.

DiagnIntervImaging2016;97:779—88.

[4]HirjiZ,HunjunJS,ChoudurHN.Imagingofthebursae.JClin

ImagingSci2011;1:22.

[5]HillCL,GaleDR,ChaissonCE,SkinnerK,KazisL,GaleME,etal.

Periarticular lesions detected on magnetic resonance

imag-ing:prevalenceinkneeswithandwithoutsymptoms.Arthritis

Rheum2003;48:2836—44.

[6]Viegas FC, Aguiar RO, Gasparetto E, Marchiori E, Trudell

DJ, Haghighi P, et al. Deep and superficial infrapatellar

bursae:cadavericinvestigationofregionalanatomyusing

mag-neticresonanceafterultrasound-guidedbursography.Skeletal

Radiol2007;36:41—6.

[7]ChatraPS.Bursaearoundthekneejoints.IndianJRadiol

Imag-ing2012;22:27—30.

[8]Niitsu M. Magnetic resonance imaging of the knee. Berlin:

Springer-Verlag;2013.

[9]ShabshinN,SchweitzerME,MorrisonWB.Quadricepsfatpad

edema:significanceonmagneticresonanceimagesoftheknee.

SkeletalRadiol2006;35:269—74.

[10]TsavalasN,KarantanasAH.Suprapatellarfat-padmasseffect:

MRIfindingsandcorrelationwithanteriorkneepain.AJRAmJ

Roentgenol2013;200:291—6.

[11]Kok HK, Donnellan J, Ryan D, Torreggiani WC. Correlation

between subcutaneouskneefatthickness and

chondromala-ciapatellaeonmagneticresonanceimagingoftheknee.Can

AssocRadiolJ2013;64:182—6.

[12]KimYM,JooYB.Patellofemoralosteoarthritis.KneeSurgRelat

Res2012;24:193—200.

[13]Haliloglu N, Inceoglu D, Sahin G. Assessment of

per-itrochanterichighT2signaldependingontheageandgender

ofthepatients.EurJRadiol2010;75:64—6.

[14]GenuA,KochG,ColinD,AhoS,PearsonE,BenSalemD.Factors

influencingtheoccurrenceofaT2-STIRhypersignalinthe

lum-bosacraladiposetissue.DiagnIntervImaging2014;95:283—8.

[15]Shi H, SchweitzerME, Carrino JA, Parker L. MR imagingof

thelumbarspine:relationofposteriorsoft-tissueedema-like

signalandbodyweight.AJRAmJRoentgenol2003;180:81—6.

[16]KuninM.Bridging septaoftheperinephric space:anatomic,

pathologic, and diagnostic considerations. Radiology

1986;158:361—5.

[17]Aizenstein RI, Wilbur AC, O’Neil HK. Interfascial and

per-inephricpathways in thespreadof retroperitoneal disease:

refined concepts based on CT observations. AJR Am J