眼科文献(1)

眼科文献IntroductionAnterior-segmentopticalcoherencetomography(AS-OCT)(VisanteOCT;CarlZeissMeditec,Dublin,Cali-fornia,USA)isanewnon-contactimagingtechniquethatprovideshigh-resolutioncross-sectionalimagesoftheanteriorsegmentoftheeye.Thistechniqueusesinfrared(1310nm)radiationtoprovidereal-timeimagesoftheanteriorsegmentandallowsobjectiveandquantitativeassessmentsofanteriorchamberangle(ACA)structures(Izattetal.1994;Koopetal.1997;Hoeraufetal.2000,2002;Radhakrishnanetal.2001,2005,2007;Wirbelaueretal.2002;Matsu-nagaetal.2004;Ishikawa2007;Lietal.2007;Nolanetal.2007;Con-soleetal.2008;Leungetal.2008;Sakataetal.2008a,2008b).Therecentlyreleasedversion2.0ofVisanteOCThasseveralfeaturesthatwerenotavailableinthepreviousver-sion.Anenhancedanteriorsegmentsinglemodeprovidesanimageaver-agedfromfourconsecutiveanterior-segmentscans.Thisreducessignalnoise,improvescontrastandprovidesbettervisualizationoffaintorindis-tinctfeatures.Anothernewfeatureistheirido-cornealtoolset,whichincludessemi-automatedquantitativedeterminationoftheACAparameters:Characteristicsandreproducibilityofanteriorchamberangleassessmentbyanterior-segmentopticalcoherencetomographyDongY.Kim,1KyungR.Sung,1SungY.Kang,1JungW.Cho,1KyoungS.Lee,1SeongB.Park,1,2SoonT.Kim1andMichaelS.Kook11DepartmentofOphthalmology,AsanMedicalCentre,CollegeofMedicine,UniversityofUlsan,Seoul,Korea2DepartmentofOphthalmology,BundangCHAMedicalCentre,CollegeofMedicine,UniversityofCHA,SungNam,KoreaABSTRACT.Purpose:Toevaluatethebasiccharacteristicsandreproducibilityofanteriorchamberangle(ACA)measurementsdeterminedbyanterior-segmentopticalcoherencetomography(AS-OCT)inopen-angleandprimaryangleclosuresuspect(PACS)patients.Methods:Thirty-nineopen-angleand18PACSpatientswereimagedforACAbyAS-OCT.Subjectsunderwentimagingofthenasal,temporalandinferiorACAunderconditionsofconstantlight,anddarkness.Foranalysis,weusedthreeACAparametershandledbytheVisanteOCTsoftware:angleopeningdistanceat500lm(AOD500),trabecular-irisspaceareaat500lm(TISA500)andanglerecessareaat500lm(ARA500).Fordeterminationofinter-sessionreproducibility,asinglewell-trainedoperator(D.Y.K.)scannedallpatientsattwodifferentvisits.Fordeterminationofinter-operatorvariabil-ity,asecondoperator(S.B.P.)acquiredanothersetofimagesindependently.Threesetsofimageswereacquiredatleast24hourapart.Results:Allparametersweresignificantlydifferentwhenmeasuredbothinlightanddarkness,andinthenasalandtemporalquadrants.TherewerenosignificantdifferencesbetweentheleftandrighteyesinthethreeACAparametersinallquadrants.Thetemporalanglewaswiderthanthenasalandinferiorangles.Allparametersofthenasal,temporalangleshadexcellentinter-sessionandinter-operatorreproducibility[intra-classcorrelationcoeffi-cient(ICC)0.796–0.981],butthesevalueswereslightlylowerforinferioranglemeasurements(ICC0.662–0.892)inbothopen-angleandPACSgroups.Conclusion:AS-OCTprovidesquantitativeandreproducibleassessmentofACA.Reproducibilitywaslowerintheinferioranglecomparedwiththenasalandtemporalangles,perhapsbecauseofvariableplacementofthescleralspur.Keywords:anteriorchamberangle–anterior-segmentOCT–narrowangle–reproducibilityActaOphthalmol.2011:89:435–441ª2009TheAuthorsJournalcompilationª2009ActaOphthalmoldoi:10.1111/j.1755-3768.2009.01714.xActaOphthalmologica2011435angleopeningdistanceat500lm(AOD500),trabecular-irisspaceareaat500lm(TISA500)andanglerecessareaat500lm(ARA500).Theseparameterscanbeusefulintheobjec-tiveandquantitativedetectionandassessmentofnarroworclosedangleswithoutdistortionfromACAconfigu-ration.Inordertoevaluateabnormalanglestructure,itisnecessarytoknowthecharacteristicsofthenormalACAconfigurationasmeasuredbymanufacturer-providedparameters.Furthermore,theabsenceoftest–ret-estvariabilityshouldbeconfirmedtovalidatetheclinicalutilityofACAparameters.Previousresearchexaminedthereli-abilityofAS-OCTmeasurementsbyuseofprototypeAS-OCTorcustomanalysissoftware(Lietal.2007;Nolanetal.2007;Radhakrishnanetal.2007;Consoleetal.2008;Leungetal.2008;Sakataetal.2008a,2008b).However,nostudieshaveexaminednormalACAcharacteristicsbasedonmanufacturer-providedparametersand,tothebestofourknowledge,therehavebeennostudiesonthereproducibilityofACAassess-mentusingthesoftwareinversion2.0oftheVisanteOCT.Inthisstudy,weevaluatedACAcharacteristicsinyounghealthyindividuals,andexploredinter-sessionandinter-opera-torvariationsinducedbytheimageacquisitionusingmanufacturer-pro-videdACAparameters.Furthermore,weevaluatedthisACAmeasurementreproducibilityinprimaryangleclosuresuspect(PACS)patients.MaterialsandMethodsDatawerecollectedprospectivelyfromyounghealthyvolunteersandPACSpatientswhounderwentexami-nationatourglaucomaclinic.Healthyparticipantswereemployeesoftheeyeclinicortheirspousesorrelatives.Informedconsentwasobtainedfromallparticipants.ThestudywasapprovedbytheInstitutionalReviewBoardoftheAsanMedicalCentre,andadheredtothetenetsoftheDec-larationofHelsinki.Weenrolled39open-angle(healthy)and18PACSpatientswhometthefollowinginclu-sioncriteria.Allparticipantshadvisualacuityofatleast20⁄40withnoevidenceofoculardisease.Gonios-copywasperformedwithaGoldmanntwo-mirrorlenswithanarrowbeamwidthofdimlightilluminationinadarkenedroom.Patientswithvisibleanteriortrabecularmeshworkatfourquadrantswereincludedintheopen-anglegroup.Patientswithapposi-tionalcontactbetweenperipheralirisandposteriortrabecularmeshwork‡270degreeswereincludedinthePACSgroup(Fosteretal.2002).Patientswithanyhistoryoftopicalorsystemicmedicationusethatcouldaffectthedrainageangleorpupillaryreflex,anyhistoryofpreviousintraoc-ularsurgery,lasertrabeculoplasty,laseriridoplastyorlaseriridotomywereexcludedfromthestudy.TherelationshipbetweensphericalequivalentandACAparameterswasdeterminedbyPearsoncorrelationanalysis.Forthepurposeofcharac-terizingnormalanglebyACAparam-eters,allhealthyparticipantswereimagedfornasal,temporalandinfe-riorACAusingAS-OCT(VisanteOCTversion2.0)undercontrolledroomlightconditions(60cd⁄m2)andindarkness(0.5cd⁄m2).TheACAwasimagedwiththe‘enhancedante-riorsegmentsingle’protocol(scanlength16mm;256A-scans)byasin-glewell-trainedoperator(D.Y.K.).Fortestinginter-sessionreproducibil-ity,thesameoperatorimagedeacheyeattwodifferentvisits.Asecondoperator(S.B.P.)imagedeacheyeaccordingtothesameprotocolwith-outanyknowledgeoftheresultsfromtwopreviousimageacquisitionsfortestinginter-operatorreproducibility.Thethreevisitstookplacewithinasingleweekatleast24hourapart.Iftheinferioranglewasobscuredbythelowereyelid,thepatientwasaskedtopulltheliddownagainsttheinfra-orbitalrimwithoutindentationoftheeyeball.EachoperatoranalysedACAsinde-pendentlyusingsoftwareprovidedbyVisanteOCTtodeterminepupilsize,anteriorchamberdepth(ACD),AOD500,TISA500andARA500withtheimagesthattheyhadtaken.First,theoperatorusedcaliperstomeasurepupilsizeandACDandthenassignedthescleralspur.Next,theVisanteOCTsoftwareautomaticallycalcu-latedAOD500andTISA500.Finally,theoperatordeterminedtheapexofanglerecess,afterwhichthesoftwarecalculatedARA500automatically.TheAOD500wasdefinedasthelineardis-tancebetweenthepointoftheinnercorneoscleralwall(whichwas500lmanteriortothescleralspur)andtheiris.TheARA500wasdefinedasthetriangularareaformedbytheAOD500.Thecornersofthetriangleweretheanglerecess(theapex),theirissurfaceandtheinnercorneoscleralwall.TheTISA500wasdefinedasthetrapezoidalareawiththefollowingboundaries:anteriorly,theAOD500;posteriorly,alinedrawnfromthescleralspurperpendiculartotheplaneoftheinnerscleralwalltotheoppos-ingiris;superiorly,theinnercorne-oscleralwall;and,inferiorly,theirissurface(Fig.1).IllustrationsoftheproceduresusedforthemeasurementofAOD500,TISA500andARA500areprovidedelsewhere(Pavlinetal.1992;Radhakrishnanetal.2005).ThedatacouldnotbedescribedbyGaussianstatistics,soweusednon-parametricmethods.ACAparameters(determinedbyAS-OCTundercon-stantlightanddarkness),thesymme-tryofallangleparametersoftherightandlefteyesandthedifferencesinACAparametersamongthethreequadrants(inferiorversusnasal,infe-riorversustemporal,nasalversustem-poral;thefirstimageobtainedduringthefirstvisitwasused)wereanalysedusingtheWilcoxonsigned-ranktest.Weusedintra-classcorrelationcoeffi-cients(ICCs)toexploreinter-sessionandinter-operatorreproducibility.TheagreementofACAparametermeasurementsbetweendifferentses-sionsanddifferentoperatorswasdeterminedbyBland–Altmanplots.Weusedrighteyeimagesforstatisti-calanalysis,exceptwhencomparisonsweremadebetweentherightandlefteyes.Statisticalanalyseswereper-formedusingspssversion15.0(SPSSInc.,Chicago,Illinois,USA).ResultsAllparticipantswereAsian(Korean).Theopen-angle(healthy)groupcom-prised14menand26women;thePACSgroupconsistedoffourmenand14women.Themean±standarddeviation(SD)agewas29.2±5.44(range23–42)yearsintheopen-angle(healthy)groupand64.2±10.1(range44–81)yearsinthePACSgroup.Themeansphericalequivalent(SE)was)2.04±2.11(range0.25)7)dioptresintheopen-anglegroup436and1.35±1.18(range)1.03.0)dioptresinthePACSgroup.ThreeACAparameters–AOD500,TISA500andARA500–atthetemporalsideshowedarelationshipwithsphericalequivalentsofallparticipants(R2=0.230,281,172;allp<0.0001)(Fig.2).ACAparametersofnormalACAinlightanddarkconditionsaredescribedinTable1.Medianpupildiameterwassignificantlysmallerinlight(3.91mm)thanindarkness(5.64mm,p<0.0001).TheACDsdidnotdiffersignificantlybetweenlightanddarkconditions.Inthenasal,temporalandinferiorquad-rants,AOD500,TISA500andARA500weresignificantlylargerinlightthanindarkness(Table1).Table2summarizestheresultsofthecomparisonoftheACAparame-tersbetweenrightandlefteyesinopen-anglepatients.Medianpupilsize(3.30mm)andACD(3.91mm)oftherighteyewerenotsignificantlydiffer-entfromthoseofthelefteye(3.30mmand3.97mm,andp=0.063and0.748,respectively).Therewerenosignificantbetween-eyediffer-encesinthethreeACAparametersinanyofthethreequadrants(Table2).Inboththeopen-angleandPACSgroups,AOD500,TISA500andARA500werelargestinthetemporalquadrant,secondlargestinthenasalquadrantandsmallestintheinferiorquadrant.Intheopen-anglegroup,allthreeparameterswerelargerinthetemporalquadrantthaninthenasalandinferiorquadrants.However,therewasnosignificantdifferencebetweenthenasalandinferiorquad-rantsinanyofthethreeparameters.InthePACSgroup,AOD500andTISA500weresignificantlylargerinthetemporalandnasalquadrantsthanintheinferiorquadrant;how-ever,therewasnosignificantdiffer-encebetweenthetemporalandnasalquadrants.ARA500revealednosignif-icantdifferenceamongthethreequad-rants(Table3).OurdeterminationsofACDandpupilsizehadexcellentinter-sessionreproducibilityinlight(ICC0.863–0.994)andindark(ICC0.934–0.998)conditionsinboththeopen-angleandPACSgroups.AllACAparametersinthenasalandtemporalquadrantsdemonstratedexcellentinter-sessionreproducibilityinlightanddarkness(ICC0.796–0.981).Intheinferiorquadrant,inter-sessionreproducibilitywasslightlylowerandhadlargerranges(ICC0.662–0.843)inbothgroups.AllACAparametersinthenasalandtemporalquadrantsexhib-itedexcellentagreementbetweenthetwooperatorsinlightanddarkness(ICC0.838–0.956),butsuchagree-mentwasslightlylowerwheninferior-quadrantdatawereanalysed(ICC0.667–0.892)(Table4).AgreementofACAparametermeasurementsbetweentwodifferentsessionsandtwodifferentoperatorswereanalysedbyBland–Altmanplots(Fig.3).DiscussionInthisstudy,weusedAS-OCTtoconfirmthatpupilsizewassmallerandACAparameters(AOD500,TISA500,ARA500)largerinlightthanindarkness.ThisindicatesthatAS-OCTmeasurementofACAparametersshouldbeperformedundercontrolledlightingconditions.AnadvantageofAS-OCTovergoni-oscopyorultrasoundbiomicroscopy(UBM)forACAassessmentisthatFig.1.Semi-automatedquantitativedeterminationoftheanteriorchamberangleparametersassessedbythebuilt-inanalysissoftwareofanterior-segmentopticalcoherencetomography:anteriorchamberdepth(ACD),angleopeningdistanceat500lm(AOD500),trabecular-irisspaceareaat500lm(TISA500)andanglerecessareaat500lm(ARA500).(A)(B)(C)0.0000.2000.4000.6000.8001.0001.2000.0000.2500.5000.7501.0001.2500.000–7.5–5–2.502.5–7.5–5–2.502.5–7.5–5–2.502.50.1000.2000.3000.4000.5000.600437imageacquisitionandsubsequentmeasurementcanbeperformedunderlight-controlledconditions,withpatientsinsittingpostures,andwith-outalterationofanglestructurebyirregularlightingandphysicalcontact.However,itisclearlynecessarytocontrollightingconditionswhenusingAS-OCT.OurcomparisonsoftherightandlefteyesofindividualparticipantsshowedthatallACAparametersweresimilarinallthreequadrants.Thissimilarity,asassessedbycom-merciallyavailableAS-OCT,hasnotbeenreportedpreviously.Ourresultssuggestthatasignificantdif-ferenceinACAparametersbetweentherightandlefteyesshouldbesuspectedasapossiblesignofpathology.DifferencesinACAparametersamongquadrants(nasal,temporalandinferior)measuredwithAS-OCTinhealthyeyeshavenotbeenexam-inedthoroughly.Inaddition,therehavebeennoquantitativecompari-sonsofACAparametersamongdif-ferentquadrants.Ourresultsindicatethatthetemporalquadrantwasthelargestamongthethreequadrantsinbothopen-angleandPACSpatients.TheseobservationsappeartobeinlinewiththeresultsofKunimatsuetal.(2005)(assessedbyUBM)andSakataetal.(2008a,2008b)(assessedusingaprototypeAS-OCT).Thesepreviousstudiesdidnotreportquanti-tativemeasurementsondifferentquadrantsbutevaluatedonlytheper-centagesofclosedanglesindifferentquadrants.Theyreportedthatthetemporalanglehadthelowestper-centageofclosedanglesandthatthenasalanglehadalowerpercentageofclosedanglesthanhadtheinferiorangle.Inotherwords,ourquantita-tivemeasurementsofACAparametersinAS-OCT,whichshowthatthetem-poralanglewaslargestandtheinfe-rioranglesmallest,aresupportiveofpreviousresultsthatshowedthetem-poralangletohavethelowestper-centageofclosedangle.Accordingtopreviousstudies,agreementbetweengonioscopicandAS-OCTassessmentispoor(Nolanetal.2007;Sakataetal.2008a,2008b).Nolanetal.(2007)speculatedthatinadvertentindentationorexces-sivelightexposureoccurringduringgonioscopyordifferentlandmarksfordefiningangleclosurebetweengonio-scopicandAS-OCTassessmentmightcontributetodisparity.Studyresultsdifferregardingtherelationshipbetweenangleclosureincidenceanddifferentquadrants,evenamongthesamegonioscopicevaluations.Inapopulation-basedstudyHeetal.(2006)reportedthattheinferioranglehadthelowestpercentageofclosedangles.However,Nolanetal.(2007)reportedthatthehighestproportionofclosedangleswasfoundintheinfe-riorquadrantonbothgonioscopyandAS-OCT.Webelievethatthelimita-tionofgonioscopicmeasurement,includingtheinherentlysubjectivenat-ureoftheexam,inadvertentindenta-tionordifferentlightingconditions,maybethesourceoftheseconflictingresults.Furthermore,asexplainedbySakataetal.(2008a,2008b),viewingTable1.Comparisonofanteriorchamberangleparametersmeasuredinopen-anglepatientsbetweenlightanddarkconditions(n=39).QuadrantParameterLightDarkp-value*MedianRangeMedianRangeACD(mm)3.32.4803.9003.3052.5003.9500.56Pupil(mm)3.912.7505.3505.643.5707.000<0.0001NasalAOD500(mm)0.6510.1841.1950.5170.2831.09<0.0001TISA500(mm2)0.2340.0570.4440.1790.0970.448<0.0001ARA500(mm2)0.270.0620.5450.2080.0980.498<0.0001TemporalAOD500(mm)0.7550.2221.5900.6680.2901.349<0.0001TISA500(mm2)0.2650.0990.5650.2360.1020.473<0.0001ARA500(mm2)0.310.1170.7800.2610.1130.581<0.0001InferiorAOD500(mm)0.5930.171.4970.5080.0651.148<0.0001TISA500(mm2)0.2240.0890.5140.1810.0550.391<0.0001ARA500(mm2)0.2670.0920.7830.2080.0690.500<0.0001*CalculatedusingtheWilcoxonsigned-ranktest.ACD,anteriorchamberdepth;AOD500,angleopeningdistanceat500lm;TISA500,trabecular-irisspaceareaat500lm;ARA500,anglerecessareaat500lm.Table2.Comparisonofanteriorchamberangleparametersmeasuredinopen-anglepatientsbetweentherightandlefteyes(n=39).QuadrantParameterRightLeftp-value*MedianRangeMedianRangeACD(mm)3.32.4803.9003.32.5903.9800.063Pupil(mm)3.912.7505.3503.972.4305.3800.748NasalAOD500(mm)0.6510.1841.1950.6620.3071.1540.812TISA500(mm2)0.2340.0570.4440.2360.1130.4430.505ARA500(mm2)0.270.0620.5450.2880.1290.6150.106TemporalAOD500(mm)0.7550.2221.5900.7290.2841.4790.219TISA500(mm2)0.2650.0990.5650.2620.1100.5570.548ARA500(mm2)0.310.1170.7800.310.1360.8630.723InferiorAOD500(mm)0.5930.171.4970.6330.2671.3930.125TISA500(mm2)0.2240.0890.5140.2330.1020.5380.153ARA500(mm2)0.2670.0920.7830.2830.1110.8190.267*CalculatedusingtheWilcoxonsigned-ranktest.ACD,anteriorchamberdepth;AOD500,angleopeningdistanceat500lm;TISA500,trabecu-lar-irisspaceareaat500lm;ARA500,anglerecessareaat500lm.438temporalandnasalanglesduringgon-ioscopycanbedifficultandsubjective.ItisapparentthatfurtherstudyofquantitativeACAmeasurementsindifferentquadrantsofalargenumberofpatientsiswarranted.Currently,AS-OCTdoesnotallowcontinuousmeasurementofa360-degreeangle.Therefore,multiplescans(includingverticalandhorizontalcross-sectionalimaging)arerequiredtoevaluateACAconfigurationbecausedifferentquadrantsmighthavedifferentACAwidths.Wefoundverygoodinter-sessionandinter-operatorreproducibilityfornasal,temporalandinferiorquad-rants,examinedeitherinlightorindarkness.ApreviousstudywithaprototypeAS-OCTandcustomsoft-warereportedthatthelong-termintra-operatorreproducibilityofAODinthenasalangle(asassessedbyICC)was0.91and0.95inlightanddarkness,respectively;thecorrespond-ingfiguresforTISAwere0.89and0.92.Thesevaluesareclosetoours(0.93and0.94,and0.88and0.93,respectively)(Lietal.2007).Radha-krishnanetal.(2007)reportedvaluesof0.84and0.93,and0.85and0.92,respectively;thesefiguresarealsoquitesimilartoours.Ourinter-sessionandinter-operatorreproducibilitieswerehigherwhennasal-andtempo-ral-quadrantdatawereanalysed,comparedwithinferior-quadrantinformationinboththeopen-angleandPACSgroups;thisisinaccor-dancewithpreviousreports(Radha-krishnanetal.2007;Sakataetal.2008a,2008b).Theinferior-quadrantreproducibilitymaybelowerbecauseofvariablescleralspurplacement:operator-dependentinputwasmoredifficultintheinferiorquadrantbecauseofdifficultyinacquiringhigh-qualityimages.Sakataetal.(2008a,2008b)reportedthatfewerscleralspursweredetectedintheinferiorquadrantthaninthenasalortempo-ralquadrant.Althoughdifficultyintheplacementoftheinferior-quadrantscleralspur,especiallyineyeswithnarrowangles,seemstobeeasedbytheenhancedqualityofoursystem,reproducibilitywasnonethelesspoor-estininferior-quadrantexaminations.Scanqualityisamajordeterminantofscleralspurlocationassignment,somanufacturer-recommendedqualitycriteriaforimagesmightbeusefulasTable3.Comparisonofanteriorchamberangleparametersinthreequadrantsmeasuredinopen-angleandprimaryangleclosuresuspect(PACS)patients.ParametersOpenangle(n=39)PACS(n=18)InferiorNasalTemporalp-value§InferiorNasalTemporalp-value§AOD500(mm)Median0.5930.6510.7550.163*<0.0001<0.0001à0.1460.2560.2860.005*0.0060.628àRange0.1701.4970.1841.1950.2221.5900.0490.2020.0400.4570.0440.768TISA500(mm2)Median0.2240.2340.2650.339*<0.0001<0.0001à0.0510.0860.1090.017*0.0010.308àRange0.0890.5140.0570.4440.0990.5650.0270.0920.0100.1770.0100.257ARA500(mm2)Median0.2670.270.310.840*<0.0001<0.00010.0860.1090.2560.176*0.8200.126àRange0.0920.7830.0620.5450.1170.7800.0400.1500.0100.2570.0150.270p-value:*comparisonbetweeninferiorandnasalquadrants;comparisonbetweeninferiorandtemporalquadrants;àcomparisonbetweennasalandtemporalquadrants.§CalculatedusingtheWilcoxonsigned-ranktest.AOD500,angleopeningdistanceat500lm;TISA500,trabecular-irisspaceareaat500lm;ARA500,anglerecessareaat500lm.439guidelinesforacceptanceofscanimages.Toimprovetest–retestrepro-ducibilitybypreciseplacementofscleralspurs,ahigher-qualityimageoftheanteriorsegmentangleisrequired.Useofthe‘enhancedhighresolution’protocol(scanlength10mm;512A-scans)insteadofthe‘enhancedanteriorsegmentsingle’protocol(scanlength16mm;256A-scans)maybehelpful.Wealsoexpectthatfullyautomatedmeasure-mentofACAparametersusingbuilt-insoftwaremayincreasetest–retestvariability.However,thesepostula-tionsremaintobevalidatedinfuturestudies.AlimitationofourstudyisthatwedidnotcomparethegonioscopicandAS-OCTassessment.Wedidsobecauseourprimarygoalwastoeval-uateACAcharacteristicsinopenangles(weexploredinter-sessionandinter-operatorvariationsinducedbytheimageacquisitionintheopen-angleandPACSgroupsusingmanu-facturer-providedACAparameters)andbecausesuchcomparativestudieshadalreadybeenpublished(Nolanetal.2007;Sakataetal.2008a,2008b).Asecondlimitationisthatarelativelysmallnumberofnarrow-anglepatientswereincluded.Wesug-gestthatourinvestigationbeextendedtoincludemorenarrow-anglepatientsinfuturestudy.Wealsosuggestfur-therstudyoflong-termmeasurementreproducibility,whichisimportantforlongitudinalfollow-upofACAchangeina