68Ga-Labeled Prostate-specific Membrane Antigen Ligand Positron Emission Tomography/Computed Tomography for Prostate Cancer: A Systematic Review and Meta-analysis

Review

–

Prostate

Cancer

68

_Ga-Labeled

_{Prostate-specific}

_Membrane

_Antigen

_Ligand

Positron

Emission

Tomography/Computed

Tomography

for

Prostate

Cancer:

A

Systematic

Review

and

Meta-analysis

Finn

E.

von

Eyben

*

,

Maria

Picchio

,

Rie

von

Eyben

,

Handoo

Rhee

,

Glenn

Bauman

a_{CenterofTobaccoControlResearch,Odense,Denmark;}b_{DepartmentofNuclearMedicine,IRCCSSanRafaeleScientificInstitute,Milan,Italy;}c_Department ofRadiationOncology,StanfordUniversity,SanFrancisco,CA,USA;d_{DepartmentofUrology,PrincessAlexandraHospital,Brisbane,Australia;}e_Department ofRadiationOncology,UniversityofWesternOntarioinLondon,ON,Canada

a v ai l a b l e a t w w w . s c i e n c e d i r e c t . c o m j o u r n al h o m e p a g e : w w w . e u r o p e an u r o l o g y . c o m / e u f o c u s Articleinfo Articlehistory: AcceptedNovember3,2016 AssociateEditor: JamesCatto Keywords: Prostatecancer Prostate-specificmembrane antigen

Positronemissiontomography Systematicreview

Meta-analysis

Abstract

Context:68_{Galliumprostate-specificmembraneantigen(PSMA)ligand}68_{Ga-HBED-CC-PSMA}

(68_{Ga-PSMA)isapromisingradiotracerforpositronemission tomography(PET)/computed}

tomography(CT)ofprostatecancer.

Objective: Toconductameta-analysistoevaluatedetectionrate,diagnostictestaccuracy,and

adverseeffectsof68_{Ga-PSMAPET/CTorPET/magneticresonanceimaging(MRI)forstagingof}

prostatecancerandforrestagingofrisingprostate-specificantigen(PSA)afterinitialtreatment.

Evidence acquisition: Followingthe PreferredReporting ItemsforSystematicreviews and

Meta-Analysis(PRISMA)guidelines,oursystematicreviewsearchedforarticlesinPubMedand EMBASEdatabasesfrom2012toJuly2016.Thereferencestandardwaspathologyafterbiopsy orsurgery.The analysesusedarandomeffectmodel andahierarchicalsummaryreceiver operatingcharacteristicmodel.

Evidencesynthesis: Fifteen68_{Ga-PSMAPET/CTstudieswith1256patientsmettheinclusion}

criteria.SevenstudiesofstagingPET/CTorPET/MRIdetectedaregionalsiteofcancerfor203of 273patients(74%).NinestudiesofrestagingPET/CTdetectedsitesofrecurrencein799of 983patients(81%)witha50%detectionrate(74of147patients)forrestagingPSAof0.2–0.49 ng/mlanda53%detectionrate(56of195patients)forrestagingPSAof0.50–0.99ng/ml. Staging68_{Ga-PSMAPET/CTinthestudieshadhigherdetectionratesofsitesintheprostatebed}

than restaging 68_{Ga-PSMAPET/CT (mean57% vs 14%, p=0.031, ttest). Bothstaging and}

restaging68_{Ga-PSMAPET/CTfoundthatasubgroupofthepatientshadmetastaticsitesin}

pelviclymphnodesordistantorgans.EightstudiesofstagingPET/CTundertookhistologic correlations.Weperformedprostate-segment-basedanalysisspecificallyregardingthe prima-rycancerlesion forfourof thesestudies,andpatient-basedanalysisspecificallyregarding pelviclymphnodemetastasesforfourotherstudies.Thepooledsensitivitiesforstaginginthe twogroupsofstudieswere70%and61%,andthepooledspecificitieswere84%and97%.Noneof thestudiesreportedcomplicationsfromthePET/CTimaging.

Conclusions: 68_{Ga-PSMAPET/CThasclinicalrelevancetodetectsitesofrecurrenceforpatients}

withPSArecurrenceafterradicalprostatectomy(RP)withPSAlevelslessthan1.0ng/ml.

Patientsummary: Cholinepositronemissiontomography(PET)/computedtomography(CT)

candetectsitesofrecurrentprostatecancerinanearlierphaseofprostate-specificantigen (PSA)recurrencethanbonescansandCTscans,butcholinePET/CTisrarelypositiveforpatients withrestagingPSAlevelsunder1ng/ml.Anewradiotracercalled68_{Ga-PSMAforPET/CTwas}

abletodetectsitesofrecurringcancerinupto50%ofpatientswhohadanearlyriseinPSA exceeding0.5ng/mlafterinitialradicalprostatectomy.Thepublishedstudiesdidnotreport adverseeffectsof68_{Ga-PSMAPET/CTimaging.}

#2016PublishedbyElsevierB.V.onbehalfofEuropeanAssociationofUrology.

* Correspondingauthor.Birkevej17,DK-5230OdenseM,Denmark.Tel.+4666145862; Fax:+4666145862.

E-mailaddress:fi[email protected](F.E.vonEyben).

http://dx.doi.org/10.1016/j.euf.2016.11.002

1. Introduction

Prostate cancer is the most frequent cancer in men in Western societies, and in men the cancer mortality is secondtothatforlungcancer[1].Localizedprostatecancer ismainlytreatedwithradicalprostatectomy(RP),external beamradiotherapy(EBRT),orbrachytherapy,butupto a third of patients develop a recurrence [2,3]. A rise of prostate-specificantigen(PSA)istypicallythefirst indica-tionofrecurrenceandiscalledPSA-onlyrecurrencebecause patients with PSA <10 ng/ml typically have negative findings with conventional computed tomography (CT) scansand99m_{Tcbonescans.Morerecently,hybridcholine}

positron-emission tomography (PET)/CT allowed reliable detectionofsitesofrecurrenceatPSAlevels>1–2 ng/ml

[4,5]. Guidelinesrecommenda potentiallycurative treat-ment of PSA recurrence after RP in the form of salvage radiotherapyfortheprostate bed(SRT)withoutguidance fromimaging.SRTgivesthebestresultswhenitisstarted while patients have restaging PSA of 0.2–0.5 ng/ml. Typically up to half of patients develop a second PSA recurrenceduringfollow-upafterSRT,andthedevelopment andapplicationofnewandmoresensitivePETprobesto guidesalvagetreatmentisafieldforongoinginvestigations toimprovesalvagetreatment.

Most prostate cancer cells express prostate-specific membrane antigen (PSMA), also denoted glutamate car-boxypeptidase2orN-acetyl-L-aspartyl-L-glutamate pepti-dase1(NAALAD1)[6].AGermangroupdevelopeda small-molecule inhibitor for PSMA, Glu-NH-CO-NH-Lys (Ahx)-{68_Ga-(N,N0

-bis-[2-hydroxy-5-(carboxyethyl)benzyl]ethylenediamine-N,N0_{-diacetic acid}}

(68_{Ga-HBED-CC-PSMA or} 68_{Ga-PSMA-11) [7], referred to}

here as68_{Ga-PSMA.In 2012,theGerman groupreported}

promising findings using this molecule as a PET/CT radiotracerforpatientswithprostatecancer[8],andlater meta-analyses confirmed the findings andindicated that

68

Ga-PSMA PET/CT detects prostate cancer better than radiolabeledcholinePET/CT[5,9].

2. Evidenceacquisition

2.1. Researchquestion

Weaimedtosummarize studiesofstagingandrestaging

68

Ga-PSMA PET/CT orPET/MRIfor patients withprostate cancer regarding detection of localized or metastatic prostate cancer. A second objective was to summarize imaging test accuracy of the new PET/CT method using pathologyafterbiopsyorsurgeryasthereferencestandard. Athird objectivewas to summarizeimaging-related side effectsfrom68Ga-PSMAPET/CTorPET/MRI.

2.2. Searchstrategy

Our systematic review and meta-analysis followed the Preferred Reporting Items for Systematic Reviews and Meta-Analysis(PRISMA)guidelines[10].InMay2016,two authors (FEvE and GB) registered a protocol for the

systematic review in the PROSPERO register (CRD 42016039690). Our systematic review included original research studies of staging or restaging with 68_Ga-PSMA

PET/CT or PET/MRI. The two authors undertook an electronic search inPubMed andEMBASEdatabases. The PubMedsearchusedmedicalsubjectheading(MeSH)terms andfreetext words: ((‘‘prostaticneoplasm*’’) [MeSH]OR (‘‘prostate cancer*’’)) AND ((‘‘positron emission tomogra-phy’’) [MeSH] OR (‘‘PET’’)) AND ((‘‘prostate membrane specificantigen’’)[MeSH]OR(‘‘PSMA’’))AND((‘‘*Gallium’’) [MeSH]) OR (‘‘*Ga’’)). Further, we searched for ongoing studiesinthedatabaseClinicalTrials(ClinicalTrials.gov).

Thetworeviewersindependentlyscreenedthetitlesand abstracts of the reports and selected original research articles publishedinEnglish.Ourreviewincludedstudies onpatientswithprostatecancerusing68Ga-PSMAPET/CT or PET/MRI for initial staging of prostate cancer or for restaging with rising PSA after the initialtreatment. We excludedarticlespublishedbefore2012,thefoundingyear for68_{Ga-PSMA,reviews,comments,andstudiesof}

labora-tory results, studies of neoplasms apart from prostate cancer,studiesofradiotracersapartfrom68Ga-PSMA,and studiesthatfocusedonthebioavailabilityoftheradiotracer. Further, we excluded studies that onlyreported patients withapositive68Ga-PSMAPET/CTbecausetheyspuriously wouldhaveincreasedthepooleddetectionrates[9],studies thatonlyundertook68_{Ga-PSMAPET/CTforpatientswitha}

negative choline PET/CT because the criterion implied selection bias, and studies with 20 patients owing to concerns regarding selection and publication bias and imprecision.Themeta-analysisalsoexcludedstudiesthat combined staging with restaging, apart from one study consisting of >200 patients where the smallest patient grouprepresented<10%ofallpatients,andonestudythat analyzedthetwopatientgroupsseparately.Whereacenter hadpublishedseveralarticlesofitsexperiencewith68

Ga-PSMAPET/CT,webasedoursummaryofthetotalnumberof examinedpatientsonthearticlewiththemostpatients,and oursummaryofdiagnostictestaccuracyonanotherarticle becauseonlythe replicatearticle reportedthediagnostic information.

2.3. Outcomemeasures

Wecalculatedthedetectionrateasthenumberofpatients withdetectedsitesinrelationtothetotalnumberofimaged patients[11].Wecalculatedimagingtestaccuracyforthe detectionoflesionsintheprostateandpelviclymphnodes based on a reference standard with histopathology after biopsy or RP and pelvic lymph node dissection. We summarized the sideeffects following the imaging with PET/CTorPET/MRIastheywerereportedinthestudies.

2.4. Datacollection

Fromthestudies,bothreviewersindependentlyextracted the radiation activity of the 68_{Ga-PSMA radiotracer, the}

uptake time between injection of the radiotracer and imaging, detection criteria, blinding of nuclear medicine

physiciansforclinicalinformation,andblindingof pathol-ogistsforthe68_{Ga-PSMAPET/CTfindings.Both reviewers}

alsoregisteredthenumberofpatients,median/meanageof thepatients,stagingorrestaging,initialtreatment,median/ meanPSAatthetimeofPET,anddetectionrateandimaging testaccuracywith68_{Ga-PSMAPET/CT.Oneofthereviewers}

contactedallprimaryinvestigatorsformoreinformation.In addition,fornineofourstudieswecomparedourextracted informationwiththatofameta-analysispublishedbyan independentteam[9].

2.5. Qualityassessment

Weassessed riskof bias in thestudies according to the QualityAssessmentofDiagnosticAccuracyStudies (QUA-DAS)-2,asperformedinpreviousreviews[11–13].

2.6. Meta-analyticalmethods

Ourmeta-analysisappliedparametricstatisticto summa-rizemeansSDsofclinicalcharacteristics,andusedtteststo compare clinical characteristics between two groups of patients. Preplanned subgroup analysis of patients with

68_{Ga-PSMA PET/CT at staging and restaging investigated}

whetherthetwosubgroupsdifferedinoveralldetectionrates andintheregionalpatternofdetectedsites.Wesummarized the sensitivity and specificity of the 68_{Ga-PSMA PET/CT in}

studies that used pathology after biopsy or surgery as referencestandard.Ourmeta-analysisusedarandomeffect model and a hierarchical summary receiver operating characteristic (HSROC) model, carried out with Metandi software and Stata 14.0 (StataCorp, College station, TX, USA).Metandi provides apooledestimateof the summary pointwithanestimateofthe95%confidenceinterval(CI)for thesummarypointtogetherwitha95%predictionareaforthe

combined sensitivity and specificity in afuture study. A p

value<0.05indicatedstatisticalsignificance.

3. Evidencesynthesis

3.1. Bibliographicsearch

Searching in PubMed and EMBASE gave 257 reports including42duplicates(Figure1).Byscreeningtitlesand abstracts of 215 unique reports, the two reviewers independently excluded reviews, comments and replies, case reports,articles published before 2012, and studies employing radiotracers other than 68Ga-PSMA. The two reviewers read the full text of 37 articles and selected articles according to some of our exclusion criteria. The reading left 25 studies for qualitative analyses. Our quantitative analyses excluded all but two studies that combinedstagingandrestaging68_{Ga-PSMAPET/CT,andall}

studiesconsistingof20patients.

3.2. Descriptionofthestudiesincluded

The meta-analysisselected15studies with1256patients (Table 1) [14–28]. Threestudies were prospectivecohort studies[18,23,26],sevenstudieswereretrospectivestudies ofconsecutiveselectedpatients[15–17,19–21,25],andfive studieswereretrospectivestudiesofnon-randompatients

[14,22,24,28]. Mean of the median/mean age for the patients was 673 yr(range,62–71 yr).Fourteenstudies only used 68_{Ga-PSMA PET/CT whereas one study reported}

both68_{Ga-PSMAPET/CTand}68_{Ga-PSMAPET/MRI[20].Mean}

ofthemedian/meantotalradiationactivityof68Ga-PSMAwas 17227MBq(range,146–236MBq)in11studies[14–18,22– 24,27,28], whereas one study reported the mean radiation activitynormalizedforbodyweightas1.9MBq/kg[20].Mean

42 duplicates 215 reports without duplicates

178 reviews, case reports, and studies of other radiotracers 85 reports through search in PubMed 172 reports through search in EMBASE

12 studies excluded after reading of full text 25 studies included for qualitative

synthesis

37 studies assessed for eligibility by full text reading

15 studies included for quantitative analysis

10 studies with

staging and restaging and with small sample size

Fig.1–PRISMA(PreferredReportingItemsforSystematicReviewsandMeta-Analysis)flowdiagramrepresentingtheselectionofstudiesinthis systematicreview.

ofthemedian/meanuptaketimefor68Ga-PSMAPET/CTwas 6113min(range,45–90min).

Thestudies evaluatedsitesbyavisualestimateofthe maximum standard uptake value (SUVmax), and many

studiesincorporatedareferencetissueSUVmaxandreported

asiteasbeingpositivewhenSUVmaxwashigherthanthatof

thereferencetissue.Overall,1002of1256imagedpatients hadapositivesitedetectedby68_{Ga-PSMAPET/CTorPET/}

MRI. Ten studies, including one replicate study, used a histologyreference standard[17–23,26,28,29].Two other studies [14,15] used histology or clinical judgment and follow-up as reference standard, whereas four further studies did not report a reference standard

[16,24,25,27].Noneofthestudiesreportedsideeffectsof the68_{Ga-PSMAPET/CTorPET/MRIimaging.}

3.3. Qualityassessment

Figure 2 summarizes our evaluation of the 15 studies regardingriskofbiasasindicatedbyQUADAS-2analysis.

3.4. StagingwithPET/CTorPET/MRI

Sevenstudiesreportedinitialimagingwith68_Ga-PSMAPET/

CT or PET/MRI before definitive treatment [17– 20,23,26,28]. Six studies examined only 68Ga-PSMA PET/ CT,whereasonestudyexaminedboth68_{Ga-PSMAPET/CT}

(35 patients) and 68_{Ga-PSMA PET/MRI (95 patients)}

[20].The mean ofthereportedmedian/mean PSA values inthestudieswas9.32.4ng/ml(range,6.1–11.8ng/ml).

68_{Ga-PSMA PET/CT or PET/MRI detected sites in 203 of}

273 patients (74%),with 163 (60%)patients demonstrating asiteintheprostatebed,12(4%)patientsdemonstratingasite inpelviclymphnodes,and28(10%)patientsdemonstrating sitesinmorethanoneregion.

Fourstudies[18,23,28,29]undertooklesion-based anal-ysis and reportedimaging test accuracy ofstaging 68

Ga-PSMAPET/CTtodelineateintraproprostaticcancerlesions based on ahistopathology referencestandard.One study usedbiopsyasthereferencestandard[28],andtheother threestudiesusedhistopathologyafterRPandlymphnode dissection as the reference standard [18,23,29]. Pooled sensitivityamongthesestudieswas70%(95%CI:53–83%) andpooledspecificitywas84%(95%CI:24–99%).Figure3

summarizestheHSROCcurvefordetectionofintraprostatic

Table1–Characteristicsofthestudies

Study Patients (n) Indication forPET/CT Median/meanage atdiagnosis(yr)

Median/meanPSAand rangeatPET/CT(ng/ml) PETprotocol Median/mean radiationactivity (MBqor*MBq/kg) Median/mean uptaketime (min) Afshar-Oromieh[14] 319 S+R 68 4.6(0.1–41395) 161 60 Ceci[15] 70 R 67 1.7(0.2–32) 146 60 Eiber[16] 248 R 70 2.0(0.2–59) 155 54 Budaus[17] 30 S 62 8.8(1.4–376) 165 NR Fendler[18] 21 S NR NR 192 58 Herlemann[19] 34 S+R 67 35.1(0.3–363) NR 60 Maurer[20] 130 S 66 11.6(6.9–24.5) 1.76* 60 Pfister[21] 28 R 64 2.4(0.04–8) NR 45 Rauscher[22] 48 R 71 1.3(0.8–2.6) 154 57 Rhee[23] 20 S 62 6.1(3.5–45) 150 60 Sachpekidis[24] 31 R 71 2.0(0.1–130) 236 85 VanLeeuwen[25] 70 R 67 0.2(0.05–1.0) NR 45 VanLeeuwen[26] 30 S 65 8.1(5.2–10.1) 236 85 Verburg[27] 155 R 70 4.0(0–2000+) 190 60 Zamboglou[28] 22 S 69 11.6(NR) 172 60 Totalnumber 1256

NR=notreported;R=restagingPET/CT;S=stagingPET/CT.

Patient flow Reference standard Index test Patient selection 0 20 40 60 80 100 Risk of bias (%) 0 20 40 60 80 100 Reference standard Index test Patient selection Applicability

Fig.2–QualityAssessmentofDiagnosticAccuracyStudies(QUADAS)-2 evaluationofthe15studies.Regardingbias:proportionofstudieswith highriskofbias(red),unclearriskofbias(yellow),andlowriskofbias (green).Regardingconcernsforapplicability:proportionofstudieswith highconcernsregardingapplicability(red),unclearconcernsregarding applicability(yellow),andlowconcernsregardingapplicability(green).

lesions in the studies examining correlations with the histopathology reference standard. Four other studies undertookpatient-basedanalysisandreported68Ga-PSMA PET/CT imaging test staging accuracy specifically for sampledpelviclymphnodemetastases[17,19,20,26].Pooled sensitivityforlymphnodedetectionwas61%(95%CI:47– 72%) and pooled specificity was 97% (95% CI: 85–99%).

Figure4summarizestheHSROCcurvefordetectionofpelvic lymphnodemetastasesinthestudies,andFigure5showsa Venndiagramthatalsosummarizeddiagnostics regarding pelviclymphnodemetastases.

3.5. RestagingwithPET/CT

Ninestudiesreportedrestagingwith68Ga-PSMAPET/CTfor patients with persisting and rising PSA after initial treatment. Sevenstudies undertook only restaging 68

Ga-PSMA PET/CT[15,16,21,22,24,25,27],whereastwostudies included staging and restaging 68_{Ga-PSMA PET/CT}

[14,19].Themeanofthemean/medianrestagingPSAlevels was 2.31.4 ng/ml (range 0.21–4.6 ng/ml). Two studies reported PSArecurrenceafterRP[16,26]whereastheother sevenstudiesreportedPSArecurrenceafterbothRPandEBRT. Inthese studiesthemain treatmentwasRPfor87%of the patients(450of515)andEBRTfor13%ofthepatients(65of 515).

Overall,forrestagingofPSA-onlyrecurrence,68Ga-PSMA PET detected sites of recurrence in 799 of 983 imaged patients (81%). The studies reportedthe regionalsitesof recurrencefor615of755patients(82%).Ofthesepatients, 79(10%)patientshadsitesintheprostatebed,164(22%) patientshadsitesinpelviclymphnodes,100(13%)patients had sites in distant organs, and 272 (36%) had sites of recurrence in several regions. In eight studies

[14,16,24,25,27,30–32], 74 of 147 patients (50%) with restagingPSAlevelsof0.20–0.49ng/mlhadpositivesites ofrecurrence,ashad56of105patients(53%)withrestaging PSAof0.50–0.99ng/ml.Fourofthestudies examinedthe PSA levels according the detected sites. PET-positive patients had significantly higher PSA than PET-negative patients [14,15,24,27]. Inone ofthestudies, PET-positive metastatic lesions inlymph nodes hada largerdiameter thanPET-negativemetastaticlesions[26].

Twostudiesundertookpatient-basedanalysisregarding

68_{Ga-PSMAPET/CT imagingtestaccuracy for restagingof}

lymph node metastases versus a histologic reference standard[21,22].Thesensitivitieswere87%and93%,and the specificities were 93% and 100%. A third study

Specificity Sensitivity 1.0 0.8 0.6 0.4 0.2 0 1.0 0.8 0.6 0.4 0.2 0

Fig.3–Hierarchicalsummaryreceiveroperatingcharacteristic(ROC) analysisofstagingusingpositronemissiontomographyimagingwith 68_{Ga-PSMAforlesion-basedanalysisregardingtestaccuracyfor} intraprostaticlesions.Thesizeofthecirclesshowsthesizeofthe studies,thefulllineshowstheROCcurve,thesquareshowsthe summaryoperatingpoint,theredstippledlineshowsthe95% confidenceregionforthesummarypoint,andtheblackstippledline showsthe95%predictionregionasaforecastofthesensitivityand specificityofafuturestudy.68

Ga-PSMA=68

Ga-labeledligandfor prostate-specificmembraneantigen.

27% TN 58% PSMA PET/CT positive FP Histopathology cancer positive FN 12% TP 3% 68 Ga 27% Histopathology cancer positive FN 12% TP

Fig.5–Venndiagramfordetectionwithstaging68_Ga-PSMAPETof pelviclymphnodemetastases.ThecirclesindicatePET-positiveand histopathology-positivepatients.FN=falsenegative;FP=falsepositive; TN=truenegative;TP=truepositive.68

Ga-PSMA=68

Ga-labeledligand forprostate-specificmembraneantigen.

Specificity Sensitivity 0 0.2 0.4 0.6 0.8 1.0 1.0 0.8 0.6 0.4 0.2 0

Fig.4–Hierarchicalsummaryreceiveroperatingcharacteristicofstaging 68

Ga-PSMApositronemissiontomographyforpatient-basedanalysis regardingimagingtestaccuracyforlymphnodemetastases.Symbols arethesamesymbolsasinFigure3.68_Ga-PSMA=68_{Ga-labeledligand} forprostate-specificmembraneantigen.

undertookonly lesion-basedanalysisandreported histo-logicverification for42 patientsregardingseparatelocal, regional,andsoft-tissuesites[14].Thirteenofthepatients hadtrue-positivefindings,threepatientshadfalse-positive findings,19patientshadacombinationoftrue-positiveand true-negative sites, three patients had a combination of true-positiveandfalse-negativesites,andfourpatientshad true-negativefindings.

68_{Ga-PSMA PET/CT for restaging at the time of PSA}

recurrenceinthestudieshadhigheroveralldetectionrates thanfor 68_{Ga-PSMAPET/CT staging inthe studies before}

definitivetreatment,butthedifferencewasnotstatistically significant (mean 78% vs mean 69%, p=0.52, t test). By contrast,detectionratesforsitesintheprostatebedwere significantlyhigher with staging 68_{Ga-PSMA PET/CT than}

with restaging 68Ga-PSMA PET/CT (mean 57% vs 14%,

p=0.031,ttest,Figure6).

4. Discussion

Thismeta-analysishasaddedinsightintotheuseof68 Ga-PSMA PET/CT and PET/MRI for patients with prostate cancer. All studies reported examinations with PET/CT, and only one study included a subgroup of patients examined with PET/MRI.The studies useda protocolfor

68_{Ga-PSMAPET/CTwitharadiationactivitygenerallyinthe}

range 130–170 MBq, an uptake time of approximately 60min, and interpretation of the imaging based on SUVmax.ForstagingPET/CTorPET/MRI,thedetectionrate

was70–80%.Forrestaging PET/CT,therestagingPSAwas positivelyassociatedwiththedetectionrate.Thedetection ratewas50%evenforrestagingPSAlevelsof0.2–0.49ng/ ml,53%forrestagingPSAof0.5–0.99ng/ml,andwasfurther increasedforhigherrestagingPSAlevels.Bothstagingand restaging PSMA PET/CT imaging were able to distinguish betweensingle sitesin the prostate bed, regionallymph nodes,anddistantorgans,andsitesinmorethanoneofthe regions. The pooled sensitivity for primary or regional

cancerwas61–70%andthepooledspecificitywas84–97%. Thestudiesdidnotreportadverseeffectsfromtheimaging. Forsitesintheprostatebed,staging68_{Ga-PSMAPET/CT}

beforetheinitialtreatmentoftheprimaryprostatecancer hadahigherdetectionratethanrestaging68_Ga-PSMAPET/

CT aftertheinitialtreatment.Thiswasinpartbecauseof debulking of the primary prostate cancer by the initial treatment,mostoftenRP.Ameta-analysisofcholinePET/CT foundthesamedifferencebetweenstagingandrestaging choline PET/CT [4]. Similarly, both the present meta-analysis of 68_{Ga-PSMA PET/CT and the meta-analysis of}

choline PET/CT found a subgroup of patients who had metastaticsitesinpelviclymphnodesanddistantorgans both atstagingandrestaging. Suchidentificationofsites with PET/CTcouldguidetreatmentafter surgeryformen withpersistingorrecurrentPSAindicatedbyPSA monitor-ing. Theconsistencybetweenstudieswiththetwo radio-tracersforPET/CTsuggeststhatthefindingsarereal[33]. Thissystematicreviewhassummarizeddetectionrates andimaging testaccuracy reportedintheliteratureuntil July 2016. Our systematic review applied more rigid selection criteria than two previous systematic reviews

[5,9].Forthisreasontheprevioussystematicreviewsand ourpresentsystematicreviewhadanoverlapofonlynine studies, andoverlapofonlyonestudythatused histopa-thologyasthereferencestandard.Thedifferentselections of studies in the three systematic reviews may have contributedtothefactthatoneoftheprevioussystematic reviewsfounda40%pooleddetectionrateinrestagingwith

68_{Ga-PSMAPET/CT[9].}

Use of PSMA PET/CT for prostate canceris expanding rapidlyandwidely.Thishasconsequencesformanagement. The internationalTNMtumorclassificationsystem classi-fies patients with PSA recurrence as M0 if they have no evidenceofmetastaseson imaging,andasM1if imaging showsevidenceofmetastases.Thusthedistinctionbetween M0 and M1 depends on the selection and validation of imagingmethodsusedfortherestaging.

Asanalternative toSRT undertakenwithoutguidance from imaging, a previous meta-analysis [5] proposed an algorithmthatintegratedPSMA PET/CTintheimagingof prostatecancer.However,becauseuptohalfthepatients with PSA-only recurrence treated with SRT obtain long-term biochemicalrecurrence-freesurvival, restagingwith

68_{Ga-PSMA PET/CT can only improve outcome for a}

subgroup ofat-riskpatientswithPSA-onlyrecurrence. As showninourmeta-analysis,restagingwith68_Ga-PSMAPET/

CT of such patients with restaging PSA <2 ng/ml might detect sites of recurrence that could be treated with targeted treatment with curative intent. We therefore proposethatsalvagetreatmentisindividualizedandguided by68_{Ga-PSMAPET/CTforthesepatientsalso.}

Our meta-analysis may have impact on research regarding 68Ga-PSMA PET/CT for patients with prostate cancer. Threeongoing trials are recruiting patients. Two trials addressstagingwith68_{Ga-PSMAPET:Evaluationof}

Gallium-HBED-CC-PSMA Imaging in Prostate Cancer patients (PSMA PET) (NCT02611882), and 68_Ga-PSMA

PET/MRIinFindingTumorsinPatientswithIntermediate

Detection rate (% ) 80 60 40 20 0 T N M Comb

Fig.6–Detectionratesforregionalsiteswithstagingandrestaging 68

Ga-PSMAPET/CT.Greybarsshowtheproportionofpatientswith staging68

Ga-PSMAPET/CTorPET/MRI,andorangebarsshowthe proportionofpatientswithrestaging68

Ga-PSMAPET/CT.68 Ga-PSMA=68

Ga-labeledligandforprostate-specificmembraneantigen; PET=positronemissiontomography;CT=computedtomograpy; MRI=magneticresonanceimaging;Comb=sitesinmorethanoneof theregions;M=sitesindistantorgans;N=sitesinpelviclymphnodes; T=siteintheprostatebed.

or High-Risk Prostate Cancer Undergoing Surgery (NCT02678351). A third trial examines restaging with

68_{Ga-PSMAPET/CT:68Ga-PSMA PET-CT Scanfor Diagnosis}

andManagementofProstateCancer(PSMA)(NCT02282137). OtherstudiescomparePSMAandcholineasradiotracersfor restagingPET/CT.

Our review has limitations. We only evaluated 68 Ga-PSMAasaradiotracerforPET/CTorPET/MRIalthoughother PSMAradiotracersarealsobeinginvestigated[34].Because of the smallnumber ofstudies, their heterogeneity, and potentialselectionandpublicationbiases,external valida-tionoftheestimatesofdetection,sensitivity,andspecificity rates will be needed. The diagnostic accuracy estimates werepredominantlybasedonstagingwith68_Ga-PSMAPET/

CT and PET/MRI, whereas we believe that the main indicationforimagingisrestagingofPSA-onlyrecurrence. Inaddition,thereviewdidnotincludeanalysisofstudies that compared 68_{Ga-PSMA PET/CT with other imaging}

modalities,orofstudiesthatreportedtreatmentguidedby

68_{Ga-PSMAPET/CT.}

5. Conclusions

Based onpublished reportsofstaging68Ga-PSMAPET/CT and PET/MRI, we found a sensitivity of 61–70% and a specificityof84–97%.Restaging68Ga-PSMAPET/CT hada detectionrateof50%foranearlyriseinPSA.

Authorcontributions:FinnE.vonEybenhadfullaccesstoallthedatain thestudyandtakesresponsibilityfortheintegrityofthedataandthe accuracyofthedataanalysis.

Studyconceptanddesign:F.vonEyben,Bauman.

Acquisitionofdata:F.vonEyben,Bauman.

Analysisandinterpretationofdata:F.vonEyben,Picchio,R.vonEyben, Rhee,Bauman.

Draftingofthemanuscript:F.vonEyben.

Criticalrevisionofthemanuscriptforimportantintellectualcontent:F.von Eyben,Picchio,R.vonEyben,Rhee,Bauman.

Statisticalanalysis:F.vonEyben,R.vonEyben.

Obtainingfunding:None.

Administrative,technical,ormaterialsupport:None.

Supervision:F.vonEyben,Picchio,Bauman.

Other:None.

Financialdisclosures:FinnE.vonEybencertifiesthatall conflictsof interest, including specific financial interests and relationships and affiliationsrelevanttothesubjectmatterormaterialsdiscussedinthe manuscript(eg,employment/affiliation,grantsorfunding, consultan-cies,honoraria,stockownershiporoptions,experttestimony,royalties, orpatentsfiled,received,orpending),arethefollowing:None.

Funding/Supportandroleofthesponsor:

References

[1] ParkinDM,FerlayJ,CuradoMP,etal.Fiftyyearsofcancerincidence: CI5I–IX.IntJCancer2010;127:2918–27.

[2] CaryKC, Punnen S, Odisho AY, etal.Nationally representative trendsandgeographicvariationintreatmentoflocalizedprostate cancer:theUrologicDiseasesinAmericaproject.ProstateCancer ProstaticDis2015;18:149–54.

[3] HanM, PartinAW, PoundCR, EpsteinJI, WalshPC. Long-term biochemical disease-free and cancer-specific survival following anatomic radical retropubic prostatectomy. The 15-year Johns Hopkinsexperience.UrolClinNorthAm2001;28:555–65.

[4] vonEybenFE,KairemoK.Meta-analysisof11_C-cholineand18

F-cholinePET/CTformanagementofpatientswithprostatecancer. NuclMedCommun2014;35:221–30.

[5] EvangelistaL,BrigantiA,FantiS,etal.Newclinicalindicationsfor

18_F/11_{C-choline,newtracersforpositronemissiontomographyand}

apromisinghybriddeviceforprostatecancerstaging:asystematic reviewoftheliterature.EurUrol2016;70:161–75.

[6] MeaseRC,FossCA,PomperMG.PETimaginginprostatecancer: focusonprostate-specificmembraneantigen.CurrTopMedChem 2013;13:951–62.

[7] EderM,SchaferM,Bauder-WustU,etal.68_{Ga-complexlipophilicity}

andthetargetingpropertyofaurea-basedPSMAinhibitorforPET imaging.BioconjugChem2012;23:688–97.

[8] Afshar-OromiehA,HaberkornU,EderM,EisenhutM,Zechmann CM.[68_{Ga]-labelled PSMAligand assuperior PETtracer for the}

diagnosisofprostatecancer:comparisonwith18_{F-FECH.EurJNucl}

MedMolImaging2012;39:1085–6.

[9] PereraM,PapaN,ChristidisD,etal.Sensitivity,specificity,and predictorsofpositive68Ga-prostate-specificmembrane antigen positron emission tomography in advanced prostate cancer: a systematicreview andmeta-analysis.EurUrol.Inpress. http:// dx.doi.org/10.1016/j.eururo.2016.06.021.

[10] LiberatiA,AltmanDG,TetzlaffJ,etal.ThePRISMAstatementfor reportingsystematicreviews andmeta-analysesofstudiesthat evaluate healthcareinterventions:explanationandelaboration. AnnInternMed2009;151,W65-94.

[11] FantiS,MinozziS,CastellucciP,etal.PET/CTwith11_C-cholinefor

evaluation of prostate cancerpatients with biochemical recur-rence: meta-analysisandcritical reviewofavailabledata. EurJ NuclMedMolImaging2016;43:55–69.

[12]WhitingPF,RutjesAW,WestwoodME,etal.QUADAS-2:arevised toolforthequalityassessmentofdiagnosticaccuracystudies.Ann InternMed2011;155:529–36.

[13] vonEybenFE,KiljunenT,KangasmakiA,KairemoK,vonEybenR, Joensuu T. Radiotherapy boost for the dominant intraprostatic cancerlesion–asystematicreviewandmeta-analysis.Clin Geni-tourinCancer2016;14:189–97.

[14]Afshar-OromiehA,AvtziE,GieselFL,etal.Thediagnosticvalueof PET/CTimagingwiththe68_{Ga-labelledPSMAligandHBED-CCin}

the diagnosisofrecurrentprostatecancer.Eur JNuclMedMol Imaging2015;42:197–209.

[15] CeciF,UprimnyC,NilicaB,etal.68_{Ga-PSMAPET/CTforrestaging}

recurrentprostatecancer:whichfactorsareassociatedwithPET/CT detectionrate?EurJNuclMedMolImaging2015;42:1284–94.

[16] EiberM,MaurerT,SouvatzoglouM,etal.Evaluationofhybrid68

Ga-PSMAligandPET/CTin248patientswithbiochemicalrecurrence afterradicalprostatectomy.JNuclMed2015;56:668–74.

[17] BudausL,Leyh-BannurahSR,SalomonG,etal.Initialexperienceof

68_{Ga-PSMAPET/CTimaginginhigh-riskprostatecancerpatients}

priortoradicalprostatectomy.EurUrol2016;69:393–6.

[18] FendlerWP,SchmidtDF,WenterV,etal.68_{Ga-PSMA-HBED-CCPET/}

CTdetectslocationandextentofprimaryprostatecancer.JNucl Med2016;57:1720–5.

[19]HerlemannA,WenterV,KretschmerAetal.,68Ga-PSMApositron emission tomography/computed tomography provides accurate stagingoflymphnoderegionspriortolymphnodedissectionin patientswithprostatecancer.EurUrol.Inpress. http://dx.doi.org/ 10.1016/j.eururo.2015.12.051.

[20] MaurerT,GschwendJE,RauscherI,etal.Diagnosticefficacyof

con-ventional imaging for lymph node staging of 130 consecutive patientswith intermediate to high riskprostate cancer. J Urol 2016;195:1436–43.

[21]PfisterD,PorresD, HeidenreichA,etal.Detection ofrecurrent prostatecancerlesionsbeforesalvagelymphadenectomyismore accuratewith68_{Ga-PSMA-HBED-CCthanwith}18

F-fluoroethylcho-linePET/CT.EurJNuclMedMolImaging2016;43:1410–7.

[22] RauscherI,MaurerT,BeerAJ,etal.Valueof68_{Ga-PSMAHBED-CC}

PET for the assessment of lymph node metastasesin prostate cancer patientswith biochemical recurrence: comparisonwith histopathology after salvage lymphadenectomy. J Nucl Med 2016;57:1713–9.

[23] RheeH,ThomasP,ShepherdB,etal.Prostatespecificmembrane antigenpositronemissiontomographymayimprovethediagnostic accuracyofmultiparametricmagneticresonanceimagingin local-izedprostatecancer.JUrol2016;196:1261–7.

[24] SachpekidisC,EderM,KopkaK,etal.68_{Ga-PSMA-11dynamicPET/}

CTimaginginbiochemicalrelapseofprostatecancer.EurJNucl MedMolImaging2016;43:1288–99.

[25] vanLeeuwenPJ,StrickerP,HrubyG,etal.68_{Ga-PSMAhasahigh}

detectionrateofprostatecancerrecurrenceoutsidetheprostatic fossainpatientsbeingconsideredforsalvageradiationtreatment. BJUInt2016;117:732–9.

[26]vanLeeuwenPJ,EmmettL,HoB,etal.Prospectiveevaluationof 68gallium-PSMApositronemissiontomography/computerized to-mographyforpreoperativelymphnodestaginginprostatecancer. BJUInt.Inpress. http://dx.doi.org/10.1111/bju.13540.

[27]VerburgFA, PfisterD, HeidenreichA,etal.Extentofdisease in recurrent prostate cancer determined by [68_{Ga]PSMA-HBED-CC}

PET/CTinrelationtoPSAlevels,PSAdoublingtimeandGleason score.EurJNuclMedMolImaging2016;43:397–403.

[28] ZamboglouC,WieserG,HenniesS,etal.MRIversus68_Ga-PSMA

PET/CTforgrosstumourvolumedelineationinradiationtreatment planningofprimaryprostatecancer.EurJNuclMedMolImaging 2016;43:889–97.

[29] EiberM,Weirich G,Holzapfel K,etal.Simultaneous Ga-PSMA HBED-CCPET/MRIimprovesthelocalizationofprimaryprostate cancer.EurUrol.Inpress.http://dx.doi.org/10.1016/j.eururo.2015. 12.053.

[30] DietleinM,KobeC,KuhnertG,etal.Comparisonof[18_F]DCFPyLand

[68_{Ga]-PSMA-HBED-CC for PSMA-PET imaging in patients with}

relapsedprostatecancer.MolImagingBiol2015;17:575–84.

[31] BluemelC,KrebsM,PolatB,etal.68_{Ga-PSMA-PET/CTinpatients}

withbiochemical prostate cancerrecurrence andnegative 18

F-Choline-PET/CT.ClinNuclMed2016;41:515–21.

[32] MorigiJJ,StrickerPD,vanLeeuwenPJ,etal.Prospectivecomparison of18_{F-fluoromethylcholineversus}68_{Ga-PSMAPET/CTinprostate}

cancerpatientswhohaverisingPSAaftercurativetreatmentand arebeingconsideredfortargetedtherapy. JNuclMed2015;56: 1185–90.

[33] vonEybenFE,KairemoK.Acquisitionwith11_C-cholineand18

F-fluorocholinePET/CTforpatientswithbiochemicalrecurrenceof prostatecancer:asystematicreviewandmeta-analysis.AnnNucl Med2016;30:385–92.

[34] RoweSP,GorinMA,AllafME,etal.PETimagingofprostate-specific membrane antigenin prostate cancer:current state ofthe art and future challenges. Prostate Cancer Prostatic Dis 2016;19: 223–30.