TITLE:EvaluatingRoleofMulti-ParametricMRIandPI-RADSv2inIncreasingtheOddsof
IdentifyingAggressiveProstateCancerinMenwithTRUSBiopsyGleasonPattern≤3
Faina Shtern1, Fiona Fennessy2, Mukesh Harisinghani3, Elmira Hassenzadeh2, Adam S. Kibel2, Quoc-Dien
Trinh2,MichelleS.Hirsch2,ZhengZhang4,DavidZurakowski5,andClareM.Tempany2
1AdMeTech
Foundation, 2Brigham and Women’s Hospital, 3Massachusetts General Hospital, 4Brown
University,and5BostonChildren’sHospital
INTRODUCTION:Mostmenundergoingscreeningwithprostatespecificantigen(PSA)haveindolent,lowgrade (Gleason Pattern, or GP ≤ 3) prostate cancer (PC) on trans-rectal ultrasound (TRUS) biopsies.1-4
However,TRUSbiopsiesunder-sampleprostatetissueandthus,under-estimateclinicallysignificantPC,as
defined by more aggressive, higher grade (GP ≥ 4) disease, which is more likely to progress, cause
metastasesanddeath.1-4Asmanyas36to47%ofmenwithTRUSbiopsyGP≤3PCareupgradedtoGP≥4
on post-operative pathology (POP).2-4.Based on the recent extensive evidence, NIH consensus conference
highlighted the need to advance diagnostic tools for improving target definition, tissue sampling and
predictionofclinicallysignificantPCasa“healthresearchpriority.”1
SPECIFIC AIM: The aim of this study was to determine whether multi-parametric (mp) MRI and related
ProstateImaging–ReportingandDataSystemVersion2(PI-RADS™v2)canimprovepredictionofPOPGP≥
4inmenwithapre-operativeTRUSbiopsyGP≤3.
BACKGROUND AND RATIONALE: A recent long-term, large-scale multi-center clinical trial showed that
increasingPSAandTRUSbiopsyGP≥4(bothPrimaryandSecondary)arethecriticalfactorsinlong-term
PC mortality after radical prostatectomy (RP).2,3 Valid concerns about under-grading with TRUS biopsy
haveledtoextensiveover-treatment,whichhasemergedasanationalpublichealthpriority.1-4Toaddress
thischallenge,therehasbeenaconcertedefforttoexpediteadvancementofmpMRI,whichhasbeenshown
toimprovedetection(andexclusion)ofclinicallysignificantdiseaseinpreliminarystudies.5-8Recently,PIRADS™ v2 standardization was developed, with the primary goal to improve risk stratification even
further.5-9ThisstudywasconductedtoevaluatetheroleofmpMRIandPI-RADS™v2incorrectlyupgrading
TRUSbiopsyGP≤3toPOPGP≥4.
METHODS AND MATERIALS: Thisinstitutionalreviewboard-approved,retrospectivestudyaccrued366
menaged40to80yearsoldwhohadabnormalPSA(>4ng/mL)and/orabnormaldigitalrectalexamand
underwentTRUSbiopsyandRPattheBrighamandWomen’sHospitalbetween2008and2014.Thetwo
study cohorts included: 1) Patients who had mpMRI, which was performed at least 6 weeks after TRUS
biopsy and within 6 months prior to RP (N=190); and 2) Patients who did not have mpMRI prior to RP
(N=176).TRUSBiopsyGP≤3wasfoundin62meninthefirstcohortandin99meninthesecondcohort.
MpMRI was defined as “negative” for GP ≥ 4 if PI-RADS v2 scores were 1-3, and “positive” if PI-RADS v2
scoreswere4-5.POPservedasareferencestandard.
Availabilityofcompletestandarddiagnosticclinicalinformation(includingPSAlevel,POPreports
and TRUS biopsy results) was required for both cohorts. All mpMRI exams included in the study were
acquired at 3 Tesla (3T) and met technical standards for image acquisition10 and clinical criteria for
acceptableimagequality.
Two readers independently reviewed all mpMRI de-identified studies and assigned PI-RADS v2
score at a patient level, based on the dominant lesion. In cases of disagreement between readers, a third
radiologistservedasanadjudicatoranddeterminedafinaloverallassessment.Allreviewerswereblinded
toclinicalinformation.
InpatientswhohadTRUSBiopsyGP<3,wecomparedtheoddsofupgradingtoPOPGP>4with
andwithoutmpMRI,includingmenwithpositive(N=38)andnegativePI-RADSv2scores(N=24).Oddsof
“upgrading” to GP ≥ 4 were analyzed using multivariable logistic regression withSAS software (version
9.3,ASInstitute,Cary,NC).11
RESULTS: A. Patient Population. MorepatientsintheMRIgroupwereAfricanAmerican,andmoremen
inthecontrolgroupdeclinedtoidentifytheirraceand/orethnicity(seeTable1):
Variable
mpMRI(N=190)
NoMRI–Controls(N=176)
Age,years,mean±SD
58.7±6.9
59.8±6.5
Race,no.(%)
Caucasian
164(86)
143(81)
AfricanAmerican
16(9)
7(4)
Hispanic
4(2)
3(2)
Other/declinedtoanswer
6(3)
23(13)
TNMStage,median(range)
2(1-3)
2(1-3)
PSA,ng/mL,median(IQR)
5.3(4.2-8.1)
5.0(4.0-6.8)
*Statisticallysignificant.SD=standarddeviation,IQR=interquartilerange.
PValue
0.123
0.002*
0.251
0.069
TABLE1:Patientpopulationsinthetwostudycohorts.
B.TheOddsofUpgradingTRUSGP<3toPOP>GP4WithandWithoutmpMRI
(irrespectiveofPI-RADSv2scores)
Post-opGP>4(Upgrading)
Post-opGP<3
•
Statisticallysignificant.
mpMRI(N=62)
30(48%)
32(52%)
NoMRI–Controls(N=
99)
30(30%)
69(70%)
PValue
0.021*
TABLE 2: Detection of POP GP > 4 among patients with TRUS biopsy GP < 3 is significantly higher with
mpMRI (irrespectively of PI-RADS v2 score) compared to no MRI (48% vs. 30%, P = 0.021). This
demonstrates an absolute improvement of 18% using mpMRI (95% CI: 3% to 33%). Logistic regression
indicatesanunadjustedoddsover2timeshigherwithmpMRIvs.controls(oddsratio:2.2,95%CI:1.2-4.2,
likelihood ratio test = 5.3, P = 0.021). Multivariable regression analysis confirmed that adjusting for age,
race, TNM stage and PSA, the odds of “upgrading” to POP GP > 4 are over 2 times greater with mpMRI
(adjustedoddsratio:2.1,95%CI:1.2-4.1,likelihoodratiotest=4.5,P=0.034).
C.TheOddsofUpgradingTRUSGP<3toPOP>GP4WithandWithoutPositivempMRI
(PI-RADSv2scores4-5)
Post-opGP>4(Upgrading)
Post-opGP<3
*Statisticallysignificant.
PositivempMRI(PI-RADS4-5)
(N=38)
25(66%)
13(34%)
NoMRI-Controls
(N=99)
30(30%)
69(70%)
PValue
0.0002*
TABLE 3: Detection of POP GP > 4 among patients with TRUS biopsy GP < 3 is significantly higher with
positive mpMRI compared to no MRI (66% vs. 30%, P = 0.0002). This demonstrates an absolute
improvement of 36% using mpMRI (95% CI: 17% to 51%). Logistic regression indicated an unadjusted
odds of yield over 4 times higher with positive mpMRI vs. no MRI (odds ratio: 4.4, 95% CI: 2.0-9.8,
likelihoodratiotest=14.3,P=0.0002).Multivariableregressionanalysisconfirmedthatadjustingforage,
race,TNMstageandPSA,theoddsofyieldofPOPGP>4arenearly5timesgreaterwithpositivempMRI
(adjustedoddsratio:4.8,95%CI:1.9-12.0,likelihoodratiotest=11.7,P=0.001).
D.TheOddsofUpgradingTRUSGP<3toPOP>GP4WithandWithoutNegativempMRI
Post-opGP>4(Upgrading)
Post-opGP<3
NegativeMRI
(PI-RADS1-3)
(N=24)
5(21%)
19(79%)
NoMRI-Controls
BWH2008-2014
(N=99)
30(30%)
69(70%)
PValue
0.454
TABLE4:YieldofPOPGP>4amongpatientswithTRUSbiopsyGleason<3wasnotsignificantlychanged
using negative mpMRI (PI-RADS v2 1-3) compared to no MRI (21% vs. 30%, P = 0.454). This actually
demonstrated a lower rate of “upgrading” by 9% using MRI. Logistic regression indicated an unadjusted
odds 0.6 times higher with negative PI-RADS vs. no MRI controls (odds ratio: 0.6, 95% CI: 0.2-1.8,
likelihoodratiotest=0.89,P=0.345).Multivariableregressionanalysisafteradjustingforage,race,TNM
stageandPSA,indicatedanon-significantoddsof“upgrading”usingMRIwithnegativePI-RADS(adjusted
oddsratio:0.4,95%CI:0.1-1.3,likelihoodratiotest=2.6,P=0.105).
E. Summary: WehavecomparedtheyieldofaggressivePOPGP>4withandwithoutmpMRIinpatients
withpre-operativeTRUSbiopsyGP<3.TherateofPOPGP>4was30%inthecontrolgroup(noMRI)and
48%withMRI(Table5).InmenwhohadmpMRI(irrespectiveofPI-RADSv2score),theoddsofcorrectly
upgradingTRUSGP<3toPOP>GP4arepredictedatover2timesgreatercomparedtothecontrolgroup.
InmenwhohadpositivempMRI(PI-RADSv2score4-5),therateofPOP>GP4isincreasedevenfurther
to 66%, with the odds of correctly upgrading TRUS GP < 3 to POP > GP 4predicted to be nearly 5-fold
greatercomparedtothecontrolgroup.
In this study population, negative mpMRI (PI-RADS v2 score 1-3) is associated with a statistically
insignificantchangeintheyieldofPOPGP>4comparedtothecontrolcohort(21%vs.30%,respectively).
YieldofAggressivePC
(POPGP>4)
MRI(AnyPI-RADSv2)
48%*
PositiveMRI(PI-RADSv2score4-5)
66%*
NegativeMRI(PI-RADSv2score1-3)
21%
Controls(noMRI)
30%
*Statisticallysignificant.
TABLE5:Summary.
DISCUSSION: MostmenundergoingPSAscreeninghaveindolentPC(GP<3)onTRUSbiopsies,thougha
significant fraction of these cases gets upgraded on POP. We have shown that in men diagnosed with
indolentdiseaseonpre-operativeTRUSbiopsy,positivePI-RADSv2scoremayincreasetheprobabilityof
correctlyupgradingTRUSGP<3toPOP>GP4nearly5times.TheeffectofmpMRI(irrespectiveofPIRADSv2score)onincreasedoddsofupgradingindolentPConTRUSbiopsytoaggressivediseaseonPOP
mostlikelyreflectsastrongeffectofpositivePI-RADSv2scores.
These results are in alignment with recent data indicating a promising role of mpMRI and/or PIRADS v2 in improving prediction of high-grade PC on POP and its discrimination from low-grade disease
comparedtoTRUSbiopsyinasurgicalpopulation.7.8
CONCLUSION:OurstudyshowsthatmpMRIandpositivePI-RADSv2scorescanimproveriskstratification
andselectionofpatientsforsurgicaltreatmentcomparedtoTRUSbiopsies.Whilethesedataindicatethat
mpMRI and PI-RADS v2 may significantly alter patient management strategy and address an important
public health challenge in PC evaluation and care, further prospective, multi-center clinical trials are
required. In the interim, TRUS biopsy will continue to be needed as a part of a standard diagnostic
evaluationfortreatmentdecisions.
REFERENCES
1) Ganz PA, et. al. NIH State-of-the-Science Statement: Role of active surveillance in the management of
menwithlocalizedprostatecancer.NIHConsensusStateScienceStatements2011Dec5-7;28(1):1-27
2) Stephenson,A.J.,M.W.Kattan,J.A.Eastham,F.J.Bianco,Jr.,O.Yossepowitch,A.J.Vickers,E.A.Klein,D.P.
Wood, and P.T. Scardino, Prostate cancer-specific mortality after radical prostatectomy for patients
treatedintheprostate-specificantigenera.JClinOncol,27(26):p.4300-5.2009.(online)availablefrom
http://jco.ascopubs.org/content/27/26/4300.full.pdf+html
3) Eggener,S.E.,P.T.Scardino,P.C.Walsh,M.Han,A.W.Partin,B.J.Trock,Z.Feng,D.P.Wood,J.A.Eastham,
O. Yossepowitch, D.M. Rabah, M.W. Kattan, C. Yu, E.A. Klein, and A.J. Stephenson, Predicting 15-year
prostatecancerspecificmortalityafterradicalprostatectomy. J Urol, 185(3): p. 869-75. 2011 (online)
availablefromhttp://jco.ascopubs.org/content/27/26/4300.full.pdf+html
4) Epstein,J.I.,Z.Feng,B.J.Trock,andP.M.Pierorazio,Upgradinganddowngradingofprostatecancerfrom
biopsy to radical prostatectomy: incidence and predictive factors using the modified Gleason grading
system and factoring in tertiary grades. Eur Urol, 61(5): p. 1019-24. 2012. (online) available from
http://www.sciencedirect.com/science/article/pii/S0302283812001686
5) Schoots,IG,et.al.MagneticResonanceImaginginActiveSurveillanceofProstateCancer:ASystematic
Review.EuropeanUrology2015.67(4):627–636
6) Barentsz,J.O.,etal.,ESURprostateMRguidelines2012.EurRadiol,2012.22(4):746-57.
7) Vos,EK,et.al.,MultiparametricMagneticResonanceImagingforDiscriminatingLow-GradeFromHighGradeProstateCancer.Invest.Radiology2015.50(8):490-7
8) Park,SY,et.al.,ProstateCancer:PI-RADSVersion2HelpsPreoperativelyPredictClinicallySignificant
Cancers.Radiology2016.280(1):108-16
9) Weinreb, JC, PI-RADS Prostate Imaging–Reporting and Data System: 2015, Version 2. European
Urology2016.69(1):16
10) Hegde,JV,et.al.MultiparametricMRIofprostatecancer:anupdateonstate-of-the-arttechniquesand
their performance in detecting and localizing prostate cancer. J Magn Reson Imaging. 2013 May;
37(5):1035-54.PMID:23606141;PMC3741996.
11) KatzMH.Multivariableanalysis:Apracticalguideforcliniciansandpublichealthresearchers,3rded.
NewYork:CambridgeUniversityPress;2011:118-39.