value of 3-t multiparametric magnetic resonance...

Value of 3-T Multiparametric Magnetic Resonance Imaging andMagnetic ResonanceYGuided Biopsy for Early Risk Restratification

in Active Surveillance of Low-Risk Prostate CancerA Prospective Multicenter Cohort Study

Caroline M.A. Hoeks, MD, PhD,* Diederik M. Somford, MD, PhD,Þþ Inge M. van Oort, MD, PhD,ÞHenk Vergunst, MD, PhD,þ Jorg R. Oddens, MD,§ Geert A. Smits, MD,|| Monique J. Roobol, PhD,¶

Meelan Bul,MD, PhD,¶ ThomasHambrock,MD, PhD,* J. AlfredWitjes,MD, PhD,Þ Jurgen J. Futterer,MD, PhD,*Christina A. Hulsbergen-van de Kaa, MD, PhD,# and Jelle O. Barentsz, MD, PhD*

Objectives: The objective of this study was to evaluate the role of 3-Tmultiparametric magnetic resonance imaging (MP-MRI) and magneticresonanceYguided biopsy (MRGB) in early risk restratification of patients onactive surveillance at 3 and 12 months of follow-up.Materials and Methods: Within 4 hospitals participating in a large activesurveillance trial, a side study was initiated. Pelvic magnetic resonance imaging,prostate MP-MRI, and MRGB were performed at 3 and 12 months (latter prostateMP-MRI andMRGBonly) after prostate cancer diagnosis in 1 of the 4 participatinghospitals. Cancer-suspicious regions (CSRs) were defined on prostate MP-MRIusing Prostate Imaging Reporting And Data System (PI-RADS) scores.

Risk restratification criteria for active surveillance discontinuance were (1)histopathologically proven magnetic resonance imaging suspicion of node/bonemetastases and/or (2) a Gleason growth pattern (GGP) 4 and/or 5 and/or cancermultifocality (Q3 foci) in MRGB specimens of a CSR on MP-MRI.Results: From 2009 to 2012, a total of 64 of 82 patients were consecutivelyand prospectively included and underwent MP-MRI and a subsequent MRGB.At 3 and 12months of follow-up, 14% (9/64) and 10% (3/30) of the patients wererisk-restratified on the basis of MP-MRI and MRGB. An overall CSR PI-RADSscore of 1 or 2 had a negative predictive value of 84% (38/45) for detection of anyprostate cancer and 100% (45/45) for detection of a GGP 4 or 5 containing canceruponMRGB, respectively. ACSR PI-RADS score of 4 or higher had a sensitivityof 92% (11/12) for detection of a GGP 4 or 5 containing cancer upon MRGB.Conclusions: Application of MP-MRI and MRGB in active surveillance maycontribute in early identification of patients with GGP 4 or 5 containing cancersat 3 months of follow-up. If, during further follow-up, a PI-RADS score of 1 or2 continues to have a negative predictive value for GGP 4 or 5 containing cancers,a PI-RADS standardized reported MP-MRI may be a promising tool for the se-lection of prostate cancer patients suitable for active surveillance.

Key Words: active surveillance, magnetic resonance imaging, MR-guidedbiopsy, prostate cancer

(Invest Radiol 2014;49: 165Y172)

P revalence of low-risk prostate cancer has increased because ofwidespread prostate-specific antigen (PSA) testing.1 Patients

with low-risk prostate cancer (Table 1) are prone to overtreatmentand its complications, which can undermine a patient’s quality oflife.1,2 To avoid overtreatment, active surveillance (AS) is an ac-cepted treatment alternative for patients with low-risk prostatecancer.3

In general, a PSA of 10 ng/mL or less, a PSA density lesserthan 0.15 ng/mL, clinical stages T1 to T2a, a Gleason score (GS) of6 or lower without a Gleason growth pattern (GGP) 4 or 5, and 33%of cancer-positive transrectal ultrasound biopsy (TRUS-Bx) cores orless with 50 volume-percent of cancer per core or less are used ascriteria to identify patients with low-risk cancer for inclusion in ASprotocols.4 However, because of differences in the latter definitionbetween different study protocols, a patient’s eligibility for AS mayvary considerably (4%Y82%) between studies.5

Active surveillance is a management option that applies to pa-tients with presumed low-risk prostate cancer, who are followed byregular PSA measurements, digital rectal examinations, and annuallyrepeated systematic TRUS-Bx. In general, PSA kinetics, upgrading toa higher GGP (4 or 5), and volume progression are used as criteria fordisease progression.4 However, because of TRUS-Bx undersamplingupon inclusion rather than of true cancer progression, 20% to 30%of AS patients harbor cancers containing a GGP 4 or 5 or a cancer witha volume greater than 0.5 mL.4,6,7 Early identification of these patients,who were incorrectly considered suitable for AS, may be essential tomaintain the opportunity for appropriate curative treatment. The detec-tion of a GGP 4 or 5 and/or a larger cancer volume and/or multifocality ofa GS of 3 + 3 cancer or less7 results in restratification of these prostatecancer patients into a higher risk category. Risk restratification impliesdiscontinuation of AS and conversion to radical treatment.

Magnetic resonanceYguided biopsy (MRGB) has shown to im-prove identification of patients having cancers with a GGP 4 or 5 becauseof a better highest GGP concordance (88%) with radical prostatectomyspecimens compared with TRUS-Bx (55%, P = 0.001).8 The latter iscaused by better localization and targeting of the most aggressive area ofa cancer-suspicous region (CSR) on diffusion-weighted magnetic reso-nance imaging (MRI).8Y10 Only a few studies have related MRI results toAS outcome.11Y14

To our knowledge, MRGB has not previously been evaluatedat AS inclusion. Our hypothesis was that combined multiparametricMRI (MP-MRI) and MRGB may improve current TRUS-BxYbasedselection of patients for AS by early detection of patients harboringcancers of a larger volume or cancers containing a higher GGP.Therefore, our purpose was to evaluate the value of 3-T MP-MRIand MRGB for early risk restratification of patients on AS at 3 andat 12 months of follow-up.

ORIGINAL ARTICLE

Investigative Radiology & Volume 49, Number 3, March 2014 www.investigativeradiology.com 165

Received for publication July 13, 2013; and accepted for publication, after revision,September 17, 2013.

From the Departments of *Radiology, and †Urology, Radboud University NijmegenMedical Centre; ‡Department of Urology, Canisius Wilhelmina Hospital, Nijmegen;§Department of Urology, Jeroen Bosch Hospital, Den Bosch; ||Department ofUrology, Alysis Zorggroep, Arnhem; ¶Department of Urology, ErasmusUniversity Medical Centre, Rottterdam; and #Department of Pathology,Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands.

Conflicts of interest and sources of funding: Supported by a Dutch Cancer SocietyGrant (KUN 2007-3971).

The authors report no conflicts of interest.Reprints: Caroline M.A. Hoeks, MD, PhD, Department of Radiology, Meander

Medical Centre, Amersfoort, POBox 1502, 3800 BMAmersfoort, The Netherlands.E-mail: [email protected].

Copyright * 2014 by Lippincott Williams & WilkinsISSN: 0020-9996/14/4903Y0165

Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

mailto:[email protected]

Part of our patient population has been reported earlier.15 Thelatter article describes the value of apparent diffusion coefficient (ADC)values of MRGB diffusion-weighted imaging (DWI) scans for pros-tate cancer differentiation in patients with prostate cancer on AS. Thecurrent study investigates the overall outcome of the incorporation ofMP-MRI and MRGB in AS at both 3 and 12 months of follow-upand on its consequences for patient management.

MATERIALS AND METHODSWithin 4 centers participating in a large AS trial (NTR1718,

http://www.trialregister.nl), a prospective side study (NTR2006) wasinitiated in consecutively and prospectively included patients fromAugust 2009 to March 2012. Informed consent from the patients wasobtained for the study as well as for the side study, and institutionalreview boards of the participating hospitals approved our side study.Inclusion and exclusion criteria are depicted in Appendix 1. In ourside study, the patients on AS underwent pelvic MRI and MP-MRI inthe second month and MRGB in the third month after initial cancerdiagnosis upon systematic TRUS-Bx (time-point zero). Initial systematicTRUS-Bx consisted of 9 to 10 cores sampling both the transition andperipheral zones. Multiparametric magnetic resonance imaging andMRGB were repeated at 12 months of follow-up. All initial and repeatedMP-MRI examinations, all initial and repeated MRGB procedures, and,as will be further elaborated in the Follow-up section below, all repeatedTRUS-Bx examinations were performed in one of the 4 centers.

Magnetic Resonance ImagingPelvic MRI for lymph node and bone staging was followed by

MP-MRI of the prostate, consisting of T2-weighted MRI, DWI, anddynamic contrast-enhanced imaging (DCE-MRI) according to theEuropean Society of Uroradiology (ESUR) guidelines.16 Imaging wasperformed on a 3-Tmagnetic resonance (MR) system (Trio Tim; Siemens,Erlangen, Germany) using a pelvic phased array and an endorectal coil(Medrad, Pittsburgh, PA) filled with 40 mL of perfluorcarbon (Fomblin;Solvay-Solexis, Milan, Italy). The DCE-MRI was performed by initialacquisition of proton-density weighted images, followed by spoiledT1-weighted gradient echoes during fast (2.5 mL/s) intravenous injection

of 0.1 mmol of gadoterate meglumine (Dotarem; Guerbet, Paris, France)per kilogram of body weight. Magnetic resonance imaging parametersare presented in Table 1.

Magnetic Resonance Imaging InterpretationAn experienced radiologist (J.O.B.) with 19 years of experience

in prostate MRI evaluated all MP-MRI examinations on in-houseYdeveloped software while disposing of clinical patient data.17

Using the software, T2-weighted MRI, DWI, and DCE-MRI wereinterpreted simultaneously.17 The Prostate Imaging Reporting AndDataSystem (PI-RADS) was used to define CSRs.16 Every CSR was scoredon a 1-to-5Ypoint scale for T2-weighted MRI, DWI, and DCE-MRIseparately. Subsequently, an overall 5-point score, based on the wholeMP-MRI examination, was given for every CSR.16 The 5-point scalewas defined as (1) highly unlikely, (2) unlikely, (3) equivocal, (4) likely,and (5) highly likely presence of clinically significant prostate cancer.Prostate cancer staging was performed in compliance with establishedcriteria such as neurovascular bundle asymmetry, obliteration of therectoprostatic angle, irregular bulging of the prostatic contour, tumorsignal intensity within the periprostatic fat, and overt extracapsular tu-mor.18 When no CSR could be defined by MP-MRI, AS was continuedwithout performing MRGB.

Magnetic ResonanceYGuided Prostate BiopsyAll MRGB procedures were performed by an experienced

radiologist (C.M.A.H.) with 3 years of experience in one of the 4 referralcenters on the same MR scanner. This radiologist performed MRGB ofevery predefined CSR on a 3-T scanner in a separate examinationsession (MAGNETOM Skyra; Siemens, Erlangen, Germany).19

Magnetic resonanceYguided biopsy was performed for every CSR,regardless of CSR PI-RADS scores. Acquisition parameters arepresented in Table 1.

Earlier defined CSRs were re-identified on T2-weighted anddiffusion-weighted MRI. An endorectally inserted needle guide wasrepositioned to aim at a CSR using sagittal and axial gradient echosequences. Once needle guide positioning was adequate, biopsieswere taken by inserting an 18-gauge needle biopsy gun (In vivo,

TABLE 1. Parameters of MP-MRI and MRGB

Protocol SequenceTR,ms TE, ms

FlipAngle,degrees

SliceThickness,

mmField of View,mm � mm

MatrixSize

Voxel Size,mm � mm � mm

b-Values,s/mm2

TemporalResolution, s

Multiparametric MRI lymph nodes and bone structures

3D T2WI TSE coronal 1390 100 100 1.0 320 � 320 320 � 320 1.0 � 1.0 � 1.0 n.a. n.a.

T1WI TSE coronal 500 11 120 3.0 384 � 384 320 � 256 1.5 � 1.5 � 3.0 n.a. n.a.

WBDWI EPI 6500 71 n.a. 3.0 385 � 385 154 � 154 2.5 � 2.5 � 3.0 600 n.a.

WBDWI EPI 6200 66 n.a. 3.0 385 � 385 154 � 154 2.5 � 2.5 � 3.0 50 n.a.

Endorectal multiparametric MRI local prostate

T2WI Axial (TSE) 4280 99 120 3.0 180 � 178 448 � 448 0.4 � 0.4 � 3.0 n.a. n.a.

Coronal 3590 98 120 3.0 192 � 96 384 � 384 0.5 � 0.5 � 3.0 n.a. n.a.

Sagittal 4290 98 120 3.0 192 � 134 384 � 384 0.5 � 0.5 � 3.0 n.a. n.a.

DWI SSEPI Axial 2600 90 n.a. 3.0 204 � 204 136 � 136 1.5 � 1.5 � 3.0 0/50/500/800 n.a.

PD GE Axial 3D 800 1.51 14 3.0 192 � 192 128 � 128 1.5 � 1.5 � 3.0 n.a. n.a.

DCE-MRI Spoiled GE axial 3D 36 1.4 10 3.0 192 � 192 128 � 128 1.5 � 1.5 � 3.0 n.a. 3.4

MR-guided biopsy

T2WI TSE axial 3620 103 120 4.0 256 � 256 320 � 320 0.8 � 0.8 � 4.0 n.a. n.a.

DWI EPI axial 3300 60 n.a. 3.6 260 � 211 160 � 120 2.2 � 1.6 � 3.6 0/100/400/800 n.a.

SSFP GE axial and sagittal 4.48 2.24 70 3.0 280 � 280 256 � 256 1.1 � 1.1 � 3.0 n.a. n.a.

3D indicates 3-dimensional; DCE-MRI, dynamic contrast-enhanced magnetic resonance imaging; DWI, diffusion-weighted magnetic resonance imaging; EPI, echo-planarimaging sequence; GE, gradient echo;MR,magnetic resonance; n.a., not applicable; PD, proton densityweighted imaging; SSEPI, single-shot echo-planar imaging; SSFP, steady-statefreeprecession;T2WI,T2-weightedmagnetic resonance imaging;TE, echo time;TR, repetition time;TSE, turbo-spin echo;WBDWI,whole-bodydiffusionweighted imaging sequence.

Hoeks et al Investigative Radiology & Volume 49, Number 3, March 2014

166 www.investigativeradiology.com * 2014 Lippincott Williams & Wilkins


http://www.trialregister.nl

Schwerin, Germany) through the needle guide.19 Axial and sagittalgradient echo sequences were repeated, with the needle situated in theprostate, to confirm needle sampling of a CSR.

Risk Restratification at 3 Months of Follow-upRisk restratification, that is, stratification into a higher pros-

tate cancer risk category, was based on (1) pelvic MRI suggestion ofnode/bone metastases of prostate cancer, which was histopathologi-cally proven,20 or (2) MRGB histopathology specimens (of CSRs)containing (a) cancers with GS composed of a GGP20 4 or 5 or (b)multifocality of 3 or more foci GS of 3 + 3 cancer or lower, including

the foci in the initial TRUS-Bx. The latter criterion of cancer multi-focality was applied to evaluate the number of additionally detectedcancer foci by MRGB and to compare it with the TRUS-Bx riskrestratification criterion of more than 2 cores with prostate cancer inthe AS study.7 An MRGB focus located contralateral to the initialTRUS-Bx cancer location or a focus in the apex versus the base andvice versa was considered a separate cancer focus. Risk-restratifiedpatients were no longer eligible for AS and were referred to undergocurative treatment.

To evaluate cancer volume using MRGB and TRUS-Bx speci-mens, retrospectively, the maximal cancer core length (MCCL) was

FIGURE 1. Study flow diagram showing patient selection. CSR indicates cancer suspicious region; MRI, magnetic resonanceimaging; MP-MRI, multiparametric MR imaging; MRGB, MR guided prostate biopsy; MR, magnetic resonance; CSR, cancersuspicious region on magnetic resonance imaging; TRUS-Bx, systematic transrectal ultrasound biopsy.

Investigative Radiology & Volume 49, Number 3, March 2014 MRI and MR-Guided Biopsy in Active Surveillance

* 2014 Lippincott Williams & Wilkins www.investigativeradiology.com 167


measured. This was defined as the longest biopsy core specimen cancercore length taken from 1 CSR. An MCCL of 6 mm or longer is relatedto a cancer volume of 0.5mLor greater in radical prostatectomy specimensusing schematic mapping biopsy.21

Follow-upAfter 11 to 12 months of follow-up, repeated MP-MRI of the

local prostate, with imaging parameters identical to those of the initialMP-MRI, was performed. On the basis of the repeated MP-MRI, anadditional repeated MRGB, with indication and procedure similar tothose of the initial MRGB, was performed in a second separate imagingsession. After the repeated MRGB, a repeated TRUS-Bx session wasperformed on the same day by an experienced nurse practitioner, whowas blinded to the initial and follow-up MP-MRI results.

Repeated TRUS-Bx consisted of a schematic 10 core biopsyscheme, including sextant biopsy of the peripheral zone and 2 biopsiesof the transition zone on both sides of the prostate. The TRUS-Bx riskrestratification consisted of prostate cancer presence in more than2 cores or a GS of 6 (3 + 3) or higher.7,21 Risk restratification criteria forrepeated MP-MRI and repeated MRGB were according to the criteriaas described previously.

HistopathologyAll biopsy samples were evaluated by 1 genitourinary patholo-

gist (C.A.H.) with 20 years of experience, who was blinded to priorhistopathology results. Gleason grading was performed according tothe modified consensus22 of the International Society of UrologicalPathology in 2005. Prostatitiswas defined as the presence of intraprostaticinflammatory infiltrates.23

Statistical AnalysisPatient risk restratification rates were determined for the ini-

tial (3 months) and repeated MRGB and for the repeated TRUS-Bx(12 months). Parametric continuous variables were reported as mean T95% confidence interval, and nonparametric continuous variableswere reported as median and the interquartile range (IQR). Receiveroperating characteristic curves were applied to compare different MP-MRI techniques using the area under the curve (Az). Analyses wereperformed using PASW Statistics version 18 (SPSS, Inc, Hong Kong,China). The threshold for significance was set at P G 0.05.

RESULTS

Initial Risk Restratification (3 Months)Of the 82 AS patients, 66 patients were included in our side

study and underwent MP-MRI. Two patients requested to be ex-cluded from the protocol before MRGB, leaving 64 patients for thestudy. Patient selection is described in Figure 1. Patient characteris-tics of these 64 patients are shown in Table 2. One additional patientwas excluded because of an MRI result suggestive of a bone metas-tasis in his third lumbar vertebra, which, upon biopsy, appeared to bea metastasis from malignancy of unknown origin.

Magnetic resonanceYguided biopsy was performed in 62 ofthe 63 remaining patients. In 1 patient, MRGB was not performedbecause MP-MRI did not show a CSR and this patient continued AS.In the remaining 62 patients, a median of 2 (IQR, 1Y2) CSRs wereidentified with a median of 4 acquired MRGB cores (IQR, 3Y5) perpatient. A patient example is illustrated in Figure 2.

Fourteen percent (9/63) of the 63 patients were risk-restratifiedon the basis of MP-MRI and MRGB and thus underwent radicaltreatment on the basis of MRGB specimens containing cancers with aGGP 4 or 5 (n = 7) and/or 3 or more foci GS of 3+3 cancer or lower(n = 2). The MRGB and MCCL results are presented in Table 3.

Six patients were lost to follow-up because diversion to radicaltreatment was not initiated according to the predefined criteria (based

on only 2 foci GS 3+3 cancer [n = 4] or based on suspicion ofextracapsular extension on MRI without confirmation of extracapsularextension on MR-guided biopsy [n = 2]).

The remaining 48 patients continued AS. Eighteen of thesepatients had MRGB results according to our predefined AS criteriaand 30 patients had a cancer-negative MRGB specimen.

Risk Restratification at 12 Months of Follow-upIn 37 of the 48 remaining AS patients (77%), a follow-up of

12 months was available. Of these 37 patients, 7 patients did notundergo the repeated MP-MRI and MRGB examinations because ofimpossible repeated examinations caused by the following: shoulderinjury (n = 1), pain after MRGB (n = 1), complications of chemo-therapy for lymphoma (n = 2), or because of patient request (n = 3).

Follow-up MRGB and MCCL results for the remaining30 patients are presented in Table 4. Forty-seven percent of thesefollow-up patients (14/30) were risk-restratified on the basis ofMRGB and/or TRUS-Bx. In 13% (4/30) of the patients, risk stratifi-cation was based on TRUS-Bx only; in 10% (3/30), on MRGB onlyand in 23% (7/30), on both modalities. At 12 months of follow-up, theapplication of MP-MRI and MRGB yielded an additional 10% (3/30)of risk-restratified patients missed by TRUS-Bx.

The 14 risk-restratified patients at 12 months of follow-upwere not identified earlier on the initial combined MP-MRI andMRGB. Retrospectively, in 4 of these patients, the CSR was detectedon the initial MP-MRI, but the initial MRGB did not sample prostatetissue (n = 2) or missed small cancers (repeated MRGB MCCL

TABLE 2. Patient Characteristics of 64 MP-MRI Patients

Characteristic All Included Patients (n = 64)

Age, median (IQR), y 65.7 (62.1Y70.1)

PSA, mean (CI), ng/mL 6.5 (5.99Y6.93)

PSA density, mean(CI), ng/mL per mL

0.1 (0.12Y0.14)

Prostate volume, median(IQR), mL

45.8 (38.0Y66.1)

No. previous negative TRUS-Bxsessions, median (range)

0 (0Y7)

Total no. TRUS-Bx cores atdiagnosis, median (IQR)

10 (9Y10)

TRUS-Bx to MRI interval,median (IQR), mo

2.1 (1.6Y2.7)

TRUS-Bx to MRGB interval,median (IQR), mo

2.7 (2.0Y3.3)

Characteristic All included patients no., percentage(fraction), [95% confidence interval]

Clinical stage

T1c 76.6 (49/64), [64.8Y85.4]

T2a 18.8 (12/64), [10.9Y30.1]

T2b 3.1 (2/64), [0.2Y11.3]

T2c 1.6 (1/64), [0.0Y9.1]

Positive TRUS-Bxcores at diagnosis,

1 67.2 (43/64), [55.0Y77.5]

2 32.8 (21/64), [22.5Y45.0]

Gleason score at diagnosis,

3+3=6 93.8 (60/64), [84.6Y98.0]

Lower 6.3 (4/64), [2.0Y15.4]

AS indicates active surveillance; CI, 95% confidence interval; IQR,interquartile range; MP-MRI, multiparametric magnetic resonance imaging;MRGB, magnetic resonanceYguided prostate biopsy; PSA, prostate-specificantigen; TRUS-Bx, systematic transrectal ultrasound biopsy.




1.5Y2 mm) (n = 2). In the other 10 of the 14 patients, presumably,small lesions (G0.5 mL) may have not been scored on the initialMP-MRI (repeated MRGB MCCL, 0.3Y4.5 mm).

For 14 of the 30 patients with an initial cancer-negativeMRGB, repeated MP-MRI and repeated MRGB were available.The negative predictive value (NPV) of a cancer-negative MRGB for

risk restratification at repeated examinations was 79% ([11/14], 95%confidence interval, 52%Y93%).

Results for MRI PI-RADS ScoresThe MP-MRI evaluation on both 3 and 12 months resulted in a

total of 168 CSRs. Because this study started at early PI-RADS

FIGURE 2. A 61-year-old man on AS with a PSA level of 7.1 ng/mL, a PSA density of 0.19 ng/mL per mL, and a clinical stage T1C.This patient was diagnosed with a GS of 6 (3þ 3) prostate cancer in 5 vol% in 1 of 12 cores in the right base of the peripheral zone.Multiparametric MRI and MRGB, consisting of 4 cores only, were performed within 3 months after the diagnosis. A, AxialT2-weighted turbo spin echo image (repetition time [TR], 4280milliseconds; echo time [TE], 99 milliseconds): a low signal intensitywith a lenticular shape and an erased charcoal sign (white arrows) is present in the ventral transition zone at the level of themid-prostate. B, Axial ADC map of DWI (single-shot echo planar imaging; TR, 2600 milliseconds; TE, 90 milliseconds; b values,0/50/500 and 800 s/mm2) at the level of the mid-prostate. A low ADC value of 0.50 � 10j3 mm2/s, suspicious for prostate cancer,was present on the right side of the ventral transition zone (dotted line). C, Axial overlay of Ktrans parameter in DCE-MRI(3-dimensional spoiled gradient echo; TR, 36 milliseconds; TE, 1.4 milliseconds; temporal resolution, 3.4 seconds), as calculated bythe Tofts model, on the axial T2-weighted turbo spin echo image (TR, 4280 milliseconds; TE, 99 milliseconds). Areas of increasedcontrast enhancement are present in large parts of the prostate. Also, in the right ventral prostate (dotted line), increasedenhancement is present. Enhancement in this region was suggestive of prostate cancer because of wash-out: a decline at the endof the relative gadolinium contrast-to-time curve (D), its location, and asymmetry. E, Axial angulated balanced gradient echoimage (TR, 4.48 milliseconds; TE, 2.24 milliseconds) of the needle position in the lesion presented in A to C directly after thebiopsy. The lesion (green dotted line) can be appreciated in the prostate (blue dotted line). The middle of the needle artifact isrepresented by a white line and is in the middle of the lesion. The needle guide (white arrows) is also depicted. The MRGBspecimen (total of only 4 cores) contained a GS of 4 + 3 = 7 prostate cancer in 80 vol%. This patient’s management wassubsequently redirected toward definitive therapy, which consisted of external beam radiotherapy.




implementation, PI-RADS scores were available for 155 CSRs only.Overall, CSR PI-RADS scores at 3 and 12 months of follow-up arepresented in Table 5.

Seventy-eight percent (121/155) of the CSRs were located in theperipheral zone, 15% (23/155) were located in the transition zone, and7% (11/155) were situated at the border of the peripheral and transitionzones or seminal vesicles. The MRGB specimens showed cancer in31% (48/155) of the CSRs. Cancer-negative MRGB specimens mainlycontained prostatitis in 41% (44/107) and healthy prostate tissue in38% (41/107). The Az values for the detection of any cancer prostatecancer and for the detection of a GGP 4 or 5 containing cancer, usingoverall PI-RADS scores, were 0.73 (0.65Y0.82) and 0.81 (0.70Y0.92),respectively.

An overall CSR PI-RADS score of 1 or 2 had an NPVof 84%(38/45) for the detection of any cancer and an NPVof 100% (45/45)for the detection of a GGP 4 or 5 containing cancer upon MRGB. ACSR PI-RADS score of 4 or 5 had a sensitivity of 75% (36/48) and92% (11/12) for the detection of, respectively, cancer and a GGP 4 or5 containing cancer upon subsequent MRGB.

DISCUSSIONThe results of our side study show that, in an AS cohort, the

initial application of MP-MRI and MRGB identifies patients withcancers containing higher GGPs (GGP 4 or 5), which were missedwith TRUS-Bx in 14% (9/64). At 12 months of follow-up, MRGBfurther improves patient risk restratification rates compared with therepeated TRUS-Bx with another 10% (3/30).

At 3 months of follow-up, our 11% rate of upgrading to aGGP 4 or 5 was lower compared with the 16% to 17% upgrading in2 series reporting upon immediate repeated TRUS-guided biopsyafter inclusion in an AS protocol.12,24 The higher rate of one of theseseries might be contributed to their overall population of patients withhigher-risk prostate cancer because they considered patients with 3 orfewer cancer-positive cores eligible for AS.12 Also, it should be notedthat we used less cores (a median of 4) in MRGB, compared with the12-core immediate repeated TRUS-guided biopsy schemes in thementioned series.

At 12 months of follow-up, the results of the combinedMP-MRIand MRGB diagnosed additional 3 patients for risk restratification,which were missed with TRUS-Bx and would therefore remain on AS.Thesewere all caused by cancermultifocality. The overall risk restratificationat 12 months of follow-up was based on a GGP 4 or 5 in 43% (6/14); in-crease in cancer volume (GS, e3 + 3), in 43% (6/14); cancer multifocality, in36% (5/14). The detection of cancers with a GGP 4 or 5 up to 1 to 2 yearsafter inclusion is assumed to be based on incorrect inclusion of cancerscontaining a higher GGP, rather than on actual tumor progression.6 Retro-spectively, on the basis of our results so far, 13 patients with a cancereventually containing a GGP 4 or 5 were incorrectly included in AS. At3 months of follow-up, MP-MRI and MRGB did not identify 46% (6/13)of these patients, who were detected at 12 months of follow-up. A pos-sible explanation for this can be deducted from Tables 3 and 4. At3 months of follow-up, 6 of 7 cancers had an MCCL of 6 mm or longer

TABLE 3. Reasons for Initial Patient Risk Restratification andConversion to Treatment at 3 Months of Follow-up

MRGB Results forRisk RestratifiedPatients No.

MRGB Core SpecimenMCCL, Mean (95%

Confidence Interval), mm*

MRGB GGP 4 or 5 7 10.6 (7.4Y13.7)

Q6 mm: n = 6

MRGB multifocality 2 4.0 (0.0Y7.9)

Q6 mm: n = 1

Total of riskrestratified patients,%†

9 (9/63 = 14%) Q6 mm: n = 7

The calculation of MRGB maximal cancer core length is based on thehighest MRGB maximal cancer core length for every patient.

*A biopsy core specimen maximal cancer core length of 6 mm or longer equalsa prostate cancer volume of 0.5 mL or greater in radical prostatectomy specimens.

†The total of 63 patients consists of the 62 patients undergoing magneticresonanceYguided biopsy and 1 patient not undergoing magnetic resonanceYguided biopsy because of the lack of CSRs.

GGP indicates Gleason grade pattern; GS, Gleason score; MCCL, maximalcancer core length; MP-MRI, multiparametric magnetic resonance imaging;MRGB, magnetic resonanceYguided prostate biopsy.

TABLE 4. Reasons for Patient Risk Restratification and Conversion to Treatment at Repeated Examinations at 12 Months of Follow-up

Repeat MRGB Results for Risk-Restratified Patients No. MRGB MCCL, Mean (95% Confidence Interval), mm*

Both MRGB and TRUS-Bx GGP 4 and/or 5 4 5.3 (3.8Y6.8)

Q6 mm: n = 2

MRGB GGP 4 and/or 5 and TRUS-Bx GS e3 + 3 cancer in 92 cores 1 4.4 (n.a.)

Only TRUS-Bx GGP 4 and/or 5 1 2.7 (n.a.)

TRUS-Bx GS e3+3 cancer in 92 cores and MRGB multifocality 2 6.5 (5.5Y7.5)

Q6 mm: n = 2

Only MRGB multifocality, (2 foci, n = 1)† 3 4.2 (1.3Y7.0)

Q6 mm: n = 1

Only TRUS-Bx GS e3 + 3 cancer in 92 cores 3 5.7 (2.8Y8.6)

Q6 mm: n = 1

Total (% of total repeated MRGB), n = 30 14 (47) 5.0 (3.0Y6.0)

Q6 mm: n = 6

The calculation of MRGB maximal cancer core length is based on the highest MRGB maximal cancer core length for every patient.

*A biopsy core specimen maximal cancer core length of 6 mm or longer equals a prostate cancer volume of 0.5 mL or greater in radical prostatectomyspecimens.

†Not conforming to the predefined risk restratification criteria.

GGP indicates Gleason growth pattern; GS, Gleason score; MCCL, maximal cancer core length; MP-MRI, multiparametric magnetic resonance imaging;MRGB, magnetic resonanceYguided prostate biopsy; n.a., not applicable; TRUS-Bx, systematic transrectal ultrasound biopsy.




corresponding to a tumor volume21 greater than 0.5 mL, whereas, at12 months of follow-up, only 2 of 6 cancers had an MCCL of 6 mmor longer. Therefore, the smaller volume of the (higher GGP withinthese relatively small) tumors may have influenced missing 46% ofGGP 4 or 5 containing cancers at 3 months of follow-up.

A probably equally important finding as the early identificationof aggressive cancer is the high NPV of MP-MRI and MRGB in AS,which may improve the selection of patients suitable for AS. Multi-parametric magnetic resonance imaging had a sensitivity of 92% for thedetection of GGP 4 or 5 containing cancers in case of higher PI-RADSscores (Q4) and an NPVof 100% for the detection of cancers containinga 4 or 5 GGP in case of PI-RADS scores 1 or 2. These results arecomparable with those of Vargas et al,25 who reported an NPVof 96% to100% and a sensitivity of 87% to 96% for biopsy upgrading in case of apredefined MR imaging score of 2 or less and 5 or higher for cancerpresence. Furthermore, an initial cancer-negative MRGB specimen hadan NPVof 79% for risk restratification at 12 months of follow-up. Thisfinding may indicate that a cancer-negative initial MRGB may be apromising prognostic parameter for AS patient selection.

Although our population is small and although our study wasnot designed to validate the PI-RADS scoring system, our results doshow that standardized MP-MRI using PI-RADS is a promisingtechnique for the differentiation between patients suitable for AS andpatients with cancers containing a GGP 4 or 5, the latter needingradical treatment. These results may support reproducibility and re-liability of our results on a larger scale.

Our results for prostate cancer detection accuracy using MP-MRI and MRGB in patients on AS are difficult to compare withthose of the literature: other studies on MRI implementation in AS didnot use MP-MRI and/or MRGB.11,13,14,26,27 Our Az values of 0.73 and0.81 for the detection of any cancer and GGP 4 or 5 containing cancerwere quite reasonable considering the expected prevalence of pre-dominantly lower (GGP 2 to 3) comprising cancers in this selectedAS patient population. Lower GGP containing cancers are known to havelower detection rates comparedwith cancers, which contain a higherGGP.28

Because our study is the first to evaluate MRGB in AS, we ap-plied low threshold criteria for CSR determination on MP-MRIfollowed by biopsy of all CSRs, also including regions with low can-cer suspicion (PI-RADS 1Y3). This resulted in a high number of pa-tients (48%) with cancer-negative CSRs upon MRGB. With increasingMRI experience in AS patients, false-positive results may be reducedby increasing the biopsy threshold to a higher PI-RADS score.

Limitations of the presented series are its small sample sizeand limited follow-up. Furthermore, because MP-MRI and MRGBare new techniques within AS, our risk restratification criteria maynot have been optimal yet because they are partly based on existentcriteria used for random systematic TRUS-Bx. The TRUS-Bx riskrestratification criterion of the number of cancer-positive cores isdesigned to estimate total cancer volume using a random systematicbiopsy technique. Magnetic resonanceYguided biopsy is a guidedbiopsy technique in which the latter criterion for tumor volume es-timation cannot be applied. Moreover, MP-MRI prostate tumor vol-umetrics are not very accurate because tumor delineation is limitedby false-positive (prostatitis, fibrosis) and false-negative (sparselygrowing cancers and lower cancer GS) imaging results.29,30 Prostate

cancer is a multifocal disease. Using MRGB, we acquired extra in-formation on the amount and locations of prostate cancer foci;however, currently, we cannot effectively use this criterion because itsprognostic value remains unknown.

Another limitation might be that, although we used a structuredsemiobjective scoring system for MP-MRI readings, our readings wereperformed by a highly experienced radiologist, which may have limitedthe general applicability of our results. To perform prostate MP-MRIand MRGB, any center needs experience in performing both prostateMP-MRI and MRGB of the prostate.

An important clinical implication of our study is that, upon ap-plication ofMRI in AS patients, acquisition of histopathology of a CSRis required because of potential false-positive CSRs. Lack of histo-pathologic confirmation of a CSR may explain poor results for (MP-)MRI as a predictive tool for AS outcome in other studies.11,13,14,26,27

Currently, no other studies that incorporate both MP-MRI and MRGBat inclusion in AS protocols exist. Therefore, future research shouldfocus on including both MP-MRI and MRGB in AS protocols for low-risk prostate cancer.

Other important subjects that warrant further investigation arethe validation of MRGB in AS populations and the influence of theexperience of the radiologist performing the MRGB procedure on theMRGB outcome. Magnetic resonanceYguided biopsy has not beenapplied earlier in patients on AS, who represent the lower part of therisk spectrum of patients with prostate cancer. Furthermore, in gen-eral, the effectiveness of any radiological procedure depends on theexperience of the involved radiologist, and this statement probablyalso holds true for MRGB.

In conclusion, the application of MP-MRI and MRGB in ASmay contribute in early identification of patients with cancerscontaining a higher GGP (4 or 5) at 3 months of follow-up. If, duringfurther follow-up, PI-RADS scores of 1 to 2 continue to have an NPVfor cancers containing a GGP 4 or 5, standardized reported MP-MRIusing PI-RADS may be a promising tool for the selection of patientssuitable for AS.

ACKNOWLEDGMENTSThe authors thank Gijs de Lauw, nurse practitioner, for his

contribution in the biopsy data acquisition.

REFERENCES1. Bangma CH, Roemeling S, Schroder FH. Overdiagnosis and overtreatment of

early detected prostate cancer. World J Urol. 2007;25:3Y9.2. Wei JT, DunnRL, Sandler HM, et al. Comprehensive comparison of health-related

quality of life after contemporary therapies for localized prostate cancer. J ClinOncol. 2002;20:557Y566.

3. Wilt TJ, Brawer MK, Jones KM, et al. Radical prostatectomy versus observationfor localized prostate cancer. N Engl J Med. 2012;367:203Y213.

4. Dall’Era MA, Cooperberg MR, Chan JM, et al. Active surveillance for early-stage prostate cancer: review of the current literature. Cancer. 2008;112:1650Y1659.

5. Dall’Era MA, Albertsen PC, Bangma C, et al. Active surveillance for prostatecancer: a systematic review of the literature. Eur Urol. 2012;62:e95Ye106.

6. Duffield AS, Lee TK, Miyamato H, et al. Radical prostatectomy (RP) findingsin patients who fail active surveillance of prostate cancer. Modern Pathol.2009;22:749.

TABLE 5. Overall PI-RADS Scores for Cancer-Suspicious Regions at 3 and 12 Months of Follow-up

Time Point (Months After Inclusion) PI-RADS1 PI-RADS 2 PI-RADS 3 PI-RADS 4 PI-RADS 5 Total

3 9 18 21 32 13 93

12 7 11 17 21 6 62

Total 16 29 38 53 19 155

PI-RADS indicates Prostate Imaging Reporting And Data System.




7. Bul M, Zhu X, Valdagni R, et al. Active surveillance for low-risk prostatecancer worldwide: the PRIAS study. Eur Urol. 2013;63:597Y603.

8. Hambrock T, Hoeks C, Hulsbergen-van de KC, et al. Prospective assessment ofprostate cancer aggressiveness using 3-T diffusion-weighted magnetic reso-nance imaging-guided biopsies versus a systematic 10-core transrectal ultra-sound prostate biopsy cohort. Eur Urol. 2012;61:177Y184.

9. Maas MC, Futterer JJ, Scheenen TW. Quantitative evaluation of computed highb value diffusion-weighted magnetic resonance imaging of the prostate. InvestRadiol. 2013;48:779Y786.

10. Hoeks CM, Vos EK, Bomers JG, et al. Diffusion-weighted magnetic resonanceimaging in the prostate transition zone: histopathological validation usingmagnetic resonance-guided biopsy specimens. Invest Radiol. 2013;48:693Y701.

11. Margel D, Yap SA, Lawrentschuk N, et al. Impact of multiparametric endorectalcoil prostate magnetic resonance imaging on disease reclassification amongactive surveillance candidates: a prospective cohort study. J Urol. 2012;187:1247Y1252.

12. Berglund RK, Masterson TA, Vora KC, et al. Pathological upgrading and upstaging with immediate repeat biopsy in patients eligible for active surveillance.J Urol. 2008;180:1964Y1967.

13. Cabrera AR, Coakley FV, Westphalen AC, et al. Prostate cancer: is inapparenttumor at endorectal MR and MR spectroscopic imaging a favorable prognosticfinding in patients who select active surveillance? Radiology. 2008;247:444Y450.

14. Ploussard G, Xylinas E, Durand X, et al. Magnetic resonance imaging does notimprove the prediction of misclassification of prostate cancer patients eligiblefor active surveillance when the most stringent selection criteria are based onthe saturation biopsy scheme. BJU Int. 2011;108:513Y517.

15. Somford DM, Hoeks CM, Hulsbergen-van de Kaa CA, et al. Evaluation ofdiffusion-weighted MR imaging (DWI) at inclusion in an active surveillanceprotocol for low-risk prostate cancer. Invest Radiol. 2013;48:152Y157.

16. Barentsz JO, Richtenberg J, Clements R, et al. ESUR Prostate MR Guidelines2012. Eur Radiol. 2012;22:746Y757.

17. Vos PC, Hambrock T, Barenstz JO, et al. Computer-assisted analysis of peripheralzone prostate lesions using T2-weighted and dynamic contrast enhanced T1-weighted MRI. Phys Med Biol. 2010;55:1719Y1734.

18. Futterer JJ, Heijmink SW, Scheenen TW, et al. Prostate cancer: local staging at3-T endorectal MR imagingVearly experience. Radiology. 2006;238:184Y191.

19. Hambrock T, Futterer JJ, Huisman HJ, et al. Thirty-two-channel coil 3T mag-netic resonance-guided biopsies of prostate tumor suspicious regions identifiedon multimodality 3T magnetic resonance imaging: technique and feasibility.Invest Radiol. 2008;43:686Y694.

20. van den Bergh RC, Vasarainen H, van der Poel HG, et al. Short-term outcomesof the prospective multicentre ‘‘Prostate Cancer Research International: ActiveSurveillance’’ study. BJU Int. 2010;105:956Y962.

21. Ahmed HU, Hu Y, Carter T, et al. Characterizing clinically significant prostatecancer using template prostate mapping biopsy. J Urol. 2011;186:458Y464.

22. Epstein JI, Allsbrook WC Jr, Amin MB, et al. The 2005 International Society ofUrological Pathology (ISUP) Consensus Conference on Gleason Grading ofProstatic Carcinoma. Am J Surg Pathol. 2005;29:1228Y1242.

23. Nickel JC, True LD, Krieger JN, et al. Consensus development of a histo-pathological classification system for chronic prostatic inflammation. BJU Int.2001;87:797Y805.

24. King AC, Livermore A, Laurila TA, et al. Impact of immediate TRUS rebiopsyin a patient cohort considering active surveillance for favorable risk prostatecancer. Urol Oncol. 2013;31:739Y743.

25. Vargas HA, Akin O, Afaq A, et al. Magnetic resonance imaging for predictingprostate biopsy findings in patients considered for active surveillance of clin-ically low risk prostate cancer. J Urol. 2012;188:1732Y1738.

26. Guzzo TJ, Resnick MJ, Canter DJ, et al. Endorectal T2-weighted MRI does notdifferentiate between favorable and adverse pathologic features in men withprostate cancer who would qualify for active surveillance. Urol Oncol. 2012;30:301Y305.

27. Fradet V, Kurhanewicz J, Cowan JE, et al. Prostate cancer managed with activesurveillance: role of anatomic MR imaging and MR spectroscopic imaging.Radiology. 2010;256:176Y183.

28. Ikonen S, Karkkainen P, Kivisaari L, et al. Magnetic resonance imaging ofprostatic cancer: does detection vary between high and low Gleason score tumors?Prostate. 2000;43:43Y48.

29. Schlemmer HP, Corvin S. Methods for volume assessment of prostate cancer.Eur Radiol. 2004;14:597Y606.

30. Langer DL, van der Kwast TH, Evans AJ, et al. Intermixed normal tissue withinprostate cancer: effect on MR imaging measurements of apparent diffusion co-efficient and T2Vsparse versus dense cancers. Radiology. 2008;249:900Y908.

APPENDIX 1. Active Surveillance Inclusion and Exclusion CriteriaWith an Additional Exclusion Criterion Used in Our Substudy

Inclusion criteria

Histopathologically proven adenocarcinoma of the prostate

Men should be fit for curative treatment.

PSA level at diagnosis e10.0 ng/mL

PSA density G0.2 ng/mL per mL

Clinical stage T1C or T2

Gleason score of e3 + 3

1 or 2 biopsy cores invaded with cancer

Participants must be willing to attend the follow-up visits.

Exclusion criteria

Men who cannot or do not want to be operated or irradiated

A former therapy for prostate cancer

Additional exclusion criterion substudy

Contraindications to MRI or to gadolinium-based contrast agents

MRI indicates magnetic resonance imaging; PSA, prostate-specific antigen.




value of 3-t multiparametric magnetic resonance...

Documents