EURURO-6255; No. of Pages 8 EUROPEAN UROLOGY XXX (2015) XXX–XXX

available at www.sciencedirect.com journal homepage: www.europeanurology.com

Prostate Cancer

Prospective Randomized Trial Comparing Magnetic Resonance Imaging (MRI)-guided In-bore Biopsy to MRI-ultrasound Fusion and Transrectal Ultrasound-guided Prostate Biopsy in Patients with Prior Negative Biopsies Christian Arsov a,*, Robert Rabenalt a, Dirk Blondin b, Michael Quentin b, Andreas Hiester a, Erhard Godehardt c, Helmut E. Gabbert d, Nikolaus Becker e, Gerald Antoch b, Peter Albers a, Lars Schimmo¨ller b a

Department of Urology, Heinrich-Heine University, Du¨sseldorf, Germany; b Department of Diagnostic and Interventional Radiology, University Dusseldorf,

Medical Faculty, D-40225 Du¨sseldorf, Germany;

c

Division of Statistics, Department of Cardiovascular Surgery, Heinrich-Heine University, Du¨sseldorf,

Germany; d Department of Pathology, Heinrich-Heine University, Du¨sseldorf, Germany; e Division of Cancer Epidemiology, German Cancer Research Center Heidelberg, Heidelberg, Germany

Article info

Abstract

Article history: Accepted June 9, 2015

Background: A significant proportion of prostate cancers (PCas) are missed by conventional transrectal ultrasound-guided biopsy (TRUS-GB). It remains unclear whether the combined approach using targeted magnetic resonance imaging (MRI)-ultrasound fusion–guided biopsy (FUS-GB) and systematic TRUS-GB is superior to targeted MRIguided in-bore biopsy (IB-GB) for PCa detection. Objective: To compare PCa detection between IB-GB alone and FUS-GB + TRUS-GB in patients with at least one negative TRUS-GB and prostate-specific antigen 4 ng/ml. Design, setting, and participants: Patients were prospectively randomized after multiparametric prostate MRI to IB-GB (arm A) or FUS-GB + TRUS-GB (arm B) from November 2011 to July 2014. Outcome measurements and statistical analysis: The study was powered at 80% to demonstrate an overall PCa detection rate of 60% in arm B compared to 40% in arm A. Secondary endpoints were the distribution of highest Gleason scores, the rate of detection of significant PCa (Gleason 7), the number of biopsy cores to detect one (significant) PCa, the positivity rate for biopsy cores, and tumor involvement per biopsy core. Results and limitations: The study was halted after interim analysis because the primary endpoint was not met. The trial enrolled 267 patients, of whom 210 were analyzed (106 randomized to arm A and 104 to arm B). PCa detection was 37% in arm A and 39% in arm B (95% confidence interval for difference, –16% to 11%; p = 0.7). Detection rates for significant PCa (29% vs 32%; p = 0.7) and the highest percentage tumor involvement per biopsy core (48% vs 42%; p = 0.4) were similar between the arms. The mean number of cores was 5.6 versus 17 (p < 0.001). A limitation is the limited number of patients because of early cessation of accrual. Conclusions: This trial failed to identify an important improvement in detection rate for the combined biopsy approach over MRI-targeted biopsy alone. A prospective comparison between MRI-targeted biopsy alone and systematic TRUS-GB is justified.

Associate Editor: Giacomo Novara Keywords: Prostate cancer Multiparametric magnetic resonance imaging Magnetic resonance imagingultrasound fusion–guided prostate biopsy Magnetic resonance imaging– guided in-bore prostate biopsy Randomized trial

* Corresponding author. Department of Urology, Heinrich-Heine University, Moorenstrasse 5, 40225 Du¨sseldorf, Germany. Tel. +49 211 8108607; Fax: +49 211 8116440. E-mail address: [email protected] (C. Arsov). http://dx.doi.org/10.1016/j.eururo.2015.06.008 0302-2838/# 2015 European Association of Urology. Published by Elsevier B.V. All rights reserved.

Please cite this article in press as: Arsov C, et al. Prospective Randomized Trial Comparing Magnetic Resonance Imaging (MRI)guided In-bore Biopsy to MRI-ultrasound Fusion and Transrectal Ultrasound-guided Prostate Biopsy in Patients with Prior Negative Biopsies. Eur Urol (2015), http://dx.doi.org/10.1016/j.eururo.2015.06.008

EURURO-6255; No. of Pages 8 2

EUROPEAN UROLOGY XXX (2015) XXX–XXX

Patient summary: Our randomized study showed similar prostate cancer detection rates between targeted prostate biopsy guided by magnetic resonance imaging and the combination of targeted biopsy and systematic transrectal ultrasound-guided prostate biopsy. An important improvement in detection rates using the combined biopsy approach can be excluded. # 2015 European Association of Urology. Published by Elsevier B.V. All rights reserved.

1.

Introduction

criteria were known PCa, a contraindication to MRI or biopsy, and prior MRI-guided biopsy. The study was approved by our institutional review

The rate of prostate cancer (PCa) detection for a first systematic transrectal ultrasound-guided biopsy (TRUS-GB) is typically 30–50% [1,2]. However, a substantial proportion of men with initially negative TRUS-GB still harbor PCa and present with persistently increasing prostate-specific antigen (PSA) values. These patients pose a diagnostic dilemma. Repeated prostate biopsies using ultrasound-based techniques show PCa detection rates of 11–47% [3]. Nevertheless, clinically significant PCa can be missed even after several repeat TRUS-GBs. This applies especially to patients with anteriorly located tumors, which are frequently underdiagnosed by TRUS-GB [4,5]. Therefore, contemporary guidelines recommend the use of multiparametric MRI (mpMRI) of the prostate and MRI-guided targeted biopsy if a suspicion of PCa persists in spite of prior negative prostate biopsies [6]. In recent years, two different MRI-guided biopsy techniques have been established: direct MRIguided biopsy, which is performed in-bore (IB-GB); and MRI-ultrasound fusion–guided biopsy (FUS-GB), which combines MRI-guided targeted biopsy and systematic TRUS-GB within one biopsy session [7–9]. In the re-biopsy setting, PCa detection rates of 9.5–59% have been reported for MRI-guided biopsy approaches [3]. However, direct comparisons of novel biopsy approaches are still pending. Owing to the rapid increase in the use of MRI-guided biopsy, the question of whether MRI-guided biopsy yields comparable PCa detection rates is of utmost importance. The present prospective randomized study compares PCa detection rates between IB-GB (up to six cores) and FUS-GB combined with systematic 12-core TRUS-GB (up to 18 cores) in patients with at least one negative TRUS-GB and persistent suspicion of PCa.

board, and written informed consent was obtained from all patients.

2.3.

Study endpoints

The primary endpoint was the overall PCa detection rate. Secondary endpoints included (1) the distribution of the highest Gleason scores; (2) the detection rate for significant tumors, defined as cancers with Gleason score 7; (3) the positivity rate for biopsy cores; (4) the number of biopsy cores needed to detect one PCa or one significant PCa; (5) the percentage tumor involvement per biopsy core; and (6) for arm B, comparison of FUS-GB and systematic 12-core TRUS-GB.

2.4.

Imaging

Patients underwent mpMRI performed at 3 T (Magnetom Trio; Siemens Healthcare, Erlangen, Germany) with a six-channel phased-array body coil. The imaging protocol included T1- and T2-weighted imaging (T2WI), diffusion-weighted imaging (DWI), and dynamic contrastenhanced imaging (DCE-MRI) [11,12]. Up to three intraprostatic lesions per patient were scored on a 5-point Likert scale for each MRI sequence (T2WI, DWI, DCE-MRI). The single scores were summed up to a score of 3–15 points. After April 2012, once the Prostate Imaging Reporting and Data System (PI-RADS) classification had been published, a protocol amendment was submitted and the scoring followed the PI-RADS classification. PI-RADS single scores (PSsingle) and PI-RADS sum scores (PSsum) were assessed for each intraprostatic lesion [13]. Scoring was performed by three radiologists (M.Q., D.B., L.S.) who had 2–4 yr of experience in prostate MRI at the start of the study.

2.5.

Randomization

Patients were randomly assigned in a 1:1 ratio to arm A or to arm B if they had at least one intraprostatic lesion with a total score 10 (up to April 2012) or a PSsum 10 (after April 2012). A computer-generated list of random numbers was used for allocation of patients. This list was password-protected in a computer database. Patients were allocated by

2.

Patients and methods

2.1.

Study design

an independent research nurse to ensure that the randomization group could not be predicted.

2.6.

Interventions

NCT02220517) has been published previously [10]. A Consolidated

2.6.1.

MR in-bore guided biopsy (arm A)

Standards of Reporting Trials flow chart for the study is shown in Fig. 1.

For IB-GB, patients were placed in a prone position and a needle guide

The detailed design of this prospective study (ClinicalTrials.gov

From November 2011 to July 2014, 275 consecutive patients referred to

fixed to a portable biopsy device (DynaTRIM; Invivo, Gainesville, FL, USA)

our institution by office urologists were assessed for eligibility. Patients

was introduced rectally [12]. Lesions were reidentified by direct visual

were randomly allocated at a ratio of 1:1 to IB-GB (arm A) or FUS-GB

correlation using the initial mpMRI (marks on T2WI and DWI). The MRI

plus systematic 12-core TRUS-GB (arm B) if mpMRI was suspicious for PCa.

patient table was moved out of the bore for needle placement. Needle-in

Overall, 210 out of 224 randomized patients (94%) were analyzed.

control scans were performed in two different planes (axial and coronal/ sagittal half-Fourier acquisition single-shot turbo spin-echo sequences).

2.2.

Study population

Two targeted cores were taken from each lesion using an MRIcompatible, 18G, fully automatic biopsy gun (Invivo). Biopsies were

Patients were eligible to participate in the study if they had at least one

carried out by two radiologists (M.Q., L.S.) who had performed more than

negative TRUS-GB and persistent serum PSA levels 4 ng/ml. Exclusion

50 procedures at the start of the study.

Please cite this article in press as: Arsov C, et al. Prospective Randomized Trial Comparing Magnetic Resonance Imaging (MRI)guided In-bore Biopsy to MRI-ultrasound Fusion and Transrectal Ultrasound-guided Prostate Biopsy in Patients with Prior Negative Biopsies. Eur Urol (2015), http://dx.doi.org/10.1016/j.eururo.2015.06.008

EURURO-6255; No. of Pages 8 3

EUROPEAN UROLOGY XXX (2015) XXX–XXX

Assessed for eligibility (n = 275)

Enrollment

Excluded before mpMRI (n = 8) • Did not meet inclusion criteria (n = 6) • Declined to parcipate (n = 2) mpMRI (n = 267) Excluded before randomizaon (n = 43) • Did not have an intraprostac lesion with PSsum ≥10 (n = 43) Randomized (n = 224)

Allocaon

Allocated to Arm A (n = 111) (MRI-guided in-bore biopsy) • Received allocated intervenon (n = 106) • Did not receive (complete) allocated intervenon (n = 5) - 2 with deteriorated health condion - 2 received incomplete biopsy (circulatory disorder, anal fissure) - 1 withdrew consent following randomizaon

Allocated to Arm B (n = 113) (MRI-US fusion–guided plus systemac 12-core TRUS-GB) • Received allocated intervenon (n = 104) • Did not receive (complete) allocated intervenon (n = 9) - 4 refused systemac TRUS-GB - 2 did not receive complete 12-core TRUS-GB - 1 withdrew consent following randomizaon - 1 received incomplete targeted biopsy - 1 received no targeted biopsy

Analysis Analyzed (n = 106)

Analyzed (n = 104)

Fig. 1 – Consolidated Standards of Reporting Trials flow chart for the study. mpMRI = multiparametric magnetic resonance imaging; PSsum = Prostate Imaging Reporting and Data System sum score; US = ultrasound; TRUS-GB = transrectal ultrasound-guided biopsy.

Combined FUS-GB with systematic TRUS-GB (arm B)

rate of 0.05. On the basis of a pilot study [14] we estimated that 38% of all

In arm B, a Urostation device (Koelis, La Tronche, France) was used for

enrolled patients would have at least one intraprostatic lesion with a

all biopsies. This device has a real-time time-tracking mechanism;

total score 10, and thus would be randomized after mpMRI.

2.6.2.

however, to confirm that the biopsy needle position was correct, the

Randomized patients were expected to have a true PCa prevalence of

biopsy gun had to be maintained in the prostate for a few seconds after

70%. Assuming a dropout rate of 10% and a false-positive mpMRI rate of

firing. Biopsy in arm B comprised transrectal targeted FUS-GB (two cores

20% (1 – specificity), we planned to enroll approximately 480 patients.

from each lesion) followed by blinded systematic 12-core TRUS-GB using

An interim analysis was planned after enrollment of 50% of patients to

an 18G fully automatic biopsy gun (Bard Medical, Karlsruhe, Germany).

recalculate the sample size. Differences between groups including 95%

As the systematic TRUS-GB was performed by the same operator,

confidence intervals (CIs) were computed using generalized linear

blinding was guaranteed by a standardized biopsy plan that included

models for independent groups and a generalized estimating equation to

lateral and midlobar cores at the base, middle, and apex of each prostate

take repeated testing into account (SAS PROC GENMOD) on the basis of a

lobe. Biopsies were carried out by two urologists (C.A., A.H.) who had

binomial or Poisson distribution. Statistical analysis was performed

performed more than 30 FUS-GB and more than 100 TRUS-GB

using IBM SPSS Statistics 22 for Windows (IBM, Armonk, NY, USA) and

procedures at the start of the study.

SAS (SAS Institute, Cary, NC, USA).

2.7.

Histology

Gleason scoring was performed by a single pathologist (H.E.G.) with more than 30 yr of experience and followed the recommendations of the 2005 consensus conference of the International Society of Urological Pathology. Gleason score and cancer involvement were recorded for each biopsy core.

2.8.

Sample size calculation and statistical analysis

The study was powered at 80% to detect a 60% overall PCa detection rate in arm B compared to a 40% overall rate in arm A, with a type 1 error

3.

Results

3.1.

Patients and mpMRI findings

The study was halted in July 2014 because interim analysis showed that the study would not meet the primary endpoint. At the time of termination, a total of 267 patients were included and consequently underwent mpMRI. Forty-three patients were excluded from randomization because of low-risk PCa (PSsum