165 Preoperative MRI improves the surgical outcome in patients with non-palpable prostate cancer in...

E U R O P E A N U R O L O G Y X X X ( 2 0 1 5 ) X X X – X X X

EURURO-6107; No. of Pages 10

Prostate Cancer

Does Preoperative Magnetic Resonance Imaging Reduce the Rate

of Positive Surgical Margins at Radical Prostatectomy in a

Randomised Clinical Trial?

Erik Rud a,b,*, Eduard Baco c, Dagmar Klotz d, Kristin Rennesund c, Aud Svindland d,b,Viktor Berge c, Eskild Lundeby c, Nicolai Wessel c, Jon-Roar Hoff c, Rolf Eigil Berg c,Lien Diep e, Heidi B. Eggesbø a, Lars Magne Eri c,b

a Department of Radiology and Nuclear Medicine, Oslo University Hospital, Oslo, Norway; b Institute of Clinical Medicine, University of Oslo, Oslo, Norway;c Department of Urology, Oslo University Hospital, Oslo, Norway; d Department of Pathology, Oslo University Hospital, Oslo, Norway;e Department of Biostatistics, Epidemiology and Health Care Economics, Oslo University Hospital, Oslo, Norway

ava i lable at www.sc iencedirect .com

journa l homepage: www.europea nurology.com

Article info

Article history:

Accepted February 27, 2015

Keywords:

RCT

Randomised controlled trial

MRI

Preoperative

Prostate cancer

Surgical margins

Surgical benefit

Abstract

Background: Magnetic resonance imaging (MRI) has the potential to help the surgeontailor radical prostatectomy (RP) more accurately according to the location and extent ofthe tumour and thereby reduce the rate of positive surgical margins (PSMs).Objective: To evaluate the benefit of performing MRI prior to RP.Design, setting, and participants: This single-institution randomised trial included438 patients between December 2009 and June 2012 who were scheduled for robot-assisted laparoscopic prostatectomy. The study was registered (ClinicalTrials.gov iden-tifier NCT01347320).Intervention: Patients were preoperatively randomly assigned to non-MRI or MRI groups.Outcome measurements and statistical analysis: The primary end point was the differ-ence in the PSM rates between the two groups. Secondary end points were the ratesof PSMs in clinical subgroups. Summary statistics were extracted from descriptiveanalyses, chi-square, or Fisher exact test, and logistic regression was used to analysethe data according to the intention-to-treat principle.Results and limitations: A total of 216 patients were randomised to non-MRI; 222 wererandomised to MRI. There were 49 cases (23%) of PSMs in the non-MRI group and43 cases (19%) in the MRI group (p = 0.4). The relative and absolute risk reduction was15% and 4%, respectively. Patients with cT1 constituted 55% of the cohort, in which therate of PSMs was 27% in the non-MRI group and 16% in the MRI group (p = 0.035). Therelative and absolute risk reduction was 41% and 11%, respectively. A limitation wassuboptimal communication between the radiologist and urologist.Conclusions: MRI prior to RP did not reduce the overall risk for PSMs in this patientcohort. However, at subgroup analysis we observed a possible benefit of MRI in patientswith cT1.Patient summary: This study could not demonstrate a definite benefit of performingmagnetic resonance imaging before surgery for all patients. However, there was apossible improved result in patients in which physical examination could not detectthe cancer.

# 2015 European Association of Urology. Published by Elsevier B.V. This is an open

access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/

* Corresponding author. OTel. +47 934 45 684.E-mail address: p.e.rud@m

Please cite this article in press as: Rud E, et al. Does PreoperaSurgical Margins at Radical Prostatectomy in a Randomisedj.eururo.2015.02.039

http://dx.doi.org/10.1016/j.eururo.2015.02.0390302-2838/# 2015 European Association of Urology. Published by ElsevieBY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

by-nc-nd/4.0/).

slo University Hospital, Box 4959, Nydalen, 0424 Oslo, Norway.

edisin.uio.no (E. Rud).

tive Magnetic Resonance Imaging Reduce the Rate of Positive Clinical Trial? Eur Urol (2015), http://dx.doi.org/10.1016/

r B.V. This is an open access article under the CC

http://dx.doi.org/10.1016/j.eururo.2015.02.039

http://creativecommons.org/licenses/by-nc-nd/4.0/


mailto:p�.�[email protected]





E U R O P E A N U R O L O G Y X X X ( 2 0 1 5 ) X X X – X X X2


1. Introduction

The role of magnetic resonance imaging (MRI) before

radical prostatectomy (RP) is still under discussion, and

routine MRI is not considered mandatory [1]. No random-

ised controlled trials (RCTs) have been conducted to

evaluate the role of MRI prior to RP.

Ideally, MRI should detect all significant cancer foci and

give an accurate T classification for optimal surgical

planning. Studies have demonstrated a 80–94% detection

rate for the clinically dominant tumour, the so-called index

tumour [2–4]. The index tumour is associated with

extraprostatic extension (EPE) when relevant, the highest

Gleason score (GS), or the largest volume, in that order of

priority [5–7]. When present, the secondary and tertiary

tumour is the two next highest ranking tumours according

to the same criteria. The sensitivity of MRI for predicting

T3 is limited, ranging from 13% to 95%, depending on patient

selection and the method used [8,9]. The laterality of T3 is

rarely reported, although it is highly relevant when

planning treatment [5,10].

Preoperative detection of T3 disease is clinically impor-

tant because T3 increases the risk of positive surgical

margins (PSMs), that is, cancer present on the surface of the

surgically removed prostate gland. PSMs are an indicator of

nonradical surgery and therefore an important parameter

when assessing the surgical quality of RP [11].

The aim of this RCT was to assess the ability of MRI before

RP to reduce the rate of PSMs.

2. Patients and methods

2.1. Setting

This single-institution RCT was approved by the local regional

committee for medical and health ethics (S-09143c2009/2183) and

registered at ClinicalTrial.gov (NCT01347320). All patients signed a

letter of consent.

2.2. Study design

All patients scheduled for robot-assisted laparoscopic prostatectomy

(RALP) at Oslo University Hospital, Aker, from December 1, 2009, to June

31, 2012, were candidates for the trial. Table 1 displays the baseline data

of the patient cohort. During the study period, 675 patients were

scheduled for RALP; 485 met the inclusion criteria, and 438 (90%) were

included for intention-to-treat analyses (Fig. 1).

2.3. Exclusion criteria

Exclusion criteria were previous relevant MRI of the prostate, contra-

indications to MRI, and hip prosthesis.

2.4. Randomisation

one envelope with another, and the envelope was disposed of after each

allocation. A logbook registered every procedure and was controlled

several times by the responsible statistician. The envelopes were

prepared by the centre of Biostatistics and Epidemiology, Research

Support Service, at Oslo University Hospital and kept accessible only to

E.R., who also performed the randomisation. Patients were scheduled for

MRI within 2 wk.

2.5. Sample size estimation

The rate of PSMs is highly associated with the incidence of pT3, and when

planning the study, the incidence of pT3 was approximately 45%. For that

reason, we expected the rate of PSMs in the non-MRI group to lie

between 35% and 40%. A 20% difference in the rate of PSMs between the

non-MRI and MRI groups was considered clinically relevant. With a

significance level of 0.05, 91–151 in each group was required to provide

80% power with a two-tailed test for two independent proportions,

assuming no dropout.

2.6. Magnetic resonance imaging

The MRI radiologist had 2 yr of experience with prostate imaging at the

start of the study. The radiologist was not systematically blinded to

prostate-specific antigen (PSA), cT stage, and GS. All patients in the MRI

group underwent 1.5-T MRI (Siemens, Erlangen, Germany) and six-

channel body matrix (Siemens). The time between biopsy and MRI was

11 � 7 wk (mean plus or minus standard deviation; range: 0–71) and

between MRI and surgery was 1 � 4 wk (range: 0–25). No patients were

excluded due to haemorrhagic artefacts or poor quality of the MRI.

The sequences included 0.9-mm isotropic three-dimensional T2 and

axial diffusion-weighted images (DWIs) using b50, b1000, and b2000. An

apparent diffusion coefficient map was prepared from b50 and b1000.

Table 1 – Baseline characteristics of 438 patients randomized tothe non–magnetic resonance imaging (MRI) or MRI groups

Non-MRI (n = 216) MRI (n = 222)

Age, yr, mean � SD 63 � 6 62 � 6

PSA, ng/ml, median (IQR) 8.2 (6.2–18.3) 7.8 (4.9–11.3)

n % n %

cT stage

1c 114 53 125 56

2a 41 19 48 22

2b 41 19 25 11

2c 14 6.5 14 6.3

3a 6 2.8 10 4.5

Gleason score in biopsy

6 79 37 70 32

7a 79 37 89 40

7b 20 9.3 27 11

8 37 17 28 12

9 1 0.5 8 3.6

D’Amico risk groups

Low 55 25 57 26

Medium 103 48 114 51

High 58 27 51 23

IQR = interquartile range; MRI = magnetic resonance imaging; PSA =

prostate-specific antigen; SD = standard deviation.

Allocation (1:1) of patients to either non-MRI or MRI was based on a

permuted block randomisation procedure with undisclosed and variable

blocking factor. After consent was received, each patient was designated

a unique, unchangeable sequentially numbered randomisation number,

matching an opaque and sealed envelope, containing the corresponding

intervention code (eg, non-MRI or MRI). It was not possible to replace


The MRI acquisition and postprocessing were previously reported in

detail [12]. We used criteria for tumour detection and staging as

previously specified [2,5,13–16]. Both T2 and DWI were used for

detection and staging.

All patients also underwent axial and coronal T1-weighted imaging

of the pelvis and the lumbosacral spine to evaluate possible lymph node




E U R O P E A N U R O L O G Y X X X ( 2 0 1 5 ) X X X – X X X 3


and skeletal metastases. These findings are not discussed in this paper.

The MRI examination was completed within 20 min.

2.7. Communication with urologists prior to surgery

A schematic tumour map displaying the location and extent of the

tumours (indicated in red), including possible EPE and seminal vesicle

invasion, was drawn manually for every patient (Fig. 2) and shown to

the surgeon before surgery in each case. In addition, the MRI information

was registered in the electronic patient record system. Postoperatively,

the surgeon completed a form, in which he reported whether, and in

what way, MRI had changed the surgical approach.

2.8. Surgery

RALP was limited to four experienced urologists at the department.

The urologists had performed an average of 103 RALP procedures (range:

78–165), and a total of 415 laparoscopic prostatectomies (range:

as a standard bilateral extrafascial dissection or as an extended excision.

The operator registered (1) the surgical plan before knowing the MRI

results, (2) the surgical plan after knowing the MRI, and (3) the means by

which the surgery actually was performed. We did not register whether

MRI resulted in more extensive surgery anteriorly or at the bladder neck.

The planned surgical approach was based on the preoperative

parameters available that included transurethral ultrasound (TRUS),

digital rectal examination (DRE), GS, tumour extent in biopsies, and

patient’s wish. When available, MRI was also taken into consideration.

2.9. Pathologic preparation and examination

Whole-mounted histologic sections were used as the reference standard.

Specimen handling was previously described in detail [2]. The surgical

margins were classified as positive when cancer tissue was present on

the inked surface of the prostate specimen. The sector associated with

each PSM was registered on a tumour map (Fig. 3). If a specific tumour

affected the same sector as the site of the PSM, it was classified as

Number of patients referred to RALP during trial period

(n = 675)

Excluded ( n = 189)

Previously performed MRI (112) Contraindications to MRI (23) Did not wish to participate (10) Treated at other hospital (9) Unknown (35)

Patients included for intention-to-treat analysis (n = 222)

Underwent MRI (203) Did not undergo MRI (19)

Contraindications to MRI (5) Change in surgical date (6) Did not meet (3) Other (6)

MRI (n = 245)

Excluded (n = 25) Withdrawn consent (3) No surgery was performed

Referred to radiation or hormonal therapy (9) Cancelled surgery (10) Other (3)

no MRI ( n = 240)

Patients included for intention-to-treat analysis (n = 216)

Underwent MRI (2) Previous MRI became known Referred to MRI and rebiopsy

Did not undergo MRI (214)

Allocation

RALP

Randomised (n = 485)

Enrolment

Excluded (n = 23) Withdrawn consent (8) No surgery was performed

Skeletal metastases on MRI (4) Referred to radiation therapy (1) Cancelled surgery (9)

Underwent surgery in another country (1)

Fig. 1 – Trial profile using the Consolidated Standards of Reporting Trials 2010 guidelines.MRI = magnetic resonance imaging; RALP = robot-assisted laparoscopic prostatectomy.

296–585) at the start of the study.

Figure 2 illustrates the possible approaches for the surgical dissection.

The operations were classified as bilateral nerve-sparing surgery (BNS),

unilateral nerve-sparing surgery (UNS), or non–nerve-sparing surgery

(NNS). Nerve-sparing surgery included both intra- and interfascial

dissection, as described by Walz et al [10]. The NNS could be performed


responsible for the PSM. Each tumour and sector associated with PSMs

was classified as detected by MRI or not.

2.10. Clinical, radiologic, and pathologic T classification

The staging results of the clinical examination and MRI were previously

reported in detail [5] and are only briefly reported in Table 2.






Preoperative DRE and TRUS, performed by urologists, classified all

patients as having clinically localised (cT1c or cT2a–c) or nonlocalised

disease (cT3a–b). Due to a low number of cT3a–b cases, these patients

were combined with cT2a–c in the statistical analyses. Correspondingly,

MRI (rT) and histopathology (pT) classified all patients as localised (T2)

or nonlocalised (T3).

2.11. Statistical analyses

The rate of PSMs in the non-MRI and MRI groups was the primary end

point. Clinical subgroups were defined according to the T classification,

D’Amico risk groups, GS, and surgical procedures.

The Pearson chi-square or Fisher exact test assessed differences in

the rate of PSMs between the non-MRI and MRI groups and in clinical

subgroups.

The proportions of PSMs and odds ratios (ORs) with 95% confidence

intervals (CIs) were used to describe the differences. Both absolute and

relative differences (risk ratio) are reported.

Differences in the ORs of the clinical subgroups were examined by

the interaction term between the non-MRI and MRI group variable and

3. Results

A total of 438 patients were included in this study, in which

216 were randomised to the non-MRI group and 222 were

randomised to the MRI group (Fig. 1). Table 1 shows that the

groups were well balanced.

3.1. Magnetic resonance imaging tumour detection results

MRI tumour detection rates and staging results of this

cohort were reported in two previous papers and are only

mentioned briefly [2,5]. MRI detected 92% of the index

tumour; it did not detect the index tumour in 16 patients.

However, in 5 of these 16, MRI detected the second largest

tumour with identical GS as the index tumour. In 11 patients

with specimen GS 6 and 7, no tumour was found on MRI.

3.2. Surgical procedures

Basis

Middle

Apex

Nerve-sp

aring exci

sion

Non-nerve-sp

aring exci

sion

Extended exci

sion

NVB

Basis

Right Le�

Middle Ant. fibromuscular stroma

Transi�onzone

Ant. lat.

Post. med. Post. lat.

Apex

Fig. 2 – The 30-sector tumour map used for registration of all tumours at magnetic resonance imaging and histology served as the basis for correlationanalysis. All suspected tumours were marked in red, and any extraprostatic extension was indicated. The map was then presented to the surgeon beforesurgery. The three possible planes of dissection are indicated with stippled lines. Extended and non–nerve-sparing surgery was combined for statisticalanalyses. Figure adapted by Kari C. Toverud from Girouin et al. Eur Radiol. 2007;17:1498–509, with permission from Springer. # Kari C. Toverud.Ant. = anterior; lat. = lateral; med. = medial; NVB = neurovascular bundle; Post. = posterior.

the clinical subgroup variable using logistic regression. All analyses

were performed according to the intention-to-treat principle. Two-sided

p values were given and not corrected for multiple testing.

IBM SPSS Statistics v.22 (IBM Corp., Armonk NY, USA) and MedCalc

statistical software v.14.8.1 (MedCalc Software, Ostend, Belgium) were

used for statistical analyses.


The surgeons reported that MRI altered the surgical procedure

on one or both sides in 59 patients (27%), of which 30 (51%)

were cT1, 29 (49%) were cT2, and 49 (83%) were rT3. All

changes were in the direction of a more radical excision. There

was no difference in mean age, PSA, and GS in biopsy of






patients with adjusted surgery compared with those with

nonadjusted surgery. BNS was performed in 35 of 90 patients

(39%) classified as rT2 versus 15 of 110 (14%) for rT3

(p < 0.001).

The overall rate of BNS, UNS, and NNS was 30%, 34%, and

35%, respectively. The absolute rate of BNS was 6.7% (95% CI,

�2 to 16; p = 0.12) lower in the MRI group compared with

the non-MRI group (Table 3). This difference was mainly

due to a 13% (95% CI, 1–25; p = 0.024) lower rate of BNS in

patients with cT2–3 (Fig. 4).

3.3. Overall risk of positive surgical margins

There were 49 cases (23%) of PSMs in the non-MRI group

and 43 (19%) in the MRI group (Table 4). The relative and

absolute reduction was 15% (0.85; 95% CI, 0.6–1.2) and 4%

(95% CI, �4 to 10; p = 0.4), respectively. The number needed

to treat was 25, and two extra nerve resections (1.7%) were

performed for each PSM avoided.

3.4. Risk of positive surgical margins in patient subgroups

Basic Basic

Right Right

Ant. fibromuscularstroma

Ant. fibromuscularstroma

Transi�onzone

Transi�onzone

Ant. lat.

Post. med. Post. lat. Post. lat.Post. med.

Ant. lat.

Le� Le�

Middle Middle

Apex Apex

No MRI MRI

Fig. 3 – Distribution of positive surgical margin (PSM) sites is marked with small circles. The PSM sites ventral to the basal urethra represent thebladder neck. Both the apex and the base are at high risk for PSMs. Figure adapted by Kari C. Toverud from Girouin et al. Eur Radiol. 2007;17:1498–509, with permission from Springer. # Kari C. Toverud.Ant. = anterior; lat. = lateral; med. = medial; MRI = magnetic resonance imaging; Post. = posterior.

Table 2 – The performance of magnetic resonance imaging anddigital rectal examination/transrectal ultrasound for detecting pT3disease

Sensitivity, % (95% CI) Specificity, % (95% CI)

MRI

All patients 73 (63–81) 65 (54–74)

cT1 56 (41–70) 70 (57–81)

cT2–3 88 (76–95) 55 (36–72)

DRE/TRUS 4 (2–8) 98 (95–99)

CI = confidence interval; DRE/TRUS = digital rectal examination/transurethral

ultrasound; MRI = magnetic resonance imaging.


MRI did not reduce the rates of PSMs in cT2–3, whereas

in cT1, the relative and absolute reduction was 41% (0.59;

95% CI, 0.4–0.9) and 11% (95% CI, 2–22; p = 0.035),

respectively (Tables 3 and 4). There was no benefit of

MRI in any other subgroups.






Rates of PSMs in patients with adjusted and nonadjusted

surgery were 25% and 17%, respectively (8% difference;

95% CI, �4 to 22; p = 0.17). Most patients with PSMs

underwent UNS or NNS; 36 of 49 (73%) in the non-MRI

group and 36 of 43 (84%) in the MRI group (11% difference;

95% CI, �8 to 29; p = 0.2). For patients classified as rT3, the

rate of PSMs was 25%; 89% of these patients underwent

either UNS or NNS (Fig. 5).

3.5. Distribution of positive surgical margins and the

associated tumour

There were 126 sites of PSMs in 92 specimens (Fig. 6). Sixty

patients had a single PSM site; 32 had more than one. The

index tumour was the responsible tumour in 89% of all cases

4. Discussion

The present study is the first RCT to evaluate the benefit of

MRI prior to RALP. We could not demonstrate a significant

reduction in the rate of PSMs for either the whole cohort or

patients with pT3. However, we found a 41% lower rate of

PSMs in patients with cT1 (p = 0.035), although this was not

statistically significant in the context of multiple testing.

The rate of PSMs varies significantly in the literature, and

our results are within acceptable limits [17,18].

4.1. Effect of magnetic resonance imaging on surgical approach

MRI changed the surgical strategy in 29%, all in the direction

of more radical excision. Other studies have reported

Table 3 – Positive surgical margin rate with respect to surgical procedure, cT stage, and Gleason score in biopsy

Non-MRI (n = 216) MRI (n = 222)

Total PSM Total PSM

n % n % n % n % p p*

Bilateral nerve-sparing surgery

DRE 0.15

cT1 45 49 11 24 46 51 5 11 0.089

cT2–3 28 67 2 7.1 14 33 2 14 0.5

Gleason score in biopsy 0.6

6 34 53 7 21 30 47 2 7 0.11

7a 26 54 3 12 22 46 5 23 0.3

7b 8 62 2 25 5 38 0 0 0.2

8 5 63 1 20 3 34 0 0 0.4

9 NA NA NA 0 NA NA NA

All patients 73 33 13 18 60 25 7 12 0.3

Unilateral nerve-sparing surgery

DRE 0.9

cT1 29 42 6 21 40 58 7 16 0.7

cT2–3 42 51 9 21 40 49 7 19 0.7


6 29 58 4 14 23 42 1 4 0.3

7a 26 43 7 27 35 57 8 23 0.7

7b 6 30 1 17 14 70 3 17 0.8

8 10 56 3 30 8 44 2 25 0.8

9

All patients 71 47 15 21 80 53 14 18 0.6

Non–nerve-sparing surgery

DRE 0.086

cT1 40 51 14 35 39 49 8 19 0.15

cT2–3 32 43 7 23 43 57 14 30 0.3


6 16 48 3 19 17 52 2 12 0.6

7a 27 46 7 26 32 54 10 30 0.7

7b 6 40 3 50 9 60 5 50 0.8

8 22 58 8 36 16 42 3 19 0.2

9 1 11 0 0 8 89 2 25 0.6

All patients 72 47 21 29 82 53 22 27 0.7

DRE = digital rectal examination; MRI = magnetic resonance imaging; NA = not available; PSM = positive surgical margins.* The p value for interaction.

of PSMs, and 60% were associated with EPE at the same

location.

MRI detected 92% of all tumours related to PSMs. The

base and apex were at higher risk for PSMs (Fig. 3). The risk

of ventral PSMs in the apex was higher in the non-MRI

group (12%; 95% CI, �3 to 23; p = 0.072).


that MRI affected the surgery in a radical direction in only

33–39% [19–21]. None of these studies had a control arm

without MRI, making comparison difficult.

Most of the patients with PSMs had undergone UNS or

NNS in both the non-MRI and MRI group. Also, when MRI

suggested T3, 89% of those with PSMs had undergone either






UNS or NNS (Fig. 5). This suggests that even more extended

excisions are required to reduce the rate of PSMs.

4.2. Rate of positive surgical margins in patients with cT1

Surprisingly, the greatest benefit of MRI was seen in patients

with cT1. This was unexpected because the sensitivity of MRI

for predicting pT3 in this group was poor (Table 2), and

surgical procedures were not significantly different between

the non-MRI and MRI group (Fig. 4). This suggests that the

primary value of MRI was to provide the surgeon with

accurate information about tumour location and extension,

whereas staging was of secondary importance.

4.3. Possible mechanisms for improved outcome in cT1

This study cannot explain the exact mechanisms for the

reduced rate of PSMs in patients with cT1. However, one

possibility may be excellent tumour visualisation on MRI,

also in patients with cT1, whereas DRE and TRUS are normal.

This probably favours selection of the proper surgical

procedure, and it may have prevented dissection too close

to the index tumour or changed the surgery in ways not

registered in this study. For instance, in case of a ventral

cancer with extraprostatic disease, it is possible to perform a

combination of BNS and extended excision ventrally.

4.4. Importance of the index tumour and positive surgical

margins site

The index tumour was the source of PSMs in 89% of the cases

(Fig. 6). This illustrates the relevance of locating the index

tumour before surgery and is in accordance with a

2012 study [22]. However, despite the fact that MRI

detected 98% of these tumours, the rate of PSMs did not

improve for the whole cohort. If the MRI findings are

Fig. 4 – In patients with cT1, the relative risk of positive surgical margins(PSMs) was 41% lower in the magnetic resonance imaging (MRI) group,and there was a reduction in PSMs regardless of the surgical procedure.There was no significant difference in the surgery between the twogroups. Unexpectedly, in patients with cT2–3, the overall PSM rateincreased by 6% (95% confidence interval, S6 to 18; p = 0.4) in the MRIgroup, despite the fact that unilateral nerve-sparing surgery and non–nerve-sparing surgery were performed 13% more frequently in the MRIgroup.BNS = bilateral nerve-sparing surgery; MRI = magnetic resonanceimaging; NNS = non–nerve-sparing surgery; PSM = positive surgicalmargins; UNS = unilateral nerve-sparing surgery.

Table 4 – Rate of positive surgical margins in the non–magnetic resonance imaging (MRI) and MRI groups

Non-MRI (n = 216) MRI (n = 222)

Total, n PSM, n % Total, n PSM, n % p OR 95% CI *p

All patients 216 49 23 222 43 19 0.4 0.81 0.5–1.3

DRE 0.029

cT1 114 31 27 125 20 16 0.035 0.51 0.3–0.9

cT2–3 102 18 18 97 23 24 0.3 1.53 0.7–2.9

D’Amico risk classification 0.8* 0.8

Low 55 7 13 57 3 5.3 0.38 0.1–1.6

Medium 103 24 23 114 25 22 0.93 0.5–1.7

High 58 18 31 51 15 29 0.41 0.4–2.1

pT stage 0.7* 0.7

2 111 9 8.1 103 5 4.9 0.58 0.2–1.8

3 105 40 38 119 38 32 0.76 0.4–1.3

Gleason score in specimen 0.8* 0.8

6 59 9 15 57 7 12 0.78 0.3–2.3

7a 83 17 20 87 14 16 0.75 0.3–1.6

7b 44 11 25 48 14 29 1.21 0.5–3.1

8 23 9 39 22 6 27 0.58 0.2–2.1

9 7 3 43 8 2 25 0.44 0.1–4.0

CI = confidence interval; DRE = digital rectal examination; MRI = magnetic resonance imaging; OR = odds ratio.

OR <1 indicates a reduced risk of positive surgical margins; OR >1 indicates an increased risk.* The p value for interaction: A significant interaction p value indicates that the ORs of the subgroups are significantly different.

Please cite this article in press as: Rud E, et al. Does Preoperative Magnetic Resonance Imaging Reduce the Rate of PositiveSurgical Margins at Radical Prostatectomy in a Randomised Clinical Trial? Eur Urol (2015), http://dx.doi.org/10.1016/j.eururo.2015.02.039





included in the planning of surgical approach to a larger

extent, it represents a considerable potential for improve-

ment in the future.

The MRI group experienced a lower rate of PSMs in the

ventral apex (Fig. 3), although the prognostic value of this is

European Society of Urogenital Radiology recommendation

[25]. Nevertheless, we achieved similar detection and

staging results as those who have used the current

recommendation [2,4,5,26,27]. This may indicate that a

simple protocol is truly sufficient, although no studies have

Fig. 6 – The associated tumour for each case of positive surgical margins (PSMs) and whether magnetic resonance imaging (MRI) detected it or not.Also shown is whether the PSMs were associated with extraprostatic extension in the same area. MRI identified most tumours and sectors associatedwith PSMs, suggesting that the surgeons did not sufficiently consider the MRI findings.EPE = extraprostatic extension; MRI = magnetic resonance imaging; PSM = positive surgical margins.

Fig. 5 – Bilateral nerve-sparing surgery was chosen in 15 of 110 patients (14%) when magnetic resonance imaging (MRI) suggested T3 disease. Forunknown reasons, the surgeon chose not to consider the MRI findings. Six of the patients turned out to have pT3 in which two had positive surgicalmargins (PSMs); the remaining nine were pT2 and one had PSM. In 95 of 110 patients (86%), the surgical procedure harmonised to a larger extent withthe MRI finding; however, 24 of 27 patients (89%) with PSMs were found in this group. This illustrates that the degree of non–nerve-sparing excisionwas not sufficient.BNS = bilateral nerve-sparing surgery; MRI = magnetic resonance imaging; NNS = non–nerve-sparing surgery; NSM = negative surgical margins;PSM = positive surgical margins; UNS = unilateral nerve-sparing surgery.

controversial [23,24].

4.5. Magnetic resonance imaging protocol

This study used a simple MRI protocol, without contrast

medium and endorectal coil, compared with the current


compared the different protocols on the same cohort, and

direct comparison of results is difficult. Alternatively, our

high rate of large tumours with pT3 resulted in extraordi-

nary results.

From a surgical point of view, it is critical to stress that

MRI tumour detection is highly sensitive in most studies;






staging remains the weakest element. For this reason,

urologists are encouraged to consider tumour location more

than rT classification when deciding surgical strategy.

4.6. Limitations

This study has some possible limitations. First, the study

was underpowered to demonstrate a significant effect of

MRI because the observed rate of PSMs was approximately

half of the expected rate. Because all subgroup analyses

were not defined a priori, one should interpret the findings

in cT1 with caution and instead conduct future studies.

Second, more patients were referred to radiation therapy in

the non-MRI group after randomisation. This may be an

indicator of more advanced disease, likely to result in PSMs.

Consequently, it may have reduced the effect of MRI. Third,

the radiologist was not fully blinded to clinical data. This

might have affected the MRI results and consequently

patient handling. Fourth, there was suboptimal communi-

cation between the radiologist and the urologist during this

study. This is probably the most important reason why the

difference in the rate of PSMs between the two groups was

not larger. We believe the radiologists and the urologist

should coinvestigate MRI images prior to surgery because

only reading a written report or looking at a schematic

drawing is suboptimal when planning surgery.

5. Conclusions

We showed that the index tumour, usually detected by MRI,

is the most frequent cause of PSMs. We did not find an

overall reduction in the rate of PSMs in patients randomised

to MRI, although subgroup analysis revealed a possible

benefit in patients with cT1. Our study suggests more

radical resection should be performed in the index tumour

region and that future studies should focus on MRI-guided

RP. In the MRI group, slightly more nerve resections were

performed, which can be considered a potential negative

effect of MRI. More studies are needed to evaluate the role of

MRI before prostatectomy.

Author contributions: Erik Rud had full access to all the data in the study

and takes responsibility for the integrity of the data and the accuracy of

the data analysis.

Study concept and design: Rud, Eggesbø, Eri.

Acquisition of data: Rud, Klotz, Svindland, Wessel, Berg, Berge, Hoff,

Rennesund, Lundeby.

Analysis and interpretation of data: Rud, Eri, Baco.

Drafting of the manuscript: Rud, Eri, Baco.

Critical revision of the manuscript for important intellectual content: Rud,

Eri, Baco, Eggesbø.

Statistical analysis: Rud, Diep.

Obtaining funding: None.

(eg, employment/affiliation, grants or funding, consultancies, honoraria,

stock ownership or options, expert testimony, royalties, or patents filed,

received, or pending), are the following: None.

Funding/Support and role of the sponsor: None.

References

[1] Heidenreich A, Bastian PJ, Bellmunt J, et al. EAU guidelines on

prostate cancer. Part 1: screening, diagnosis, and local treatment

with curative intent. Update 2013. Eur Urol 2014;65:124–37.

[2] Rud E, Klotz D, Rennesund K, et al. Detection of the index tumor and

tumor volume in prostate cancer using T2w and DW MRI alone. BJU

Int 2014;114:E32–42.

[3] Turkbey B, Mani H, Aras O, et al. Correlation of magnetic resonance

imaging tumor volume with histopathology. J Urol 2012;188:

1157–63.

[4] Le JD, Tan N, Shkolyar E, et al. Multifocality and prostate cancer

detection by multiparametric magnetic resonance imaging:

correlation with whole-mount histopathology. Eur Urol 2015;67:

569–76.

[5] Rud E, Klotz D, Rennesund K, et al. Preoperative magnetic resonance

imaging for detecting uni- and bilateral extraprostatic disease in

patients with prostate cancer. World J Urol. In press.

[6] van der Kwast TH, Amin MB, Billis A, et al. International Society of

Urological Pathology (ISUP) Consensus Conference on Handling

and Staging of Radical Prostatectomy Specimens. Working group

2: T2 substaging and prostate cancer volume. Mod Pathol 2010;

24:16–25.

[7] Baco E, Ukimura O, Rud E, et al. Magnetic resonance imaging-

transectal ultrasound image-fusion biopsies accurately character-

ize the index tumor: correlation with step-sectioned radical

prostatectomy specimens in 135 patients. Eur Urol 2015;67:

787–94.

[8] Yacoub JH, Oto A, Miller FH. MR imaging of the prostate. Radiol Clin

North Am 2014;52:811–37.

[9] Hricak H, Choyke PL, Eberhardt SC, Leibel SA, Scardino PT. Imaging

prostate cancer: a multidisciplinary perspective. Radiology 2007;

243:28–53.

[10] Walz J, Burnett AL, Costello AJ, et al. A critical analysis of the current

knowledge of surgical anatomy related to optimization of cancer

control and preservation of continence and erection in candidates

for radical prostatectomy. Eur Urol 2010;57:179–92.

[11] Boorjian SA, Karnes RJ, Crispen PL, et al. The impact of positive

surgical margins on mortality following radical prostatectomy

during the prostate specific antigen era. J Urol 2010;183:1003–9.

[12] Rud E, Baco E, Eggesbø HB. MRI and ultrasound-guided prostate

biopsy using soft image fusion. Anticancer Res 2012;32:3383–9.

[13] Sala E, Akin O, Moskowitz CS, et al. Endorectal MR imaging in the

evaluation of seminal vesicle invasion: diagnostic accuracy and

multivariate feature analysis. Radiology 2006;238:929–37.

[14] Yu KK, Hricak H, Alagappan R, Chernoff DM, Bacchetti P, Zaloudek

CJ. Detection of extracapsular extension of prostate carcinoma with

endorectal and phased-array coil MR imaging: multivariate feature

analysis. Radiology 1997;202:697–702.

[15] Outwater EK, Petersen RO, Siegelman ES, Gomella LG, Chernesky

CE, Mitchell DG. Prostate carcinoma: assessment of diagnostic

criteria for capsular penetration on endorectal coil MR images.

Radiology 1994;193:333–9.
Administrative, technical, or material support: None.
Supervision: Rud, Eggesbø.

Other (specify): None.

Financial disclosures: Erik Rud certifies that all conflicts of interest,

including specific financial interests and relationships and affiliations

relevant to the subject matter or materials discussed in the manuscript


[16] Kim CK, Choi D, Park BK, Kwon GY, Lim HK. Diffusion-weighted MR

imaging for the evaluation of seminal vesicle invasion in prostate

cancer: initial results. J Magn Reson Imaging 2008;28:963–9.

[17] Yossepowitch O, Briganti A, Eastham JA, et al. Positive surgical

margins after radical prostatectomy: a systematic review and

contemporary update. Eur Urol 2014;65:303–13.


http://refhub.elsevier.com/S0302-2838(15)00204-3/sbref0140


























































[18] Novara G, Ficarra V, Mocellin S, et al. Systematic review and meta-

analysis of studies reporting oncologic outcome after robot-

assisted radical prostatectomy. Eur Urol 2012;62:382–404.

[19] Hricak H, Wang L, Wei DC, et al. The role of preoperative endorectal

magnetic resonance imaging in the decision regarding whether to

preserve or resect neurovascular bundles during radical retropubic

prostatectomy. Cancer 2004;100:2655–63.

[20] McClure TD, Margolis DJA, Reiter RE, et al. Use of MR imaging

to determine preservation of the neurovascular bundles at robot-

ic-assisted laparoscopic prostatectomy. Radiology 2012;262:

874–83.

[21] Labanaris A, Zugor V, Takriti S, et al. The role of conventional and

functional endorectal magnetic resonance imaging in the decision

of whether to preserve or resect the neurovascular bundles during

radical retropubic prostatectomy. Scand J Urol Nephrol 2009;43:

25–31.

[22] Karavitakis M, Ahmed HU, Abel PD, Hazell S, Winkler MH. Margin

status after laparoscopic radical prostatectomy and the index

lesion: implications for preoperative evaluation of tumor focality

in prostate cancer. J Endourol 2012;26:503–8.

[23] Dev HS, Wiklund P, Patel V, et al. Surgical margin length and

location affect recurrence rates after robotic prostatectomy. Urol

Oncol. In press. http://dx.doi.org/10.1016/j.urolonc.2014.11.005

[24] Fontenot PA, Mansour AM. Reporting positive surgical margins

after radical prostatectomy: time for standardization. BJU Int 2013;

111:E290–9.

[25] Barentsz JO, Richenberg J, Clements R, et al. ESUR prostate MR

guidelines 2012. Eur Radiol 2012;22:746–57.

[26] Somford DM, Hamoen EH, Futterer JJ, et al. The predictive value of

endorectal 3 Tesla multiparametric magnetic resonance imaging

for extraprostatic extension in patients with low, intermediate and

high risk prostate cancer. J Urol 2013;190:1728–34.

[27] Brajtbord JS, Lavery HJ, Nabizada-Pace F, Senaratne P, Samadi DB.

Endorectal magnetic resonance imaging has limited clinical ability

to preoperatively predict pT3 prostate cancer. BJU Int 2011;107:

1419–24.

Please cite this article in press as: Rud E, et al. Does Preoperative Magnetic Resonance Imaging Reduce the Rate of PositiveSurgical Margins at Radical Prostatectomy in a Randomised Clinical Trial? Eur Urol (2015), http://dx.doi.org/10.1016/j.eururo.2015.02.039





















http://dx.doi.org/10.1016/j.urolonc.2014.11.005
















165 Preoperative MRI improves the surgical outcome in patients with non-palpable prostate cancer in...

Documents

Transcript of 165 Preoperative MRI improves the surgical outcome in patients with non-palpable prostate cancer in...