URO ONCO

Welcome, this website is intended for all international healthcare professionals in uro-oncology. By clicking the link below you are declaring and confirming that you are a healthcare professional.

You are here

What Is the Negative Predictive Value of Multiparametric Magnetic Resonance Imaging in Excluding Prostate Cancer at Biopsy? A Systematic Review and Meta-analysis from the European Association of Urology Prostate Cancer Guidelines Panel

European Urology, Volume 72, Issue 2, August 2017, Pages 250-266

Abstract

Context

It remains unclear whether patients with a suspicion of prostate cancer (PCa) and negative multiparametric magnetic resonance imaging (mpMRI) can safely obviate prostate biopsy.

Objective

To systematically review the literature assessing the negative predictive value (NPV) of mpMRI in patients with a suspicion of PCa.

Evidence acquisition

The Embase, Medline, and Cochrane databases were searched up to February 2016. Studies reporting prebiopsy mpMRI results using transrectal or transperineal biopsy as a reference standard were included. We further selected for meta-analysis studies with at least 10-core biopsies as the reference standard, mpMRI comprising at least T2-weighted and diffusion-weighted imaging, positive mpMRI defined as a Prostate Imaging Reporting Data System/Likert score of ≥3/5 or ≥4/5, and results reported at patient level for the detection of overall PCa or clinically significant PCa (csPCa) defined as Gleason ≥7 cancer.

Evidence synthesis

A total of 48 studies (9613 patients) were eligible for inclusion. At patient level, the median prevalence was 50.4% (interquartile range [IQR], 36.4–57.7%) for overall cancer and 32.9% (IQR, 28.1–37.2%) for csPCa. The median mpMRI NPV was 82.4% (IQR, 69.0–92.4%) for overall cancer and 88.1% (IQR, 85.7–92.3) for csPCa. NPV significantly decreased when cancer prevalence increased, for overall cancer ( r = –0.64, p < 0.0001) and csPCa ( r = –0.75, p = 0.032). Eight studies fulfilled the inclusion criteria for meta-analysis. Seven reported results for overall PCa. When the overall PCa prevalence increased from 30% to 60%, the combined NPV estimates decreased from 88% (95% confidence interval [95% CI], 77–99%) to 67% (95% CI, 56–79%) for a cut-off score of 3/5. Only one study selected for meta-analysis reported results for Gleason ≥7 cancers, with a positive biopsy rate of 29.3%. The corresponding NPV for a cut-off score of ≥3/5 was 87.9%.

Conclusions

The NPV of mpMRI varied greatly depending on study design, cancer prevalence, and definitions of positive mpMRI and csPCa. As cancer prevalence was highly variable among series, risk stratification of patients should be the initial step before considering prebiopsy mpMRI and defining those in whom biopsy may be omitted when the mpMRI is negative.

Patient summary

This systematic review examined if multiparametric magnetic resonance imaging (MRI) scan can be used to reliably predict the absence of prostate cancer in patients suspected of having prostate cancer, thereby avoiding a prostate biopsy. The results suggest that whilst it is a promising tool, it is not accurate enough to replace prostate biopsy in such patients, mainly because its accuracy is variable and influenced by the prostate cancer risk. However, its performance can be enhanced if there were more accurate ways of determining the risk of having prostate cancer. When such tools are available, it should be possible to use an MRI scan to avoid biopsy in patients at a low risk of prostate cancer.

Take Home Message

Prostate cancer prevalence was highly variable among patients referred to prebiopsy multiparametric magnetic resonance imaging (mpMRI). As the negative predictive value depends on the prevalence, patients submitted to mpMRI must be reliably risk stratified before defining who may omit prostate biopsy when mpMRI is negative.

Keywords: Prostate cancer, Multiparametric magnetic resonance imaging, Prostate biopsy, Risk stratification.

1 Introduction

A correlation with radical prostatectomy specimens has demonstrated that multiparametric magnetic resonance imaging (mpMRI) has excellent sensitivity in detecting prostate cancer (PCa) with a Gleason score of ≥7 1 2 3 . As a result, prostate mpMRI is increasingly used in patients with a suspicion of PCa to localise abnormal areas before biopsy. A large body of literature has shown that targeted biopsies of suspicious lesions seen on mpMRI (TBx) improved the detection of clinically significant PCa (csPCa), at least in the repeat biopsy setting 4 5 6 . As a result, it is now recommended that an mpMRI is performed before repeat biopsy to allow TBx of suspicious lesions in addition to standard biopsies [7] .

Some authors have recently suggested that, besides improving csPCa detection, mpMRI could also be used as a triage test so that patients with negative mpMRI findings could obviate biopsy. Such a strategy remains highly controversial [8] and depends upon the negative predictive value (NPV) of mpMRI. Therefore, the European Association of Urology Prostate Cancer Guidelines Panel undertook this systematic review and meta-analysis to assess the NPV of mpMRI in patients with a suspicion of PCa and, thus, its potential role in eliminating unnecessary prostate biopsy.

2 Evidence acquisition

2.1 Objective

Our primary aim was to systematically evaluate the performance of negative prebiopsy prostate mpMRI in predicting a negative biopsy result for overall PCa and csPCa in biopsy-naïve men and in men with previously negative biopsies. A further objective was to explore and define factors that may contribute to relevant thresholds in order to provide guidance for future studies.

2.2 Data acquisition and search strategy

The review was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement [9] . The review protocol was published in PROSPERO database ( http://www.crd.york.ac.uk/PROSPERO; registration number CRD42015021929). Databases searched included the Embase and OVID Medline databases, the Cochrane database of systematic reviews, and the Cochrane Central Register for Clinical Trials, covering from January 1, 2000 to February 13, 2016. Systematic or standard prostate biopsies were used as reference standards, with positive or negative cases of PCa being determined by histopathological examination. The detailed search strategy is presented in Supplement 1.

2.3 Inclusion and exclusion criteria

Included studies focused on men who were assessed for suspected PCa by mpMRI before undergoing prostate biopsy. Studies enrolling both biopsy-naïve men and men who had undergone previous negative biopsies were included. Prebiopsy prostate mpMRI was considered the index test and comprised T2-weighted imaging (T2WI) and at least one functional imaging technique (diffusion-weighted imaging [DWI], dynamic contrast-enhanced imaging [DCEI], or magnetic resonance spectroscopic imaging [MRSI]). For inclusion, studies had to report on false negatives and true negatives, in order to calculate NPV (ie, results of systematic/standard prostate biopsies when the mpMRI was negative). When available, false positive and true positive findings were also noted to calculate the positive predictive value (PPV) and the cancer prevalence. There was restriction neither on the biopsy technique (transrectal or transperineal) nor on the number of biopsy cores. Studies using radical prostatectomy specimens as reference standards were excluded, as were studies evaluating men with histologically proven PCa. Studies with less than 50 participants were excluded. No language restrictions were applied.

2.4 Data collection and data extraction

Two reviewers (P.C.M. and T.V.D.B.) independently screened all abstracts and full-text articles for eligibility. Disagreement was resolved by consensus or reference to an independent third party (L.M.). All screening was performed using a predefined eligibility form.

Using a data extraction form developed a priori, the same two reviewers independently extracted data concerning study methodology, patient characteristics, technical characteristics of the MR scanners, mpMRI protocol, mpMRI scoring system, definition of positive mpMRI, biopsy protocol, and definition of csPCa. Any discrepancies concerning data extraction were resolved by consensus or reference to an independent arbiter (O.R. or T.B.L.).

2.5 Assessment of risk of bias

To assess the risk of bias (RoB), all included reports were reviewed using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool for diagnostic accuracy studies [10] .

2.6 Data synthesis and analysis

Outcome data regarding false negative and true negative values of mpMRI before prostate biopsy were recorded as reported by authors. When not available, data were indirectly derived from specificity, sensitivity, and prevalence values reported by authors using an online Bayesian statistics calculator ( http://www.medcalc.com/bayes.html ). Descriptive statistics were used to summarise baseline characteristics and outcomes, including median and interquartile range (IQR) for estimates of NPV across studies. A correlation between mpMRI NPV and a positive biopsy rate was established using the Pearson's correlation coefficient.

A meta-analysis was undertaken to calculate pooled NPV and PPV. To ensure appropriate clinical homogeneity of the studies included in the meta-analysis, we selected only the studies enrolling biopsy-naïve patients and/or patients with a history of negative biopsy, and fulfilling the following criteria that were defined a priori: (1) reference standard consisting of prostate biopsy with at least 10 samples on all patients; (2) mpMRI protocol comprising at least T2WI and DWI; (3) mpMRI results presented as a five-level score, using a subjective Likert scale or the Prostate Imaging Reporting Data System (PI-RADS) score [11] ; (4) definition of positive mpMRI as a score ≥3/5 or ≥4/5; and (5) results reported on a per patient basis. In addition, only studies defining csPCa as Gleason ≥7 cancers were selected for the meta-analysis assessing the mpMRI NPV for csPCa. A bivariate random-effects approach was employed using the Midas package in Stata 12 (StataCorp LP, College Station, TX, USA). Since the NPV decreases and the PPV increases as the prevalence increases, post-test probability estimates of NPV and PPV were reported for the given values of the prevalence based on Bayes’ theorem.

For other studies not included in the meta-analysis based on the criteria described above, a narrative synthesis of the data was performed. To explore and define clinical heterogeneity, subgroups were analysed at patient level based on the following variables: biopsy-naïve versus previous negative biopsy; patients with positive versus negative digital rectal examination (DRE); mpMRI performed with an endorectal versus without an endorectal coil; transrectal ultrasound (TRUS) versus template transperineal (TTP) biopsy approach; and ≤16 cores versus >16 cores as the reference standard. Studies reporting mpMRI NPV for patients with a prostate-specific antigen (PSA) level of ≤10 ng/ml were also reported separately.

3 Evidence synthesis

3.1 Quantity of evidence identified

The study selection process is depicted in the PRISMA flow diagram ( Fig. 1 ). A total of 2980 abstracts were retrieved. After abstract screening and removal of duplicates, 240 articles were eligible for full text screening, of which 48 studies were eligible for inclusion 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 .

gr1
 

Fig. 1
Preferred Reporting Items for Systematic Reviews and Meta-analysis flow chart. Bx = biopsy; CAD = computer-aided diagnosis; mpMRI = multiparametric magnetic resonance imaging; RP = radical prostatectomy; T2WI = T2-weighted imaging.

 

3.2 Quality of studies

Out of the 48 included studies, 42 were single-centre and six were multicentre studies. Thirty-four studies were prospective and six were retrospective, whilst the design of the rest was unclear. RoB assessment using QUADAS-2 was performed for each of the individual studies ( Fig. 2 A and B). Overall, the RoB was highly heterogeneous across studies for all criteria, except for the reference standard domain, in which RoB was low in most studies.

gr2
 

Fig. 2
(A) Assessment of the risk of bias for included studies. (B) Risk of bias summary graph. RoB = risk of bias.

 

3.3 Characteristics of studies

The 48 studies comprised a total of 9613 men who underwent prostate mpMRI followed by biopsy. The study and patient baseline characteristics are presented in Table 1 . The patient population consisted of biopsy-naive men in nine studies, men with at least one previous negative biopsy in 16 studies, and both biopsy-naïve men and men with a history of previous negative biopsy in nine studies. In 14 studies, the biopsy history of the patients was unclear.

Study Study design Period Population Pt Nb Mean/median* age (yr) Mean/median* PSA (ng/ml) Mean/median* prostate volume (cm 3 ) Magnetic field strength MR pulse sequences Endorectal coil MR score Definition of positive MRI Reference standard
Hauth (2015) [25] Prospective 2011–2013 FB 94 64 (43–83) 9 (3–31) 51 (17–140) 1.5 T T2WI/DWI/DCEI/MRSI No PI-RADS v1 ≥3 11–13-core TRUS SBx + TBx
Lamb (2015) [36] Retrospective 2013–2013 FB & PNB 173 NR NR NR 1.5 T T2WI/DWI NR NR NR 12-core TRUS SBx
Brock (2015) [15] Prospective 2013–2014 PNB 168 64* (IQR 59–70) 9.2* (IQR 6.7–13.4) 55.4* (IQR 42–80) 3 T T2WI/DWI/DCEI No PI-RADS v1 NR 12/24-core TRUS SBx + TBx
Grenabo Bergdahl (2016) [22] Prospective 2013–2014 FB & PNB 83 69.3* (IQR 69–69.6) 1.6* (IQR 0.9–2.7) NR 3 T T2WI/DWI/DCEI/MRSI No PI-RADS v1 ≥3 10-core TRUS SBx + TBx
Wang (2015) [58] Prospective 2002–2009 FB & PNB 586 70 (26–91) 11.11* (0.02–9794) NR 1.5 T T2WI/DWI/DCEI/MRSI Yes PI-RADS v1 ≥3 TRUS SBx
Pepe (2015) [41] Prospective 2011–2014 PNB 100 64* 8.6* (4.2–10) NR 3 T T2WI/DWI/DCEI/MRSI No PI-RADS v1 ≥4 TPBx + TBx
Panebianco (2015) [39] Prospective 2011–2014 FB (61.62%) & PNB (38.34%) 925
(1140 total cohort)
NR NR NR 3 T T2WI/DWI/DCEI Yes PI-RADS v1 NR 14-core TRUS SBx;
45-core sat TPBx + TBx
Radtke (2015) [46] Prospective 2013–2013 FB (63.3%) & PNB (36.7%) 294 64* (60–71) 7.3 (6.0) 47 (37.5) 3 T T2WI/DWI/DCEI No PI-RADS v1 PI-RADS ≥2;
PI-RADS ≥3;
PI-RADS ≥4;
PI-RADS = 5;
24-core TPBx + TBx
Itatani (2014) [27] Retrospective 2004–2007 NR 193 68.9 (8.4); 70* (47–89) 11.8 (15.9); 7.9* (1.22–159) NR 1.5 T T2WI/DWI/DCEI No 1–5 scale (Likert) ≥3 12–14-core TRUS SBx
Porpiglia (2014) [44] Prospective 2011–2013 PNB 170 65* (60–70) 6.9* (5.2–9.8) 42* (36–48) 1.5 T T2WI/DWI/DCEI Yes Dichotomous Positive: at least 2/3 MR seq. with suspicious findings 18–24-core TRUS SBx (volume dependent)
Thompson (2014) [56] Prospective 2012–2013 FB (88%) & PNB (12%) 150 62.4* (IQR 55.0–66.4) 5.6* (IQR 4.5–7.5) 40* (IQR 30–57) 1.5 T (47%) & 3 T (53%) T2WI/DWI/DCEI No PI-RADS v1 ≥3 Median of 30 TPBx (volume dependent)
Pokorny (2014) [43] Prospective 2012–2013 FB 223 63* (IQR 57–68) 5.3* (IQR 4.1–6.6) 41* (IQR 30–59) 3 T T2WI/DWI/DCEI No PI-RADS v1 ≥3 (primary);
≥4
12-core TRUS SBx + TBx
Petrillo (2014) [42] Prospective 2009–2010 NR 136 NR NR NR 1.5 T T2WI/DWI/MRSI Yes 1–5 scale (Likert) ≥3 12–16-core TRUS SBx (volume + PSA dependent)
Javali (2014) [29] Retrospective 2002–2011 NR 140 Control: 62.4 (10.5); Study: 62.9 (12.1) Control: 6.8 (2.3); Study: 6.87 (2.6) Control: 44 (14.2); Study: 43 (18.4) 1.5 T T2WI/MRSI Yes Dichotomous Cit/[Cho + Cr] < 1.2 6-core TRUS SBx ( n = 69),
12-core TRUS SBx ( n = 119)
Abd-Alazeez (2014) [13] Prospective 2007–2011 FB 129 62* (41–82) 5.8* (1.2–20) 40* (16–137) 1.5 & 3 T T2WI/DWI/DCEI No PI-RADS v1 ≥3 (primary);
≥4
20-core TPBx
Abd-Alazeez (2014) [12] Retrospective NR PNB 54 64* (39–75) 10* (2–23) 53* (19–136) 1.5 & 3 T T2WI/DWI/DCEI No PI-RADS v1 ≥3 (primary);
≥4
≥20-core TPBx + TBx ( n = 15)
Matsuoka (2014) [37] Prospective 2007–2012 NR 135 67* (50–80) 7.0* (2.9–19.8) 25.4* (12.7–90.2) 1.5 T T2WI/DWI No 1–5 scale (Likert) ≥3 14-core TRUS SBX
Junker (2013) [30] Retrospective 2011–2013 PNB 73 62 (7.4) 7.0* (5.1–12.9) 45* (34–61) 3 T T2WI/DWI/DCEI No PI-RADS PI-RADS ≥10 and ≥11 for all PCa
PI-RADS ≥13 for significant PCa
10-core TRUS SBx + TBx
Busetto (2013) [16] Prospective 2010–2012 PNB 163 66.4 (5.3) 6.8 (1.6) NR 3 T T2WI/DWI/DCEI/MRSI Yes NR NR 10-core TRUS SBx + TBx
Rais-Bahrami (2013) [47] Prospective 2007–2012 NR 583 61.3 (8.4) 9.9 (13.1) NR 3 T T2WI/DWI/DCEI/MRSI Yes 1–4 scale (Likert) ≥2
≥3
12-core TRUS SBx + TBx
Kuru (2013) [34] Prospective 2010–2011 FB (51%) & PNB (49%) 347 65.3 (42–82) 9.85 (0.5–104) 48.7 (9–180) 3 T T2WI/DWI/DCEI No 1–3 scale (Likert) ≥2 12–36-core TRUS SBx (volume dependent) + TBx
Ferda (2013) [20] Prospective NR NR 164 (49–74) (4.2–123) NR 3 T T2WI/DWI/DCEI/MRSI No NR In house TRUS SBx
Ganie (2013) [21] NR 2007–2009 PNB 87 NR NR NR 1.5 T T2WI/MRSI Yes MRSI Cho/Cit ratio In house 6 core TRUS SBx + MRI TBx
Vinet (2013) [57] Prospective 2009–2011 FB 69 NR 5.2* (3.2–28) NR 1.5 T (35 pts) & 3 T (34 pts) T2WI/DWI/DCEI No 1–5 scale (Likert) ≥3 12-core TRUS SBx + TBx
Numao (2013) [38] Prospective 2006–2010 FB 351 65* (59–70) 6.3* (4.9–9.1) 32* (24–42) 1.5 T T2WI/DWI/DCEI (no DCEI in 42 pts) No 1–5 scale (Likert) ≥3 3D 26-core (12 TRUS SBx + 14 TPBx – 203 pts);
3D 14-core (6 TRUS SBx + 8 TPBx – 102 pts);
TPBx 14 core (46 pts)
Belas (2012) [14] Prospective 2010–2011 FB 71 66* (47–76) 7* (4–10) 45* (15–150) 1.5 T T2WI/DWI/DCEI No TZ: 0–4 scale
PZ: 0–10 scale
TZ: >2
PZ: >6
12-core TRUS SBx + TBx
Ibrahiem (2012) [26] Prospective 2008–2009 FB 100 65.03 (7.13) 26.3 (24.2) 60.09 (28.7) 1.5 T T2WI/DWI No NR In house 12-core TRUS SBX
Sciarra (2012) [51] Prospective 2008–2011 PNB 84 64.09 (46–76) 7.07 (4.2–15.5) NR 3 T T2WI/DWI/DCEI/MRSI Yes NR NR 10-core TRUS Bx + TBx
Portalez (2012) [45] Prospective 2011 PNB 129 64.7 (47–79) 9.6 (2.7–40) 51.1 (12–192) 1.5 T T2WI/DWI/DCEI Mixed 1–5 scale (Likert)
PI-RADS
Likert ≥3
PI-RADS ≥9
10–12-core TRUS SBx + TBx
Watanabe (2012) [59] Prospective 2004–2008 NR 1448 72 (7.5) NR NR 1.5 T T2WI/DWI No NR ADC value ≤1.35 × 10 −3 mm 2 /s 8-core TRUS SBx + TBx
Tamada (2011) [54] Retrospective 2006–2009 NR 50 70 (40–84) 6.84* (4.06–9.94) NR 1.5 T T2WI/DWI/DCEI No NR In house 12-core TRUS SBx
Choi (2011) [18] NR 2009–2010 NR 51 67.16 (56–90) 14.16 (1.02–38.9) 42.98 (13.8–77.3) 3 T T2WI/DWI NR NR NR 10–12-core TRUS SBx + TBx
Iwazawa (2011) [28] Retrospective 2008–2009 NR 178 68.8 (41–86) 20.51 (4.04–568.5) NR 1.5 T T2WI/DWI/DCEI NR 1–4 scale (Likert) NR 10–12-core TRUS SBx (TBx included in SBx chart)
Rouse (2011) [48] Prospective 2005–2007 PNB 114 63.6 (41–83) 13.4 (0–228) NR 1.5 T T2WI/DWI/DCEI NR 1–5 scale (Likert) ≥3 TRUS SBx
Haffner (2011) [23] Prospective 2005–2009 FB 555 64* (47–83) 6.75* (0.18–100) 46* (15–200) 1.5 T T2WI/DCEI No 1–5 scale (Likert) ≥3 10-core TRUS SBx + TBx
Panebianco (2010) [40] Prospective 2007–2009 PNB 150 61.2 (46–78) 9.42 (3.91) 41.17 (7.47) 1.5 T T2WI/DCEI/MRSI Yes NR In house 10-core TRUS SBx (TBx included in SBx chart)
Roy (2010) [49] Not specified 2011–2009 FB (53%) & PNB (47%) 103 63 (52–78) 7* NR 3 T T2WI/DWI/DCEI Yes NR NR 8-core TRUS SBx + TBx
Testa (2010) [55] Not specified 2007 PNB 54 63.9 (52–76) 11.4 (3–42) 59.3 (30–150) 1.5 T T2WI/MRSI Yes 1–3 scale (Likert) ≥2 12-core TRUS SBx + TBx
Sciarra (2010) [52] Prospective 2007–2009 PNB 110 63.5 (49–74) NR NR 1.5 T T2WI/DCEI/MRSI Yes NR In house 10-core TRUS SBx + TBx
Kitajima (2010) [31] Prospective 2008–2009 NR 53 69* (56–84) 11.1* (4.2–112.1) NR 3 T T2WI/DWI/DCEI No 1–5 scale (Likert) ≥3 20-core TPBx
Labanaris (2010) [35] Prospective 2004–2008 PNB 260 NR NR NR 1 T T2WI/DWI/DCEI Yes Dichotomous In house 18-core TRUS SBx
Kumar (2009) [32] NR NR NR 61 65.3 (9.3) 16.5 (0.21–155) NR 1.5 T T2WI/MRSI Yes NR (Cit/(Cho + Cr) ≤ 1.2 TRUS Bx
Schmuecking (2009) [50] NR NR FB & PNB 67 68 11.5 NR 1.5 T T2WI/DCEI No NR NR 20-core Bx
Cheikh (2009) [17] Retrospective 2005–2008 PNB 93 63.2 (52–74) 9.63 (1.6–40) NR 1.5 T T2WI/DCEI No Dichotomous In house 12-core TRUS SBx + TBx
Cirillo (2008) [19] Prospective 2004–2006 PNB 54 65.5 (5.2) 10.8 (7.5) NR 1.5 T T2WI/MRSI Yes Dichotomous In house 10-core TRUS SBx + TBx
Kumar (2007) [33] Prospective 2003–2005 NR 83 NR NR NR 1.5 T T2WI/MRSI Yes NR NR 12-core TRUS SBx + TBx
Squillaci (2005) [53] Prospective 2004–2005 NR 65 NR NR NR 1.5 T T2WI/DCEI/MRSI No 1–3 scale (Likert) ≥2 10-core TRUS SBx + TBx
Hara (2005) [24] Prospective 2003–2004 FB 90 67.2 (NR) NR NR 1.5 T T2WI/DCEI No 1–3 scale (Likert) ≥2
= 3
14-core TRUS SBx
FB = first biopsy; IQR = interquartile range; PNB = previous negative biopsy; Pt = patient; Nb = number; MR = magnetic resonance; MRI magnetic resonance imaging; T2WI = T2-weighted imaging; DWI = diffusion-weighted imaging; DCEI = dynamic contrast-enhanced imaging; MRSI = magnetic resonance spectroscopic imaging; NR = not reported; Bx = biopsy; TRUS SBx = transrectal ultrasound-guided standard biopsy; TPBx = transperineal template biopsy; TBx = targeted biopsy; PCa = prostate cancer; PI-RADS = Prostate Imaging Reporting Data System; PSA = prostate-specific antigen; PZ = peripheral zone; TZ = transition zone.
Figures between parenthesis correspond to standard deviations; intervals between brackets correspond to full ranges or interquartile ranges when specified (IQR). The asterisk sign indicates median value (as opposed to mean value).

Table 1Baseline characteristics of included studies

 

The magnetic field strength was 1, 1.5, and 3 T in one, 28, and 15 studies, respectively. Four studies used both 1.5 and 3 T MR systems. DWI and DCEI were used in 36 and 35 studies, respectively. Nineteen studies also added MRSI. An endorectal coil was used in 18 studies. The definition of positive mpMRI varied across studies. The PI-RADS v1 score was used in 12 studies, a five-level subjective (Likert) score was used in eight studies, and one study reported data based on the two scoring systems. In-house criteria were used in 13 studies for defining positive mpMRI, and five studies used a dichotomous definition. Nine studies did not report on the criteria for positive mpMRI. No study used the PI-RADS v2 score.

Regarding the reference standard, TRUS-guided biopsies were used in 39 studies, TTP biopsies in six studies, and mixed TRUS-guided and TTP biopsies in two studies. In one study, the biopsy approach was unclear. The number of cores per biopsy procedure was ≤16 in 30 studies, >16 in nine studies, and variable among patients in three studies. For six studies, the number of biopsy cores taken was unclear.

3.4 NPV of prebiopsy mpMRI

At patient level, the median biopsy positivity rate (ie, cancer prevalence) was 50.4% (IQR, 36.4–57.7%) for overall cancer and 32.9% (IQR, 28.1–37.2%) for csPCa ( Table 2 ). The median mpMRI NPV was 82.4% (IQR, 69.0–92.4%) for overall cancer and 88.1% (IQR, 85.7–92.3) for csPCa. NPV significantly decreased when cancer prevalence increased, both for overall cancer ( r = –0.64, p < 0.0001) and csPCa ( r = –0.75, p = 0.032; Fig. 3 ). In addition, NPV was highly dependent on the definition used for csPCa, with differences of up to 21% when several definitions were used in the same dataset [12 13 38 47 48] .

Study Overall PCa prevalence (%) Reporting level Multiparametric MRI performance for PCa detection Definition of csPCa csPCa prevalence Multiparametric MRI performance for csPCa detection
      TN FN TP FP NPV PPV     TN FN TP FP NPV PPV
Hauth (2015) [25] 45.7 Patient 6 1 42 45 85.7% 48.3% Low grade: GS ≤ 6
High grade: GS ≥ 7
NR NR NR NR NR NR NR
    Lesion 59 13 55 73 81.9% 43%                
Lamb (2015) [36] 65.9 Patient 23 22 92 36 51.1% 71.9% GS ≥ 7 50.9% 31 14 71 57 68.9% 55.5%
Brock (2015) [15] 42.3 Patient 17 7 56 88 70.8% SBx: 38.8%
TBx: 22.2%
Duo: 44.4%
Epstein: GS > 6 and/or max CCL ≥50% 24.4% SBx: 21 SBx: 3 SBx: 38
TBx: 27
duo: 47
SBx: 106
TBx: 117
duo: 97
87.5% SBx: 26.4%
TBx: 18.8%
duo: 32.6%
Grenabo Bergdahl (2016) [22] 33.7 Patient 36 7 19 21 83.7% 47.5% Clinically insignificant PCa: T1c, PSAd < 0.15, GS < 7, ≤2 positive cores, and unilateral cancer NR NR NR NR NR NR NR
Wang (2015) [58] 58 Patient 165 8 332 81 95.4% 80.4% NR NR NR NR NR NR NR NR
Pepe (2015) [41] 37 Patient 23 8 29 40 74.2% 42% GS ≥ 7 or GS 6 with CCL ≥ 50% NR NR NR NR NR 100% 55.8%
Panebianco (2015) [39] 74.7 Patient Group A (satBx): 104 Group A (satBx): 43 Group A: 186 Group A: 22 Group A (satBx): 70.7% Group A: 89.4% NR 60% Group A: 147 Group A: 0 Group A: 183 Group A: 25 Group A: 100% Group A: 88%
      Group B (TRUSGB NR TRUS G satBx): 93 Group B (TRUSGB NR TRUS G satBx): 37 Group B: 417 before and 425 after satBx Group B: 23/440 and 15/440 after satBx. Group B (TRUSGB–TRUS G satBx): 71.5% Group B: 94.8% before and 96.6% after satBx   71.9% Group B: 130 Group B: 0 Group B:
410
Group B: 30 Group B: 100% Group B: 93.2%
Radtke (2015) [46] 51 Patient ≥2/5: 80
≥3: 103
≥4: 138
= 5: 142
≥2/5: 18
≥3/5: 38
≥4/5: 78
= 5/5: 126
≥2/5: 132
≥3/5: 112
≥4/5: 72
= 5/5: 24
≥2/5: 64
≥3/5: 41
≥4/5: 6
= 5/5: 2
≥2/5: 81.6%
≥ 3/5: 73.1%
≥ 4/5: 63.9%
= 5/5: 53%
≥2/5: 67.4%
≥ 3/5: 73.2%
≥ 4/5: 92.3%
= 5/5: 92.3%
GS ≥ 7 29.3% ≥2/5: 91
≥3/5: 124
≥4/5: 183
= 5/5: 203
≥2/5: 7
≥3/5: 17
≥4/5: 33
= 5/5: 65
≥2/5: 79
≥3/5: 69
≥4/5: 53
= 5/5: 21
≥2/5: 117
≥3/5: 84
≥4/5: 25
= 5/5: 5
≥2/5: 92.2%
≥3/5: 87.9%
≥4/5: 84.7%
= 5/5: 75.8%
≥2/5: 40.3%
≥3/5: 45.1%
≥4/5: 68%
= 5/5: 80.7%
Itatani (2014) [27] 13 Patient 168 25 NR NR 87% NR (1) PSAd ≥ 0.1 or GP 4/5 or ≥ 3/6 pos cores; max CCL < 50%
(2) PSAd ≥ 0.15 or GP 4/5 or max core length invasion < 3 mm (min 6 cores)
NR NR NR NR NR NR NR
Porpiglia (2014) [44] 30.6 Patient 107 5 47 11 95.5% 81% NR NR NR NR NR NR NR NR
Thompson (2014) [56] 61.3 Patient 35 16 76 23 68.6% 76.7% Moderate or high risk
Moderate risk: GS 7 (GP4 >5%) and <50% of positive cores or GS 6–7 (GP4 ≤ 5%) and either ≥30% of positive cores or max core length invasion ≥8 mm
High risk: GS ≥ 7 (GP4 > 5%); ≥50% + cores or GS ≥ 8
34% 49 2 49 50 96% 49.5%
Pokorny (2014) [43] 63.7 Patient ≥3/5: 56
≥4/5: 74
≥3/5: 25
≥4/5: 40
≥3/5: 101
≥4/5: 86
≥3/5: 41
≥4/5: 23
≥3/5: 69.1%
≥4/5: 64.9%
≥3/5: 71.1%
≥4/5: 78.9%
NR NR NR NR NR NR NR NR
Petrillo (2014) [42] 18.4 Patient 56 4 21 55 93% 28% NR NR NR NR NR NR NR NR
Javali (2014) [29] 16.4 Patient 49 1 22 68 98% 24.4% NR NR NR NR NR NR NR NR
Abd-Alazeez (2014) [13] 54.7 Lobe ≥3/5: 33 ≥3/5: 14 ≥3/5: 127 ≥3/5: 84 ≥3/5: 70% ≥3/5: 60% Def 1: GS ≥ 4 + 3 and/or max CCL ≥ 6 mm
Def 2: GS ≥ 3 + 4 and/or max CCL ≥ 4 mm
Def 3: GS ≥ 4 + 3
Def 4: GS ≥ 3 + 4
Def 5: max CCL ≥ 6 mm
Def 6: max CCL ≥ 4 mm
Def 1: NR
Def 2: 44.2%
Def 3: 23.2%
Def 4: 36%
Def 5: 32.9%
Def 6: 39.5%
Def 1: 46
Def 2: 42
Def 3: 47
Def 4: 43
Def 5: 46
Def 6: 43
Def 1: 1
Def 2: 5
Def 3: 0
Def 4: 4
Def 5: 1
Def 6: 4
Def 1: 45
Def 2: 72
Def 3: 13
Def 4: 50
Def 5: 39
Def 6: 59
Def 1: 166
Def2: 139
Def 3: 198
Def 4: 161
Def 5: 172
Def 6: 152
Def 1: 98
Def 2: 89
Def 3: 100
Def 4: 92
Def 5: 98
Def 6: 91
Def 1: 21
Def 2: 34
Deg 3: 6
Def 4: 24
Def 5: 19
Def 6: 28
      ≥4/5: 87 ≥3/5: 62 ≥3/5: 79 ≥3/5: 30 ≥3/5: 58% ≥3/5: 73%     Def 1: 140
Def 2: 124
Def 3: 148
Def 4: 133
Def 5: 141
Def 6: 131
Def 1: 9
Def 2: 25
Def 3: 1
Def 4: 16
Def 5: 8
Def 6: 18
Def 1: 37
Def 2: 52
Def 3: 12
Def 4: 38
Def 5: 32
Def 6: 45
Def 1: 72
Def 2: 57
Def 3: 97
Def 4: 71
Def 5: 77
Def 6: 64
Def 1: 94
Def 2: 83
Def 3: 99
Def 4: 89
Def 5: 95
Def 6: 88
Def 1: 34
Def 2: 48
Def 3: 11
Def 4: 35
Def 5: 30
Def 6: 42
Abd-Alazeez (2014) [12] 47.2 Lobe ≥3/5: 26 ≥3/5: 13 ≥3/5: 38 ≥3/5: 31 ≥3/5: 66% ≥3/5: 55% Def 1: GS ≥ 4 + 3 and/or max CCL ≥ 6 mm
Def 2: GS ≥ 3 + 4 and/or max CCL ≥ 4 mm
Def 3: GS ≥ 4 + 3
Def 4: GS ≥ 3 + 4
Def 5: max CCL ≥ 6 mm
Def 6: max CCL ≥ 4 mm
Def 1: 18.5%
Def 2: 31.5%
Def 3: 4.6%
Def 4: 21.3%
Def 5: 16.7%
Def 6: 25%
Def 1: 37
Def 2: 31
Def 3: 39
Def 4: 36
Def 5: 37
Def 6: 32
Def 1: 2
Def 2: 8
Def 3: 0
Def 4: 3
Def 5: 2
Def 6: 7
Def 1: 18
Def 2: 26
Def 3: 5
Def 4: 20
Def 5: 16
Def 6: 20
Def 1: 51
Def 2: 43
Def 3: 64
Def 4: 49
Def 5: 53
Def 6: 49
Def 1: 95
Def 2: 79
Def 3: 100
Def 4: 92
Def 5: 95
Def 6: 82
Def 1: 26
Def 2: 38
Def 3: 7
Def 4: 29
Def 5: 23
Def 6: 29
      ≥4/5: 49 ≥4/5: 25 ≥4/5: 26 ≥4/5: 8 ≥4/5: 66% ≥4/5: 76%     Def 1: 70
Def 2: 63
Def 3: 73
Def 4: 68
Def 5: 70
Def 6: 64
Def 1: 4
Def 2: 11
Def 3: 1
Def 4: 6
Def 5: 4
Def 6: 10
Def 1: 16
Def 2: 23
Def 3: 4
Def 4: 17
Def 5: 14
Def 6: 17
Def 1: 18;
Def 2: 11
Def 3: 30
Def 4: 17
Def 5: 20
Def 6: 17
Def 1: 94
Def 2: 85
Def 3: 99
Def 4: 92
Def 5: 94
Def 6: 86
Def 1: 47
Def 2: 67
Def 3: 12
Def 4: 50
Def 5: 41
Def 6: 49
Matsuoka (2014) [37] 64.8 Lobe 46 49 149 26 48.4% 85.1% NR NR NR NR NR NR NR NR
Junker (2013) [30] 53.4 Patient ≥10/15: 21
≥11/15: 28
≥10/15: 4
≥11/15: 12
≥10/15: 35
≥11/15: 27
≥10/15: 13
≥11/15: 6
≥10/15: 84%
≥11/15: 70%
≥10/15: 72.9%
≥11/15: 81.8%
GS ≥ 4 + 3 13.7% ≥13: 54 ≥13: 2 ≥13: 8 ≥13: 9 ≥13: 96.4% ≥13: 47%
Busetto (2013) [16] 41.7 Patient 59 7 61 36 89% 63% NR NR NR NR NR NR NR NR
Rais-Bahrami (2013) [47] 54.2 Patient ≥2/4 : 80 ≥2/4: 275 ≥2/4: 187 ≥2/4: 275 ≥2/4: 66% ≥2/4: 59% GS ≥ 7:
GS ≥ 8
31.7% NR NR NR NR GS ≥ 7: 91%
GS ≥ 8: 91%
GS ≥ 7: 38%
GS ≥ 8: 18%
      ≥3/4: 251 ≥3/4: 76 ≥3/4: 16 ≥3/4: 76 ≥3/4: 51% ≥3/4: 82%     NR NR NR NR GS ≥ 7: 75%
GS ≥ 8: 91%
GS ≥ 7: 67%
GS ≥ 8: 41%
Kuru (2013) [34] 57.6 Patient 80 14 67 186 85.1% 73.5% NR NR NR NR NR NR NR NR
Ferda (2013) [20] 51.2 Patient 52 2 82 28 96.3% 74.5% NR NR NR NR NR NR NR NR
Ganie (2013) [21] 74.7 Patient 13 3 62 9 59.1% 95.4% NR NR NR NR NR NR NR NR
Vinet (2013) [57] 49.3 Patient ≥3/5: 11 ≥3/5: 6 ≥3/5: 28 ≥3/5: 24 ≥3/5: 64.7% ≥3/5: 53.8% NR NR NR NR NR NR NR NR
      ≥4/5: 22 ≥4/5: 11 ≥4/5: 23 ≥4/5: 13 ≥4/5: 66.7% ≥4/5: 63.8% NR NR NR NR NR NR NR NR
Numao (2013) [38] 45 Patient 137 57 101 56 70.6% 64.3% Def 1: GS ≥4 + 3 and/or ≥20% positive cores and/or max CCL ≥5 mm
Def 2: GS ≥ 3 + 4 and/or ≥20% + cores and/or max CCL ≥5 mm
Def 3: GS ≥ 3 + 4 and/or ≥20% + cores
Def 1: 33.3%
Def 2:37.8%
Def 3: 35.8%
Def 1: 74
Def 2: 65
Def 3: 69
Def 1: 83
Def 2: 92
Def 3: 88
Def 1: 160
Def 2: 153
Def 3: 156
Def 1: 34
Def 2: 41
Def 3: 38
Def 1: 82.4%
Def 2: 78.8%
Def 3: 80.4%
Def 1: 58.8%
Def 2: 58.5%
Def 3: 56%
Belas (2012) [14] 53.5 Patient 22 12 23 13 64.7% 63.8% NR NR NR NR NR NR NR NR
Ibrahiem (2012) [26] 73.9 Patient 14 11 57 10 56% 85.1% NR NR NR NR NR NR NR NR
Sciarra (2012) [51] 34.5 Patient 41 4 25 14 91.1% 64.1% NR NR NR NR NR NR NR NR
Portalez (2012) [45] 48.1 Lesion Likert: 357 Likert: 50 Likert: 81 Likert: 50 Likert: 95% Likert: 38% Max CCL >3 mm and/or GG 4/5 NR NR NR NR NR NR NR
      PI-RADS: 404 PI-RADS: 47 PI-RADS: 34 PI-RADS: 47 PI-RADS: 95% PI-RADS: 58%                
Watanabe (2012) [59] 48.1 Patient 485 73 624 266 86.9% 70.1% NR NR NR NR NR NR NR NR
Tamada (2011) [54] 70 Patient 12 6 29 3 67% 91% NR NR NR NR NR NR NR NR
    Region 277 48 55 20 85% 73%                
Choi (2011) [18] 70.6 Patient 9 5 31 6 64.2% 75.7% NR NR NR NR NR NR NR NR
Iwazawa (2011) [28] 40.5 Region 887 86 232 219 91.1% 51.4% NR NR NR NR NR NR NR NR
Rouse (2011) [48] 33.7 Sextant 145 11 74 22 92.9% 77.1% GS ≥ 7:
Def 1: ≥3 mm
Def 2: ≥5 mm
Def 1: 26.6%
Def 2: 26.2%
Def 1: 153
Def 2: 153
Def 1: 3
Def 2: 3
Def 1: 64
Def 1: 63
Def 1: 32
Def 2: 33
Def 1: 98.1%
Def 2: 98.1%
Def 1: 68.1%
Def 2: 67%
  59.6 Patient 24 14 54 22 63.2% 71.1%   Def 1: 41.2%
Def 2: 36.8%
Def 1: 30
Def 2: 31
Def 1: 4
Def 2: 3
Def 1: 43
Def 2: 39
Def 1: 4
Def 2: 3
Def 1: 88.2%
Def 2: 91.2%
Def 1: 53.8%
Def 2: 48.8%
Haffner (2011) [23] 54.4 Patient 154 50 240 111 75.4% 68.3% NR NR NR NR NR NR NR NR
Panebianco (2010) [40] 42.7 Patient NR NR NR NR 95.1% 88.2% NR NR NR NR NR NR NR NR
Roy (2010) [49] 55.9 Patient NR NR NR NR 71% 75% NR NR NR NR NR NR NR NR
Testa (2010) [55] 40.7 Patient NR NR NR NR 79.3% 64% NR NR NR NR NR NR NR NR
Sciarra (2010) [52] 50 Patient 61 4 66 9 93.8% 88% NR NR NR NR NR NR NR NR
Kitajima (2010) [31] 56.6 Patient 311 19 80 14 92.2% 85.1% NR NR NR NR NR NR NR NR
Labanaris (2010) [35] 73.9 Patient 17 73 96 74 81.11% 56.47% NR NR NR NR NR NR NR NR
Kumar (2009) [32] 21.7 Patient 39 3 10 8 92.8% 55% NR NR NR NR NR NR NR NR
Schmuecking (2009) [50] NR Lobe NR NR NR NR 96% 61% NR NR NR NR NR NR NR NR
Cheikh (2009) [17] 24.7 Patient 36 12 11 34 80% 22.9% NR NR NR NR NR NR NR NR
Cirillo (2008) [19] 31.5 Patient 19 0 17 18 100% 48.6% NR NR NR NR NR NR NR NR
Kumar (2007) [33] 13.3 Patient 39 0 11 33 100% 25% NR NR NR NR NR NR NR NR
Squillaci (2005) [53] 50.8 Patient 29% 8% 25% 3% 89% 80% NR NR NR NR NR NR NR NR
Hara (2005) [24] 41.5 Patient ≥2/3: 40
= 3/3: 47
≥2/3: 4
= 3/3: 8
≥2/3 30
= 3/3: 26
≥2/3: 8
= 3/3: 1
≥2/3: 90%
= 3/3: 85%
≥2/3: 78.9%
= 3/3: 96.3%
NR NR NR NR NR NR NR NR
PCa = prostate cancer; MRI = magnetic resonance imaging; TN = true negative; FN = false negative; FP = false positive; TP = true positive; PPV = positive predictive value; NPV = negative predictive value; csPCa = clinically significant prostate cancer; max = maximum; CCL = cancer core length; PSAd = PSA density; GG = Gleason grade; GS = Gleason score; SBx = systematic biopsy; TBx = targeted biopsy; NR = not reported; PI-RADS = Prostate Imaging Reporting Data System; PSA = prostate-specific antigen.

Table 2Diagnostic performance of prebiopsy multiparametric MRI using biopsy findings as reference standard

gr3
 

Fig. 3
Negative predictive value of prebiopsy multiparametric MRI as a function of cancer prevalence (blue crosses: overall prostate cancer; red crosses: clinically significant prostate cancer). The blue line is the correlation line for overall prostate cancer; the red dotted line is the correlation line for clinically significant prostate cancer. mpMRI = multiparametric magnetic resonance imaging; MRI = magnetic resonance imaging.

 

Cancer prevalence tended to be higher and mpMRI NPV lower in the biopsy-naïve group as compared with the repeat biopsy group, in men with positive DRE as compared with men with negative DRE and when an endorectal coil was not used ( Table 3 ). There were no clear differences in the prevalence and NPV of the other analysed subgroups (TRUS-guided vs TTP biopsy, biopsy procedures with ≤16 cores vs >16 cores; Table 3 ). However, comparisons must be interpreted with care, due to the small number of studies in some subgroups. In patients with a PSA level of ≤10 ng/ml, the median NPV for overall PCa was 86.3% (IQR, 73.3–93.6%) for a median cancer prevalence of 35.4% (IQR, 27.6–42.5%).

  Nb of studies Median PCa prevalence Median mpMRI NPV Nb of studies Median csPCa prevalence Median mpMRI NPV
Biopsy-naïve patients 8 51.4% (45.5–56.7) 69.9% (64.2–78) 1 35.8% (NA) 80.4% (NA)
Repeat biopsy 14 42% (35.1–52.6) 82.6% (75.5–93.1) 3 24.4% (19.1–32.8) 88.2% (87.9–92.3)
TRUS-guided biopsy 36 49.7% (34.3–57.7) 84.6% (68.6–92.8) 4 28.1% (21.7–36.5) 89.3% (82.9–92.4)
TTP biopsy 4 53.8% (47.5–57.8) 73.6% (72–78.7) 2 31.6% (30.5–32.8) 92% (89.9–94)
Biopsy with ≤16 cores 28 48.7% (39.2–54.8) 81.9% (66.8–89.3) 5 28.1% (21.8–36.5) 89.3% (82.9–92.4)
Biopsy with >16 cores 5 56.6% (51–61.3) 81.1% (73.1–92.2) 2 31.7% (30.5–32.8) 92% (89.9–93.9)
Positive DRE 1 73.9% (NA) 56% (NA) 0
Negative DRE 6 36% (34.6–46.8) 82.7% (74.2–93.1) 0
Endorectal coil 17 41.7% (30.6–55.9) 92.8% (79.3–95.4) 1 31.7% (NA) 91% (NA)
No endorectal coil 22 50.9% (41.7–56.1) 77.7% (69.5–86.6) 7 34% (26.9–46.1%) 87.9% (78.2–92.1)
PCa = prostate cancer; csPCa = clinically significant prostate cancer; NPV = negative predictive value; TRUS = transrectal ultrasound; TTP = template transperineal; DRE = digital rectal examination; PSA = prostate-specific antigen; NA = not applicable; mpMRI = multiparametric magnetic resonance imaging; MRI = magnetic resonance imaging; Nb = number.
Intervals in parenthesis are interquartile ranges.

Table 3Reported ranges of negative predictive values for prebiopsy multiparametric MRI

 

3.5 Meta-analysis

3.5.1 NPV and PPV for overall PCa

Eight studies reported NPV at patient level for overall PCa and fulfilled the inclusion criteria for meta-analysis ( Table 4 ) [22 25 38 41 43 46 56 57] .

    Study Prevalence a (%) Neg MRI (%) NPV (%) PPV (%) Spe (%) Se (%)
All PCa Score ≥ 3/5 Grenabo Bergdahl (2016) [22] 31.3 66.9 83.7 47.5 63.2 73.1
    Numao (2013) [38] 45 29.4 70.6 64.3 71.0 63.9
    Hauth (2015) [25] 45.7 14.3 85.7 48.3 11.8 97.7
    Vinet (2013) [57] 49.3 24.6 64.7 53.8 31.4 82.4
    Radtke (2015) [46] 51 27 73 73.2 71.5 74.7
    Thompson (2014) [56] 61.3 31.4 68.6 76.8 60.3 82.6
    Pokorny (2014) [43] 63.7 30.9 69.1 82.4 69.1 82.4
  Score ≥ 4/5 Pepe (2015) [41] 37 31 74.2 42 36.5 78.4%
    Vinet (2013) [57] 49.3 47.8 66.7 63.9 62.9 67.6%
    Radtke (2015) [46] 51 73.5 63.9 92.3 95.8 48%
Gleason ≥7 PCa Score ≥ 3/5 Radtke (2015) [46] 29.3 27 87.9 45.1 59.6 80.2%

a Prevalence of overall prostate cancers (10 first lines) or Gleason ≥7 cancers (last line).

PCa = prostate cancer; MRI = magnetic resonance imaging; Neg MRI = proportion of negative magnetic resonance imaging; NPV = negative predictive value; PPV = positive predictive value; Se = sensitivity; Spe = specificity.

Table 4Prebiopsy multiparametric MRI results in the series selected for the meta-analysis

 

Seven studies used a score of ≥3/5 for defining positive mpMRI ( Fig. 4 A and B) [22 25 38 43 46 56 57] . Fig. 4 C shows the conditional probability plot of 1 – NPV and PPV as a function of overall PCa prevalence. Table 5 shows NPV and PPV estimates for the given values of PCa prevalence.

gr4
 

Fig. 4
Forest plot showing the (A) NPV and (B) PPV of prebiopsy multiparametric MRI for overall prostate cancer in the seven studies selected for meta-analysis that used a cut-off score of ≥3/5 for defining positive MRI. Studies have been ranked according to cancer prevalence (left column). Intervals in the right column are 95% CIs of the (A) NPV or (B) PPV. As NPV and PPV vary with cancer prevalence, combined estimates of NPV and PPV have not been provided. (C) Conditional probability plot showing the estimation of the combined NPV and PPV in the seven studies, as a function of the prevalence of overall prostate cancer. The x axis (prior probability) indicates the overall prostate cancer prevalence. The y axis (posterior probability) indicates either PPV (dashed line, upper quadrant) or 1 – NPV (dotted line, lower quadrant). CI = confidence interval; MRI = magnetic resonance imaging; NPV = negative predictive value; PPV = positive predictive value.

PCaPrev PPV NPV
0.30 0.43 (0.34–0.53) 0.88 (0.77–0.99)
0.40 0.54 (0.45–0.64) 0.82 (0.70–0.94)
0.50 0.64 (0.55–0.73) 0.76 (0.64–0.88)
0.60 0.73 (0.65–0.80) 0.67 (0.56–0.79)
0.70 0.81 (0.75–0.87) 0.57 (0.47–0.67)
0.75 0.84 (0.79–0.89) 0.51 (0.42–0.59)
PCaPrev = prevalence of prostate cancer; PPV = positive predictive value; MRI = magnetic resonance imaging; NPV = negative predictive value.
Intervals in parenthesis are 95% confidence intervals. A score of ≥3/5 was used to define positive MRI.

Table 5Positive and negative predictive estimates for prebiopsy multiparametric MRI as a function of prostate cancer prevalence (meta-analysis)

 

Only three studies used a score of ≥4/5 for defining positive mpMRI ( Table 4 ) [41 46 57] , and a formal meta-analysis could not be performed.

3.5.2 NPV and PPV for Gleason ≥7 cancers

Only one study reporting NPV at patient level for Gleason ≥7 cancers met the selection criteria for inclusion in the meta-analysis. It reported NPV and PPV of 87.9% and 45.1%, respectively, for a prevalence of 29.3% ( Table 4 ) [46] .

3.6 Discussion

3.6.1 Principal findings

We observed a large variability in reported NPV. Many factors, such as differences in mpMRI protocols, definition of negative mpMRI, or biopsy protocols, can explain this variability. However, two major causes of variability must be pointed out. First, the cancer prevalence was highly variable, ranging at patient level from 13% to 74.7% for overall PCa, and from 13.7% to 50.9% for csPCa. This variability was observed in both the biopsy-naïve and the repeat biopsy setting. As NPV depends on prevalence, this had a major impact on reported NPV ( Fig. 3 ). Second, the definition of csPCa was highly variable from one series to another, and differences of up to 21% could be observed in NPV when different definitions of csPCa were used in the same dataset [12 13 38 47 48] .

To account for clinical heterogeneity and to further explore the clinical relevance of the results, we carefully selected studies for inclusion in the meta-analysis based on stringent criteria. Particularly, we included only studies that: (1) had biopsy protocols with at least 10 cores, since it is no longer recommended to obtain less than 10 cores per biopsy; (2) used DWI, which is the most informative technique, at least for cancers in the peripheral zone [60] ; and (3) reported mpMRI findings using a five-level score, so that negative findings could be better defined. We accepted studies using a subjective (Likert) scale because experienced readers obtained equivalent [45 61 62] or better [63] results with the Likert score than with the PI-RADS v1 score. Owing to the large variations of NPV induced by differences in definitions of csPCa, we did not include different definitions in the meta-analysis since this would have introduced unacceptable clinical heterogeneity in the results, possibly resulting in erroneous and biased estimates. We, therefore, a priori restricted the definition of csPCa to cancers with a Gleason score of ≥7, given the low lethal potential of Gleason 6 cancers [64] and the lack of consensus among pathologists on the best method to measure biopsy core invasion length [65 66] .

In this more homogeneous group of studies, the prevalence range was still large (31.3–63.7%). As a result, we modelled the evolution of NPV (and PPV) as a function of overall PCa prevalence. Unfortunately, we could not duplicate this for csPCa since only one study reporting NPV for Gleason ≥7 cancers met the inclusion criteria for meta-analysis.

3.6.2 Reference standard

We included only studies that reported the results of systematic/standard biopsy in patients with negative mpMRI and used the systematic/standard biopsy as a reference standard. It is well known that TRUS-guided biopsy harbours both random and systematic errors, as evidenced by the high rates of positivity of immediate repeat biopsy after a first series of negative biopsies [67 68] , and as confirmed recently by the PROMIS trial [69] . Therefore, using TRUS-guided biopsy as a reference standard may have overestimated the NPV of mpMRI. However, studies using radical prostatectomy specimens as a reference standard have already reported mpMRI detection rates in relation to PCa Gleason score and volume [1] . In this review, we intended to address the more pragmatic question as to whether a negative mpMRI could predict a negative subsequent biopsy. This is an important question because if the NPV of mpMRI was sufficiently high in comparison with the reference standard of systematic/standard biopsies, then in practice a negative mpMRI result could indeed avoid the need for prostate biopsy. Therefore, studies reporting only biopsy results when the mpMRI was positive (eg, obtained through MRI-targeted, guided, or fusion biopsies with added systematic biopsies) were not included in this review.

3.6.3 Impact on clinical practice and research

It is now well established that mpMRI is a sensitive tool for detecting aggressive PCa 1 2 3 69] . However several reasons preclude its broad use as a triage test before biopsy.

Firstly, the population referred to prostate biopsy is not standardised. The large range of reported prevalence for overall PCa and csPCa suggests substantial heterogeneity in the way patients are selected for biopsy. Owing to this heterogeneity, we did not provide a pooled estimate for mpMRI NPV. The role of mpMRI as a triage test before prostate biopsy should be evaluated in the broader context of the selection of patients with a suspicion of (aggressive) PCa. In a recent retrospective study of 514 patients, mpMRI NPV for Gleason ≥7 cancers was 91% when the PSA density was ≤0.2 ng/ml/ml, and only 71% when the PSA density was >0.2 ng/ml/ml ( p = 0.003) [70] . In another series of 288 biopsy-naïve patients, no csPCa (Gleason score ≥7 or maximum cancer core length ≥4 mm) was found in 44 patients with a PSA density of <0.15 ng/ml/ml and a PI-RADS v2 score of <3/5 [71] . We believe that such prestratification of the risk of csPCa is an interesting way for rationalising the use of mpMRI before biopsy. Patients found at very low risk would be spared both mpMRI and biopsy. Patients at a low risk—for whom mpMRI would have an NPV high enough to be used as a triage test—could avoid biopsy in case of negative mpMRI. Patients at a higher risk would need biopsy even in case of negative mpMRI. Many tools can be used to risk stratify the population of patients referred to biopsy, ranging from simple parameters such as PSA density to more complicated risk calculators [72 73] . The impact of these tools on the NPV of prebiopsy mpMRI needs to be carefully evaluated, both in the biopsy-naïve and in the repeat biopsy setting. For the moment, it is impossible to make any recommendations on the best way to risk stratify patients before referring them for mpMRI.

Secondly, the large variability in the definition of csPCa precludes any definitive conclusion on the ability of mpMRI to rule out aggressive cancer. The issue of the most appropriate definition of csPCa on biopsy is complex, since biopsy results may accurately reflect neither tumour burden nor aggressiveness. Nonetheless, there is an urgent need to standardise the histological definition(s) of csPCa, to allow meaningful comparisons between studies.

Thirdly, the specificity of mpMRI remains moderate, and there is a substantial proportion of false positives in the lesions scored 3/5 or 4/5 [1 74 75] , even with the new PI-RADS v2 score [76] . In a series of 62 patients with 116 lesions biopsied under magnetic resonance/ultrasound fusion, the overall cancer detection rates for PI-RADS v2 scores of 3/5 and 4/5 were only 15.8% and 29.8%, respectively [77] . In theory, a triage test used to rule out a disease needs to be highly sensitive for this disease. However, if its specificity is too low, it will be clinically useless since most patients will be positive, whether they have the disease or not. Therefore, if mpMRI is to be used as a triage test in the future, there is a need to improve its specificity. This could be achieved by a continuous refinement of scores [78] . Promising results in characterising csPCa have also been reported with a quantitative analysis [79] .

Finally, all published studies were conducted in specialised centres. The broad use of mpMRI as a triage test assumes good interobserver reproducibility. Unfortunately, interobserver reproducibility of existing scoring systems remains moderate [62 63 80] even with the use of the PI-RADS v2 score [80 81] . Studies evaluating on a large scale the reproducibility of mpMRI findings between expert and nonexpert centres are currently lacking.

3.6.4 How this review compares with other reviews

Three systematic reviews (including two meta-analyses) regarding the role of mpMRI in localised PCa have been published recently 4 5 6 . Crucially, all three reviews focused exclusively on the sensitivity of mpMRI-targeted, guided, or fusion biopsies in diagnosing overall PCa and csPCa, using TRUS-guided prostate biopsies as reference standards. The impact of systematic biopsies on the outcome was not addressed in any of the reviews, within either the index test or the reference standard. Our review had a totally different research question and objective, focusing on NPV of mpMRI to see if a negative mpMRI can avoid the need for a prostate biopsy. As MRI-targeted/guided/fusion biopsies are not relevant if the mpMRI was negative for cancer, it can be argued that the three reviews assessed a different index test altogether. As such, we believe that the findings of this review are novel and unique, and pave the way for further focused clinical studies.

3.6.5 Strengths and limitations

The current study represents the first systematic review addressing the role of mpMRI as a triage test before biopsy. The review elements were developed in conjunction with a multidisciplinary panel of experts (EAU Prostate Cancer Guidelines Panel), which included a patient representative, and the review was performed robustly in accordance with recognised standards. However, it is limited by the major heterogeneity of the existing literature in patient population, study design, and definitions of positive mpMRI and csPCa. It highlighted further areas of research that could help in defining the best use of mpMRI in the early detection of aggressive PCa in the future.

4 Conclusions

Although mpMRI can detect aggressive PCa with excellent sensitivity, a definitive conclusion on its role as a triage test before prostate biopsy will be possible only when three main issues are addressed. Firstly, because NPV depends on prevalence, and because overall PCa and csPCa prevalence was highly variable in the published series, it becomes mandatory to define the optimal way to pre-evaluate the risk of csPCa in patients with a suspicion of PCa. Depending on the risk category, mpMRI could then be used to obviate biopsies or not. Secondly, there is a need for consensus definitions of csPCa on biopsy findings to allow interstudy comparisons. Thirdly, although efforts have been made to standardise mpMRI technical protocols and interpretation in the past few years [11 60 76] , there is still a crucial need to improve mpMRI specificity and inter-reader reproducibility.

This systematic review was performed under the auspices of:

  • -

    The European Association of Urology Guidelines Office Board

  • -

    The European Association of Urology Prostate Cancer Guideline Panel

 

Author contributions: Olivier Rouvière 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: Rouvière, Mottet, Cornford.

Acquisition of data: Rouvière, Moldovan, Van den Broeck, Yuan.

Analysis and interpretation of data: Rouvière, Moldovan, Van den Broeck, Lam.

Drafting of the manuscript: Rouvière, Moldovan, Van den Broeck, Lam.

Critical revision of the manuscript for important intellectual content: Bellmunt, van den Bergh, Bolla, Briers, Van den Broeck, Cornford, Cumberbatch, De Santis, Fossati, Gross, Henry, Joniau, Lam, Matveev, Moldovan, van der Poel, van der Kwast, Rouvière, Schoots, Wiegel, Mottet.

Statistical analysis: None.

Obtaining funding: None.

Administrative, technical, or material support: Yuan, Sylvester, Marconi.

Supervision: None.

Other: None.

Financial disclosures: Olivier Rouvière 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 (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: Bellmunt is a company consultant for Janssen, Astellas, Pierre Fabre, Genentech, Merck, Ipsen, Pfizer, Novartis, and Sanofi Aventis. He has received research support from Takeda, Novartis, and Sanofi, and received travel grants from Pfizer and Pierre Fabre. Bolla has received company speaker honoraria from Ipsen and Astellas, honoraria or consultation fees from Janssen, and fellowship and travel grants from Janssen, AstraZeneca, and Astellas. Briers has received grant and research support from Ipsen, European Association of Urology, and Bayer; is an ex officio board member for Europa UOMO; is an ethics committee and advisory group member for REQUITE; is a member patient advisory board member for PAGMI; and is a member of SCA and EMA PCWP. Cornford is a company consultant for Astellas, Ipsen, and Ferring. He receives company speaker honoraria from Astellas, Janssen, Ipsen, and Pfizer; participates in trials from Ferring; and receives fellowships and travel grants from Astellas and Janssen. De Santis is a company consultant for GlaxoSmithKline, Janssen, Bayer, Novartis, Pierre Fabre, Astellas, Amgen, Eisai Inc., ESSA, Merck, and Synthon; has received company speaker honoraria from Pfizer, Takeda, Sanofi Aventis, Shionogi, Celgene, and Teva OncoGenex; has participated in trials for Pierre Fabre, Astellas, Exelixis, Bayer, and Roche; has received fellowship and travel grants from Bayer, Novartis, Ferring, Astellas, Sanofi Aventis, and Janssen; has received grant and research support from Pierre Fabre; has received honoraria from AstraZeneca; and is associated with Amgen. Joniau is a company consultant for Astellas, Ipsen, Bayer, Sanofi, and Janssen; has received company speaker honoraria from Astellas, Amgen, Bayer, Sanofi, Janssen, and Ipsen; has participated in trials for Astellas, Janssen, and Bayer; has received fellowship and travel grants from Astellas, Amgen, Bayer, Sanofi, Janssen, Ipsen, and Pfizer; and has received grant and research support from Astellas, Bayer, and Janssen. Matveev has received company speaker honoraria from Sanofi and Astellas, has participated in trials for Astellas, Pfizer and Novartis. Lam is a company consultant for and has received company speaker honoraria from Pfizer, GSK, Astellas, and Ipsen. van der Poel is a company consultant for Intuitive Surgical, has participated in trials for Astellas and Steba Biotech, and has received grant and research support from Astellas. Rouvière is a company consultant for EDAP-TMS, Bracco, and Philips; has received company speaker honoraria from EDAP-TMS, Bracco, and Philips; and has participated in trials for EDAP-TMS and Bracco. Wiegel has received company speaker honoraria from Astellas, Takeda, Hexal, Ipsen, Janssen-Cilac, and Ferring. Mottet has received grant and research support from Takeda Pharmaceutical, Millenium, Astellas, Pierre Fabre, Sanofi, and Pasteur, and has received honoraria or consultation fees from Takeda Pharaceutical, Millenium, Jansen, Astellas, BMS, Bayer, Ipsen, Ferring, Novartis, Nuclétron, Pierre Fabre, Sanofi, and Zeneca. van den Bergh, Van den Broeck, Cumberbatch, Fossati, Gross, Henry, van der Kwast, Sylvester, Yuan, Schoots, Moldovan, and Marconi have nothing to disclose.

Funding/Support and role of the sponsor: None.

Appendix A Supplementary data

Attached file

Joint first co-authors.

References

  • [1] F. Bratan ,E. Niaf ,C. Melodelima ,et al. Influence of imaging and histological factors on prostate cancer detection and localisation on multiparametric MRI: a prospective study. Eur Radiol. 2013;23 :2019-2029 Crossref
  • [2] J.Y. Kim ,S.H. Kim ,Y.H. Kim ,H.J. Lee ,M.J. Kim ,M.S. Choi. Low-risk prostate cancer: the accuracy of multiparametric MR imaging for detection. Radiology. 2014;271 :435-444 Crossref
  • [3] B. Turkbey ,P.A. Pinto ,H. Mani ,et al. Prostate cancer: value of multiparametric MR imaging at 3 T for detection—histopathologic correlation. Radiology. 2010;255 :89-99 Crossref
  • [4] M. Valerio ,I. Donaldson ,M. Emberton ,et al. Detection of clinically significant prostate cancer using magnetic resonance imaging-ultrasound fusion targeted biopsy: a systematic review. Eur Urol. 2015;68 :8-19 Crossref
  • [5] I.G. Schoots ,M.J. Roobol ,D. Nieboer ,C.H. Bangma ,E.W. Steyerberg ,M.G. Hunink. Magnetic resonance imaging-targeted biopsy may enhance the diagnostic accuracy of significant prostate cancer detection compared to standard transrectal ultrasound-guided biopsy: a systematic review and meta-analysis. Eur Urol. 2015;68 :438-450 Crossref
  • [6] O. Wegelin ,H.H. van Melick ,L. Hooft ,et al. Comparing three different techniques for magnetic resonance imaging-targeted prostate biopsies: a systematic review of in-bore versus magnetic resonance imaging-transrectal ultrasound fusion versus cognitive registration. Is there a preferred technique?. Eur Urol. 2017;71 :517-531 Crossref
  • [7] N. Mottet ,J. Bellmunt ,M. Bolla ,et al. EAU-ESTRO-SIOG guidelines on prostate cancer. (, 2016)
  • [8] I.G. Schoots. Omission of systematic transrectal ultrasound guided biopsy from the MRI targeted approach in men with previous negative prostate biopsy might still be premature. Ann Transl Med. 2016;4 :205 Crossref
  • [9] D. Moher ,A. Liberati ,J. Tetzlaff ,D.G. Altman ,P. Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med. 2009;151 :264-269 W64 Crossref
  • [10] P.F. Whiting ,A.W. Rutjes ,M.E. Westwood ,et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med. 2011;155 :529-536 Crossref
  • [11] J.O. Barentsz ,J. Richenberg ,R. Clements ,et al. ESUR prostate MR guidelines 2012. Eur Radiol. 2012;22 :746-757 Crossref
  • [12] M. Abd-Alazeez ,H.U. Ahmed ,M. Arya ,et al. The accuracy of multiparametric MRI in men with negative biopsy and elevated PSA level—can it rule out clinically significant prostate cancer?. Urol Oncol. 2014;32 :45 45.e17–45.e22
  • [13] M. Abd-Alazeez ,A. Kirkham ,H.U. Ahmed ,et al. Performance of multiparametric MRI in men at risk of prostate cancer before the first biopsy: a paired validating cohort study using template prostate mapping biopsies as the reference standard. Prostate Cancer Prostatic Dis. 2014;17 :40-46
  • [14] O. Belas ,J. Klap ,F. Cornud ,et al. IRM mutiparamétrique de la prostate avant biopsies: la fin des biopsies systématisées?. Prog Urol. 2012;22 :583-589 Crossref
  • [15] M. Brock ,C. von Bodman ,J. Palisaar ,W. Becker ,P. Martin-Seidel ,J. Noldus. Detecting prostate cancer: a prospective comparison of systematic prostate biopsy with targeted biopsy guided by fused MRI and transrectal ultrasound. Dtsch Ärztebl Int. 2015;112 :605-611
  • [16] G.M. Busetto ,E. De Berardinis ,A. Sciarra ,et al. Prostate cancer gene 3 and multiparametric magnetic resonance can reduce unnecessary biopsies: decision curve analysis to evaluate predictive models. Urology. 2013;82 :1355-1362 Crossref
  • [17] A.B. Cheikh ,N. Girouin ,M. Colombel ,et al. Evaluation of T2-weighted and dynamic contrast-enhanced MRI in localizing prostate cancer before repeat biopsy. Eur Radiol. 2009;19 :770-778 Crossref
  • [18] M.S. Choi ,Y.S. Choi ,B.I. Yoon ,et al. The clinical value of performing an MRI before prostate biopsy. Korean J Urol. 2011;52 :572-577 Crossref
  • [19] S. Cirillo ,M. Petracchini ,P. Della Monica ,et al. Value of endorectal MRI and MRS in patients with elevated prostate-specific antigen levels and previous negative biopsies to localize peripheral zone tumours. Clin Radiol. 2008;63 :871-879 Crossref
  • [20] J. Ferda ,J. Kastner ,M. Hora ,et al. A role of multifactorial evaluation of prostatic 3T MRI in patients with elevated prostatic-specific antigen levels: prospective comparison with ultrasound-guided transrectal biopsy. Anticancer Res. 2013;33 :2791-2795 Crossref
  • [21] F. Ganie ,M. Wani ,F. Shaheen ,et al. Endorectal coil MRI and MR-spectroscopic imaging in patients with elevated serum prostate specific antigen with negative TRUS transrectal ultrasound guided biopsy. Urol Ann. 2013;5 :172-178 Crossref
  • [22] A. Grenabo Bergdahl ,U. Wilderäng ,G. Aus ,et al. Role of magnetic resonance imaging in prostate cancer screening: a pilot study within the Göteborg randomised screening trial. Eur Urol. 2016;70 :566-573 Crossref
  • [23] J. Haffner ,L. Lemaitre ,P. Puech ,et al. Role of magnetic resonance imaging before initial biopsy: comparison of magnetic resonance imaging-targeted and systematic biopsy for significant prostate cancer detection. BJU Int. 2011;108 :E171-E178 Crossref
  • [24] N. Hara ,M. Okuizumi ,H. Koike ,M. Kawaguchi ,V. Bilim. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is a useful modality for the precise detection and staging of early prostate cancer. Prostate. 2005;62 :140-147 Crossref
  • [25] E. Hauth ,H. Hohmuth ,C. Cozub-Poetica ,S. Bernand ,M. Beer ,H. Jaeger ,M.R.I. Multiparametric. of the prostate with three functional techniques in patients with PSA elevation before initial TRUS-guided biopsy. Br J Radiol. 2015;88 :20150422 Crossref
  • [26] E.I. Ibrahiem ,T. Mohsen ,A.M. Nabeeh ,Y. Osman ,I.A. Hekal ,M. Abou El-Ghar. DWI-MRI: single, informative, and noninvasive technique for prostate cancer diagnosis. Sci World J. 2012;2012 :973450
  • [27] R. Itatani ,T. Namimoto ,S. Atsuji ,et al. Negative predictive value of multiparametric MRI for prostate cancer detection: outcome of 5-year follow-up in men with negative findings on initial MRI studies. Eur J Radiol. 2014;83 :1740-1745 Crossref
  • [28] J. Iwazawa ,T. Mitani ,S. Sassa ,S. Ohue. Prostate cancer detection with magnetic resonance imaging: is dynamic contrast-enhanced imaging necessary in addition to diffusion-weighted imaging?. Diagn Interv Radiol. 2011;17 :243-248 Crossref
  • [29] T.D. Javali ,D.K. Dwivedi ,R. Kumar ,N.R. Jagannathan ,S. Thulkar ,A.K. Dinda. Magnetic resonance spectroscopy imaging-directed transrectal ultrasound biopsy increases prostate cancer detection in men with prostate-specific antigen between 4-10 ng/mL and normal digital rectal examination: MRSI-directed TRUS biopsy in prostate cancer . Int J Urol. 2014;21 :257-262 Crossref
  • [30] D. Junker ,G. Schäfer ,M. Edlinger ,et al. Evaluation of the PI-RADS scoring system for classifying mpMRI findings in men with suspicion of prostate cancer. BioMed Res Int. 2013;2013 :252939
  • [31] K. Kitajima ,Y. Kaji ,Y. Fukabori ,K.-I. Yoshida ,N. Suganuma ,K. Sugimura. Prostate cancer detection with 3 T MRI: comparison of diffusion-weighted imaging and dynamic contrast-enhanced MRI in combination with T2-weighted imaging. J Magn Reson Imaging. 2010;31 :625-631 Crossref
  • [32] V. Kumar ,N.R. Jagannathan ,R. Kumar ,et al. Potential of 1H MR spectroscopic imaging to segregate patients who are likely to show malignancy of the peripheral zone of the prostate on biopsy . J Magn Reson Imaging. 2009;30 :842-848 Crossref
  • [33] V. Kumar ,N.R. Jagannathan ,R. Kumar ,et al. Transrectal ultrasound-guided biopsy of prostate voxels identified as suspicious of malignancy on three-dimensional1H MR spectroscopic imaging in patients with abnormal digital rectal examination or raised prostate specific antigen level of 4-10 ng/ml . NMR Biomed. 2007;20 :11-20 Crossref
  • [34] T.H. Kuru ,M.C. Roethke ,J. Seidenader ,et al. Critical evaluation of magnetic resonance imaging targeted, transrectal ultrasound guided transperineal fusion biopsy for detection of prostate cancer. J Urol. 2013;190 :1380-1386 Crossref
  • [35] A.P. Labanaris ,K. Engelhard ,V. Zugor ,R. Nützel ,R. Kühn. Prostate cancer detection using an extended prostate biopsy schema in combination with additional targeted cores from suspicious images in conventional and functional endorectal magnetic resonance imaging of the prostate. Prostate Cancer Prostatic Dis. 2010;13 :65-70 Crossref
  • [36] B.W. Lamb ,W.S. Tan ,A. Rehman ,et al. Is prebiopsy MRI good enough to avoid prostate biopsy?. A cohort study over a 1-year period. Clin Genitourin Cancer. 2015;13 :512-517 Crossref
  • [37] Y. Matsuoka ,N. Numao ,K. Saito ,et al. Combination of diffusion-weighted magnetic resonance imaging and extended prostate biopsy predicts lobes without significant cancer: application in patient selection for hemiablative focal therapy. Eur Urol. 2014;65 :186-192 Crossref
  • [38] N. Numao ,S. Yoshida ,Y. Komai ,et al. Usefulness of pre-biopsy multiparametric magnetic resonance imaging and clinical variables to reduce initial prostate biopsy in men with suspected clinically localized prostate cancer. J Urol. 2013;190 :502-508 Crossref
  • [39] V. Panebianco ,F. Barchetti ,A. Sciarra ,et al. Multiparametric magnetic resonance imaging vs. standard care in men being evaluated for prostate cancer: a randomized study. Urol Oncol. 2015;33 17.e1–7
  • [40] V. Panebianco ,A. Sciarra ,M. Ciccariello ,et al. Role of magnetic resonance spectroscopic imaging ([1H]MRSI) and dynamic contrast-enhanced MRI (DCE-MRI) in identifying prostate cancer foci in patients with negative biopsy and high levels of prostate-specific antigen (PSA). La Radiol Med. 2010;115 :1314-1329 Crossref
  • [41] P. Pepe ,A. Garufi ,G. Priolo ,M. Pennisi. Can 3-Tesla pelvic phased-array multiparametric MRI avoid unnecessary repeat prostate biopsy in patients with PSA <10 ng/mL? . Clin Genitourin Cancer. 2015;13 :e27-e30 Crossref
  • [42] A. Petrillo ,R. Fusco ,S.V. Setola ,et al. Multiparametric MRI for prostate cancer detection: performance in patients with prostate-specific antigen values between 2.5 and 10 ng/mL . J Magn Reson Imaging. 2014;39 :1206-1212 Crossref
  • [43] M.R. Pokorny ,M. de Rooij ,E. Duncan ,et al. Prospective study of diagnostic accuracy comparing prostate cancer detection by transrectal ultrasound–guided biopsy versus magnetic resonance (MR) imaging with subsequent MR-guided biopsy in men without previous prostate biopsies. Eur Urol. 2014;66 :22-29 Crossref
  • [44] F. Porpiglia ,F. Russo ,M. Manfredi ,et al. The roles of multiparametric magnetic resonance imaging, PCA3 and Prostate Health Index—which is the best predictor of prostate cancer after a negative biopsy?. J Urol. 2014;192 :60-66 Crossref
  • [45] D. Portalez ,P. Mozer ,F. Cornud ,et al. Validation of the European Society of Urogenital Radiology Scoring System for prostate cancer diagnosis on multiparametric magnetic resonance imaging in a cohort of repeat biopsy patients. Eur Urol. 2012;62 :986-996 Crossref
  • [46] J.P. Radtke ,T.H. Kuru ,S. Boxler ,et al. Comparative analysis of transperineal template saturation prostate biopsy versus magnetic resonance imaging targeted biopsy with magnetic resonance imaging-ultrasound fusion guidance. J Urol. 2015;193 :87-94 Crossref
  • [47] S. Rais-Bahrami ,M.M. Siddiqui ,B. Turkbey ,et al. Utility of multiparametric magnetic resonance imaging suspicion levels for detecting prostate cancer. J Urol. 2013;190 :1721-1727 Crossref
  • [48] P. Rouse ,G. Shaw ,H.U. Ahmed ,A. Freeman ,C. Allen ,M. Emberton. Multi-parametric magnetic resonance imaging to rule-in and rule-out clinically important prostate cancer in men at risk: a cohort study. Urol Int. 2011;87 :49-53 Crossref
  • [49] C. Roy ,R. Pasquali ,A. Matau ,G. Bazille ,H. Lang. The role of diffusion 3-Tesla MRI in detecting prostate cancer before needle biopsy: multiparametric study of 111 patients. J Radiol. 2010;91 :1121-1128 Crossref
  • [50] M. Schmuecking ,C. Boltze ,H. Geyer ,et al. Dynamic MRI and CAD vs. choline MRS: where is the detection level for a lesion characterisation in prostate cancer?. Int J Radiat Biol. 2009;85 :814-824 Crossref
  • [51] A. Sciarra ,V. Panebianco ,S. Cattarino ,et al. Multiparametric magnetic resonance imaging of the prostate can improve the predictive value of the urinary prostate cancer antigen 3 test in patients with elevated prostate-specific antigen levels and a previous negative biopsy. BJU Int. 2012;110 :1661-1665 Crossref
  • [52] A. Sciarra ,V. Panebianco ,M. Ciccariello ,et al. Value of magnetic resonance spectroscopy imaging and dynamic contrast-enhanced imaging for detecting prostate cancer foci in men with prior negative biopsy. Clin Cancer Res. 2010;16 :1875-1883 Crossref
  • [53] E. Squillaci ,G. Manenti ,S. Mancino ,et al. MR spectroscopy of prostate cancer. Initial clinical experience. J Exp Clin Cancer Research. 2005;24 :523-530
  • [54] T. Tamada ,T. Sone ,H. Higashi ,et al. Prostate cancer detection in patients with total serum prostate-specific antigen levels of 4-10 ng/mL: diagnostic efficacy of diffusion-weighted imaging, dynamic contrast-enhanced MRI, and T2-weighted imaging . Am J Roentgenol. 2011;197 :664-670 Crossref
  • [55] C. Testa ,R. Schiavina ,R. Lodi ,et al. Accuracy of MRI/MRSI-based transrectal ultrasound biopsy in peripheral and transition zones of the prostate gland in patients with prior negative biopsy. NMR Biomed. 2010;23 :1017-1026 Crossref
  • [56] J.E. Thompson ,D. Moses ,R. Shnier ,et al. Multiparametric magnetic resonance imaging guided diagnostic biopsy detects significant prostate cancer and could reduce unnecessary biopsies and over detection: a prospective study. J Urol. 2014;192 :67-74 Crossref
  • [57] M. Vinet ,V. Vlaeminck-Guillem ,O. Rouvière ,et al. Le score PCA3 et l’IRM prostatique permettent-ils de sélectionner les patients candidats a une première série de biopsies prostatiques?. Prog Urol. 2013;23 :121-127 Crossref
  • [58] R. Wang ,H. Wang ,C. Zhao ,et al. Evaluation of multiparametric magnetic resonance imaging in detection and prediction of prostate cancer. PLoS One. 2015;10 :e0130207 Crossref
  • [59] Y. Watanabe ,A. Terai ,T. Araki ,et al. Detection and localization of prostate cancer with the targeted biopsy strategy based on ADC map: a prospective large-scale cohort study. J Magn Reson Imaging. 2012;35 :1414-1421 Crossref
  • [60] J.O. Barentsz ,J.C. Weinreb ,S. Verma ,et al. Synopsis of the PI-RADS v2 guidelines for multiparametric prostate magnetic resonance imaging and recommendations for use. Eur Urol.. 2016;69 :41-49 Crossref
  • [61] A.B. Rosenkrantz ,R.P. Lim ,M. Haghighi ,M.B. Somberg ,J.S. Babb ,S.S. Taneja. Comparison of interreader reproducibility of the Prostate Imaging Reporting and Data System and Likert scales for evaluation of multiparametric prostate MRI. AJR Am J Roentgenol. 2013;201 :W612-W618 Crossref
  • [62] R. Renard-Penna ,P. Mozer ,F. Cornud ,et al. Prostate Imaging Reporting and Data System and Likert scoring system: multiparametric MR imaging validation study to screen patients for initial biopsy. Radiology. 2015;275 :458-468 Crossref
  • [63] T. Vache ,F. Bratan ,F. Mege-Lechevallier ,S. Roche ,M. Rabilloud ,O. Rouviere. Characterization of prostate lesions as benign or malignant at multiparametric MR imaging: comparison of three scoring systems in patients treated with radical prostatectomy. Radiology. 2014;272 :446-455 Crossref
  • [64] S.E. Eggener ,K. Badani ,D.A. Barocas ,et al. Gleason 6 prostate cancer: translating biology into population health. J Urol. 2015;194 :626-634 Crossref
  • [65] S. Karram ,B.J. Trock ,G.J. Netto ,J.I. Epstein. Should intervening benign tissue be included in the measurement of discontinuous foci of cancer on prostate needle biopsy? Correlation with radical prostatectomy findings. Am J Surg Pathol. 2011;35 :1351-1355 Crossref
  • [66] T.H. Van der Kwast. Re: should intervening benign tissue be included in the measurement of discontinuous foci of cancer on prostate needle biopsy? Correlation with radical prostatectomy findings. Eur Urol. 2012;61 :220 Crossref
  • [67] H. Singh ,E.I. Canto ,S.F. Shariat ,et al. Predictors of prostate cancer after initial negative systematic 12 core biopsy. J Urol. 2004;171 :1850-1854 Crossref
  • [68] B.M. Mian ,Y. Naya ,K. Okihara ,F. Vakar-Lopez ,P. Troncoso ,R.J. Babaian. Predictors of cancer in repeat extended multisite prostate biopsy in men with previous negative extended multisite biopsy. Urology. 2002;60 :836-840 Crossref
  • [69] H.U. Ahmed ,A. El-Sater Bosaily ,L.C. Brown ,et al. Diagnostic accuracy of multi-parametric MRI and TRUS biopsy in prostate cancer (PROMIS): a paired validating confirmatory study. Lancet. 2017;389 :815-822 Crossref
  • [70] N.L. Hansen ,T. Barrett ,B. Koo ,et al. The influence of prostate-specific antigen density on positive and negative predictive values of multiparametric magnetic resonance imaging to detect Gleason score 7-10 prostate cancer in a repeat biopsy setting. BJU Int. 2017;119 :724-730 Crossref
  • [71] S. Washino ,T. Okochi ,K. Saito ,et al. Combination of PI-RADS score and PSA density predicts biopsy outcome in biopsy naive patients. BJU Int. 2017;119 :225-233 Crossref
  • [72] K.S. Louie ,A. Seigneurin ,P. Cathcart ,P. Sasieni. Do prostate cancer risk models improve the predictive accuracy of PSA screening?. A meta-analysis. Ann Oncol. 2015;26 :848-864 Crossref
  • [73] H.A. van Vugt ,R. Kranse ,E.W. Steyerberg ,et al. Prospective validation of a risk calculator which calculates the probability of a positive prostate biopsy in a contemporary clinical cohort. Eur J Cancer. 2012;48 :1809-1815 Crossref
  • [74] P. Mozer ,M. Roupret ,C. Le Cossec ,et al. First round of targeted biopsies using magnetic resonance imaging/ultrasonography fusion compared with conventional transrectal ultrasonography-guided biopsies for the diagnosis of localised prostate cancer. BJU Int. 2015;115 :50-57 Crossref
  • [75] H. Liddell ,R. Jyoti ,H.Z. Haxhimolla. mp-MRI prostate characterised PIRADS 3 lesions are associated with a low risk of clinically significant prostate cancer—a retrospective review of 92 biopsied PIRADS 3 lesions. Curr Urol. 2015;8 :96-100 Crossref
  • [76] J.C. Weinreb ,J.O. Barentsz ,P.L. Choyke ,et al. PI-RADS Prostate Imaging – Reporting and Data System: 2015, Version 2. Eur Urol. 2016;69 :16-40 Crossref
  • [77] F.V. Mertan ,M.D. Greer ,J.H. Shih ,et al. Prospective evaluation of the Prostate Imaging Reporting and Data System Version 2 for prostate cancer detection. J Urol. 2016;196 :690-696 Crossref
  • [78] A.B. Rosenkrantz ,A. Oto ,B. Turkbey ,A.C. Westphalen. Prostate Imaging Reporting and Data System (PI-RADS), Version 2: a critical look. AJR Am J Roentgenol. 2016;206 :1179-1183 Crossref
  • [79] A. Hoang Dinh ,C. Melodelima ,R. Souchon ,et al. Quantitative analysis of prostate multiparametric MR images for detection of aggressive prostate cancer in the peripheral zone: a multiple imager study. Radiology. 2016;280 :117-127 Crossref
  • [80] A.B. Rosenkrantz ,L.A. Ginocchio ,D. Cornfeld ,et al. Interobserver reproducibility of the PI-RADS Version 2 Lexicon: a multicenter study of six experienced prostate radiologists. Radiology. 2016;280 :793-804 Crossref
  • [81] B.G. Muller ,J.H. Shih ,S. Sankineni ,et al. Prostate cancer: interobserver agreement and accuracy with the revised Prostate Imaging Reporting and Data System at multiparametric MR imaging. Radiology. 2015;277 :741-750 Crossref