The Impact of Magnetic Resonance Imaging on Prediction of Extraprostatic Extension and Prostatectomy Outcome in Patients with Low-,Intermediate- and High-Risk Prostate Cancer: Try to Find a Standard

Abstract

Purpose:

To investigate the value of multiparametric magnetic resonance imaging (mpMRI) and to predict extracapsular extension (ECE), seminal vesicle (SV) infiltration, and a negative surgical margin (SM) status at radical prostatectomy (RP) for different prostate cancer (PC) risk groups.

Patients and Methods:

In the study, 805 men underwent 3 tesla mpMRI without endorectal coil before MRI/transrectal ultrasonography-fusion guided prostate biopsy. MRIs were analyzed using the prostate imaging reporting and data system. The cohort was classified into risk groups according to National Comprehensive Cancer Network (NCCN) criteria. Of 132 men who subsequently underwent RP, pathologic stage and SM status at RP were used as reference. Retrospectively, we investigated a European Society of Urogenital Radiology (ESUR) score for ECE and SV-infiltration. Statistical analyses included regression analyses, receiver operating characteristics (ROC), and Youden Index to assess an ESUR-score cutoff.

Results:

Area under the curve in ROC curve analyses was 0.82 for ESUR-ECE score to detect pT_3a-disease and 0.77 for ESUR-SV score for pT_3b. Using a cutoff of 4 for ECE and of 2 for SV, the positive predictive value of the ECE-score for harboring pT₃ was 50.0%, 90.0%, and 88.8% for the low-, intermediate- and high-risk cohort. Retrospectively, the use of the ESUR-ECE score preoperatively would have changed the initial surgical plan, according to NCCN criteria, in 31.1% of patients. In the high-risk subgroup, 9/35 (25.7%) patients were correctly assessed as not harboring pT₃ by imaging (ECE score <4), and would have allowed secure robot-assisted radical prostatectomy and nerve-sparing surgery (NSS). When T₃ suspicion on preoperative MRI would be taken into account, intraoperative frozen-sections (IFS) might avoid positive SM in 12/18 high-risk patients and an oncologic secure NSS in 8/20 intermediate-risk patients.

Conclusion:

Prediction of pT₃ disease is crucial to plan NSS and to achieve negative SM in RP. Standardized ECE scoring on mpMRI is an independent predictor of pT₃ and may help to plan RP with oncologic security, even in high-risk patients. In addition, it allows more accurate selection of a subgroup of patients for systematic and MRI-guided IFS.

Introduction

Multiparametric MRI (mpMRI) has shown promising results in the detection and staging of localized prostate cancer (PC).^1,2 In addition, recent studies have reported excellent concordance in Gleason Score (GS) and index lesion prediction for mpMRI and MRI-targeted prostate biopsies compared with radical prostatectomy (RP) specimens.³

Risk group stratification for localized PC, however, as defined by National Comprehensive Cancer Network (NCCN) criteria, incorporates serum prostate-specific antigen (PSA)-level, the GS of the biopsy specimen, and the clinical T-stage based on digital rectal examination (DRE) and lacks mpMRI inclusion.⁴

This concept often leads to a nonnerve-sparing treatment strategy, although 30% to 50% of men in high-risk cohorts have organ-confined disease and thus would be suitable for minimally invasive and nerve-sparing surgery (NSS).^5,6 Somford and associates⁷ and Marcus and colleagues⁸ have been demonstrating the accurate negative and positive predictive values of extraprostatic extension (EPE) for mpMRI in low-risk and high-risk patients, according to D'Amico criteria, and the possibility of changing surgical strategy based on pre-MRI D'Amico risk-grade by mpMRI.

Because RP offers excellent oncologic outcome in patients with pT₃-disease, preoperative MRI should not only focus on prediction of EPE, but also on tailoring surgical concepts for patients and guiding surgeons to achieve negative surgical margins (SM) and nerve sparing. Recently, Petralia and coworkers⁹ demonstrated that MRI can guide intraoperative frozen sections (IFS), and patients with presurgery MRI had one-seventh the risk of positive SM relative to control patients.⁹ Marcus and colleagues⁸ and Petralia and coworkers⁹ did not use a reproducible classification in mpMRI reading of extraprostatic disease, however.

In the present study, we first retrospectively evaluated standardized reading of preoperative mpMRI using a scoring system for extracapsular extension (ECE) and seminal vesicle (SV) invasion, published by the European Society of Urogenital Radiology (ESUR), and defined a cutoff that predicted EPE best.¹⁰

Second, we retrospectively investigated whether findings at preoperative mpMRI would be of additional value for surgical planning of RP. In particular, we investigated how standardized mpMRI reading might be useful to increase the rate of negative SM in intermediate- and high-risk patients, especially in case of ECE-suspicion on MRI. In addition, we investigated the potential role of MRI in the decision to preserve neurovascular bundles (NVB) in the high-risk group, according to NCCN.⁴ We also used the ESUR scoring to define candidates for real-time histologic monitoring of the margin status.

Patients and Methods

Study population

Consecutive patients were enrolled in a prospective database assessing mpMRI with Prostate Imaging Reporting and Data System (PI-RADS) scoring and MRI/transrectal ultrasonography (TRUS)-fusion guided prostate biopsy at Heidelberg University Hospital between October 2012 and September 2014. Institutional Review Board approval was obtained (S011/2011), and all subjects provided written informed consent. Of 805 patients with MRI/TRUS-fusion guided biopsy from October 2012 on, we retrospectively analyzed 132 consecutive patients, who underwent mpMRI, MRI/TRUS-fusion guided biopsy, and subsequent RP at Heidelberg University Hospital. All patients underwent mpMRI because of an elevated PSA level and/or positive DRE (Table 1). Parts of the cohort were analyzed previously.¹¹

Table 1.

Study Population of Included Patients Including Demographics, Magnetic Resonance Imaging Prostate Imaging Reporting and Data System Scoring, Biopsy, and Radical Prosatectomy Specimen Results

Number of men included in analysis	132
Median age, years {interquartile range, IQR}	66 {60;71}
Median prebiopsy PSA level ng/mL {IQR}	8.2 {5.8;15.3}
Suspicious DRE findings (≥T₂), no. (in %)	28 (21.2)
Median prostate volume, cc {IQR}	42.0 {30.0;57.0}
Men with previous prostate biopsy, no. (% of overall)	63 (47.7)
Men without previous biopsy, no. (% of overall)	69 (52.3)
Mean PSA density±SD	0.19±0.06
Men with PIRADS 2–5 lesions on mpMRI, no. (% of overall)	120 (66.7)
Number of lesions and median per patient {IQR}	154; 1{1–2}
Number of lesions in initial biopsy	73
Number of lesions in repeat biopsy	81
Overall PIRADS score 2 lesions, no.	8
Overall PIRADS score 3 lesions, no.	31
Overall PIRADS score 4 lesions, no.	64
Overall PIRADS score 5 lesions, no.	51
Biopsies per patient, median {IQR}	28 {25–30}
Men with significant prostate cancer (Gleason score ≥3+4), no. (% of all cancers)	110 (83.3)
Men with significant prostate cancer (definition according to Epstein et al.¹⁷), no. (% of all cancers)	107 (81.1)
Men with significant prostate cancer (definition according to Wolters et al.¹⁶), no. (% of all cancers)¹⁶	112 (84.8)

IQR=interquatile range; PSA=prostate-specific antigen; DRE=digital rectal examination; SD=standard deviation; PIRADS=prostate imaging reporting and data system; mpMRI=multiparametric magnetic resonance imaging.

Patients were stratified into low-risk (PSA ≤10 ng/mL, combined GS ≤3+3, and clinical stage T_1–2a), intermediate-risk (PSA 10–20 ng/mL, and/or GS ≤4+3, and/or clinical stage T_2b–T_2c), and high-risk (PSA ≥20 ng/mL and/or combined GS ≥4+4, and/or clinical stage T₃), according to NCCN criteria.⁴ Patients in the very high-risk subgroup according to NCCN criteria (T_3–T₄, primary Gleason pattern 5 or >4 cores with GS ≥8) were excluded from the analysis.⁴

Imaging and interpretation

All mpMRI examinations were performed using a 3 tesla-system (Magnetom, Siemens, Erlangen, Germany) using a multichannel surface coil without an endorectal coil. The parameters of mpMRI sequences (T2 weighted imaging (T2w), Diffusion-weighted Imaging (DWI), Dynamic contrast enhancement [DCE]) are presented in Supplementary Table 1 (Supplementary Data available at www.liebertpub.com/end). Imaging analyses were performed or supervised by dedicated uroradiologists with 7 and >10 years of experience in prostate MRI (MR, HPS) according to the 2012 ESUR guidelines.¹⁰ The localization of the lesions was reported using a 27-regions form-sheet.¹² PI-RADS scoring was performed prospectively.

All lesions on mpMRI were retrospectively scored for ECE on T2w and for SV-invasion on T2w and DWI according to the ESUR score, by or under supervision of one dedicated uroradiologist without co-reading (MR) (Fig. 1, Table 2).¹⁰ We also assessed the sum score of both ECE and SV. Radiologists were not blinded to clinical data at initial PI-RADS scoring but at retrospective assessment of ESUR scoring for EPE.

FIG. 1.

Histopathologic correlation of European Society of Urogenital Radiology (ESUR) extracapsular extension score on mpMRI (T2w imaging) on every level.¹⁰ mpMRI=multiparametric magnetic resonance imaging.

Table 2.

European Society of Urogenital Radiology Scoring of Extraprostatic Disease ^*

Criteria	Findings	Score
1. Extracapsular extension	Abutment	1
	Irregularity	3
	Neurovascular bundle thickening	4
	Bulge, loss of capsule	4
	Measurable extracapsular disease	5
2. Seminal vesicles	Expansion	1
	Low T2 Signal	2
	Filling in of angle	3
	Enhancement and impeded diffusion	4

Modified to extracapsular extension and seminal vesicles alone for our cohort, according to Barentsz and associates.¹⁰

Surgical technique

There were 113 RP (84.9%) performed as a robot-assisted RP (RARP), 19 as a retropubic RP. The surgical procedures of RARP and retropubic RP were performed as standard institutional procedures, as previously described or based on previous widely known publications.^13,14 Each RP was performed by one of four experienced surgeons (SAP, MH, BAH, DT, each with at least 4 years of experience, having performed more than 200 RP). Because results of standardized MRI reading were assessed retrospectively, surgeons were unaware of MRI results at RP.

Pathology processing

Analyses were performed by or under supervision of a dedicated uropathologist (WR, 10 years of experience), according to the standards of the international prostate consensus group.¹⁵

Statistical analysis

The patient cohort was analyzed descriptively. For ECE score, SV score, and the sum score of ECE+SV, we calculated receiver operating characteristics (ROC) curve analyses and Youden Index to detect the Youden-selected optimal threshold of the scores for pT₃-detection (Fig. 2).¹⁰

FIG. 2.

Receiver operating characteristics curve analysis and Youden-selected threshold for prediction of pT₃ disease. ECE=extracapsular extension; SV=seminal vesicle; AUC=area under the curve; CI=confidence interval.

Differences in SM status, SV invasion, and possibility of nerve sparing for patients with ECE-score <4 versus ≥4 (Youden-selected optimal threshold) were performed using the Student t test (P value<0.05 stated as significant).

The potential of ECE and SV score as well as other clinical parameters (definitions of significance according to Wolters and coworkers¹⁶ and Epstein and colleagues,¹⁷ tumor diameter, GS, PI-RADS score, and NCCN risk groups) on the prediction of T₃ disease and SM was retrospectively assessed by logistic and linear regression analyses.⁴

Last, we analyzed the predictive potential of the ESUR ECE score on pT₃ disease stratified into risk groups using pre- and post-test probabilities and likelihood ratios, based on probability of T₃ disease from Partin tables and recent studies.^6,18,19 In particular for high-risk patients, we also analyzed the influence of a standardized reading on the Briganti nomogram for organ-confined PC in high-risk patients.²⁰

Results

Clinicopathologic and MRI data of the 132 patients (median age 66.0 years, median prebiopsy PSA level 8.2 ng/mL, median prostate volume 42 cc, PI-RADS distribution of 154 lesions) are detailed in Table 1. Table 3 shows the results of ESUR ECE score, RP technique, and histopathologic specimen stratified into NCCN risk groups. Eighteen patients had low-risk, 79 patients intermediate-risk, and 35 patients high-risk disease. Overall, a positive SM status occurred in 26.5%.

Table 3.

Influence of European Society of Urogenital Radiology Extracapsular Extension Score on T-Stage, Pre- and Post-Test-Probabilities, and Logistic Regression Analysis for Different National Comprehensive Cancer Network Risk-Groups

NCCN risk group	Low risk	Intermediate risk	High risk
Number of patients	18	79	35
Maximum pretest probability of pT₃ disease according to Partin tables and recent publications^6,18,19 in %	18	38	53
Maximum pretest probability of pT₃ disease according to Briganti et al²⁰ in %			70
MRI assessment
No. of patients with ECE score <4 suggesting T₂ disease	14	59	17
No. of patients with ECE score ≥4 suggesting T₃ disease	4	20	18
T-stage in RP specimen
pT₂	16	48	11
pT_3a	2	23	8
pT_3b	0	7	16
pT₄	0	1	0
No. of patients with ECE score <4 and pT₃ disease	1	13	10
Post-test probability of pT₃ disease for ECE <4, based on maximum pretest probability according to Partin tables and recent publications in %	11.7	23.1	38.0
Post-test probability of pT₃ disease for ECE <4, based on maximum pretest probability according to Briganti et al²⁰ in %			56.0
Negative predictive value for ECE <4	92.9	78.0	52.9
No. of patients with ECE score ≥4 and pT₃ disease	2	18	16
Post-test probability of pT₃ disease for ECE ≥4, based on maximum pretest probability according to Partin tables and recent publications in %	38.5	88.5	72.8
Post-test probability of pT₃ disease for ECE ≥4, based on maximum pretest probability according to Briganti et al²⁰ in %			84.6
Positive predictive value for ECE ≥4	50.0	90.0	88.8
Univariate logistic regression analysis for pT₃-disease for ECE <4 versus ≥4 in the same NCCN risk group	4.3 (0.2–90.9) P=0.345	31.9 (6.5–155.4) P<0.001	5.6 (0.9–32.5) P=0.05

NCNN=National Comprehensive Cancer Network; MRI=magnetic resonance imaging; ECE=extracapsular extension; RP=radical prostatectomy.

In the low-risk subgroup, all patients underwent RARP, and NSS was performed in 15 (83.3%). All low-risk patients had negative SM (Table 3). In the intermediate-risk group, 71 (89.9%) patients underwent RARP, and 93.7% underwent NSS. In this subgroup, 39.2% had pT₃ disease and 22.8% had positive SM (Fig. 3). Positive SM occurred in 16.7% with pT₂ disease. In the high-risk cohort (35 patients), overall positive SM occurred in 18.2% of the patients with pT₂ disease and in 62.5% with pT₃ disease (Fig. 3). Twenty-four (68.6%) patients underwent RARP. NSS could only be performed in 9 of 35 (25.7%), all with negative SM.

FIG. 3.

Diagram shows the number of patients stratified into different National Comprehensive Cancer Network (NCCN) risk groups and divided into T₂ (ECE-score <4) or T₃ (ECE-score ≥4) on magnetic resonance imaging (MRI). Each group is substratified into pT stage pT₂ and pT_3/4. Then nerve-sparing or nonnerve-sparing approach, margin status, and details on localization of positive surgical margins are provided. Ipsilateral means on the same side as the MRI-suspicious lesion. Contralateral means on the contralateral side of the MRI-suspicious lesion. RP=radical prostatectomy; IFS=intraoperative frozen sections.

We calculated ROC curves of the ECE, the SV and the sum score of both, as well as Youden-selected thresholds (Fig. 2). For ECE score, the AUC of the ROC curve to predict pT₃ disease was 0.82, and Youden-selected threshold was 4. The AUC for the SV score to predict pT_3b disease was 0.77, and the Youden-selected threshold was 2. For the sum score, the AUC was 0.86 to predict pT₃ disease, and the Youden-selected threshold was 6.

The influence of the ECE score at the Youden-selected cutoff of 4 on the pretest probability to harbor a pT₃ disease, based on maximum probability assessed by the Partin tables, recent publications, and for high-risk patients by a Briganti nomogram, is mentioned in Table 3.^6,18
–20

The positive (PPV) and negative predictive values (NPV) of the ECE score to predict the presence or absence of ECE are also detailed in Table 3. Because the NPV of the ECE score <4 is 52.9% in the high-risk subgroup, we adjusted the threshold in this population to increase patients' oncologic security in terms of negative SM.

Applying a reduced threshold of 3 for the NCCN high-risk cohort, 7 patients had an ESUR ECE score <3 and 28 patients harbored an ESUR ECE score ≥3. PPV and NPV for a cutoff of 3 were 92.3% and 77.8%. All seven patients with an ESUR ECE score <3 had pT₂ disease in the RP specimen. Thus, 7/35 (20.0%) had undergone a nonnerve-sparing RP because of the high-risk NCCN subgroup but were retrospectively probably eligible for NSS because of a ESUR score <3 predicting organ-confined disease.

Figures 3 and 4 demonstrate the possible merit of a standardized MRI on the surgeon's decision to perform NSS. Figure 3 gives a detailed view based on the NCCN risk group. Retrospectively, 20 intermediate-risk patients had pT_3/4 disease, and T₃ was correctly predicted on MRI. Sixteen of them underwent NSS, but 8 had positive SM and 4 of them ipsilaterally to the MRI lesion. Systematic IFS were false-negative in 75% of the patients. In contrast, 46/59 intermediate-risk patients had pT₂ disease that was correct predicted by MRI. A positive SM, however, occurred ipsilaterally to the MRI-lesion in eight patients, who underwent NSS. None of the patients had undergone IFS. In the high-risk subgroup, 12 patients had pT₃ disease and positive SM, although a non-NSS approach was performed. IFS were performed in four cases only and were false-negative.

FIG. 4.

Diagram shows the number of patients stratified into different NCCN risk groups and the potential influence of the ECE score of a preoperatively performed mpMRI on the surgical plan (retropubic RP or robot-assisted radical prostatectomy, nerve-sparing or nonnerve-sparing, IFS). SM=surgical margin.

Table 4 demonstrates the results substratified for the initial surgical plan (NSS vs non-NSS) based on the risk groups. There were 9/35 (25.7%) who had undergone a nonnerve-sparing RP because of high-risk disease, but retrospectively could probably have undergone NSS after ECE score <4 predicting organ-confined disease. Vice versa, 8/19 low- and intermediate-risk patients with pT₃ disease, who had an ECE score ≥4 and thus were suspicious of T₃ preoperatively, underwent NSS and had positive SM.

Table 4.

Logistic and Linear Regression Analyses of Potential Predictive Factors for pT₃ Disease in Radical Prostatectomy

Parameter	Regression coefficient/odds ratio (95% CI), P value
Logistic regression of ESUR ECE score <4 vs ≥4, according to Youden Index	11.8 (4.6;29.9), P<0.001
Logistic regression of ESUR seminal vesicle (SV) score <2 vs ≥2, according to Youden Index	6.3 (2.9;13.6), P<0.001
Logistic regression of ESUR sum score of ECE score and SV score <6 vs ≥6, according to Youden Index	11.4 (4.8;27.0), P<0.001
Linear regression of tumor diameter per 2 mm	2.2 (0.5;5.2), P=0.04
Linear regression of PI-RADS score per 1	2.2 (0.6;5.2), P=0.04
Linear regression of Gleason score per additional pattern	3.6 (0.7;7.0), P=0.041
Logistic regression of nonsignificant vs significant PC, according to Wolters definition	2.0 (0.7;5.5), P=0.210
Logistic regression of nonsignificant vs significant PC, according to Epstein definition	2.9 (1.1;8.0), P=0.03
Logistic regression of low- vs intermediate- and high-risk PC, according to NCCN definition	6.4 (2.8;14.7), P<0.001

CI=confidence interval; ESUR=European Society of Urogenital Radiology; PI-RADS=Prostate Imaging Reporting and Data System; PC=prostate cancer.

Regression analysis of the ESUR ECE score, SV score, the sum score, and clinical parameters is detailed in Table 4. Odds ratio (OR) of logistic regression of an ECE score ≥4 to harbor pT₃ disease was 11.8 (P<0.001), of an ESUR sum score 11.4 (P<0.001). OR of an ECE score ≥3 to predict pT₃ disease was 8.2 (P<0.001) (not shown in a Table). Also, the linear regression of tumor diameter per 2 mm (OR=2.2, P=0.04), PI-RADS score per level (OR=2.2, P=0.04), GS per additional pattern (OR=3.6, P=0.041), and logistic regression analyses of significant PC according to Epstein criteria and intermediate- and high-risk risk groups were significant predictors of pT₃.^4,16,17

Discussion

It has been well established that prevalence of ECE differs greatly between different risk groups from 11.0% to 18.0% in the low-risk group to 33.0% to 53.0% in high-risk patients, depending also on the definition of high-risk disease.^6,18,19 Previous series analyzed the predictive value of mpMRI for PC staging in risk-group stratified populations.^7,8,21
–23 In the study by Marcus and coworkers,⁸ only 20% underwent RP, which renders this study only partially comparable. The prevalence of pT₃ disease was highest in our cohort (43.9%) compared with Cornud and associates (21.0%),²² Somford and colleagues (24.7%),⁷ and Boesen and coworkers (37%).²³ For the intermediate-risk group, our prevalence (38.0%) was lower compared with Somford and colleagues (58.1%),⁷ which might explain our higher NPV (78.0% vs 57.1%). In concordance with recent publications, our NPV was 92.9% in the low-risk subgroup.^7,22

A recent study showed the pretest probability for harboring pT₃ disease in the high-risk cohort to be 53%; PPV is thus of incremental significance in the high-risk cohort.¹⁸ PPV at the Youden-selected threshold of 4 was 90.0% and 88.8% (Table 3) for intermediate- and high-risk disease and therefore comparable with Somford and colleagues,⁷ using the same ESUR scoring system. Furthermore, standardized MRI reading increased the already excellent PPV of 81% published by Cornud and associates.²²

Using ROC curve analyses, the accuracy of the ESUR ECE score (AUC=0.82, Fig. 1) for harboring extraprostatic disease was comparable with recent publications, thus demonstrating that standardized reading of mpMRI has the potential to predict EPE accurately.²³ Our results of the ECE score were comparable with Boesen and coworkers²³ and Somford and colleagues,⁷ who used the ESUR scoring and provided a good prediction with high NPV for patients in low-risk- and high PPV for patients in intermediate- and high-risk groups.

Different existing scoring systems, such as those published by Sala and coworkers²⁴ and Wang and associates²¹ from Memorial Sloan-Kettering Cancer Center, who developed a score that is analogous to the later published PI-RADS score, did not improve the AUC of ESUR ECE score of our cohort (0.70–0.87 in Sala and coworkers,²⁴ 0.76 in Wang and associates²¹).

Regarding clinical relevance, prostate MRI, in contrast to clinical parameters like the NCCN criteria, offers a localized staging.²⁵ This enables the surgeon to sculpt the extent of PC and possible ECE.²⁵ In patients undergoing open RP, MR imaging has been shown to improve the accuracy of the surgeon's decision to resect or preserve the NVB.²⁵

One alternative might be neurovascular structure-adjacent frozen-section examination.²⁶ Schlomm and colleague²⁶ described a significantly less frequent positive SM of 15% versus 22% for all stages and 21% versus 32% for pT_3a, suggesting that a systematic application of IFS has the potential to significantly increase nerve sparing and to reduce positive SM in RP.²⁶ Relevant studies, however, demonstrated a high false-negative rate, potentially resulting in unjustified NSS, and eminent time consumption.^9,27

When IFS might be used in intermediate- and high-risk patients, our results suggest to perform IFS in intermediate-risk patients, when a nerve-sparing approach is desired. When IFS would be performed on the ipsilateral side of the lesion, the rate of positive SM in patients with organ-confined disease on MRI and pT₂ might be reduced by 8/45 (17.8%). In case of suspicion of T₃ on MRI (ECE score ≥4), eight low- and intermediate-risk patients in our cohort might have undergone MRI-guided IFS and thus potentially NSS without compromising oncologic security. Here, preoperative mpMRI and assessment of the ECE score retrospectively might have changed the nerve-sparing approach or may define a subgroup of candidates, who might profit from intraoperative real-time histologic monitoring using IFS.

In the high-risk subgroup, MRI might increase the negative SM rate by guiding IFS when ECE is expected based on preoperative MRI (Table 3). These results are in concordance with Petralia and coworkers,⁹ who could also demonstrate the utility of MRI-guidance of IFS to reduce positive SM.⁹

In a setting without IFS and in a prospective manner of study, in which NSS were cancelled based on a positive ECE score ≥4 in preoperative MRI, the overall positive SM rate in our cohort would have been decreased from 26.5% to 20.5% (McNemar test P=0.013, not shown in a Table) and from 42.4% to 28.8% in the p≥T₃ cohort (McNemar test P=0.013, not shown in a Table).

Retrospectively, based on the ESUR ECE score, our study's initial surgical plan would potentially have been influenced by MRI in 31.1% (24 men in low- and intermediate- but T₃ disease on MRI and 17 men in the high-risk subgroup but T₂ disease on MRI, Tables 3 and 4) of patients. This is comparable to McClure and associates,²⁸ who report a change in surgical plan based on MRI findings in 27% of patients.

Using standardized MRI, upgrading from intermediate risk to high risk occurred more often in our cohort (19.6%) than in the study of Marcus and colleagues⁸ (9.9%), suggesting that standardized mpMRI using the ESUR score leads to more frequent upgrading from intermediate- to high-risk PC.⁸ Vice versa, in 48.6% of patients in the high-risk cohort, an ECE score <4 occurred. Consecutively, pT₂ disease in the RP specimen was detected in 25.7%, which is in concordance with the literature.⁷ In these patients, the NVB could have been spared on the basis of a standardized presurgical MRI and ensured by guided IFS. Most importantly, this change in surgical plan would not appear to be associated with an increase in positive SM. This is comparable to McClure and associates,²⁸ reporting a rate of 37.1% high-risk patients who might undergo a secure NSS.

For preoperative high-risk patients, many centers in Europe still propagate an open retropubic RP.²⁹ In the high-risk cohort, 7/35 (20%) patients had an ESUR ECE score <3 on MRI and harbored pT₂ disease. Consecutively, these patients might have a secure RARP and NSS based on a presurgical MRI. This echoes the results in a recent publication from the Mayo Clinic, considering NSS for high-risk patients but without clear evidence for nonorgan-confined disease.³⁰ For such cases, NSS, compared with wide local excision, does not compromise biochemical progression rates.³⁰ In persistent suspicion of ECE, however, obtaining an IFS might be useful to decrease the occurrence of positive SM.²⁶

Regarding the surgical approach Lightfoot and coworkers³¹ have been demonstrating that the utility of presurgical MRI can decrease positive SM in high-risk PC patients undergoing RARP.³¹ Using the ECE score, we can reflect these results and demonstrate that presurgical MRI can enable an oncologic secure mimimally invasive RARP in high-risk men.

Using pre- and post-test probabilities, we demonstrate the potential influence of mpMRI on surgery planning. In all three risk subgroups, the ECE score changed the post-test probability (Table 3), but only in the intermediate-risk and high-risk subgroups; the OR for ECE score ≥3 (high-risk subgroup) and ECE score ≥4 (intermediate-risk subgroup) was statistically significant. For the overall cohort, univariate logistic regression analysis of ECE score ≥4 for harboring pT₃ disease (OR=11.8, P<0.001) was comparable to the study of Somford and associates⁷ (OR 10.3).⁷ Also, stratification to intermediate- and high-risk subgroups, significant PC according to Epstein and colleagues,¹⁷ and linear regression analysis of tumor diameter per 2 mm, PI-RADS score per level and GS were significant predictors of T₃ disease in univariate analysis. The ESUR score, however, was the strongest predictor of ECE in our cohort and also in the cohort of Somford and associates.⁷

Limitations of our study include first, the fact that assessment of ESUR score for EPE was performed in a retrospective manner. A prospective setting, in which the surgeon is aware of the MRI results and decides to perform IFS or a change of the surgical strategy would provide the real impact of presurgical MRI. Second, another limitation might be the single-arm design of our study. A randomized trial with a control group of patients who did not undergo presurgical MRI would permit us to analyze the additional potential of preoperative MRI in the decision-making process (NSS or wide resection).

Third, we investigated the scores of the two most important clinical parameters, the ECE score and the SV invasion score, but did not include the scores for distal sphincter and bladder neck invasion. To date, however, it has not yet been validated that all parts of the ESUR score are mandatory. Furthermore, ESUR ECE score and SV score reflect both prediction of ECE and SV infiltration, like the previous studies of Wang and colleagues²¹ and Sala and associates,²⁴ therefore offering reproducibility. Fourth, we acknowledge that we suggest MRI-guided IFS for selected subgroups of patients (intermediate-risk when nerve-sparing is desired and high-risk subgroups to achieve negative SM), but that IFS were only performed in a few cases. In addition, the analysis of IFS was not prospectively or systematically done.

Last, we acknowledge that long-term follow-up is missing in our cohort. Our final outcome was pathology related to the RP specimen, not biochemical recurrence and cancer-specific survival. Follow-up studies specifically investigating long-term cancer recurrence rates are needed to determine the true significance of preoperative mpMRI findings on surgical planning. Furthermore, a prospective study, in which the surgeon is consulted preoperatively with knowledge of ECE status on MRI, to validate the retrospective results of potential additional value of presurgical MRI on surgical planning, is in preparation.

Conclusion

Standardized reading of mpMRI using ESUR ECE and SV score offers good prediction of EPE at both RARP and open RP. When using preoperative MRI, surgical planning, depending on the risk group, would have been changed in 31.1% of patients. This cumulates in allowing a secure RARP and NSS in 25.7% of patients in the NCCN high-risk group who have organ-confined disease. In addition, presurgical MRI can help to define subgroups in both intermediate- and high-risk patients, who might benefit from IFS, especially when NSS is desired or T₃ is expected. As long as PPV and NPV are still imperfect, however, MRI staging should be carefully interpreted, and IFS might be performed in case of doubt.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

Abbreviations Used

References

De Rooij

, Hamoen

, Fütterer

, et al. Accuracy of multiparametric MRI for prostate cancer detection: A meta-analysis. Am J Roentgenol, 2014; 202:343–351.

Hoeks

, Barentsz

, Hambrock

, et al. Prostate cancer: Multiparametric MR imaging for detection, localization, and staging. Radiology, 2011; 261:46–66.

Baco

, Ukimura

, Rud

, et al. Magnetic resonance imaging-transectal ultrasound image-fusion biopsies accurately characterize the index tumor: Correlation with step-sectioned radical prostatectomy specimens in 135 patients. Eur Urol, 2015; 67:787–794.

Mohler

, Kantoff

, Armstrong

, et al. Prostate cancer, version 2.2014. J Natl Compr Cancer Netw, 2014; 12:686–718.

Loeb

, Smith

, Roehl

, Catalona

. Intermediate-term potency, continence, and survival outcomes of radical prostatectomy for clinically high-risk or locally advanced prostate cancer. Urology, 2007; 69:1170–1175.

Gandaglia

, Abdollah

, Hu

, et al. Is robot-assisted radical prostatectomy safe in men with high-risk prostate cancer? Assessment of perioperative outcomes, positive surgical margins, and use of additional cancer treatments. J Endourol, 2014; 28:784–791.

Somford

, Hamoen

, Fütterer

, 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; 190:1728–1734.

Marcus

, Rossi

, Nour

, Jani

. The impact of multiparametric pelvic magnetic resonance imaging on risk stratification in patients with localized prostate cancer. Urology, 2014; 84:132–137.

Petralia

, Musi

, Padhani

, et al. Robot-assisted radical prostatectomy: Multiparametric MR imaging-directed intraoperative frozen-section analysis to reduce the rate of positive surgical margins. Radiology, 2015; 274:434–444.

10.

Barentsz

, Richenberg

, Clements

, et al. ESUR prostate MR guidelines 2012. Eur Radiol, 2012; 22:746–757.

11.

Radtke

, Kuru

, 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.

12.

Dickinson

, Ahmed

, Allen

, et al. Magnetic resonance imaging for the detection, localisation, and characterisation of prostate cancer: Recommendations from a European consensus meeting. Eur Urol, 2011; 59:477–494.

13.

Walsh

, Mostwin

. Radical prostatectomy and cystoprostatectomy with preservation of potency. Results using a new nerve-sparing technique. Br J Urol, 1984; 56:694–697.

14.

Hatiboglu

, Teber

, Hohenfellner

. Robot-assisted prostatectomy: The new standard of care. Langenbecks Arch Surg, 2012; 397:343–352.

15.

Egevad

, Srigley

, Delahunt

. International society of urological pathology consensus conference on handling and staging of radical prostatectomy specimens. Adv Anat Pathol, 2011; 18:301–305.

16.

Wolters

, Roobol

, van Leeuwen

, et al. A critical analysis of the tumor volume threshold for clinically insignificant prostate cancer using a data set of a randomized screening trial. J Urol, 2011; 185:121–125.

17.

Epstein

, Walsh

, Carmichael

, Brendler

. Pathologic and clinical findings to predict tumor extent of nonpalpable (stage T1c) prostate cancer. JAMA, 1994; 271:368–374.

18.

Yossepowitch

, Eggener

, Bianco

, et al. Radical prostatectomy for clinically localized, high risk prostate cancer: Critical analysis of risk assessment methods. J Urol, 2007; 178:493–499.

19.

Makarov

, Trock

, Humphreys

, et al. Updated nomogram to predict pathologic stage of prostate cancer given prostate-specific antigen level, clinical stage, and biopsy Gleason score (Partin tables) based on cases from 2000 to 2005. Urology, 2007; 69:1095–1101.

20.

Briganti

, Joniau

, Gontero

, et al. Identifying the best candidate for radical prostatectomy among patients with high-risk prostate cancer. Eur Urol, 2012; 61:584–592.

21.

Wang

, Hricak

, Kattan

, et al. Prediction of organ-confined prostate cancer: Incremental value of MR imaging and MR spectroscopic imaging to staging nomograms. Radiology, 2006; 238:597–603.

22.

Cornud

, Rouanne

, Beuvon

, et al. Endorectal 3D T2-weighted 1mm-slice thickness MRI for prostate cancer staging at 1.5Tesla: Should we reconsider the indirects signs of extracapsular extension according to the D'Amico tumor risk criteria?. Eur J Radiol, 2012; 81:e591–e597.

23.

Boesen

, Chabanova

, Løgager

, et al. Prostate cancer staging with extracapsular extension risk scoring using multiparametric MRI: A correlation with histopathology. Eur Radiol, 2015; 25:1776–1785.

24.

Sala

, Eberhardt

, Akin

, et al. Endorectal MR imaging before salvage prostatectomy: Tumor localization and staging. Radiology, 2006; 238:176–183.

25.

Hricak

, Wang

, Wei

, 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; 100:2655–2663.

26.

Schlomm

, Tennstedt

, Huxhold

, et al. Neurovascular structure-adjacent frozen-section Examination (Neuro-SAFE) increases nerve-sparing frequency and reduces positive surgical margins in open and robot-assisted laparoscopic radical prostatectomy: Experience after 11,069 consecutive patients. 2012; 62:333–340.

27.

Gillitzer

, Thüroff

, Fandel

, et al. Intraoperative peripheral frozen sections do not significantly affect prognosis after nerve-sparing radical prostatectomy for prostate cancer. BJU Int, 2011; 107:755–759.

28.

McClure

, Margolis

, Reiter

, et al. Use of MR imaging to determine preservation of the neurovascular bundles at robotic-assisted laparoscopic prostatectomy. Radiology, 2012; 262:874–883.

29.

Thomas

, Neisius

, Roos

, et al. [Robotic-assisted radical prostatectomy.] (Ger). Urologe A, 2015; 54:178–182.

30.

Ward

, Zincke

, Bergstralh

, et al. The impact of surgical approach (nerve bundle preservation versus wide local excision) on surgical margins and biochemical recurrence following radical prostatectomy. J Urol, 2004; 172:1328–1332.

31.

Lightfoot

, Su

, Sehgal

, et al. Positive surgical margin trends in patients with pathologic T3 prostate cancer treated with robot-assisted radical prostatectomy. J Endourol, 2015; 29:634–639.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.02 MB