In this single-institution study, 67 patients with clinical T2 or T3 disease diagnosed by MRI and who did not undergo neoadjuvant hormonal therapy were included between October 2010 and September 2012. All patients had biopsy-proven PC. Preoperative 3-T MRI was performed to determine the feasibility and extent of a nerve-sparing RP. The following data were collected: age at diagnosis, preoperative PSA level, clinical staging, pathological staging, and Gleason score from biopsy and surgical specimens. The Kobe University Institutional Review Board approved this protocol. Written informed consent was obtained from all participants before inclusion in the study.

MRI was performed by using a 3-T MR scanner (Intera Achieva, Philips Healthcare, Amsterdam, The Netherland) with a phased-array pelvic coil for signal reception. No endorectal coil was used in this study. All patients underwent sagittal, coronal, and axial oblique turbo spin-echo T2-weighted imaging, and all MRI findings were evaluated by a single radiologist (S.T.). Additionally, patients underwent echo-planar diffusion-weighted imaging (DWI) with calculation of apparent diffusion coefficients and dynamic contrast-enhanced imaging. The criteria for a positive cancer finding were as follows: 1) low-intensity imaging in both T2-weighted imaging and apparent diffusion coefficient or 2) enhancing in the early phase but washed out in dynamic imaging. An antiperistaltic agent, 0.5 mg glucagon, was administrated intravenously just before the MRI examinations, and an additional 0.5 mg was administered immediately preceding the acquisition of dynamic contrast enhanced MR. A minimum of 8 weeks was required between the date of the MRI and the previous biopsy to reduce the influence of postbiopsy change in diagnostic accuracy on the basis of Hricak's study [5], in which the median interval between MRI and biopsy was 8 weeks. Prostate biopsy was performed transrectally with 12 cores (6 sextant, 2 from the far peripheral zone [PZ], and 4 cores from the transitional zone [TZ]).

Common criteria was used to determine ECE and local staging grade. Low-intensity lesions on T2-weighted MR images within the PZ of the prostate were considered suspicious for tumor [13]. In the TZ, areas with homogeneous low signal intensity, ill-defined margins, or lack of capsule were interpreted as tumor foci. Asymmetric bulging, an irregular margin, or direct extension of the lesion in the periprostatic fat or NVB was graded as capsular penetration (stage T3a). Signs of seminal vesicle invasion included low intensity in one or both seminal vesicles (stage T3b). The radiological findings were compared with the final operative histological reports.

RARP with lymph node resection was performed with a da Vinci Surgical System (Intuitive Surgical, Sunnyvale, CA, USA) using the standard procedure [14]. Briefly, 4 robotic arms and 2 additional trocars as assistants were used in a 30-degree Trendelenburg position by a transperitoneal approach. Nerve-sparing procedures used an athermal, antegrade interfascial method with minimization of traction. The decision for nerve-sparing was based on MRI findings and preoperative International Index of Erectile Function Questionnaire-25 scores.

The prostate was serially sectioned from base to apex into different levels (depending on the size of the prostate) for histological analysis and labeled as right or left and anterior or posterior apex, midgland, and base. Seminal vesicles were also analyzed separately. All reports were reviewed to determine the presence of ECE and seminal vesicle invasion and to compare staging at the pathologic examination and MRI in each prostate lobe (2 lobes in one patient).

Diagnostic accuracy was measured as sensitivity, specificity, positive predictive value, and negative predictive value. Univariate analysis was calculated for ECE and achievement of nerve-sparing. A p-value of 0.05 or less was considered as statistically significant. A Mann-Whitney U test and a chi-square test were used to determine significant differences. Statistical analysis was conducted with XLSTAT (Addinsoft, New York, NY, USA).

The characteristics of all patients are shown in Table 1. All patients underwent 3-T MRI before RALP (Table 1). The preoperative 3-T MRI results showed that when the samples were divided by prostate side or lobe (right side or left side), 106 of 134 sides were ECE negative and 28 of 134 were ECE positive (Table 2). The representative MRI findings of the positive ECE and negative ECE sides are shown in Fig. 1. Pathologic examination of the surgical specimens in all 67 patients revealed that 50 patients (74.6%) had disease confined to the prostate (pT2) and 17 patients (25.4%) had locally advanced disease (pT3). The pathological stages were pT2a (n=11), pT2b (n=6), pT2c (n=33), pT3a (n=15), and pT3b (n=2) (Table 1).

In the MRI and pathological findings, the overall sensitivity, specificity, and positive predictive value for predicting ECE according to the findings by every prostate side and the negative predictive value were 60.0% (12 of 20 sides), 86.0% (98 of 114 sides), 42.9% (12 of 28 sides), and 92.5% (98 of 106 sides), respectively (Table 2).

On the basis of the 3-T MRI findings, nerve-sparing surgery was performed on 42 of 134 sides (31.3%). Nerve-sparing surgery was achieved in 38.7% of sides with no ECE reported by 3-T MRI. All 41 sides with negative ECE on MRI underwent nerve-sparing surgery with no positive surgical margins (100%). Table 3 shows the nerve-sparing procedure, pathological stage, and positive surgical margin rate in the MRI groups with and without ECE. All values were significantly different (Table 3).