EUROPEAN UROLOGY 65 (2014) 610 619

EUROPEAN UROLOGY 65 (2014) 610 619 available at www.sciencedirect.com journal homepage: www.europeanurology.com Prostate Cancer Comparisons of the Perioperative, Functional, and Oncologic Outcomes After Robot-Assisted Versus Pure Extraperitoneal Laparoscopic Radical Prostatectomy Guillaume Ploussard *, Alexandre de la Taille, Morgan Moulin, Dimitri Vordos, Andras Hoznek, Claude-Clément Abbou, Laurent Salomon Department of Urology, Hospital Henri Mondor, 51 Avenue du Maréchal de Lattre de Tassigny, 94010 Créteil, France Article info Article history: Accepted November 22, 2012 Published online ahead of print on December 1, 2012 Keywords: Prostate neoplasm Laparoscopy Prostatectomy Robotic surgery Continence Potency Recurrence Abstract Background: In spite of the increasing use of robot-assisted radical prostatectomy (RALP) worldwide, no level 1 evidence-based benefit favouring RALP versus pure laparoscopic approaches has been demonstrated in extraperitoneal laparoscopic procedures. Objective: To compare the operative, functional, and oncologic outcomes between pure laparoscopic radical prostatectomy (LRP) and RALP. Design, setting, and participants: From 2001 to 2011, 2386 extraperitoneal LRPs were performed consecutively in cases of localised prostate cancers. Intervention: A total of 1377 LRPs and 1009 RALPs were performed using an extraperitoneal approach. Outcome measurements and statistical analysis: Patient demographics, surgical parameters, pathologic features, and functional outcomes were collected into a prospective database and compared between LRP and RALP. Biochemical recurrence free survival was tested using the Kaplan-Meier method. Mean follow-up was 39 and 15.4 mo in the LRP and RALP groups, respectively. Results and limitations: Shorter durations of operative time and of hospital stay were reported in the RALP group compared with the LRP group ( p < 0.001) even beyond the 100 first cases. Mean blood loss was significantly lower in the RALP group ( p < 0.001). The overall rate and the severity of the complications did not differ between the two groups. In pt2 disease, lower rates of positive margins were reported in the RALP group ( p = 0.030; odds ratio [OR]: 0.396) in multivariable analyses. The surgical approach did not affect the continence recovery. Robot assistance was independently predictive for potency recovery ( p = 0.045; OR: 5.9). Survival analyses showed an equal oncologic control between the two groups. Limitations were the lack of randomisation and the short-term follow-up. Conclusions: Robotic assistance using an extraperitoneal approach offers better results than pure laparoscopy in terms of operative time, blood loss, and hospital stay. The robotic approach independently improves the potency recovery but not the continence recovery. When strict indications of nerve-sparing techniques are respected, RALP gives better results than LRP in terms of surgical margins in pathologically organ-confined disease. Longer follow-up is justified to reach conclusions on oncologic outcomes. # 2012 European Association of Urology. Published by Elsevier B.V. All rights reserved. * Corresponding author. Tel. +33 1 49 81 25 53; Fax: +33 1 49 81 25 64. E-mail address: g.ploussard@gmail.com (G. Ploussard). 0302-2838/$ see back matter # 2012 European Association of Urology. Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.eururo.2012.11.049

EUROPEAN UROLOGY 65 (2014) 610 619 611 1. Introduction Radical prostatectomy (RP) is a standard treatment for localised prostate cancer. The first laparoscopic radical prostatectomy (LRP) was performed in 1991 and thought not to be feasible because of the excessive operative time [1,2]. However, in the following years, the development of minimally invasive surgery was driven by work in Europe. Some centres can now report considerable experience and are able to standardise the technique. Experienced surgeons have described the various advantages of laparoscopy [3,4]. And yet LRP remains a technically demanding procedure and requires a learning curve [4,5]. These difficulties and the emergence of robotic assistance that improves precision led laparoscopic urologists to develop the technique of robotassisted radical prostatectomy (RALP) [6 8]. One of the purposes of the robotic assistance was to reduce the learning curve, even in laparoscopically naive surgeons, without sacrificing the oncologic standards established by the open approach [8,9]. Recent reviews and meta-analyses of the literature recently highlighted the potential benefit of RALP regarding the functional outcomes [10,11]. In a recent metaanalysis, Tewari et al. also found that total perioperative complication rates were higher for LRP than for RALP [12]. Authors highlighted that the lack of randomised controlled trials were drawbacks in all published studies. The two first controlled trials comparing LRP and RALP were recently published using a transperitoneal approach [13,14]. Authors reported better functional results in terms of potency favouring RALP in both series and only in terms of continence recovery in one series. However, only a few patients were included, limiting the power of analysis in comparing low events rates such as positive margins, operative complications, and severe incontinence. Most of these series reported short-term experience with the transperitoneal (but not the extraperitoneal) approach. The aim of our study was to compare the perioperative parameters and the functional and oncologic outcomes between pure LRP and RALP using an extraperitoneal approach and performed in a high-volume laparoscopy reference centre. 2. Patients and methods 2.1. Patient selection Between July 2001 and December 2011, 2386 consecutive men underwent a LRP including 1377 pure LRPs and 1009 RALPs. Indications of surgery were identical in the cases of pure LRP or RALP. We performed the first LRP and the first RALP in our department in 2001, and we have been performing RALP routinely since 2006. The respective rates of LRP and RALP procedures for RP were 95.6% and 4.4% before 2006, 55.9% and 44.1% during the period 2007 2009, and have been 4.6% and 95.4% since 2010. Most of the LRPs were performed by three senior surgeons (L.S., C.C.A., A.D.L.T.), and the total number of RALPs by two senior surgeons (C.C.A., A.D.L.T.) who had performed >100 LRPs before starting the RALP learning curve. Mean follow-up in our cohort was 50.4 mo (1 138 mo). Mean follow-up was 39 and 15.4 mo in the LRP and RALP groups, respectively. The study was carried out in accordance with our local good clinical practice rules. 2.2. Surgical procedure The da Vinci system including three robot arms and a single console (first generation) was used for all RALPs. We described the surgical technique andthedifferent stepsofthesurgery previously [15]. The bladder neck was incised circumferentially with an attempt to spare the it except for highrisk prostate cancers at risk of seminal vesicle invasion (assessed by preoperative magnetic resonance imaging [MRI]). A standard lymphadenectomy was performed prior to the completion of the vesicourethral anastomosis in patients whose Gleason score was >6 and/or prostate-specific antigen (PSA) was >10 ng/ml. A running vesicourethral anastomosis was then performed. At the beginning of the experience, a 3-0 polyglactin suture on a 5/8 circle tapered needle was used. For 2 yr we used a bidirectionalbarbed suture to perform the running anastomosis. The anastomosis started with a posterior reconstruction of the rhabdosphincterasdescribed byrocco. An anteriorreconstruction was performed suspending the anastomosis to the Santorini plexus. Preoperatively, potent low- or intermediate-risk patients underwent a nerve-sparing procedure. An intrafascial dissection as a nerve-sparing procedure could be proposed to very low-risk patients who were potent preoperatively. Very low-risk prostate cancer was defined by a clinical T1c cancer with favourable factors including a PSA < 10 ng/ml, a moderate extent of cancer on positive cores, few cores involved with cancer, and favourable MRI findings (it1 or small it2 cancer). The urethral catheter was usually removed on postoperative day 7 with no cystogram. 2.3. Database and statistical analysis Data were collected prospectively into a database by a clinical research assistant (MM) including preoperative clinical and biologic characteristics, surgical data, and postoperative parameters. Pathologic assessments of RP specimens by a senior pathologist were recorded. After fixation, the apex and base (3-mm-thick slices) were removed from each specimen and examined by the cone method. The prostate body was step-sectioned at 3-mm intervals perpendicular to the long axis (apicalbasal) of the gland. Positive surgical margins were defined by the presence of tumour tissue on the inked surface of the specimen. All patients prospectively completed self-administered questionnaires concerning their quality of life (European Organisation for Research and Treatment of Cancer QLQ-C30) and their voiding and sexual disorders (International Index of Erectile Function [IIEF]-5), preoperatively and at 1, 3, 6, 12, and 24 mo after RP. Potency was defined as the ability to achieve an erection sufficient for penetration (full erections or diminished erections are routinely sufficient for intercourse) with or without the use of a phosphodiesterase type 5 enzyme inhibitor (excluding cases with intracavernous injection of prostaglandin E or vacuum). Urinary continence was assessed by questionnaires and defined as the absence of pads (strict urinary continence). Continence and potency results were studied in the overall cohort regardless of the continence and potency status before surgery. Phosphodiesterase type 5 enzyme inhibitors were systematically proposed at patient discretion. The use of oral erectogenic medications was not statistically different between LRP and RALP groups. Biochemical recurrence was defined as any detectable serum PSA (>0.2 ng/ml). Perioperative complications were noted and reported according to the updated Clavien classification [16]. The qualitative data were tested using the chi-square or the Fisher test as appropriate. The quantitative data were analysed by the Mann-Whitney test. Multivariable analyses used a regression logistic model. In these multivariable analyses, we have only included the patients operated beyond the 100 first cases of each procedure and

612 EUROPEAN UROLOGY 65 (2014) 610 619 operated on by the two surgeons performing both LRP and RALP (A.D.L.T. and C.C.A.). Time-dependent variables were compared using the Kaplan-Meier method and a log-rank test. A double-sided p value <0.05 was considered statistically significant. All data were analysed using SPSS v.13.0 software (IBM Corp., Armonk, NY, USA). 3. Results 3.1. Perioperative parameters Table 1 shows the preoperative characteristics of the patient s cohort according to the surgical approach. The two groups (LRP vs RALP) were comparable in terms of PSA, body mass index, age, and clinical stage. High-grade prostate cancers involving a greater number of cores were more frequently reported in the RALP cohort ( p < 0.001). Shorter durations of operative time (129 vs 175 min) and of hospital stay (4.0 vs 5.7 d) were reported in the RALP group compared with the LRP group ( p < 0.001) even beyond the 100 first cases of each procedure (Table 2). Mean blood loss was significantly lower in the RALP group (515 vs 800 ml; p < 0.001). However, bladder catheterisation was longer in the RALP group (8.0 vs 7.2 d). The overall rate and the severity (defined by the Clavien classification) of complications did not differ between the two groups (Table 2). Detailed complications in each group are listed in Table 3. 3.2. Pathologic findings, surgical margins, and follow-up data The results of pathologic findings are listed in Table 4. Findings in RP specimens did not differ between the two Table 1 Preoperative characteristics of the overall cohort and comparisons between laparoscopic radical prostatectomy and robot-assisted radical prostatectomy subgroups Overall cohort n = 2386 LRP n = 1377 RALP n = 1009 p value Age, yr, mean (IQR) 62.7 (9.9) 62.7 62.7 1.00 BMI, kg/m 2, mean (IQR) 26.5 (4.6) 26.6 26.5 0.648 PSA, ng/ml, mean (IQR) 9.6 (4.8) 9.8 9.2 0.123 No. of positive 4.1 (3.0) 3.9 4.5 <0.001 cores, mean (IQR) Clinical stage, % T1c 81.3 81.0 81.8 T2a b 15.9 16.2 15.6 T2c T3 2.7 2.8 2.6 >T1c 18.6 19.0 18.1 0.764 Biopsy GS, % <0.001 6 63.4 65.7 60.1 7 30.9 29.4 33.0 8 10 5.7 4.9 6.9 BMI = body mass index; GS = Gleason score; LRP = laparoscopic radical prostatectomy; RALP = robot-assisted radical prostatectomy; IQR = interquartile range. groups in terms of Gleason score and ptnm stage ( p = 0.488 and 0.788, respectively). Patients operated on by RALP have smaller glands compared with their counterparts operated on by LRP (47.5 vs 53.0 g; p < 0.001). Overall, in univariable analysis, the rate of positive surgical margins was significantly higher in the RALP group (31.3 vs 26.6%; p = 0.038), especially in the case of pt3 cancer ( p = 0.030). Multivariable analysis is shown in Table 5. Classical prognostic factors were significantly associated with positive margins: PSA, ptnm stage, and Gleason score. The prostate volume was inversely correlated with the risk Table 2 Comparisons of intraoperative and postoperative data, complications between the laparoscopic radical prostatectomy and robotassisted radical prostatectomy groups Overall cohort n = 2386 LRP n = 1377 RALP n = 1009 p value Beyond the learning curve y p value Perioperative data Operative time, min, mean (IQR) 155.8 (60.0) 175.5 128.9 <0.001 <0.001 Blood loss, ml, mean (IQR) 680.3 (695.0) 800.3 515.4 <0.001 <0.001 Bladder catheterization, d, mean (IQR) 7.5 (4.0) 7.2 8.0 <0.001 <0.001 Hospital stay, d, mean (IQR) 5.0 (3.0) 5.7 4.0 <0.001 <0.001 Transfusion rate, % 3.9 4.7 2.9 0.084 0.235 Lymph node excision, % 44.4 43.8 45.4 0.488 0.441 Nerve-sparing procedure, % No 22.4 24.1 20.3 0.002 0.249 Unilateral 11.4 13.1 9.2 Bilateral 66.1 62.9 70.5 Complications Clavien 0 95.7 96.0 95.3 0.756 0.797 1 0.7 0.6 0.7 2 3.3 3.1 3.6 3 0.0 0.0 0.0 4 0.2 0.2 0.2 5 0.0 0.0 0.1 Anastomosis leakage, % 6.5 9.7 2.3 <0.001 <0.001 Anastomosis stenosis, % 1.3 1.7 0.7 0.081 0.036 LRP = laparoscopic radical prostatectomy; RALP = robot-assisted radical prostatectomy; IQR = interquartile range. Comparisons were made among the overall cohort and after excluding patients operated on during the 100 first cases of each procedure. y After excluding the first 100 LRPs and the first 100 RALPs of each surgeon.

EUROPEAN UROLOGY 65 (2014) 610 619 613 Table 3 Detailed surgical and medical complications in the two laparoscopic radical prostatectomy and robot-assisted radical prostatectomy groups LRP n = 1377 RALP n = 1009 Medical complications Death 1 Urinary infection 32 32 Fever 1 8 Phlebitis 2 Pulmonary embolus 1 2 Atelectasia 1 Pneumonia 6 4 Ill-being 5 3 Angor 1 1 Threat syndrome 3 1 Myocardial infarction 2 3 Renal insufficiency 2 3 Retina detachment 1 Overall, % 4.1 5.9 Surgical complications Haemorrhage 6 4 Rectal injury 11 3 Epigastric injury 4 3 Wound complications Haematoma 9 14 Abscess 3 9 Retzius collection 20 23 Lymphorrhea 1 2 Lymphocele 5 9 Haematuria 13 19 Anastomosis leakage 22 14 Ileus 2 4 Neurapraxia 1 1 Bowel injury 1 Overall, % 6.8 10.5 LRP = laparoscopic radical prostatectomy; RALP = robot-assisted radical prostatectomy. of positive margins ( p = 0.004). The rate of positive surgical margins increased over time and was 1.9-fold higher in the period 2010 2011 compared with the earlier period ( p = 0.045). Surgical experience continuously improved the margin status even beyond the 100 first cases of each procedure. The rate of surgical margins decreased by 2.6- fold after 500 procedures ( p < 0.001). In multivariable analysis, RALP was not associated with an increased rate of positive margins. On the contrary, we observed a trend toward lower rates of positive margins in the RALP group compared with LRP, but differences did not reach significance. This difference was significant in the subgroup of pt2 cancers revealing RALP as the favourable factor independently associated with better oncologic control of margins in organ-confined disease ( p = 0.030; odds ratio [OR]: 0.396). In pt3 cancers, the type of procedure (LRP vs RALP) did not affect the rate of surgical margins in multivariable analysis ( p = 0.619). Only the PSA level and surgical experience were independent predictors of positive margins in pt3 cancers ( p < 0.001 and p < 0.001, respectively). Survival curves stratified by the surgical approach were not significantly different in the overall cohort and in pt2, pt3a, or pt3b subgroups (log-rank tests in Table 4) showing an equal short-term oncologic control. According to the D Amico preoperative risk groups (Fig. 1), biochemical recurrence free survival curves were not significantly different between the two procedures in the low-risk (log-rank test: p = 0.672), intermediate-risk ( p = 0.928), and high-risk groups ( p = 0.413). The surgical volume had a significant impact on the recurrence-free survival in the LRP group with better oncologic outcomes after 300 procedures compared with the 300 first LRP procedures ( p = 0.027; Fig. 2). No difference was reported in the RALP group ( p = 0.132). Overall, 221 and 92 salvage treatments were reported in the LRP (16.0%) and RALP (9.1%) groups, respectively. 3.3. Continence In univariable analysis, the rate of continence was significantly in favour of RALP at each postoperative visit ((Fig. 3; p < 0.001). Results of continence recovery in the overall cohort are shown in Figure 3. When considering only patients operated on beyond the learning curve of each surgeon, differences remained significant; after 6 and 12 mo, the rate of continence was 59% and 48.8% in the cases of LRP as compared with 72% and 83.6% in the cases of RALP. In the multivariable analysis shown in Table 6, the only factor independently associated with a better continence recovery was age ( p = 0.002) at each time point. Table 6 illustrates the multivariable analysis at the 12-mo visit. Surgical experience, nerve-sparing surgery, and surgical approach (LRP vs RALP) were not independent predictors for short-term (at 1, 3, and 6 mo) or long-term (at 12 and 24 mo) continence recovery. Surgical treatment for persistent incontinence was more frequent in the LRP group as compared with the RALP group ( p < 0.001). Use of the Macroplastique injection, adjustable continence therapy balloon, suburethral sling, and artificial sphincter were reported in 3, 10, 17, and 13 cases in the LRP group versus 0, 2, 5, and 0 cases in the RALP group. 3.4. Potency In univariable analysis, the rate of potency was significantly in favour of RALP at each postoperative visit. This difference remained significant in a subgroup of patients undergoing a bilateral nerve-sparing preservation (Fig. 4; p < 0.001). After 6 and 12 mo, 20% and 31.6% of patients were potent after LRP compared with 42% and 57.7% after RALP, respectively. In the multivariable analysis shown in Table 6, age ( p = 0.001), nerve-sparing surgery ( p = 0.033; OR: 3.9), and RALP ( p = 0.045; OR: 5.9) were significant independent predictors of potency recovery 12 mo after surgery. These factors were also associated with potency at each medical visit during follow-up. Surgical experience, the surgeon, and the date of intervention were not associated with this functional outcome. The evolution of IIEF-5 over time is reported in Table 7. Only scores from patients undergoing a bilateral nervesparing surgery were reported. The scores at baseline were comparable in the two groups. After surgery, scores were

614 EUROPEAN UROLOGY 65 (2014) 610 619 Table 4 Comparisons of pathological features and follow-up data between the two groups (LRP versus RALP). Comparisons have been made among the overall cohort and after excluding patients operated on during the 100 first cases of each procedure. Biochemical recurrence has been tested using a log-rank test Overall cohort n = 2386 LRP n = 1377 RALP n = 1009 p value Beyond the learning curve * p value Pathologic findings Prostate weight, g, mean (IQR) 50.7 (24.0) 53.0 47.5 <0.001 <0.001 Specimen GS, % 6 28.0 27.0 29.2 0.488 0.997 7 61.4 62.2 60.4 8 10 10.6 10.8 10.4 Pathologic stage, % pt0 0.7 0.5 0.8 0.788 0.208 pt2a 10.5 10.2 10.9 pt2b 2.0 2.5 1.4 pt2c 45.4 45.8 44.8 pt3a 31.1 30.3 32.2 pt3b pt4 10.2 10.5 9.8 Positive surgical margins, % Overall 28.6 26.6 31.3 0.038 0.013 pt2 17.9 16.8 19.6 0.513 0.391 pt3 43.7 41.1 47.4 0.030 0.048 Follow-up data PSA failure, % <0.001 Overall 14.7 18.0 10.3 0.753 y 0.820 y pt2 6.2 7.9 3.7 0.794 y 0.502 y pt3 27.5 33.5 19.7 0.663 y 0.979 y Follow-up, mo, mean (IQR) 29.1 (36.6) 39.0 15.4 <0.001 <0.001 GS = Gleason score; LRP = laparoscopic radical prostatectomy; PSA = prostate-specific antigen; RALP = robot-assisted laparoscopic prostatectomy; IQR = interquartile range. * After excluding the first 100 LRPs and the first 100 RALPs of each surgeon. y Log-rank test. higher in the RALP group as compared with those reported in the LRP at each time point. In the overall cohort, comparisons confirmed this significant difference at each time point favouring RALP. 4. Discussion The use of the robotic system reduces both the difficulty in performing complex laparoscopic techniques and the learning curve compared with the pure laparoscopic procedure. As noted earlier, the practice of the pure LRP requires a steep learning curve [17,18]. Experience can improve pathologic and operative outcomes well beyond the initial learning curve with a plateau at 200 250 procedures [19]. The learning curve for RALP was thoroughly studied, and surgical teams were able to accomplish comparable operative times after 12 or 18 cases [20,21]. Some studies comparing pure LRP and RALP were also published and show interesting postoperative results for RALP [22 24]. It is interesting to notice that each of the parameters operative, pathologic, or functional requires a different learning curve that should be assessed separately in LRP and RALP procedures. The operative time should not be the single variable evaluated to define the learning curve and the surgeon s expertise. However, due to a lack of large randomised controlled trials, differences in patient characteristics or surgical experience might explain differences in outcomes between the surgical approaches. The two first controlled trials comparing LRP and RALP using a transperitoneal approach were recently published [13,14]. Authors reported better functional results in terms of potency favouring RALP in both series, and only in terms of continence recovery in one series. However, few patients were included, limiting the power of analysis in comparing low events rates such as positive margins, operative complications, and severe incontinence. Most of these series reported the midterm experience of the transperitoneal (but not extraperitoneal) approach. Concerning the extraperitoneal LRP, Rozet et al. showed equivalent operative, postoperative, and pathologic results when comparing the extraperitoneal and the transperitoneal approach [24]. Stolzenburg and co-workers also reported interesting results regarding extraperitoneal LRP in terms of operative parameters and surgical margin rate [23]. The extraperitoneal approach proved to be a safe and reproducible procedure, with a fast recovery after surgery [25]. Our results were in line with those published by transperitoneal surgeons. RALP offered advantages concerning the perioperative parameters in terms of hospital stay, operative time, and blood loss, even beyond the 100 first cases of each surgeon. The overall rate and the severity of complications did not differ between the two groups. However, there was a trend towards lower rates of anastomotic complications in favour of RALP even beyond the 100 first cases of each procedure. Robotic assistance offers technical parameters that may improve the anastomotic procedure: a magnified three-dimensional (3D) visual field, a greater range of instrument motion, and a minimisation of tremor. Advantages of RALP regarding decreased adverse events and severity of complications

Table 5 Predictive factors for positive surgical margins in multivariable analysis: overall cohort and pt2 subgroup Surgical margins p value OR 95% CI Overall cohort Age 0.497 PSA <0.001 Prostate volume 0.004 pt2 disease <0.001 0.478 0.37 0.62 GS 6 Ref. 1 7 <0.001 2.176 1.56 3.04 8 10 <0.001 3.665 2.29 5.86 Date of intervention Before 2006 Ref. 1 2007 2009 0.169 1.355 0.88 2.09 2010 2011 0.045 1.887 1.02 3.51 Surgeon experience beyond the learning curve 100 300 Ref. 1 300 500 0.035 0.724 0.54 0.98 >500 <0.001 0.382 0.26 0.57 Nerve preservation 0.130 1.254 0.94 1.68 Surgeon 0.407 0.878 0.65 1.19 Procedure Pure laparoscopy Ref. 1 Robot assisted 0.057 0.559 0.31 1.02 pt2 cancers Age 0.283 PSA 0.063 Prostate volume <0.001 GS 6 Ref. 1 7 <0.001 2.722 1.85 4.01 8 10 0.657 1.340 0.37 4.88 Date of intervention Before 2006 Ref. 1 2007 2009 0.253 1.426 0.78 2.62 2010 2011 0.026 2.715 1.13 6.54 Surgeon experience beyond the learning curve 100 300 Ref. 1 300 500 0.121 0.709 0.46 1.10 >500 0.001 0.387 0.22 0.68 Nerve preservation 0.137 1.484 0.88 2.50 Surgeon 0.692 0.917 0.60 1.41 Procedure Pure laparoscopy Ref. 1 Robot assisted 0.030 0.396 0.17 0.91 CI = confidence interval; GS = Gleason score; PSA = prostate-specific antigen. EUROPEAN UROLOGY 65 (2014) 610 619 615 [(Fig._1)TD$FIG] were not found. No significant difference in rectal injuries was detected. Nevertheless, RALP demonstrated a significant improvement of perioperative parameters, shortening the duration of hospitalisation. Regarding the continence rate and the time to continence, RALP performed better than LRP in univariable models as highlighted by the recent meta-analysis from Ficarra et al. [11]. Nevertheless, the prevalence of urinary incontinence after prostatectomy is highly influenced by numerous parameters such as patient characteristics, surgeon experience, or continence definitions (no pad vs safety pad). Drawbacks of a systematic analysis must be taken into consideration such as the impossibility of controlling surgical skills and individual surgeon factors [11]. The integration of all potential predictive factors for continence recovery is mandatory before drawing major conclusions. Thus, in line with the prospective trial Fig. 1 Biochemical recurrence free survival (RFS) curves stratified by the procedure (laparoscopic radical prostatectomy [LRP] vs robot-assisted radical prostatectomy [RALP]) according to the preoperative D Amico risk group: low risk (log-rank test: p = 0.672), intermediate risk ( p = 0.928), and high risk ( p = 0.413).

616 [(Fig._2)TD$FIG] EUROPEAN UROLOGY 65 (2014) 610 619 [(Fig._3)TD$FIG] Fig. 3 Continence recovery rates at each medical visit in the laparoscopic radical prostatectomy and the robot-assisted prostatectomy groups (overall cohort). The p values at each time point were preoperative ( p = 0.213), 1 mo ( p = 0.191), 3 mo ( p = 0.019), 6 mo ( p = 0.018), 12 mo ( p = 0.177), and 24 mo ( p = 0.024). Fig. 2 Biochemical recurrence free survival (RFS) curves stratified by the surgical volume of each surgeon (100 300, 300 500, >500 procedures) for the laparoscopic radical prostatectomy (LRP) and the robot-assisted radical prostatectomy (RALP) groups. Curves were significantly different after 300 procedures in the LRP group (log-rank test: p = 0.027) but not in the RALP group ( p = 0.132). published by Asimakopoulos et al., we did not find any significant improvement of continence recovery in favour of RALP after integrating all potential confounding factors in multivariable analysis [14]. The impact of anterior or posterior reconstruction on functional outcomes could not be studied in our database. However, a recent review did not find a benefit of such reconstructions on shortening the time to continence recovery [26]. A recent meta-analysis based on four studies comparing potency recovery after LRP and RALP only reported a nonstatistical trend in favour of RALP [11,27,28]. Theonly two prospective studies highlighted better functional results in terms of potency favouring RALP. However, the limited number of patients included did not allow definitive conclusions. Comparisons are also subject to debate because the incidence of potency recovery is influenced by numerous factors. Our multivariable results were in line with those published in the two level 2 studies [13,14]. The use of the robot was significantly linked with a better recovery independently of surgical experience, the surgeon, the date of the intervention, and the practice of neurovascular preservation. After 6 and 12 mo, 20% and 31.6% of patients were potent after LRP compared with 42% and 57.7% after RALP, respectively. This difference remained significant after a 2-mo followup. Our surgical technique is based on a retrograde dissection using low-intensity bipolar cautery and clips. We did not perform a strict athermal dissection that might improve potency recovery [29].However,theuseof a cautery-free technique has not definitively proved its superiority and needs further evaluation. The fourth robot arm potentially provides an adequate exposition avoiding tractions of the bundles that may hinder potency outcomes. We used a three-arm robot, and our experience with the fourarm one is limited. However, the extraperitoneal approach is surely not the best approach to use this fourth arm extensively due to the limited space of the retroperitoneum. Limitations of our report on functional outcomes were that continence and potency require a short-term follow-up. Our findings from the RALP cohort need to be confirmed. Another drawback of the study is that our evaluation of the plane of neurovascular bundle dissection is subjective and not based on the pathologic outcome. Obtaining the intrafascial plane may become more frequent and easier with the robotic approach. The first statistical analysis found higher positive margin rates in the RALP cohort compared with those reported in the LRP, especially in non organ-confined disease. Several factors with an impact on the risk of margins might explain these differences. The RALP cohort included patients at higher risk of positive margins. Those patients had a significantly smaller volume of gland, frequently showed an aggressive cancer on biopsy cores, and were more often operated on using nerve-sparing techniques. When all factors potentially linked with the risk of margins were taken into account in a multivariable

Table 6 Predictive factors for urinary continence and potency in multivariable analysis: logistic regression model assessing the factors predictive for continence and potency recovery 12 mo after surgery p value OR 95% CI EUROPEAN UROLOGY 65 (2014) 610 619 617 [(Fig._4)TD$FIG] Continence Age 0.002 PSA 0.746 Prostate volume 0.524 pt2 disease 0.393 0.782 0.44 1.37 GS 6 Ref. 1 7 0.751 0.915 0.53 1.60 8 10 0.699 1.239 0.42 3.68 Date of intervention Before 2006 Ref. 1 2007 2009 0.981 0.991 0.47 2.08 2010 2011 0.387 0.572 0.16 2.03 Surgeon experience beyond the learning curve 100 300 Ref. 1 300 500 0.230 0.678 0.36 1.28 >500 0.108 0.532 0.25 1.15 Nerve preservation 0.983 0.993 0.54 1.81 Surgeon 0.514 1.295 0.60 2.82 Procedure Pure laparoscopy Ref. 1 Robot assisted 0.253 2.079 0.59 7.29 Potency Age 0.001 PSA 0.085 Prostate volume 0.943 pt2 disease 0.630 0.802 0.33 1.97 GS 6 Ref. 1 7 0.489 1.358 0.57 3.24 8 10 0.817 1.246 0.19 7.98 Date of intervention Before 2006 Ref. 1 2007 2009 0.976 1.021 0.25 4.10 2010 2011 0.467 0.497 0.08 3.27 Surgeon experience beyond the learning curve 100 300 Ref. 1 300 500 0.085 3.029 0.86 10.71 >500 0.903 0.913 0.21 3.89 Nerve preservation 0.033 3.925 1.11 13.84 Surgeon 0.308 0.556 0.18 1.72 Procedure Pure laparoscopy Ref. 1 Robot assisted 0.045 5.933 1.04 33.82 CI = confidence interval; GS = Gleason score; OR = odds ratio; PSA = prostate-specific antigen. Fig. 4 Potency recovery rates at each medical visit in the laparoscopic radical prostatectomy and in the robot-assisted prostatectomy groups after bilateral nerve-sparing surgery (overall cohort). The p values at each time point were preoperative ( p = 0.716), 1 mo ( p < 0.001), 3 mo ( p = 0.001), 6 mo ( p < 0.001), 12 mo ( p < 0.001) and 24 mo ( p < 0.001). model, the pt3 margin rates were not statistically different for each surgical modality. We found lower positive surgical margin rates in pt2 disease for RALP compared with LRP in line with recent reviews [12,30]. Potential factors explaining this significant difference could be the traumatic manipulation of the gland during the neurovascular dissection in LRP and the magnified 3D vision improving the capsular preservation provided by robotic assistance. However, urologists performing RALP tend to expand indications of nerve preservation, exposing patients selected on nonstringent criteria to an increased risk of positive margins. Due to the widespread acceptance of active surveillance in low-risk prostate cancers, we have observed an important stage migration towards more aggressive and larger prostate cancers (pt3 and/or Gleason 8 10) in our surgical cohort for the past 5 yr. However, when strict indications of nerve-sparing techniques are respected, RALP performs better than LRP in terms of surgical margins in pathologically organ-confined disease and does not increase the rate of positive margins in non organ-confined cancers. Positive surgical margins and perioperative parameters are early outcome measures of importance in comparing surgical modalities. Our findings showed equivalent biochemical recurrence rates for RALP and LRP. Surgeon volume did not have an impact on the biochemical Table 7 Evolution of International Index of Erectile Function-5 scores over time and their comparisons after laparoscopic radical prostatectomy and robot-assisted radical prostatectomy * IIEF-5 scores Baseline Month 1 Month 3 Month 6 Month 12 Month 24 Bilateral preservation LRP 17.6 3.3 6.1 7.0 8.2 8.1 RALP 17.7 5.2 8.8 10.6 11.5 13.5 p values 0.847 <0.001 0.001 <0.001 <0.001 <0.001 Overall cohort p values 0.800 <0.001 <0.001 <0.001 <0.001 <0.001 IIEF = International Index of Erectile Function; LRP = laparoscopic radical prostatectomy; RALP = robot-assisted radical prostatectomy. * Overall cohort and subgroups of patients undergoing bilateral nerve-sparing surgery.

618 EUROPEAN UROLOGY 65 (2014) 610 619 recurrence survival in our series. However, this lack of difference based on surgeon volume might reflect a low power effect and be related to the short follow-up. Sooriakumaran et al. previously reported that lower surgeon volume was associated with an increased risk of recurrence after >5-yr follow-up [31]. Biochemical recurrences, metastases, and survival statistics require long-term follow-up to assess [9]. Although we previously reported the oncologic safety of extraperitoneal LRP, a longer followup is warranted to confirm it in RALP [25]. We would also like to emphasise that data were collected prospectively but reviewed in a retrospective manner that introduced an interpretation bias. Our statistical analyses highlighted that several factors must be taken into account when determining functional outcomes after RP. Many published studies did not include all these factors with impact, thus limiting the power of their conclusions. Level of surgical experience, changes of surgical details over time, cancer characteristics, patient characteristics, or extension of nerve-sparing surgery could play a part in differential outcomes and introduce interpretation biases when comparing surgical techniques. Systematic reviews and meta-analyses are also limited by multiple designs, and methodological factors have to be considered too. The strength of our series was that we chose to control all these potential confounding factors in multivariable analyses. 5. Conclusions Robotic assistance using an extraperitoneal approach confers better results than the pure laparoscopic procedure in terms of operative time, blood loss, and hospital stay. There was no independent impact on continence by surgical approach. The robotic approach improves potency recovery at short-term follow-up compared with pure laparoscopy. Urologists performing RALP tend to expand indications of nerve preservation exposing patients selected on nonstringent criteria to an increased risk of positive margins. However, when strict indications of nerve-sparing techniques are respected, RALP does not increasetherateofpositivemargins.thisstudydidnot lead us to a definitive recommendation for the robot assistance approach, but it does suggests improved perioperative parameters and potency recovery without compromising oncologic control. Author contributions: Guillaume Ploussard 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: Ploussard, de la Taille, Salomon. Acquisition of data: Ploussard, de la Taille, Moulin, Salomon. Analysis and interpretation of data: Ploussard, de la Taille, Salomon. Drafting of the manuscript: Ploussard. Critical revision of the manuscript for important intellectual content: Ploussard, de la Taille, Moulin, Vordos, Hoznek, Abbou, Salomon. Statistical analysis: Ploussard. Obtaining funding: None. Administrative, technical, or material support: None. Supervision: de la Taille, Abbou, Salomon. Other (specify): None. Financial disclosures: Guillaume Ploussard 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: None. Funding/Support and role of the sponsor: None. References [1] Schuessler WW, Schulam PG, Clayman RV, et al. Laparoscopic radical prostatectomy: initial short-term experience. Urology 1997;50:854 7. [2] Guillonneau B, Vallancien G. Laparoscopic radical prostatectomy: the Montsouris experience. J Urol 2000;163:418 22. [3] Rassweiler J, Schulze M, Teber D, et al. Laparoscopic radical prostatectomy: functional and oncological outcomes. Curr Opin Urol 2004;14:75 82. [4] Stolzenburg JU, Rabenalt R, Do M, et al. Endoscopic extraperitoneal radical prostatectomy: oncological and functional results after 700 procedures. J Urol 2005;174:1271 5. [5] Rassweiler J, Hruza M, Teber D, Su L-M. Laparoscopic and robotic assisted radical prostatectomy critical analysis of the results. Eur Urol 2006;49:612 24. [6] Salomon L, Levrel O, de la Taille A, et al. Radical prostatectomy by the retropubic, perineal and laparoscopic approach: 12 years of experience in one center. Eur Urol 2002;42:104 11. [7] Guillonneau B, El-fettouh H, Baumert H, et al. Laparoscopic radical prostatectomy: oncologic evaluation after 1000 cases at Montsouris institute. J Urol 2003;169:1261. [8] Rassweiler J, Schulze M, Teber D, et al. Laparoscopic radical prostatectomy with the Heilbronn technique: oncologic results in the first 500 patients. J Urol 2005;173:761 4. [9] Novara G, Ficarra V, Mocellin S, et al. Systematic review and metaanalysis of studies reporting oncologic outcome after robot-assisted radical prostatectomy. Eur Urol 2012;62:382 404. [10] Ficarra V, Novara G, Ahlering TE, et al. Systematic review and metaanalysis of studies reporting potency rates after robot-assisted radical prostatectomy. Eur Urol 2012;62:418 30. [11] Ficarra V, Novara G, Rosen RC, et al. Systematic review and metaanalysis of studies reporting urinary continence recovery after robot-assisted radical prostatectomy. Eur Urol 2012;62:405 17. [12] Tewari A, Sooriakumaran P, Bloch DA, Seshadri-Kreaden U, Hebert AE, Wiklund P. Positive surgical margin and perioperative complication rates of primary surgical treatments for prostate cancer: a systematic review and meta-analysis comparing retropubic, laparoscopic, and robotic prostatectomy. Eur Urol 2012;62:1 15. [13] Porpiglia F, Morra I, Lucci Chiarissi M, et al. Randomised controlled trial comparing laparoscopic and robot-asisted radical prostatectomy. Eur Urol 2013;63:606 14. [14] Asimakopoulos AD, Pereira Fraga CT, et al. Randomized comparison between laparoscopic and robot-assisted nerve-sparing radical prostatectomy. J Sex Med 2011;8:1503 12. [15] Ploussard G, Xylinas E, Salomon L, et al. Robot-assisted extraperitoneal laparoscopic radical prostatectomy: experience in a highvolume laparoscopy reference centre. BJU Int 2010;105:1155 60. [16] Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004;240:205 13.

EUROPEAN UROLOGY 65 (2014) 610 619 619 [17] Poulakis V, Dillenburg W, Moeckel M, et al. Laparoscopic radical prostatectomy: prospective evaluation of the learning curve. Eur Urol 2005;47:167 75. [18] Bollens R, Sandhu S, Roumeguere T, et al. Laparoscopic radical prostatectomy: the learning curve. Curr Opin Urol 2005;15:79 82. [19] Secin FP, Savage C, Abbou C, et al. The learning curve for laparoscopic radical prostatectomy: an international multicenter study. J Urol 2010;184:2291 6. [20] Patel VR, Tully AS, Holmes R, et al. Robotic radical prostatectomy in the community setting the learning curve and beyond: initial 200 cases. J Urol 2005;174:269 72. [21] Ahlering TE, Skarecky D, Lee D, et al. Successful transfer of open surgical skills to a laparoscopic environment using a robotic interface: initial experience with laparoscopic radical prostatectomy. J Urol 2003;170:1738 41. [22] Joseph JV, Vicente I, Madeb R, et al. Robot-assisted vs pure laparoscopic radical prostatectomy: are there any differences? BJU Int 2005;96:39 42. [23] Stolzenburg JU, Rabenalt R, Do M, et al. Endoscopic extraperitoneal radical prostatectomy: the University of Leipzig experience of 1,300 cases. World J Urol 2007;25:45 51. [24] Rozet F, Jaffe J, Braud G, et al. A direct comparison of robotic assisted versus pure laparoscopic radical prostatectomy: a single institution experience. J Urol 2007;178:478 82. [25] Paul A, Ploussard G, Nicolaiew N, et al. Oncologic outcome after extraperitoneal laparoscopic radical prostatectomy: midterm follow-up of 1115 procedures. Eur Urol 2010;57:267 73. [26] Rocco B, Cozzi G, Spinelli MG, et al. Posterior musculofascial reconstruction after radical prostatectomy: a systematic review of the literature. Eur Urol 2012;62:779 90. [27] Park JW, Won Lee H, Kim W, et al. Comparative assessment of a single surgeon s series of laparoscopic radical prostatectomy: conventional versus robot-assisted. J Endourol 2011;25:597 602. [28] Hakimi AA, Blitstein J, Feder M, et al. Direct comparison of surgical and functional outcomes of robotic-assisted versus pure laparoscopic radical prostatectomy: single-surgeon experience. Urology 2009;73:119 23. [29] Ahlering TE, Rodriguez E, Skarecky DW. Overcoming obstacles: nerve-sparing issues in radical prostatectomy. J Endourol 2008;22:745 50. [30] Coelho RF, Rocco B, Patel MB, et al. Retropubic, laparoscopic, and robot-assisted radical prostatectomy: a critical review of outcomes reported by high-volume centers. J Endourol 2010;24: 2003 15. [31] Sooriakumaran P, Haendler L, Nyberg T, et al. Biochemical recurrence after robot-assisted radical prostatectomy in a European single-centre cohort with a minimum follow-up time of 5 years. Eur Urol 2012;62:768 74.