Downsides of Robot-assisted Laparoscopic Radical Prostatectomy: Limitations and Complications

EUROPEAN UROLOGY 57 (2010) 735 746 available at www.sciencedirect.com journal homepage: www.europeanurology.com Platinum Priority Collaborative Review Prostate Cancer Editorials by Alexandre M. Mottrie on pp. 747 749 and by Manfred P. Wirth and Michael Froehner on pp. 750 751 of this issue Downsides of Robot-assisted Laparoscopic Radical Prostatectomy: Limitations and Complications Declan G. Murphy a,b, *, Anders Bjartell c, Vincenzo Ficarra d, Markus Graefen e, Alexander Haese e, Rodolfo Montironi f, Francesco Montorsi g, Judd W. Moul h, Giacomo Novara d, Guido Sauter i, Tullio Sulser j, Henk van der Poel k a Department of Urological Oncology, The Peter MacCallum Cancer Centre, Melbourne, Australia b Australian Prostate Cancer Research Centre, Epworth Richmond, Melbourne, Australia c Division of Urological Cancers, Department of Clinical Sciences, Lund University, Malmo, Sweden d Department of Oncological and Surgical Sciences, Urology Clinic, University of Padua, Padua, Italy e Martini-Clinic, Prostate Cancer Centre, University Medical Centre Eppendorf, Hamburg, Germany f Section of Pathological Anatomy, Polytechnic University of the Marche Region, Ancona, Italy g Universita Vita-Salute San Raffaele, Via Olgettina 60, 20132 Milan, Italy h Division of Urologic Surgery, Duke Prostate Centre, Duke University, Durham, NC, USA i Institute of Pathology, University Medical Centre Eppendorf, Hamburg, Germany j Department of Urology, University Hospital Zurich, University of Zurich, Switzerland k Department of Urology, Netherlands Cancer Institute, Amsterdam, The Netherlands Article info Article history: Accepted December 10, 2009 Published online ahead of print on December 17, 2009 Keywords: da Vinci Complications Continence Erectile function Laparoscopic Prostate cancer Radical prostatectomy Robotic Abstract Context: Robot-assisted laparoscopic radical prostatectomy (RALP) using the da Vinci Surgical System (Intuitive Surgical, Sunnyvale, CA, USA) is now in widespread use for the management of localised prostate cancer (PCa). Many reports of the safety and efficacy of this procedure have been published. However, there are few specific reports of the limitations and complications of RALP. Objective: The primary purpose of this review is to ascertain the downsides of RALP by focusing on complications and limitations of this approach. Evidence acquisition: A Medline search of the English-language literature was performed to identify all papers published since 2001 relating to RALP. Papers providing data on technical failures, complications, learning curve, or other downsides of RALP were considered. Of 412 papers identified, 68 were selected for review based on their relevance to the objective of this paper. Evidence synthesis: RALP has the following principal downsides: (1) device failure occurs in 0.2 0.4% of cases; (2) assessment of functional outcome is unsatisfactory because of nonstandardised assessment techniques; (3) overall complication rates of RALP are low, although higher rates are noted when complications are reported using a standardised system; (4) long-term oncologic data and data on high-risk PCa are limited; (5) a steep learning curve exists, and although acceptable operative times can be achieved in <20 cases, positive surgical margin (PSM) rates may require experience with >80 cases before a plateau is achieved; (6) robotic assistance does not reduce the difficulty associated with obese patients and those with large prostates, middle lobes, or previous surgery, in whom outcomes are less satisfactory than in patients without such factors; (7) economic barriers prevent uniform dissemination of robotic technology. Conclusions: Many of the downsides of RALP identified in this paper can be addressed with longer-term data and more widespread adoption of standardised reporting measures. The significant learning curve should not be understated, and the expense of this technology continues to restrict access for many patients. # 2009 European Association of Urology. Published by Elsevier B.V. All rights reserved. * Corresponding author. The Peter MacCallum Cancer Centre, St Andrews Place, Melbourne, Victoria 3002, Australia. Tel. +61 39936 8032; Fax: +61 39429 4683. E-mail address: decmurphy@doctors.net.uk (D.G. Murphy). 0302-2838/$ see back matter # 2009 European Association of Urology. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.eururo.2009.12.021

736 EUROPEAN UROLOGY 57 (2010) 735 746 1. Introduction 3. Evidence synthesis Robot-assisted laparoscopic radical prostatectomy (RALP) using the da Vinci Surgical System (Intuitive Surgical, Sunnyvale, CA, USA) enjoys a high profile, and there is considerable patient demand for this approach. RALP is now the dominant approach to radical prostatectomy (RP) in the United States and is increasing in popularity in other regions where health economic conditions permit. Patients are attracted by oft-unsubstantiated claims posted on commercial and health provider Web sites that RALP is minimally invasive and that outcomes are superior to other approaches [1,2]. Nevertheless, although randomised trials are lacking, there is reasonable evidence from reviews of case series and comparative studies to suggest that RALP is a well-tolerated, safe, and efficacious intervention for the management of localised prostate cancer (PCa) [3,4]. Although many reports of the feasibility, safety, and early functional and oncologic efficacy of RALP have been published, there are few specific reports of its limitations and complications. In this review, we evaluate the current status of RALP, with a particular focus on its limitations and complications. 2. Evidence acquisition 2.1. Literature search A Medline search of English-language literature was performed in September 2009 using the following search terms: robotic radical prostatectomy, robot-assisted radical prostatectomy, and da Vinci radical prostatectomy. Original and review articles were included, and relevant editorials were considered. All papers providing data on technical failures, complications, learning curve, or other downsides of RALP were taken into consideration. Additional papers identified in the bibliography of selected papers were included, if relevant. In total, 412 articles were identified. We reviewed them and selected those with the greatest relevance to this paper for inclusion. Sixty-eight papers were included in the final review. 3.1. Da Vinci Surgical System device failure A limitation specific to this procedure is device failure (see Table 1). However, such failures appear to be rare events, occurring in only 34 of 8240 cases (0.4%) in a multiinstitutional study [5]. Of these, 24 events were identified preoperatively, leading to cancellation of the procedure. Of the 10 device failures that developed intraoperatively, eight cases were converted to open surgery, with two converted to a conventional laparoscopic approach. In a number of smaller studies, device failure has been reported in 0.2 2.6% of cases [6,7]. Two papers have reviewed adverse events related to mechanical failure of the da Vinci Surgical System that have been reported on the Manufacturer and User Facility Device Experience (MAUDE) database of the US Food and Drug Administration (FDA). Andonian et al estimated a device failure rate of 0.38% based on 168 da Vinci system malfunctions reported between 2000 and 2007 [8]. Of these, nine (4.8%) were associated with patient injury. Regarding failures of da Vinci instrumentation, Murphy et al identified 38 system failures and 78 adverse events reported on the MAUDE database between May 2006 and April 2007 [9]. Most of the adverse events relate to either broken instrument tips or failure of electrocautery elements of the da Vinci instruments. In one case, a robotic bipolar grasper was left on the patient s abdomen, and the console surgeon inadvertently activated the device. A 2-mm superficial burn was noted on the patient s abdominal wall. No further injury was noted. This is a specific complication relating to the remote position of the operating surgeon and highlights the need for clear communication between the console surgeon and operating surgeon in these cases. Another issue with device failure is the consequence of an unrecoverable fault. If the bladder neck has not been divided, then abandoning the procedure remains an option. Otherwise, the surgeon must revert to conventional laparoscopy or convert to open surgery. Of the 38 such instances identified by Murphy et al in their review of the MAUDE database 2006 7, 32 procedures were converted to open surgery, which reflects the lack of experience with Table 1 Device failure and adverse events related to the da Vinci surgical system Author Study design System failure rate Adverse events Comment Lavery et al. [5] Multi-institutional questionnaire 34/8240 cases (0.4%) N/R The majority of cases were cancelled, as device failure was noted before the procedure Patel et al. [6] Single-institution case series 1 of 500 cases (0.2%) N/R Case converted to standard LRP Borden et al. [7] Single-institution case series 9 of 350 cases (2.6%) N/R Two procedures converted to ORP; one converted to LRP Andonian et al. [8] Murphy et al. [9] Review of MAUDE database 2000 2007 Review of MAUDE database May 2006 April 2007 0.38% 9 of 189 (4.8%) patient injury 38 reported in 1 yr (32 converted to ORP) One iliac vein injury resulting from insulation failure; one skin burn 78 78 instrument failures (3 converted to open) N/R = not reported; LRP = laparoscopic radical prostatectomy; ORP = open radical prostatectomy; MAUDE = Manufacturer and User Facility Device Experience (of the US Food and Drug Administration).

EUROPEAN UROLOGY 57 (2010) 735 746 737 conventional laparoscopic radical prostatectomy (LRP) in the United States. Although device failure is rare, the increasing penetration of robotic surgery into training programmes may lead to less availability of open radical prostatectomy (ORP) and LRP skills to deal with the consequences of such failure in the future. 3.2. Patient issues Patient selection was of particular importance in the early days of RALP, when the pioneers of this procedure were on not just a learning curve but a discovery curve. Previous abdominal or transurethral surgery, obesity, prior radiation therapy (RT), and unfavourable disease characteristics were all factors that might have dissuaded surgeons from offering a robot-assisted approach. It is now clear that in centres with adequate experience, the indications for a robotassisted approach have expanded. 3.2.1. Obesity Herman et al analysed a group of 132 men undergoing RALP, including 60 overweight and 34 obese patients with matched disease characteristics [10]. The overweight and obese patients had a longer operative duration (304 min vs 235 min; p < 0.001), greater blood loss, longer hospital stay, and higher positive surgical margin (PSM) rates (21% vs 11%; p = 0.18). Similarly, significantly higher complication rates (26.3% vs 4.9%; p =0.01), slower return to normal activity, significantly lower 6-mo continence rates (47% vs 91.4%; p =0.001)inobesemenaswellaslonger operative times (302 min vs 240 min; p =0.003), greater blood loss (396 ml vs 293 ml; p =0.008), and higher complications and PSM rates (26.5% vs 13.1%; p = 0.009) were reported by Ahlering et al. [11] and Castle et al. [12], respectively. It is of course no surprise that obesity is associated with worse outcomes following RALP: It is associated with worse outcomes following many surgical procedures, including ORP, after which obese men are more likely to suffer from wound infections (16.1% vs 4.5%; p < 0.05), urinary incontinence (25.8% vs 8.7%; p < 0.05), and anastomotic stricture (46.2% vs 12.3%; p < 0.05) when compared to nonobese men [13]. The key question that remains unanswered is whether any particular surgical approach offers a safer option for obese and morbidly obese patients with localised PCa. These patients should understand that robotic surgery does not diminish the deleterious effect of obesity on surgical outcome and should be advised in many instances to undergo a weight-reduction programme prior to surgery. 3.2.2. Previous transurethral surgery It remains unclear whether prior surgery for bladder outlet obstruction (BOO) has a negative impact on outcomes following RP. Although some authors have demonstrated longer operative times, increased complications, and higher rates of PSMs in patients undergoing ORP [14] or LRP [15] following prior surgery for BOO, Palisaar et al have recently shown that outcomes are equivalent following ORP in patients who had previously undergone transurethral resection of the prostate (TURP) [16]. It is commonly acknowledged that the dissection of the posterior bladder neck and seminal vesicles is one of the more difficult steps of RALP, particularly in the early part of the learning curve. A large middle lobe and previous transurethral surgery for BOO are additional factors that may increase the difficulty associated with this step. Hampton et al reported on 51 of 1768 patients who underwent RALP having previously had TURP [17]. Compared to patients who had not undergone previous TURP, post-turp patients had a significantly higher PSM rate (35.3% vs 17.6%; p < 0.015), and these PSMs were more likely to be located at the bladder neck [17]. Although this is not an outcome exclusive to the RALP approach, it should be borne in mind when counselling such patients for this procedure. 3.2.3. Large prostates Increasing prostate size is associated with greater blood loss, longer hospital stay, and a higher rate of complications but not with higher PSMs in ORP or LRP [18,19]. In a large RALP series, Link et al evaluated the impact of prostate size in a series of 1847 cases, demonstrating greater blood loss, longer hospital stay, and more complications (eg, urinary leak) for larger-sized prostates [20]. Moreover, Ahlering et al reported that patients with a larger prostate experienced a delayed return of potency following RALP [21]. Therefore, a robot-assisted approach confers neither an advantage nor a disadvantage for larger-sized prostates when compared to other approaches: Large prostates are surgically challenging for all approaches. 3.2.4. Salvage Because the morbidity of salvage ORP is considerably higher than primary ORP, RALP surgeons have been somewhat cautious about offering this approach to such patients. However initial series are now appearing in press. Kaouk et al reported their experience of four patients undergoing RALP following failed external-beam RT (EBRT) or brachytherapy [22]. There was no significant perioperative morbidity, and three patients were fully continent 1 mo following catheter removal. A larger group of 11 patients was reported by Boris et al, demonstrating similar findings [23]. These are small series, but they demonstrate the potential use of the robot-assisted approach for this group of patients. Satisfactory reports of salvage RALP following failed high-intensity focused ultrasound treatment or cryotherapy have been reported in small case series, and it is to be expected that reports of larger experience with longer follow-up will emerge [24,25]. More conclusive comments on the role of salvage RALP may then be possible. Overall, it appears that the issues outlined above are no longer a limitation to a robot-assisted approach to RP. These more complex cases are best avoided within the early robotic experience of any centre, and patients must be adequately counselled regarding the increased morbidity when compared to more straightforward RALPs. Patients in

738 EUROPEAN UROLOGY 57 (2010) 735 746 these circumstances should understand that a robotapproach does not negate the increased likelihood of morbidity, which they are at risk of developing compared to patients without such confounding factors. 3.3. Complications of robot-assisted radical prostatectomy The outcome data of many centres have focused on perioperative results and oncologic and functional outcome, with little detail reserved for complications. Although there is a trend towards the use of validated classification systems such as the Clavien classification [26] for reporting complications [27], an important limitation of the current published literature is the poor quality of complication reporting [28]. Martin et al identified 10 essential elements, including methods of data accruing, duration of follow-up, presence of outpatient information, definitions of complications, mortality and morbidity rates, procedure-specific complications, severity grading, length of in-hospital stay, and analysis of risk factors, that should be used to ensure accurate recording of morbidity data [29]. Sadly, these criteria are underused. Most series do not use a standardised reporting system, and complications are not clearly defined, leading to wide discrepancy in complication rates (4.4 26%; see Table 2) between those centres using a standardised system and those that do not. Three papers have focused entirely on complications relating to RALP [30 32], and two further papers [33,34] have also adopted the Clavien system for reporting complications. Table 2 summarises the findings from these publications with respect to complications. To date, only Novara et al define and report the complications in their RALP series of 415 patients following all the Martin criteria [32]. Using such standardised criteria, an overall complication rate of 21.6% was demonstrated with Clavien I II complications, accounting for 18.6% of the series. Practically speaking, it is likely that the higher the methodologic quality of the data collection, the higher the reported complication rates and the more validity that can be presumed. Therefore, use of standardised criteria for complication recording and reporting must be encouraged. It should be noted that such standardised systems have not been widely used for ORP series to date, and this is a limitation of ORP data. 3.3.1. Transperitoneal versus extraperitoneal approach Although an extraperitoneal approach to RALP has been described and remains standard in a small number of institutions [35], the vast majority of RALPs are undertaken using a transperitoneal approach. The transperitoneal approach is usually preferred because of the larger working space and reduced instrument arm collisions; however, this approach requires a steep Trendelenburg position and increases the possibility of intraperitoneal complications such as bowel injury and ileus. In practice, the reported incidence of these complications is very low. Atug et al in a nonrandomised study compared 40 extraperitoneal with 40 transperitoneal RALPs and showed no difference in operative time, estimated blood loss (EBL), complications, or PSMs and concluded that the extraperitoneal approach was feasible and should be considered as an option [36]. Table 2 Reported complications of robot-assisted laparoscopic prostatectomy classified with and without the Clavien system Author (n) Badani [33] (2766) Hu [30] (322) Fischer [31] (210) Murphy [34] (400) Novara [32] (415) Patel [83] (1500) Zorn [66] (300) Mottrie [84] (184) Joseph [53] (325) Krambeck [85] (286) Complication Clavien(**) system used Clavien system not used Overall complications (%) 12.2 22.6 26 15.75 21.6 4.4 10.6 11.9 10.1 8 Clavien III or greater (%) 0.6 2.7 8.5 5.25 3.2 Death (%) <0.01 0 0 0 0 0 N/R 0 0 N/R Blood transfusion (%) 1.5 1.6 1 2.5 5.3 0.5 1.7 0.5 1.3 5.1 Neuropraxia (%) N/R 0 N/R N/R 0.9 N/R 1.4 1.6 0.6 N/R Urine leak (%) N/R 7.5 6.5 4.5 6.7* 1.4 1.4 0.5 1.2 9.1 Bowel injury (%) N/R 0.6 1.5 1.2 1.2 0.13 0 0 0.3 N/R Anastomotic stricture (%) N/R 0.6 0.5 3.7 N/R 0.13 1.4 N/R 2.1 1.2 Thromboembolic event (%) N/R 0.6 N/R N/R 0.2 0.33 0.6 N/R 1.5 1 Urinary retention N/R 4 N/R N/R 1.4 0.3 N/R 1.6 N/R 2.8 N/R = not reported. * >5% leakage noted on day 4 cystogram. ** Clavien classification of complications occurring within 90 d of procedure: Grade I: Any deviation from normal postoperative course without the need for pharmacologic, surgical, endoscopic, or radiologic interventions. Permitted pharmaceuticals include antiemetics, antipyretics, analgesics, and diuretics. Grade II: Requiring pharmacologic treatment with drugs other than allowed for grade I complications. Blood transfusions and total parenteral nutrition also included. Grade III: Requiring surgical, endoscopic, or radiologic intervention. Grade IIIa: Intervention not under general anaesthesia. Grade IIIb: Intervention under general anaesthesia. Grade IV: Life-threatening complication requiring intensive care management. Grade IVa: Single-organ dysfunction (including dialysis). Grade IVb: Multiorgan dysfunction. Grade V: Death of a patient.

EUROPEAN UROLOGY 57 (2010) 735 746 739 3.3.2. Robot-assisted radical prostatectomy and anaesthesiology Although the steep Trendelenburg position and prolonged operative times in the early part of some series might lead to some concerns about the safety of RALP form an anaesthesiology perspective, it would appear that by and large, the procedure is well tolerated. Danic et al reviewed their first 1500 RALP cases to identify any anaesthesia-related complications and reported three corneal abrasions, one overnight ventilation because of a difficult intubation, and three pulmonary emboli [37]. Awad et al prospectively measured the intraocular pressure (IOP) before, during, and after RALP in 33 patients and documented a 13 mmhg increase in IOP towards the end of the procedure in steep Trendelenburg position [38]. However, no clinical sequelae were documented, and this therefore does not appear to be of consequence. 3.4. Oncologic outcomes 3.4.1. Specimen processing and positive surgical margin reporting Comparison of PSMs between surgical approaches and among different series relies on standardised processing of the RP specimen and uniform definition of what constitutes a PSM [39]. Table 3 summarises some recent series of RALP with regard to the reporting of their histopathologic data. It is recommended that whole-mount specimens should be the standard for specimen processing and that the presence of tumour at the inked margin constitute a PSM [40,41]. Furthermore, the precise site and extent of PSM should be recorded, as this information is an important reflection of surgical technique and may determine whether adjuvant treatment post-rp is required. Although some series report their histopathologic data in this manner [34,42], many do not. 3.4.2. Positive surgical margins PSMs are an independent risk factor for biochemical recurrence (BCR), local disease recurrence, and the need for salvage treatment [41] and may be used as a surrogate for good surgical technique [43]. There are important differences in surgical approach between ORP and RALP (eg, early dissection of the seminal vesicles and antegrade nerve sparing) that may influence the location and extent of PSMs, and the lack of tactile feedback using the robotassisted approach introduces an unknown variable when compared to ORP, particularly for patients with locally advanced disease. Overall, although there is no level 1 evidence to provide conclusive data, PSM rates for RALP compare favourably to those for ORP and LRP. In their review of studies comparing PSMs for ORP, LRP, and RALP, Ficarra et al concluded that there was a statistically significant advantage for RALP (risk ratio [RR]: 1.58; confidence interval [CI] of RR: 1.29 1.94; p < 0.00001) particularly for patients with organ-confined disease (RR: 2.23; CI of RR: 1.36 3.67; p < 0.002) [4]. 3.4.2.1. Positive surgical margins in the early experience of robot-assisted radical prostatectomy. As one might expect, many series have shown an improvement in PSMs with progressive experience [3]. However, closer inspection of the early experience of some centres reveals PSM rates that are certainly higher than one might anticipate (see Table 4). Atug et al reported their initial series of 140 patients undergoing RALP and stratified PSMs as their experience increased [44]. PSMs dropped from 45.1% in the first 33 patients to 21.2% in the next 33 patients and 11.7% in the last 34. The authors suggest that the learning curve for RALP is about 30 cases, but one could suggest that as their PSMs continued to improve in the latter third of their series, a figure of 60 cases might be more appropriate. Reporting their transition from ORP to RALP, Ahlering et al reported overall and pt2 PSM rates of 35.5% and 8.9%, respectively, in their first 45 patients [45]. Because their operative time dropped below 4 h after 12 cases, the authors concluded that the learning curve for open surgeons transitioning to RALP was 8 12 cases. However, if the learning curve were assessed in terms of PSM rates, then it is clear that more experience is required before proficiency is achieved. This group subsequently reported overall PSMs of 16.7% from cases 51 140 [46]. Weizer et al reviewed their initial 193 RALPs to examine the pattern of PSMs as their experience increased [47]. Table 3 Compliance with suggested criteria for reporting of histopathologic data in large robot-assisted laparoscopic prostatectomy series Author (n) Menon et al. [56] (2652) Link et al. [20] (1847) Patel et al. [83] (1500) Shikanov et al. [86] (1398) Chan et al. [87] (660) Murphy et al. [34] (400) Ficarra et al. [44] (322) Reported histopathologic criteria Specimen processing * N Y Y Y Y Y Y Overall and stage-specific PSM N N Y Y Y Y Y Site of PSM N N Y Y N N Y Single vs multiple PSM sites N N Y Y N N Y Focal vs nonfocal PSM N N N Y N N N PSM in nerve sparing vs N N N Y Y N Y non nerve sparing Overall PSM rate (%) 13 23.8 9.3 17 17.8 19.2 29.5 pt2 PSM rate (%) N/R N/R 4 11 11.3 9.6 10.6 N = data not included; Y = data included; PSM = positive surgical margin; N/R = not reported. * With regards to specimen processing, we would also specify whether the data included in the reports indicated optimal or suboptimal processing (whole mount, macrosections). We do both of them.

740 EUROPEAN UROLOGY 57 (2010) 735 746 Table 4 Positive surgical margin rates in the early experience of robot-assisted laparoscopic prostatectomy from selected centres Author n Overall PSM rate, % pt2 PSM rate, % Comment Atug et al. [44] 33 45.4 38.4 Overall PSM dropped to 11.7% after 66 cases Ahlering et al. [45] 45 35.5 8.9 Overall PSM rate improved to 16.7% beyond case 50 Weizer et al. [47] 193 26 N/R Extensive margins in first 15 cases Artibani et al. [90] 41 26.8 6.9 Ou et al. [48] 30 50 13.3 pt3 rate: 86.7% Sim et al. [49] 17 59 N/R Most PSMs in posterolateral region Jaffe et al. [50] 12 58 50 Overall PSM rate improved to 9% after 189 cases PSM = positive surgical margin; N/R = not reported. The proportion of extensive margins decreased from 12% in the first 15 cases to 2% beyond case 80. This decrease led the authors to conclude that the learning curve to achieve good PSM rates was about 80 cases. In three other initial series, PSM rates exceeding 50% have been reported. Ou et al reported an overall PSM rate of 50% in their first 30 RALP cases compared with 20% in their previous 30 ORPs [48]. Sim et al reported an overall PSM rate of 59% in their initial experience of 17 RALP cases performed in 2003 [49]. Jaffe et al, describing their initial 293 RALPs from Montsouris in Paris, reported an overall PSM rate of 58% in their first 12 cases [50]. This rate dropped to 9% after 189 cases, suggesting that there is a considerable PSM learning curve even for surgeons very experienced in LRP. 3.4.2.2. Predictors of positive surgical margins. To date, only Ficarra et al have reported on the predictors of PSMs [42]. The overall and pt2 PSM rates were 29.5% and 10.6%, respectively, and about two-thirds of PSMs were located in the posterolateral area. Extracapsular disease was the most powerful predictor of overall, posterolateral, and multiple PSM, while perineural invasion was predictive of PSM in organ-confined disease. Clearly, a learning curve is associated with PSMs for all surgical approaches to RP. The use of robotic assistance should not be presumed to reduce the importance of experience when evaluating PSM rates. Moreover, PSM results must always be interpreted in the context of potency and continence outcomes. 3.4.3. Overall and disease-free survival Regarding overall survival (OS) and biochemical recurrence-free survival (brfs) following RALP, long-term results are not yet available a limitation of any assessment of RALP at this time. However, in the short term, OS and brfs rates are encouraging and compare favourably to ORP and LRP [4,51]. Menon reported brfs rates of 92.7% in a cohort of 2766 patients, with a median of 22-mo (range: 6 71) follow-up [33]. Murphy et al reported a lower brfs of 86.6% in a cohort of 400 men with a similar median follow-up of 22 mo (interquartile range: 15 30) [34]. However, this population had a greater proportion of high-risk disease and a minimum 12-mo follow-up, which inevitably led to lower brfs. Longer follow-up and randomised studies will be required to establish the comparable oncologic efficacy of RALP when compared to ORP. 3.4.3.1. The need for salvage therapy. Those sceptical of LRP and RALP (minimally invasive RP [MIRP]) have seized upon the report by Hu et al that suggests that the need for salvage therapy (with EBRT or androgen-deprivation therapy) within 6 mo of MIRP is much higher than following ORP [52]. In an observational study based on Medicare data of 2702 patients operated on in the United States between 2003 and 2005, those who underwent MIRP (it is unclear how many of these were RALP cases) were more likely to undergo salvage treatment than those who underwent ORP (27.8% vs 9.1%; p < 0.001). This study in itself has many limitations in fact, a subsequent observational study from Hu et al using Medicare data showed no difference in the use of additional cancer therapies following MIRP and ORP [53] but it does expose the need for prospective data on oncologic outcome following RALP. Moreover, Chino et al reviewed their series of ORP (n = 536) and RALP (n = 368) and, on multivariate analysis to control for selection bias, did not identify any increased indication or referral for early adjuvant RT in the RALP group [54]. Debate continues over which patients benefit most from adjuvant or salvage therapy following RP. This topic is being examined by a large randomised, controlled trial coordinated by the Medical Research Council in the United Kingdom [55]. 3.5. Functional outcomes Another population-based observational study from Hu et al comparing MIRP and ORP reported a higher incidence of erectile dysfunction (26.8 vs 19.2 per 100 person-years; p = 0.009) and incontinence (15.9 vs 12.2 per 100 personyears; p = 0.02) in MIRP patients based on US NCI Surveillance Epidemiology and End Results Program data [53]. This paper has received a great deal of attention, but do these numbers represent true downsides of RALP? Because this is an observational study based on administrative data, there is great potential for unmeasured differences to confound the data. Claims-based analyses characterise neither the nature nor the severity of the conditions they code for and are considered too blunt an instrument when considering the subtleties of erectile function and continence assessment after radical surgery. In addition, the study population was limited to patients aged >65 yr, a population in which urinary and sexual dysfunction often preexist. However, although the methodologic limitations of this study are significant, it does draw attention to the

EUROPEAN UROLOGY 57 (2010) 735 746 741 important area of evaluation of functional outcomes following RP. 3.5.1. Continence outcomes The objective evaluation of urinary continence outcomes post-rp remains hindered by the lack of standardisation of outcome reporting. Although validated tools exist, they are not in popular use, and the introduction of terms such as socially dry and security liner has somewhat obscured the issue of true continence [56]. In addition, it is accepted that patient- and surgeon-reported outcomes vary considerably, yet surgeon-reported outcomes are the most commonly utilised variable. Using the criteria of no pads or one security liner to define continence, the proportion of patients reported dry at 12 mo following RALP is 90 95% [3,4,57]. Using similar criteria, the same proportion of patients are also considered dry following ORP and LRP [58,59]. However, in a prospective, nonrandomised study and using a more strict definition of no leakage or no pads at 12 mo postoperatively, Touijer et al reported a higher rate of incontinence in their LRP group (52%) when compared to their ORP group (25%) [60]. In a nonrandomised, prospective study of ORP and RALP, Ficarra et al used the validated International Consultation of Incontinence Questionnaire Urinary Incontinence (ICIQ-UI) and demonstrated continence rates of 88% and 97% ( p = 0.01) in the ORP and RALP groups, respectively [61]. Use of validated questionnaires such as the ICIQ-UI is strongly encouraged and would allow meaningful interpretation of urinary continence outcome following RP. 3.5.2. Potency outcomes As with the evaluation of urinary continence post-rp, it is difficult to evaluate the potency outcomes following RALP because of the variability of definitions used to determine potency. The most commonly used criteria to define potency are a Sexual Health Inventory for Men (SHIM) score of 21 and a score of at least 2 on question 2 of the SHIM questionnaire ( When you had erections with stimulation, how often were your erections hard enough for penetration? ). Additional confounding factors include the type of nerve-sparing technique used, whether nerve sparing is unilateral or bilateral, variation in the use of energy sources during dissection, and the use of adjuvant medications such as phosphodiesterase type 5 inhibitors. The obvious limitation regarding potency outcome following RALP also applies to other approaches for RP, namely, the lack of standardised assessment and outcome reporting. Mulhall [62] and Tal et al. [63] have recently reviewed this and suggested minimum standards that should be adhered to for reporting erectile function following RP. Table 5 lists these criteria and evaluates some recent RALP series to assess the level of detail included for assessment of erectile function. Compliance with these standards will allow meaningful assessment of potency outcomes and comparison of surgical approaches that will address some of these shortcomings in the reporting of erectile function. Menon et al, evaluating a series of patients undergoing veil of Aphrodite preservation, reported 97% potency rates in a selected group of 35 men [64]. More recently, the same group updated their technique, describing a superveil nerve-sparing approach and reporting potency rates of 94% in 85 men who underwent this approach [65]. However, these figures have not been replicated in other series, and concerns exist regarding the risks of PSMs with intrafascial dissection [66]. More recently, Novara et al, in a report complying with the Mulhall criteria, reported on a series of 208 patients treated with a bilateral intrafascial or interfascial RALP, finding a more realistic 62% 12-mo potency rate. They also report that patients age, Charlson comorbidity index, and preoperative erectile function were the most powerful predictors of erectile function recovery [67]. Table 5 Potency outcome reporting following robot-assisted laparoscopic prostatectomy stratified according to the Mulhall criteria [62] for reporting erectile function following radical prostatectomy Author (n) Shikanov et al. [88] (361) Menon et al. [65] (85) Murphy et al. [34] (400) Novara et al. [67] (215) Rodriguez et al. [89] (58) Mulhall criteria Patient comorbidity profile N N N Y N Degree to which patient selection was exercised Y Y Y Y Y Who collected the erectile function outcome data N Y N Y Y Which validated questionnaires were used Y (UCLA-PCI) Y (SHIM) Y (SHIM) Y (IIEF-6) Y (EPIC, IIEF-5) Baseline erectile function data N N N Y Y Long-term (24-mo) erectile function data N N N N Y Definition of adequate erectile function N N Y Y Y Proportion of men returning to normal N N N N N Proportion of men returning to preoperative N N N Y Y erectile function level Extent of utilisation of erectogenic medications N N N Y N Extent to which a rehabilitation strategy was used Y Y N Y N Potency outcome N QoL outcome only 94% 62% 62% 90% N = detail not included; Y = detail included; UCLA-PCI = University of California, Los Angeles, Prostate Cancer Index; SHIM = Sexual Health Inventory for Men; IIEF = International Index of Erectile Function; EPIC = Expanded Prostate Cancer Index Composite; QoL = quality of life.

742 EUROPEAN UROLOGY 57 (2010) 735 746 3.5.3. Subjective versus objective reporting Shikanov et al recently published trifecta outcomes in 380 previously potent and continent men who underwent bilateral nerve-sparing RALP with a mean follow-up of 22 mo (range: 12 54) [68]. Functional outcomes were assessed subjectively by the operating surgeons and objectively using the University of California, Los Angeles, Prostate Cancer Index (UCLA-PCI) validated questionnaire. Trifecta rates applying subjective continence and potency definitions were 34%, 52%, 71%, and 76% at 3, 6, 12, and 24 mo, respectively. The corresponding trifecta rates using objective continence and potency definitions stood at 16%, 31%, 44%, and 44%. The difference was statistically significant at each time point ( p < 0.0001). The difference between subjectively and objectively reported functional outcomes is not unique to RALP, but it is nevertheless important that this confounding variable be eliminated by the use of standardised questionnaires administered by a third party. 3.5.4. Is there a quality-of-life benefit to the robotic approach? Patient expectations are largely dependent on information derived from health professionals and, increasingly, from Internet resources not subject to regulation and oversight. Schroeck et al compared patient satisfaction in 400 patients using questionnaires following RALP and ORP and reported less satisfaction following RALP [69]. They concluded that patients who underwent RALP were more likely to be regretful and dissatisfied, possibly because of higher expectations of an innovative procedure. However, this study was limited by its cross-sectional design and the shorter follow-up in the RALP cohort. The accompanying editorial for this interesting paper drew attention to the frequently misleading information available on the Internet regarding RALP, which remains a considerable problem [2]. In a nonrandomised comparison of short-term healthrelated quality-of-life (HRQoL) scores between ORP and RALP, Miller et al only demonstrated a 1-wk difference in return-to-baseline scores in the physical domain in favour of RALP and no difference in the mental domain [70]. Further evidence to establish whether RALP truly improves HRQoL and patient satisfaction compared to ORP remains outstanding. 3.6. The learning curve One of the claimed benefits of the robot-assisted approach is that it reduces the difficulty associated with conventional LRP, reducing the learning curve (ie, operative time) to as few as 12 cases [45]. The initial series of RALP understandably featured quite long operative times (354 498 min [71,72]) as surgical teams became familiar with the technology and developed the technique. It is to be expected that as trainees emerge from fellowship programmes in robotic surgery, initial operative times will be much lower than this from the outset [73]. At present, however, many series are still reporting relatively long operative times as established open surgeons continue to transition to the robot-assisted approach. Mayer et al from the United Kingdom reported a mean operative time of 369 min for their first 50 cases as they transitioned from open surgery with some laparoscopic experience [74]. The mean EBL was 700 ml, and 12% of patients required a blood transfusion. Wilson et al reported a mean operative time of 280 min in their first 30 RALP cases, having had considerable experience with LRP prior to starting a robotic surgery programme [75]. Clearly, it would be better for patients, surgeons, and administrators if such long initial operative times could be avoided, and this should be possible by having fellowshiptrained surgeons appointed to initiate new robotic surgery programmes and to follow best practice guidelines for training and mentoring in robotic surgery [76]. In the meantime, long operative times remain a definite downside of RALP in the early experience of many units. However, the learning curve issue is clearly much more complex than simply achieving acceptable operative times, and outcomes in terms of PSMs, continence, and potency must also be considered. Vickers et al have clearly shown the importance of experience for ORP [77] and have also shown the considerably steeper learning curve for LRP when brfs is used as an outcome measure [78]. A similar analysis of the learning curve for RALP is awaited. The learning curve for continence and potency is much more difficult to quantify because of inconsistencies and subjectivity in outcome reporting for these variables across all approaches for RP. It seems reasonable to suggest that the robot-assisted approach in itself is not a limitation when trying to optimise functional outcomes following RP but that claims for superiority for this approach when compared to ORP and LRP are limited by the lack of standardised outcome reporting. 3.7. Health economic issues At the time of writing, the da Vinci Si is the latest incarnation of the da Vinci surgical robot marketed by Intuitive Surgical. The installation cost is approximately s1.8 million (s2.2 million for the dual-console version), with maintenance costs of about s100,000 per annum. Robotic instrument costs are about s1500 per RALP case. The lack of a competitor in this area has contributed to costs remaining prohibitively high for many hospitals and indeed many countries, thereby preventing equitable availability of this technology across diverse health care systems. Nevertheless, the number of robots continues to increase, and at the end of the first quarter of 2009, there were 1171 da Vinci Surgical Systems installed worldwide, including 863 in the United States and 211 in Europe. 3.7.1. Is there a health economic benefit to justify robot-assisted laparoscopic prostatectomy? Proponents of RALP and those with vested interests often claim that it leads to shorter hospitalisation, faster return to work, and other benefits that justify the expense of the robot-assisted approach. However, these claims are usually unsubstantiated and are often limited by the great variation in health economies from one country to another. Scales et al demonstrated cost-equivalence of RALP with

EUROPEAN UROLOGY 57 (2010) 735 746 743 ORP based on 10 cases per week and cost superiority based on 14 cases per week in the United States [79]. However, this is purely an economic model, and achieving 14 cases per week on a single robot is unlikely in the majority of facilities. Although Burgess et al reported a 27% reduction in operative charges once the learning curve (>50 cases) had been overcome, RALP still remained considerably more expensive than ORP [80]. Steinberg et al generated an economic model to predict the cost of prolonged operative times during the learning curve based on a review of eight RALP series in which the mean number of cases required to reach a total operative time of 3 4 h was 77 [81]. The predicted cost of this model in terms of operative time was $217,000. However, critics have pointed out that none of these reports has used standardised health economic modelling systems [82], and clearly, further work is required to establish the true health economic benefit, if any, of using a robot-assisted approach. Robotic surgery is not always adequately reimbursed in some health economies (including the United States), and these programmes have sometimes run as a loss-leader to attract business or patients have had to cover the funding gap. Neither of these situations is particularly satisfactory and may not be tenable in the long run. 3.7.2. Does robotic surgery have to be so expensive? There is little doubt that Intuitive Surgical s acquisition of Computer Motion Inc. (manufacturers of the Zeus master slave system) in 2003 effectively wiped out the only potential competitor in this market. Intuitive Surgical retain a monopoly, with no significant competition on the horizon. Although they are to be congratulated for bringing this device into clinical practice and for continuing to develop the technology, one wonders if there is not more they could do to make this device more affordable and therefore more accessible. For example, in 2008, Intuitive Surgical reported a 51% growth in revenue from instrumentation alone, totalling $293 million. One really wonders whether EndoWrist instruments have to be discarded after 10 or so uses, especially when EndoWrist training instruments often work very well for 50 sessions. Could EndoWrist instruments not be programmed for infinite use and be discarded at the surgeons discretion when their performance is suboptimal? A competitor in the field of robotic instrumentation would be most welcome as a stimulus to an otherwise monopolised market. 4. Conclusions The da Vinci Surgical System has provided an exciting new dimension to complex laparoscopic surgery and has had a profound effect on the RP market in some countries. Broadly speaking, patients seem to benefit from the minimally invasive approach, short-term oncologic results are acceptable, and functional outcomes appear at least comparable to other approaches. It should be stated that many of the downsides of RALP noted in this review may equally be applied to ORP and LRP in particular, the lack of standardised reporting of outcomes and complications. Overall, there are a number of points to bear in mind. First, the learning curve is not just about operative time. Even if <4-h competency can be achieved in 10 20 cases, there are clearly improvements in PSMs and other important outcome measures that require >80 cases to achieve. Second, the ongoing issue with nonstandardised reporting of surgical complications, histopathologic data, and functional outcomes bedevils the RP literature, particularly for the robotic surgery literature. Authors and editors need to agree and implement better guidance on the reporting of such variables to allow more transparent assessment and comparison of outcomes. We suggest that the following standardised methodologies be considered for all RALP series: (1) Martin criteria [29] for complication recording and Clavien [26] classification for complication reporting; (2) the Stanford protocol [39] for prostate specimen processing and PSM reporting; and (3) validated questionnaires to assess functional outcome such as International Index of Erectile Function (IIEF), ICIQ-UI, UCLA-PCI and compliance with the Mulhall criteria [62] for series reporting detailed potency data. Third, the huge costs of the da Vinci Surgical System mean that availability of this technology remains severely limited. Even if high-quality evidence were to suggest that RALP was superior to LRP and ORP, it is likely that RALP would remain limited to those few individuals or societies who can afford it. Either a competitor in this monopolised market or a radical rethink from Intuitive Surgical is required to break the economic barriers preventing equitable availability of this technology. Finally, is there any point asking for randomised, controlled trials of RALP versus other approaches? Despite the continuing refinement and improvement of ORP, patients and surgeons have already spoken with their feet where robotic technology is available, and recruitment into a trial where open surgery is an option seems increasingly improbable. Instead, standardised reporting of outcomes and less marketing hype would allow patients and surgeons to better understand the true value of RALP and acknowledge its limitations. Author contributions: Declan G. Murphy 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: Montorsi, Graefen. Acquisition of data: Murphy. Analysis and interpretation of data: Murphy. Drafting of the manuscript: Murphy. Critical revision of the manuscript for important intellectual content: Bjartell, Ficarra, Graefen, Haese, Montironi, Montorsi, Moul, Novara, Sauter, Sulser, van der Poel. Statistical analysis: Murphy, Novara, Ficarra. Obtaining funding: None. Administrative, technical, or material support: None. Supervision: Montorsi. Other (specify): None.

744 EUROPEAN UROLOGY 57 (2010) 735 746 Financial disclosures: I certify 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] Wirth MP, Hakenberg OW. Surgery and marketing: comparing different methods of radical prostatectomy. Eur Urol 2009;55: 1031 3. [2] Eastham JA. Robotic-assisted prostatectomy: is there truth in advertising? Eur Urol 2008;54:720 2. [3] Ficarra V, Cavalleri S, Novara G, Aragona M, Artibani W. Evidence from robot-assisted laparoscopic radical prostatectomy: a systematic review. Eur Urol 2007;51:45 56. [4] Ficarra V, Novara G, Artibani W, et al. Retropubic, laparoscopic, and robot-assisted radical prostatectomy: a systematic review and cumulative analysis of comparative studies. Eur Urol 2009;55: 1037 63. [5] Lavery HJ, Thaly R, Albala D, et al. Robotic equipment malfunction during robotic prostatectomy: a multi-institutional study. J Endourol 2008;22:2165 8. [6] Patel VR, Thaly R, Shah K. Robotic radical prostatectomy: outcomes of 500 cases. BJU Int 2007;99:1109 12. [7] Borden Jr LS, Kozlowski PM, Porter CR, Corman JM. Mechanical failure rate of da Vinci robotic system. Can J Urol 2007;14:3499 501. [8] Andonian S, Okeke Z, Okeke DA, et al. Device failures associated with patient injuries during robot-assisted laparoscopic surgeries: a comprehensive review of FDA MAUDE database. Can J Urol 2008;15:3912 6. [9] Murphy D, Challacombe B, Elhage O, Dasgupta P. Complications in robotic urological surgery. Minerva Urol Nefrol 2007;59:191 8. [10] Herman MP, Raman JD, Dong S, Samadi D, Scherr DS. Increasing body mass index negatively impacts outcomes following robotic radical prostatectomy. JSLS 2007;11:438 42. [11] Ahlering TE, Eichel L, Edwards R, Skarecky DW. Impact of obesity on clinical outcomes in robotic prostatectomy. Urology 2005;65: 740 4. [12] Castle EP, Atug F, Woods M, Thomas R, Davis R. Impact of body mass index on outcomes after robot assisted radical prostatectomy. World J Urol 2008;26:91 5. [13] van Roermund JG, van Basten JP, Kiemeney LA, Karthaus HF, Witjes JA. Impact of obesity on surgical outcomes following open radical prostatectomy. Urol Int 2009;82:256 61. [14] Colombo R, Naspro R, Salonia A, et al. Radical prostatectomy after previous prostate surgery: clinical and functional outcomes. J Urol 2006;176:2459 63. [15] Jaffe J, Stakhovsky O, Cathelineau X, Barret E, Vallancien G, Rozet F. Surgical outcomes for men undergoing laparoscopic radical prostatectomy after transurethral resection of the prostate. J Urol 2007;178:483 7. [16] Palisaar JR, Wenske S, Sommerer F, Hinkel A, Noldus J. Open radical retropubic prostatectomy gives favourable surgical and functional outcomes after transurethral resection of the prostate. BJU Int 2009;104:611 5. [17] Hampton L, Nelson RA, Satterthwaite R, Wilson T, Crocitto L. Patients with prior TURP undergoing robot-assisted laparoscopic radical prostatectomy have higher positive surgical margin rates. J Robotic Surg 2008;2:213 6. [18] Pettus JA, Masterson T, Sokol A, et al. Prostate size is associated with surgical difficulty but not functional outcome at 1 year after radical prostatectomy. J Urol 2009;182:949 55. [19] Levinson AW, Ward NT, Sulman A, et al. The impact of prostate size on perioperative outcomes in a large laparoscopic radical prostatectomy series. J Endourol 2009;23:147 52. [20] Link BA, Nelson R, Josephson DY, et al. The impact of prostate gland weight in robot assisted laparoscopic radical prostatectomy. J Urol 2008;180:928 32. [21] Ahlering TE, Kaplan AG, Yee DS, Skarecky DW. Prostate weight and early potency in robot-assisted radical prostatectomy. Urology 2008;72:1263 8. [22] Kaouk JH, Hafron J, Goel R, Haber GP, Jones JS. Robotic salvage retropubic prostatectomy after radiation/brachytherapy: initial results. BJU Int 2008;102:93 6. [23] Boris RS, Bhandari A, Krane LS, Eun D, Kaul S, Peabody JO. Salvage robotic-assisted radical prostatectomy: initial results and early report of outcomes. BJU Int 2009;103:952 6. [24] Murphy DG, Pedersen J, Costello AJ. Salvage robotic-assisted laparoscopic radical prostatectomy following failed primary high-intensity focussed ultrasound treatment for localised prostate cancer. J Robotic Surg 2008;2:201 3. [25] Rodriguez E, Skarecky D, Ahlering T. Salvage robot-assisted radical prostatectomy with pelvic lymph node dissection after cryotherapy failure. J Robotic Surg 2007;1:89 90. [26] 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. [27] Morgan M, Smith N, Thomas K, Murphy DG. Is Clavien the new standard for reporting urological complications? BJU Int 2009;104:434 6. [28] Donat SM. Standards for surgical complication reporting in urologic oncology: time for a change. Urology 2007;69:221 5. [29] Martin RC, Brennan MF, Jaques DP. Quality of complication reporting in the surgical literature. Ann Surg 2002;235:803 13. [30] Hu JC, Nelson RA, Wilson TG, et al. Perioperative complications of laparoscopic and robotic assisted laparoscopic radical prostatectomy. J Urol 2006;175:541 6. [31] Fischer B, Engel N, Fehr JL, John H. Complications of robotic assisted radical prostatectomy. World J Urol 2008;26:595 602. [32] Novara G, Ficarra V, D Elia C, Secco S, Cavalleri S, Artibani W. Prospective evaluation with standardised criteria for postoperative complications after robotic-assisted laparoscopic radical prostatectomy. Eur Urol 2010;57:363 70. [33] Badani KK, Kaul S, Menon M. Evolution of robotic radical prostatectomy: assessment after 2766 procedures. Cancer 2007;110: 1951 8. [34] Murphy DG, Kerger M, Crowe H, Peters JS, Costello AJ. Operative details and oncological and functional outcome of robotic-assisted laparoscopic radical prostatectomy: 400 cases with a minimum of 12 months follow-up. Eur Urol 2009;55:1358 67. [35] Joseph JV, Rosenbaum R, Madeb R, Erturk E, Patel HR. Robotic extraperitoneal radical prostatectomy: an alternative approach. J Urol 2006;175:945 50. [36] Atug F, Thomas R. Transperitoneal versus extraperitoneal roboticassisted radical prostatectomy: which one? Minerva Urol Nefrol 2007;59:143 7. [37] Danic MJ, Chow M, Alexander G, Bhandari A, Menon M, Brown M. Anesthesia concerns for robotic-assisted laparoscopic prostatectomy: a review of 1,500 cases. J Robotic Surg 2007;1:119 23. [38] Awad H, Santilli S, Ohr M, et al. The effects of steep Trendelenburg positioning on intraocular pressure during robotic radical prostatectomy. Anesth Analg 2009;109:473 8.