Systematic Review and Meta-analysis of Studies Reporting Oncologic Outcome After Robot-assisted Radical Prostatectomy

Transcription

1 EUROPEAN UROLOGY 62 (20) available at journal homepage: Platinum Priority Review Prostate Cancer Editorial by Peter C. Albertsen on pp of this issue Systematic Review and Meta-analysis of Studies Reporting Oncologic Outcome After Robot-assisted Radical Prostatectomy Giacomo Novara a, *, Vincenzo Ficarra a,b, Simone Mocellin a, Thomas E. Ahlering c, Peter R. Carroll d, Markus Graefen e, Giorgio Guazzoni f, Mani Menon g, Vipul R. Patel h, Shahrokh F. Shariat i, Ashutosh K. Tewari i, Hendrik Van Poppel j, Filiberto Zattoni a, Francesco Montorsi k, Alexandre Mottrie b, Raymond C. Rosen l, Timothy G. Wilson m a University of Padua, Padua, Italy; b O.L.V. Robotic Surgery Institute, Aalst, Belgium; c University of California, Irvine, Orange, CA, USA; d University of California, San Francisco Helen Diller Family Comprehensive Cancer Center, San Francisco Veterans Affairs Medical Center, San Francisco, CA, USA; e Martini- Clinic, Prostate Cancer Centre, University Hamburg-Eppendorf, Hamburg, Germany; f University Vita-Salute San Raffaele, H. San Raffaele-Turro, Milan, Italy; g Vattikuti Urology Institute, Henry Ford Hospital, Detroit, MI, USA; h Global Robotics Institute, Florida Hospital Celebration Health, Celebration, FL, USA; i Weill Medical College of Cornell University, New York, NY, USA; j University Hospital, Katholieke Universiteit Leuven, Leuven, Belgium; k University Vita-Salute San Raffaele, H. San Raffaele, Milan, Italy; l New England Research Institutes, Inc., Watertown, MA, USA; m City of Hope National Cancer Center, Duarte, CA, USA Article info Article history: Accepted May 22, 20 Published online ahead of print on June 2, 20 Keywords: Prostatic neoplasms Prostatectomy Laparoscopy Robotics Please visit europeanurology to read and answer questions on-line. The EU-ACME credits will then be attributed automatically. Abstract Context: Despite the large diffusion of robot-assisted radical prostatectomy (RARP), literature and data on the oncologic outcome of RARP are limited. Objective: Evaluate lymph node yield, positive surgical margins (PSMs), use of adjuvant therapy, and biochemical recurrence (BCR)free survival following RARP and perform a cumulative analysis of all studies comparing the oncologic outcomes of RARP and retropubic radical prostatectomy (RRP) or laparoscopic radical prostatectomy (LRP). Evidence acquisition: A systematic review of the literature was performed in August 2011, searching Medline, Embase, and Web of Science databases. A free-text protocol using the term radical prostatectomy was applied. The following limits were used: humans; gender (male); and publications dating from January 1, A cumulative analysis was conducted using Review Manager software v.4.2 (Cochrane Collaboration, Oxford, UK) and Stata 11.0 SE software (StataCorp, College Station, TX, USA). Evidence synthesis: We retrieved 79 papers evaluating oncologic outcomes following RARP. The mean PSM rate was 15% in all comers and 9% in pathologically localized cancers, with some tumor characteristics being the most relevant predictors of PSMs. Several surgeon-related characteristics or procedure-related issues may play a major role in PSM rates. With regard to BCR, the very few papers with a follow-up duration >5yr demonstrated 7-yr BCR-free survival estimates of approximately 80%. Finally, all the cumulative analyses comparing RARP with RRP and comparing RARP with LRP demonstrated similar overall PSM rates (RARP vs RRP: odds ratio [OR]: 1.21; p = 0.19; RARP vs LRP: OR: 1.; p = 0.47), pt2 PSM rates (RARP vs RRP: OR: 1.25; p = 0.31; RARP vs LRP: OR: 0.99; p = 0.97), and BCR-free survival estimates (RARP vs RRP: hazard ratio [HR]: 0.9; p = 0.526; RARP vs LRP: HR: 0.5; p = 0.141), regardless of the surgical approach. Conclusions: PSM rates are similar following RARP, RRP, and LRP. The few data available on BCR from high-volume centers are promising, but definitive comparisons with RRP or LRP are not currently possible. Finally, significant data on cancer-specific mortality are not currently available. # 20 Published by Elsevier B.V. on behalf of European Association of Urology. * Corresponding author. Department of Surgical and Oncological Sciences, Urologic Unit, Via Giustiniani 2, 350 Padua, Italy. Tel ; Fax: address: giacomonovara@gmail.com, giacomo.novara@unipd.it (G. Novara) /$ see back matter # 20 Published by Elsevier B.V. on behalf of European Association of Urology.

2 EUROPEAN UROLOGY 62 (20) Introduction Radical prostatectomy (RP) is a standard surgical treatment of clinically localized prostate cancer (PCa) [1]. Patients who undergo RP experience 15-yr cancer-specific mortality (CSM) ranging from 7% to 20%. Conversely, 15-yr biochemical recurrence (BCR)free survival estimates may be as high as 75% [2,3]. These figures were obtained in series of retropubic RPs (RRPs) performed in US and European referral centers. Robot-assisted RP (RARP) also has become very popular in the United States and Europe; it has been estimated that >75% of RPs are performed using the da Vinci platform (Intuitive Surgical, Inc., Sunnyvale, CA, USA) [4,5]. Literature and data on the oncologic outcome of RARP are limited and sparse. Very few series reported BCR rates at a follow-up duration as long as 5 yr [6,7], which was still insufficient for a comprehensive evaluation of CSM. More data are available on other outcomes that can be considered surrogates for oncologic control (eg, positive surgical margin [PSM] rates). We previously reported that, compared with RRP, RARP was associated with a significantly lower risk of overall PSMs and PSMs in pathologically confined disease, whereas statistically significant differences between RARP and laparoscopic radical prostatectomy (LRP) failed to be identified [8]. However, data on other relevant issues, such as lymph node yield during robotic lymph node dissection and use of adjuvant therapies following RARP, are sparse and controversial [8]. Because of the increasing use of RARP as well as the mounting literature in the field of oncologic outcomes of RARP and controversies in the available literature, we elected to update our previous systematic reviews. We aimed at evaluating lymph node yield after RARP, prevalence and risk factors for PSMs after RARP, surgical techniques that are able to improve PSM rates after RARP, use of adjuvant therapy after RARP, and BCR-free survival estimates following RARP. Finally, we aimed at performing a cumulative analysis of all studies comparing the oncologic outcomes of RARP and RRP or LRP. 2. Evidence acquisition To update our previous systematic reviews [8,9], we performed a literature search in August 2011 using the Medline, Embase, and Web of Science databases. The Medline search included only a free-text protocol using the term radical prostatectomy in the title and abstract fields of the records. The following limits were used: humans; gender (male); and publications dating from January 1, The searches of the Embase and Web of Science databases used the same free-text protocol, keyword, and publication dates. Two authors (G.N. and V.F.) separately reviewed the records to select RARP series as well as studies that compared RRP with LRP, RRP with RARP, and LRP with RARP, and discrepancies were resolved by open discussion. Other significant studies cited in the reference lists of the selected papers were evaluated, as were studies published after the systematic search. All noncomparative studies reporting the outcome of RARP for >0 cases were collected. The present review included only studies reporting oncologic outcomes (ie, lymph node yield, PSM rates, use of adjuvant therapies, and BCR-free survival rates). Studies published only as abstracts and reports from meetings were not included in the review. All the data retrieved from the selected studies were recorded in an electronic database. Quality control of the electronic data recording was performed on a random sample of papers (approximately 15% of the articles). All the papers were categorized according to the 2011 level of evidence for therapy studies: systematic review of randomized trials or n-of-1 trials (level 1); randomized trials or observational studies with dramatic effect (level 2); nonrandomized controlled cohort/follow-up studies (level 3);, casecontrol studies, or historically controlled studies (level 4); and mechanism-based reasoning (level 5) [13] Statistical analysis Cumulative analysis was conducted using Review Manager v.4.2 software designed for composing Cochrane Reviews (Cochrane Collaboration, Oxford, UK). Statistical heterogeneity was tested using the x 2 test. A p value <0. was used to indicate heterogeneity. Where there was a lack of heterogeneity, fixed-effects models were used for the cumulative analysis. Random effects models were used in case of heterogeneity. For continuous outcomes, the results were expressed as weighted mean differences and standard deviations; for dichotomous variables, results were given as odds ratios (OR) and 95% confidence intervals (CIs). For cumulative analysis of BCR, statistical analyses were performed using Stata 11.0 SE software (StataCorp, College Station, TX, USA). A weighted average of study-specific estimates of the hazard ratio (HR) was calculated using the inverse of variance as the weighting factor. RARP was considered the reference treatment with which either RRP or LRP was compared. The natural logarithm of HR and the corresponding standard error were used as data points for the meta-analysis. In studies performing Cox multivariable survival analysis, HR and CI were usually reported. For studies performing only univariable survival analysis, HR and 95% CI were calculated from survival curves adopting a hierarchical series of steps, as in Parmar et al. [14]. For indirect treatment comparisons, an extended version of the Bucher method [15] was used to obtain HR estimates from studies in which all three surgery types were considered but RARP was not the reference treatment. For all statistical analyses, a two-sided p < 0.05 was considered statistically significant. 3. Evidence synthesis 3.1. Quality of the studies and level of evidence Figure 1 shows the flowchart of this systematic review of the literature. We selected 9 records reporting oncologic outcomes after RARP. Two further studies (one level 2 and one level 3) published during the realization of the systematic review were added [16,17]. Forty-seven abstracts or meeting

3 384 [(Fig._1)TD$FIG] EUROPEAN UROLOGY 62 (20) Fig. 1 Flow-chart of the systematic review. RARP = robot-assisted radical prostatectomy; RRP = retropubic radical prostatectomy. reports and five duplicate publications were excluded. The remaining studies were 34 (level 4), 24 studies comparing different techniques in the context of RARP (1 study, level 2; 14 studies, level 3; 9 studies, level 4), 13 studies comparing RARP with RRP (5 studies, level 3; 8 studies, level 4), and 8 studies comparing RARP with LRP (1 study, level 2; 7 studies, level 4) Lymph node yield after robot-assisted radical prostatectomy Table 1 summarizes lymph node yield in the surgical series published between 2008 and Very few series reported data on extension of lymph node dissection, lymph node yield, prevalence of positive nodes, or complications of lymphadenectomy. Freicke et al. [19], Ham et al. [20], and Menon et al. [21] reported data on external iliac, internal iliac, and obturator lymph node dissection; they obtained mean numbers of nodes ranging from to 19 and positive node rates ranging from 11% to 24%, according to the different patient characteristics. As expected, such an extended lymph node dissection template was associated with complication rates higher than those reported in the series in which a more limited template was adopted. Freiche et al. reported a 7% complication rate, including lower lymphedema (2%), lymphocele (4%), and lymphatic fistula at a port site (1%) [19]. A single study compared transperitoneal and extraperitoneal limited lymph node dissection [25]. The study demonstrated that the lymph node yield was similar regardless of the surgical approach, suggesting that an extraperitoneal approach did not reduce the possibility of performing limited lymph node dissection. However, as expected, extraperitoneal lymph node dissection seemed to be associated with a slightly higher risk of postoperative lymphoceles than transperitoneal lymph node dissection (6% compared with 4%, respectively) [25]. Waldert et al. demonstrated that the use of FloSeal in the lymphadenectomy zone immediately after extraperitoneal lymph node resection significantly reduced the risk of lymphocele compared with standard extraperitoneal dissection (3% vs 14%, respectively) [26] Incidence and predictors of positive surgical margin rates after robot-assisted radical prostatectomy Table 2 summarizes the prevalence of PSMs in the surgical series published between 2008 and The prevalence of PSMs ranged from 6.5% to 32%, with a mean value of 15%. Many series did not report the pathologic protocol for sampling of the RARP specimens, which is a well-known factor affecting the PSM rate. Table 3 summarizes the prevalence of PSMs stratified by pathologic tumor stage in the surgical series published between 2008 and The mean PSM rate was 9%

4 Table 1 Lymph node yield in robot-assisted radical prostatectomy series First author Institution Cases, n Study design Indications PSA ng/ml, mean/ median Clinical T stage, % Biopsy Gleason score, % Patients receiving LND, % Anatomic extension of LND Operative time, min, mean/ median No. of retrieved nodes pn+, % Site of metastatic node Complications due to LND Cooperberg, 20 [18] Feicke, 2009 [19] Ham, 2009 [20] Menon, 2009 [21] Zorn, 2009 [22] Truesdale, 20 [23] Jayram, 2011 [24] University of California, San Francisco, San Francisco, CA, USA University of Zurich, Zurich, Switzerland Yonsei University College of Medicine, Seoul, Korea Vattikuti Urology Institute, Detroit, MI, USA University of Chicago, IL, USA Columbia University Medical Center, New York, NY, USA University of Chicago, IL, USA Retrospective ct: 200 ct3: PSA > ng/ml or Gleason score >6 orct>2a PSA > ng/ml or Gleason score >6 or ct3 32 Internal iliac, obturator, 7.7 ( ) Unclear Unclear with historical control 2007 Partin table predicted likelihood of nodal metastasis 01% [PSA < ng/ml, primary Gleason score 3 on biopsy (3 + 3 or 3 + 4), ct1c] 55 T2bT3 disease, Gleason score of 8, or PSA >20 ng/ml Retrospective 148 high risk PSA > ng/ml, or primary Gleason pattern >3 orct>2a T1: 67% T2: 27% T3: 36% 6: 22% 7: 65% 8: 13% 6: 67% 7: 24% 8: 9% 6: 26% 7: 25% 8: 49% T1: 75% T1: 0% All 80 All External iliac, internal iliac, obturator Obturator, external iliac, and internal iliac lymph nodes All External iliac and obturator Internal iliac and obturator nodes T1: 25% External iliac, obturator, internal iliac 9 T1: 61% T2: 38% T3: 1% Ta: 58% T2b: 4% T2c3: 38% 6: 16% 7: 68% 8: 16% 6: 28% 7: 34% 8: 78% All External iliac, obturator All External iliac, obturator Lymphoceles 1% 51 (2981) 19 (853) 16 5 internal iliac, 6 external iliac, 9 obturator fossa %: Lower lymphedema (2%); lymphocele (4%); lymphatic fistula at a port site (1%) Lymphocele 0.5%.5 (716) 8 2% All Lymphocele 2.5% EUROPEAN UROLOGY 62 (20)

5 386 EUROPEAN UROLOGY 62 (20) Table 1 (Continued ) Complications due to LND Site of metastatic node pn+, % No. of retrieved nodes Operative time, min, mean/ median Anatomic extension of LND Patients receiving LND, % Biopsy Gleason score, % Clinical T stage, % Indications PSA ng/ml, mean/ median First author Institution Cases, n Study design Lymphocele 6% Comparative studies Chung, 2011 [25] Lymphocele 4% Retrospective comparative study with historical control, level 4 Extraperineal: 155 Transperineal: 5 Yonsei University College of Medicine, Seoul, Korea Lymphocele 3% All Obturatoric and external iliac T2:75% T3: 25% Unclear Extraperitoneal RARP with FloSeal: 32 Standard extraperitoneal RARP: 1 Medical University of Vienna, Vienna, Austria Waldert, 2011 [26] Lymphocele 14% : 56% 7: 41% 8: 3% 6: 46% 7: 44% 8: % T2:70% T3: 30% LND = lymph node dissection; PSA = prostate-specific antigen; RARP = robot-assisted radical prostatectomy. All studies are level 4 evidence. (range: 423%) in pt2 cancers, 37% (range: 2950%) in pt3 cancers, and 50% (range: 4075%) in pt4 cancers. Table 4 summarizes the prevalence of PSMs stratified by PSM location in the surgical series published between 2008 and Surgical margins were positive at the prostate apex in 5% (range: 17%) of the cases, anteriorly in 0.6% (range: 0.22%), at the bladder neck in 1.6% (range: %), and posterolaterally in 2.6% (range: 221%). Finally, multifocal PSMs were detected in 2.2% (range: 29%) of the cases. Relatively few studies evaluated predictors of PSMs (Table 5). In the most extensive analysis, Ficarra et al. evaluated clinical and pathologic predictors of any PSM, posterolateral PSMs, multiple PSMs, and PSMs in pt2 cancers in a series of 322 patients [33]. Prostate volume (HR: 0.4; p = 0.002) and ct stage (HR: 2.2; p = 0.008) were the only clinical variables predictive of any PSM, whereas pt stage (HR: 11.9; p < 0.001) was the unique pathologic predictor. The ct stage (HR: 2.0; p = 0.025) and biopsy Gleason score ( p for trend: 0.019) were the only preoperative predictors of posterolateral PSMs, whereas pt stage (HR: 7.5; p < 0.001) and perineural invasion (HR: 3.4; p = 0.037) were the pathologic predictors. The pt stage (HR: 0.3; p < 0.001) and RP specimen Gleason score ( p for trend <0.001) were predictors of multiple PSMs. Finally, perineural invasion (HR: 4.1; p = 0.028) was the only predictor of PSMs in pt2 cancer. Further data were provided by a large multicenter collaboration that collected data on >8000 patients, including >00 cases with PSMs [46]. The authors found that patients body mass index (BMI), prostate-specific antigen (PSA) level, pt stage, and prostate volume (all p values <0.001) were independent predictors of any PSM, whereas Gleason score was not. Similarly, BMI, PSA, and prostate volume (all p values <0.001) were predictors of PSMs in pt2 cancers [46] Positive surgical margins in difficult cases Table 6 summarizes the data on the prevalence of PSMs in special, difficult cases, including patients with high BMI, patients with large prostate, patients who had prior abdominal surgery, patients who had surgery for benign prostatic hyperplasia before RARP, and patients with a median lobe. Nine papers addressed PSM rates in these special conditions [4957]. In all cases, these conditions were not associated with an increased risk of PSMs. Patients with larger prostates were indeed at lower risk of PSM Aspects of surgery influencing positive surgical margin rates after robot-assisted radical prostatectomy Table 7 shows studies that evaluated the association of surgeon experience with PSM rates. Zorn et al. compared PSM rates for the first 700 RARPs performed in a high-volume institution, stratifying patients into three consecutive groups (cases 1300, cases , and cases ). The overall PSM rates were stable in the three groups (19%, 19%, and 16%, respectively), although a statistically significant decrease in PSM rates in pt2 cancers was observed (15%, %, and 7%,

6 Table 2 Positive surgical margin in robot-assisted radical prostatectomy series including >0 cases published between 2008 and 2011 First author Institution Cases, n Study design Sampling protocol Pathologic stage, % Overall PSM rate, % pt2 pt3a pt3b pt4 Park, 2008 [27] Yonsei University College of Medicine, Seoul, Korea 200 Unclear Not reported Liss, 2008 [28] University of California, Irvine, Irvine, CA, USA 216 Retrospective 5-mm section Patel, 2008 [29] Global Robotic Institute, Celebration, FL, USA mm section Tewari, 2008 [30] Weill Cornell Medical College, New York, NY, USA 215 Not reported Carlucci, 2009 [31] Mount Sinai Medical Center, New York, NY, USA 700 Not reported Davis, 20 [32] University of Texas MD Anderson Cancer Center, Houston, TX, USA 178 Not reported Ficarra, 2009 [33] University of Padua, Padua, Italy mm section, Jaffe, 2009 [34] Montsouris, Paris, France 293 Not reported Martin, 2009 [35] Mayo Clinic Arizona, Phoenix, AZ, USA 509 Retrospective Not specified 22 Murphy, 2009 [36] Melbourne, Australia to 4-mm section Ploussard, 20 [37] Creteil, Paris, France 208 Not reported Shikanov, 2009 [38] University of Chicago, IL, USA mm section Coelho 20 [39] Global Robotic Institute, Celebration, FL, USA 876 Whole mount, 4-mm section Lasser, 20 [40] Warren Alpert Medical School at Brown University, Providence, RI, USA 239 Not reported Lee, 20 [41] Yonsei University College of Medicine, Seoul, Korea 307 Unclear Not reported 13 Shikanov, 20 [42] University of Chicago, IL, USA 1436 Not reported Tewari, 20 [43] Weill Cornell Medical College, New York, NY, USA 1340 Not reported Heldt, 2011 [44] Loma Linda University, Loma Linda, CA, USA 418 Retrospective Not reported 15 Jayram, 2011 [24] University of Chicago, Chicago, IL, USA 148 D Amico high risk 3-mm section 20 Lebeau, 2011 [45] PitiéSalpetriere Hospital, Paris, France 240 Not reported Patel, 2011 [46] Multiinstitutional 8095 Retrospective Whole mount, to 5-mm section Overall PSM 15 (6.532) PSM = positive surgical margin. All studies are level 4 evidence. EUROPEAN UROLOGY 62 (20)

7 Table 3 Positive surgical margin stratified by pathological stage in robot-assisted radical prostatectomy series including >0 cases published between 2008 and 2011 First author Institution Cases, n Study design Sampling protocol Pathologic stage, % PSM rate, % 388 pt2 pt3a pt3b pt4 pt2 pt3a pt3b pt4 Park, 2008 [27] Yonsei University College of Medicine, Seoul, Korea 200 Unclear Not reported Liss, 2008 [28] University of California, Irvine, Irvine, CA, USA 216 Retrospective 5-mm section Patel, 2008 [29] Global Robotic Institute, Celebration, FL, USA mm section Carlucci, 2009 [31] Mount Sinai Medical Center, New York, NY, USA 700 Not reported Davis, 20 [32] University of Texas MD Anderson Cancer Center, Houston, TX, USA 178 Not reported Ficarra, 2009 [33] University of Padua, Padua, Italy mm section, Jaffe, 2009 [34] Montsouris, Paris, France 293 Not reported Murphy, 2009 [36] Melbourne, Australia to 4-mm section Shikanov, 2009 [38] University of Chicago, Chicago, IL, USA mm section Coelho 20 [39] Global Robotic Institute, Celebration, FL, USA 876 Whole mount, 4-mm section Hong, 20 [47] George Washington University, Washington, DC, USA 469 Retrospective Not reported Lasser, 20 [40] Warren Alpert Medical School at Brown University, Providence, RI, USA 239 Not reported Patel, 2011 [46] Multiinstitutional 8095 Retrospective Whole mount, 4- to 5-mm section Overall PSM 9 (423) 37 (2950) 50 (4075) PSM = positive surgical margin. All studies are level 4 evidence. Table 4 Location of positive surgical margin in robot-assisted radical prostatectomy series including >0 cases published between 2008 and 2011 First author Institution Cases, n Study design Sampling protocol Pathological stage, % Location PSM, % Patel, 2008 [29] Carlucci, 2009 [31] Global Robotic Institute, Celebration, FL, USA Mount Sinai Medical Center, New York, NY, USA Ficarra, 2009 [33] Univeristy of Padua, Padua, Italy 322 Shikanov, 2009 [38] University of Chicago, IL, USA 1398 Coelho 20 [39] Tewari, 20 [43] Global Robotic Institute, Celebration, FL, USA Weill Cornell Medical College, New York, NY, USA Patel, 2011 [46] Multiinstitutional 8095 Retrospective pt2 pt3a pt3b pt4 Apex Anterior Bladder neck Posterolateral Multifocal 3-mm section Not reported mm section, mm section Whole mount, 4-mm section Not reported Whole mount, 4- to 5-mm section Overall PSM 5 (17) 0.6 (0.22) 1.6 () 2.6 (221) 2.2 (29) EUROPEAN UROLOGY 62 (20) PSM = positive surgical margin. All studies are level 4 evidence.

8 EUROPEAN UROLOGY 62 (20) Table 5 Predictors of positive surgical margin in robot-assisted radical prostatectomy series including >0 cases published between 2008 and 2011 Predictors of PSM Pathologic stage, % Overall PSM rate, % First author Institution Cases, n Study design Sampling protocol pt2 pt3a pt3b pt4 5-mm section PSAD, pt, pathologic Gleason score, case order, nerve sparing 216 Retrospective Liss, 2008 [28] University of California, Irvine, Irvine, CA, USA 5-mm section, Any PSM clinical variables: prostate volume and ct stage Any PSM pathologic variables: pt stage Posterolateral PSM clinical variables: ct stage, biopsy Gleason score Posterolateral PSM pathologic variables: pt stage, perineural invasion Multiple PSM pt stage, pathologic Gleason score PSM in pt2 cancers perineural invasion 322 Ficarra, 2009 [33] University of Padua, Padua, Italy Whole mount, ct/pt percentage of tumor 4-mm section Not reported 8 15 PSA, pt, pathologic Gleason score, prostate weight 876 Coelho 20 [39] Global Robotic Institute, Celebration, FL, USA 690 Marchetti, 2011 [48] University of Chicago, Chicago, IL, USA Any PSM: BMI, PSA, pt, prostate volume PSM in T2: BMI, PSA, prostate volume Whole mount, 4- to 5-mm section Patel, 2011 [46] Multiinstitutional 8095 Retrospective BMI = body mass index; PSA = prostate-specific antigen; PSAD = prostate-specific antigen density; PSM = positive surgical margin. All studies are level 4 evidence. respectively) [58]. Samadi et al. adopted a similar approach and reported stable PSM rates in the cohort and in pt2 cancers in a series of 1181 patients [59]. Two other studies compared the performance of surgeons who were fellowship trained in RARP and surgeons without RARP training [60,61]. Leroy et al. found that fellowship-trained RARP surgeons had significantly lower PSM rates than RRP surgeons moving to RARP (15% compared with 34%, p = 0.008) in the first 30 cases performed by each surgeon [61]. Similarly, Kwon et al. demonstrated that fellowship-trained surgeons outperformed surgeons who had previously performed >25 LRPs in terms of PSM rates (24% compared with 35%, p = 0.05) and PSMs at the prostate apex (8% compared with 21%, p = 0.003) [60]. Table 8 summarizes all the studies evaluating the impact of different modifications in RARP technique on PSM rates. Technical details in virtually all steps of RARP were addressed. PSM rates were not affected by the adoption of the transperitoneal compared with the extraperitoneal approach [25] or by preservation of the bladder neck [62]. Shikanov et al. demonstrated similar PSM rates in patients undergoing extrafascial and interfascial dissection of the neurovascular bundle [63]. However, significant differences were found in the location of PSMs in the two approaches. PSMs at the level of midprostate (37% vs 11%; p < 0.001) and lateral PSMs (29% vs 11%; p = 0.001) were more common in those patients receiving interfascial dissection, whereas bladder neck margins were more common in those having extrafascial dissection (6% vs 34%; p 0.01). Tewari et al. recently reported on a risk-stratified approach to nervesparing, with patients receiving different dissection of the layers of periprostatic fascial according to their risk of extraprostatic extension [16]. Specifically, patients with ct1c cancer, primary Gleason pattern 3, PSA < ng/ml, <5% of cancer in the biopsy, and negative magnetic resonance imaging (MRI) had the greatest degree of nerve sparing (so-called grade 1), with dissection of the lateral pelvic fascia (LPF) just outside the prostatic capsule. Patients with ct2 cancer, primary Gleason pattern 3, PSA < ng/ml, 622% of cancer in the biopsy, and negative MRI had grade 2 nerve sparing (dissection of LPF just outside the layer of veins of the prostate capsule and incision of the Denonvilliers fascia, leaving deeper layers on the rectum). Patients with primary Gleason pattern 4, PSA 19 ng/ml, 2349% of cancer in the biopsy, more than two positive cores, and apical tumor on MRI had grade 3 nerve sparing (inciding through the outer compartment of the LPF and excising all layers of Denonvilliers fascia). Finally, patients with Gleason score 8 or PSA >20 ng/ml, >50% of cancer in the biopsy, extraprostatic extension on MRI, or all positive cores had a nonnerve-sparing procedure (grade 4). Using such a risk-adjusted approach, the authors also demonstrated very low risk of PSM in the patients receiving the greatest degree of nerve sparing (PSM rates of % in grade 1, 8% in grade 2, 7% in grade 3, and 9% in grade 4; p =0.636). Three studies evaluated the effect of specific techniques for control of the dorsal venous complex (DVC) on PSMs.

9 390 Table 6 Positive surgical margin after robot-assisted radical prostatectomy in special, difficult cases First author Institution Cases, n Study design Sampling protocol Pathologic stage, % Overall PSM rate, % PSM rate, % pt2 pt3a pt3b pt4 pt2 pt3a pt3b pt4 Wiltz, 2009 [49] Moskovic, 20 [50] Zilberman, in press [51] University of Chicago, Chicago, IL, USA Mount Sinai Medical Center, New York, NY, USA Duke University Medical Center, Durham, NC, USA Patient BMI BMI <25: 216 comparative 3-mm section BMI 2530: BMI >30: BMI <25: 270 comparative Not reported BMI 2530: BMI >30: BMI <25: 0 comparative Not reported 17 BMI : BMI : BMI >35: PV Link, 2008 [52] City of Hope, Duarte, CA, USA PV <30: 69 comparative 5-mm section 35* 3049: * 5069: * 70: * Allaparthi, 20 [53] Turfts University, Brighton, PV <30 cm 3 : comparative Whole mount, * MS, USA PV 3049 cm 3 : mm section * PV 5079 cm 3 : * PV 80 cm 3 : * Martinez, 20 [54] University of Western Ontario, PV < 40 cm 3 : 75 comparative Not reported London, Ontario, Canada PV 4060 cm 3 : PV >60 cm 3 : Skolarus, 20 [55] University of Michigan Health PV <50 cm 3 : 582 Retrospective comparative study, Not reported 19* System, Ann Arbor, MI, USA PV 500 cm 3 : 279 level 4 11* PV >0 cm 3 :24 0 Huang, 2011 [56] Harvard Medical School, Prostate size 2441 g: 221 comparative study, Not reported Boston, MA, USA Prostate size 4250 g: 240 level Prostate size 5162 g: Prostate size >62 g: EUROPEAN UROLOGY 62 (20) Ginzburg, 20 [57] Huang, 2011 [56] Huang, 2011 [56] Connecticut Health Center, Farmington, CT, USA Harvard Medical School, Boston, MA, USA Harvard Medical School, Boston, MA, USA Prior surgery Prior abdominal surgery: 251 Retrospective comparative study, Not reported No prior surgery: 588 level Prior BPH surgery: 59 comparative study, Not reported 15 No prior BPH surgery: 892 level 3 13 Presence of median lobe Median lobe: 42 comparative study, Not reported No median lobe: 909 level 3 14 BMI = body mass index; BPH = benign prostatic hyperplasia; PV = prostate volume.

10 Table 7 Positive surgical margin after robot-assisted radical prostatectomy according to surgeon experience First author Institution Cases, n Study design, level of evidence Sampling protocol Pathologic stage, % Overall PSM rate, % Location PSM, % PSM rate, % pt2 pt3a pt3b pt4 Apex Anterior Bladder neck Posterolateral Multifocal pt2 pt3a pt3b pt4 Zorn, 2009 [58] Samadi, 20 [59] Kwon, 20 [60] Leroy, 20 [61] University of Chicago, Chicago, IL, USA Mount Sinai Medical Center, New York, NY, USA Kaiser Permanente Los Angeles Medical Center, Los Angeles, CA, USA Mayo Clinic Florida, Jacksonville, FL, USA Cases 1300 Cases Cases Case 1590 Case Case RARP by 4 fellowshiptrained surgeons 165 RARP by 8 robotically naïve, LRP surgeons 60 RARP by 2 fellowship-trained urologists 90 RARP by 3 open surgeons moving to RARP comparative comparative comparative comparative PSM = positive surgical margin; RARP = robot-assisted radical prostatectomy. * Statistically significant. Whole mount, 3-mm section No. of case performed Not reported * * 7 * Prior surgical experience Not reported * 8 * * 21 * Not reported 15 * 34 * EUROPEAN UROLOGY 62 (20)

11 Table 8 Positive surgical margin after robot-assisted radical prostatectomy (RARP) in different RARP surgical techniques First author Institution Cases, n Study design, level of evidence Sampling protocol Pathologic stage, % Overall PSM Location PSM, % PSM rate, % pt2 pt3a pt3b pt4 rate, % Apex Anterior Bladder neck Postero lateral Multifocal pt2 pt3a pt3b pt4 392 Chung, 2011 [25] Yonsei University College of Medicine, Seoul, Korea Extraperineal: 155 Transperineal: 5 Retrospective comparative study with historical control series, level 4 Transperitoneal vs extraperitoneal approach Not reported Freire, 2009 [62] Harvard Medical School, Boston, MA, USA Bladder neck preservation: 348 Standard: 271 Bladder neck preservation comparative Not reported Shikanov, 2009 [63] Kowalczyk, 2011 [64] Tewari, 2011 [16] Guru, 2009 [65] Wu, 20 [66] Lei, 2011 [67] University of Chicago, Chicago, IL, USA Harvard Medical School, Boston, MA, USA Weill Cornell Medical College, New York, NY, USA Roswell Park Cancer Institute, Buffalo, NY, USA Northwestern University, Chicago, IL USA Harvard Medical School, Boston, MA, USA Extrafascial NS: 1 Intrafascial NS: 703 NS without coutertraction: 342 NS with coutertraction: 268 Grade 1 NS: 619 Grade 2 NS: 363 Grade 3 NS: 154 Grade 4 NS: 62 Incision of the DVC without ligation: 145 Incision of the DVC after ligation: 158 DVC suture 67 DVC staple ligation 95 Athermal division and selective suturing of DVC: 240 Nonselective suture ligation followed by athermal division: 303 comparative study, level 3 comparative study with historical control series, level 4 comparative comparative comparative study with historical control series, level 4 comparative study, level 3 Neurovascular bundles dissection 3-mm section Not reported Not reported DVC control Not reported Not reported 18 * Not reported * 8 * 6 * 13 * * 1 * * EUROPEAN UROLOGY 62 (20) Tewari, 20 [68] Weill Cornell Medical College, New York, NY, USA Anatomic retroapical technique of synchronous (posterior and anterior) urethral transection: 209 Standard: 1665 case series with historical control, level 4 Urethral dissection Not reported * 4 * Patel, 2009 [69] Global Robotic Institute, Celebration, FL, US Periurethral suspension stitch: 237 Standard: 94 comparative Anterior and posterior reconstruction 4-mm section

12 EUROPEAN UROLOGY 62 (20) Location PSM, % PSM rate, % Pathologic stage, % Overall PSM pt2 pt3a pt3b pt4 rate, % Apex Anterior Bladder Sampling protocol Multifocal pt2 pt3a pt3b pt4 Postero lateral neck Not reported Not reported Not reported Guru et al. compared incision of the DVC without ligation to incision following ligation, demonstrating that incision without ligation was associated with a significantly lower risk of apical PSMs (2% and 8%, p = 0.02) [65]. Lei et al. compared athermal division and selective suturing of the DVC with nonselective suture ligation followed by athermal division and failed to demonstrate any statistically significant difference in overall PSM rates (% and %, p =1)or apical PSM rates (1% and 3%, p = 0.361). However, athermal division and selective suturing of the DVC were associated with significantly higher early urinary continence [67]. Wu et al. compared zero polyglactin suture of the DVC with staple ligation and found that staple ligation was followed by a significantly lower risk of overall PSMs (18% and 6%, p = 0.02) and apical PSMs (13% and 2%, p = 0.005) [66]. To reduce apical PSMs, Tewari et al. recently proposed an original technique for synchronous posterior and anterior urethral transection using a retroapical approach [68]. The posterior wall of urethra is incised after visualization of the apical-urethral junction and membranous urethra by way of a retroapical approach before incision of the anterior wall. Such an approach was associated with a 1.4% apical PSM rate, which is significantly lower than the 4% rate observed with the conventional urethral transection by way of an anterior approach [68]. Finally, four studies evaluated the impact of anterior, posterior, or complete anterior and posterior reconstruction of the vesicourethral junction [17,6971] and failed to demonstrate any impact on PSM rates Adjuvant therapy after robot-assisted radical prostatectomy Table 8 (Continued ) First author Institution Cases, n Study design, level of evidence case series with historical control, level 4 Standard:214 Anterior reconstruction: 303 Total anatomic restoration: 1383 Tan, 20 [70] Weill Cornell Medical College, New York, NY, USA with historical control, level 4 Anterior reconstruction: 87 Standard: 142 Johnson, 2011 [71] University of Michigan, Ann Arbor, MI, USA randomized study, level 2 Hurtes, 20 [17] Multicenter Anterior and posterior reconstruction: 39 Standard: 33 DVC = dorsal venous complex; NS = nerve sparing; PSM = positive surgical margin. Statistically significant. * Table 9 summarizes the use of adjuvant therapies in the surgical series published between 2008 and Very few series from referral centers reported on the issue. The reported use of adjuvant therapies ranged from 0.5% to 23%, with a mean value as low as 4% Biochemical recurrencefree survival rates after robotassisted radical prostatectomy Table summarizes BCR rates in the surgical series published between 2008 and Two series reported outcome at a follow-up duration of 60 mo [6,7]. Menon et al. reported on 1384 patients treated at the Vattikuti Urology Institute (Detroit, MI, USA) [6]. Adopting 0.2 ng/ml as the definition of PSA recurrence, at a median follow-up duration of 60 mo, the researchers found 3-, 5-, and 7-yr BCR rates as high as 90%, 87%, and 81%, respectively, with 95.5% cancer-specific survival. Among preoperative variables, the authors found that PSA, biopsy Gleason score, perineural invasion in the biopsy specimen, and D Amico risk group were independent predictors of BCR. Similarly, in a model that also included pathologic variables from the RP specimens, the authors demonstrated that PSA, pathology Gleason score, pt stage, surgical margin status, and presence of lymphovascular invasion were all independent predictors of BCR. Similar figures were reported by Suardi

13 394 EUROPEAN UROLOGY 62 (20) Table 9 Use of adjuvant therapies in robot-assisted radical prostatectomy series First author Institution Cases, n Study design Adjuvant therapy, % Follow-up, mo PSA recurrence, % Murphy 2009 [36] Melbourne, Australia Menon, 20 [6] Vattikuti Urology Institute, Detroit, MI, USA Novara, 2011 [72] University of Padua, Padua, Italy Jayram, 2011 [24] University of Chicago, Chicago, IL, USA 148 D Amico high risk Overall 4 PSA = prostate-specific antigen. All studies are level 4 evidence. et al. in a smaller series of 184 patients treated at the O.L.V. Robotic Surgery Institute, Aalst, Belgium, and evaluated at a median follow-up of 67.5 mo [7]. Three papers reported on the oncologic outcome of patients with high-risk PCa identified by PSA > ng/ml, Gleason score 8, or pt stage pt3b [75]; pathologic Gleason score 8 or 9 [76]; or D Amico high-risk group [24]. Although limited by short follow-up duration, these studies reconfirmed that RARP is followed by a relatively high risk of recurrence in such patients compared with low-risk patients. Table 11 summarizes the studies reporting the effects of patient characteristics, surgeon experience, and surgical technique on BCR rates. With regard to patient characteristics, prostate volume, prior abdominal surgery, and BMI turned out to be unrelated to oncologic outcome [50,52,57]. With regard to surgeon experience, expressed as the number of RARP cases performed, Zorn et al. [58] and Samadi et [59] demonstrated in two large series that the risk of PSA recurrence was quite stable over 700 and 00 cases, respectively. However, in both studies, followup duration was relatively short, and BCR-free survival analyses adjusted for covariates were not provided. Leroy et al. compared the oncologic outcomes of 60 RARPs performed by two fellowship-trained urologists with the outcomes of 90 RARPs performed by three open surgeons adopting RARP [61]. The comparison showed that the 3-mo PSA recurrence was higher for the patients operated on by the nonfellowship-trained surgeons (%) compared with fellowship-trained surgeons (2%) ( p = 0.056). Finally, with regard to surgical technique, a single paper evaluated the BCR rate following RARP with conventional vesicourethral anastomosis, anterior reconstruction, or total anterior and posterior reconstruction in a large series of 1900 patients treated at Weill Medical College of Cornell University, New York Presbyterian Hospital (New York, NY, USA). The evaluation failed to identify a specific effect of this surgical step on the risk of BCR [70] Cumulative analysis of studies comparing robot-assisted radical prostatectomy with retropubic radical prostatectomy or laparoscopic radical prostatectomy Table summarizes the comparative studies evaluating RRP and RARP that report PSM rates. Cumulative analyses showed only nonstatistically significant differences in overall PSM rates following RRP and RARP (21% and 20%; OR: 1.21; 95% CI, ; p = 0.19) (Fig. 2). Figure 3 shows a forest plot comparing PSM rates in pt2 cancers following RRP and RARP. PSM rates in pt2 cancers were similar following RRP and RARP (% and 11%; OR: 1.25; 95% CI, ; p = 0.31) (Fig. 3). Table 13 summarizes the comparative studies evaluating LRP and RARP that report PSM rates. PSM rates were similar following LRP and RARP (18% and 18%; OR: 1.; 95% CI, ; p =0.47)(Fig. 4). Similarly, PSM rates in pt2 cancers following LRP and RARP were overlapping (11% and %; OR: 0.99; 95% CI, ; p =0.97) (Fig. 5). Table 14 summarizes BCR-free survival estimates in comparative studies evaluating RRP, LRP, and RARP. Three studies reported BCR following RRP or RARP [86,95,96]; two studies compared RRP, LRP, and RARP [81,89]. No significant differences in BCR-free survival rates were demonstrated between RRP and RARP (HR: 0.9; 95% CI, ; p = 0.526) or between LRP and RARP (HR: 0.5; 95% CI, ; p = 0.141). 4. Discussion The data of the present systematic review suggest that extended lymph node dissection yielding a reasonably high number of lymph nodes is feasible during RARP, although such extended dissections might be associated with a risk of perioperative complications that is more than marginal. In the evaluated RARP series, the mean PSM rate was 15% in all comers and 9% in pathologically localized cancers, with tumor characteristics (ie, according to different studies, PSA, pt stage, Gleason score, and prostate volume) being the most relevant predictors of PSMs. Several surgeon-related characteristics (eg, caseload, type of RARP training, and prior surgical experience) or procedure-related issues (eg, type of nerve-sparing approach, technique for DVC control) may play a major role in PSM rates, although the literature data on those issues were not conclusive. Very few data are available on the use of adjuvant therapies, suggesting that a limited percentage of patients received such treatments following RARP. With regard to BCR, the very few papers with a follow-up duration >5 yr demonstrated 7-yr BCR-free survival estimates of approximately 80%, with the classic pathologic features (pt stage and RARP specimen Gleason score) being associated with the BCR. Finally, all the cumulative analyses comparing RARP with RRP and comparing RARP with LRP demonstrated

14 Table Biochemical recurrence rates in robot-assisted radical prostatectomy series First author Institution Cases, n Study design Followup, mo Adjuvant therapy Definition of BCR, ng/ml BCR rate Predictors of BCR CSS rate, % Carlucci, 2009 [31] Mount Sinai Medical Center, New York, NY, USA 700 Murphy, 2009 [36] Melbourne, Australia 395 Shikanov, 2009 [38] Shikanov, 2009 [73] Menon, 20 [6] Novara, 20 [72] University of Chicago, Chicago, IL, USA University of Chicago, Chicago, IL, USA Vattikuti Urology Institute, Detroit, MI, USA University of Padua, Padua, Italy Xylinas, in press [74] Creteil, Paris, France 500 Suardi, 20 [7] OLV Clinic, Aalst, Belgium 184 Retrospective Engel, 20 [75] Wambi, 20 [76] Jayram, 2011 [24] George Washington University Hospital, Washington, DC, USA Vattikuti Urology Institute, Detroit, MI, USA University of Chicago, Chicago, IL, USA 73 high-risk cases (PSA > ng/ml, specimen Gleason score 8, pt stage pt3b) 368 specimen Gleason score D Amico high risk Not reported PSA 0.2 Recurrence at followup: 2% 22 % PSA yr BCR: 74% Not reported 22 Not reported PSA 0.2 Recurrence at followup: 4% PSA, pt Gleason score, PSM, and PSM length PSA yr BCR: 96% 2-yr BCR: 91% 60 0 PSA >0.2 3-yr BCR: 90% 5-yr BCR: 87% 7-yr BCR: 81% Preoperative: D Amico risk group, perineural invasion in biopsy Pathologic model: PSA, pt, Gleason score, PSM, lymphovascular invasion 14 4% PSA >0.2 1-yr BCR: 94% 24 0 PSA > 0.2 Recurrence at followup % 69.3 Not reported High-risk patients 32 Not reported BCR = biochemical recurrence; CSS = cancer-specific survival; PSA = prostate-specific antigen; PSM = positive surgical margin. All studies are level 4 evidence. PSA > yr BCR:91% 5-yr BCR:84% 7-yr BCR:81% PSA > yr BCR: 77% 2-yr BCR: 69% 3-yr BCR: 59% 23 (46) 0 PSA > 0.2 Gleason 8 5yr BCR: 36 5 Gleason 9 5yr BCR: % PSA > 0.05 Recurrence at followup 21% Seminal vesicle involvement Gleason score pt, pn, pathologic Gleason score, tumor volume EUROPEAN UROLOGY 62 (20)

15 396 Table 11 Biochemical recurrence rates in robot-assisted radical prostatectomy series according to surgeon experience, patient characteristics, and details of surgical technique First author Institution Cases, n Study design, level of evidence Follow-up, mo Adjuvant therapy Definition of BCR, ng/ml Outcome Patient characteristics Link, 2008 [52] City of Hope, Duarte, CA, USA Prostate volume <30: : : : 327 Ginzburg, 20 [57] Moskovic, 20 [50] Zorn, 2009 [58] Samadi, 20 [59] Leroy, 20 [61] Connecticut Health Center, Farmington, CT, USA Mount Sinai Medical Center, New York, NY, USA University of Chicago, Chicago, IL, USA Mount Sinai Medical Center, New York, NY, USA Mayo Clinic Florida, Jacksonville, FL, USA Prior abdominal surgery: 251 No prior abdominal surgery: 588 BMI <25: 270 BMI 2530: 600 BMI > 30: 242 Cases 1300 Cases Cases Case 1590 Case Case RARP by 2 fellowship-trained urologists 90 RARP by 3 open surgeons moving to RARP comparative comparative comparative Surgeon experience comparative comparative comparative 34.7 (5.670) 28.3 (58) 27.7 (1155) 28.9 (59) Not reported PSA yr BCR: 94% 1-yr BCR: 91% 1-yr BCR: 93% 1-yr BCR: 94% Not reported PSA > 0.2 Recurrence at follow-up: 5% Recurrence at follow-up: 6% Not reported PSA > 0.2 Recurrence at follow-up: 3% Recurrence at follow-up: 4% Recurrence at follow-up: 7% Not reported PSA > 0.1 Recurrence at follow-up: 8% Recurrence at follow-up: 8% Recurrence at follow-up: 9.5% Not reported PSA > 0.2 Recurrence at follow-up: 5% Recurrence at follow-up: 5% Recurrence at follow-up: 6% Recurrence at follow-up: 4% 3 Not reported PSA > 0.15 Recurrence at follow-up: 2% Recurrence at follow-up: % EUROPEAN UROLOGY 62 (20) Surgical technique Tan, 20 [70] Weill Cornell Medical College, New York, NY, USA Standard:214 Anterior reconstruction: 303 Total anatomic Restoration: 1383 with historical control, level 4 Not reported Not reported Not reported Recurrence at follow-up: 6% Recurrence at follow-up: 4% Recurrence at follow-up: 4% BCR = biochemical recurrence; BMI = body mass index; PSA = prostate-specific antigen; RARP = robot-assisted radical prostatectomy.

16 EUROPEAN UROLOGY 62 (20) Table Positive surgical margin rates in the comparative studies evaluating retropubic radical prostatectomy and robot-assisted radical prostatectomy Level of evidence First author Cases, n Overall PSM, % pt2 PSM, % 3 Ficarra, 2009 [77] 5 RRP 3 RARP Di Pierro, 2011 [78] 75 RRP 75 RARP Kim, 2011 [79] 235 RRP 528 RARP 4 Caballero-Romeu, 2008 [80] 62 RRP 60 RARP Drouin, 2009 [81] 83 RRP 71 RARP Laurila, 2009 [82] 84 RRP 88 RARP Ou, 2009 [83] 30 RRP 30 RARP White, 2009 [84] 50 RRP 50 RARP Breyer, 20 [85] 695 RRP 293 RARP Barocas, 20 [86] 491 RRP 1413 RARP Doumerc, 20 [87] 502 RRP 2 RARP Lo, 20 [88] 20 RRP 20 RARP Magheli, 2011 [89] 522 RRP 522 RARP PSM = positive surgical margin; RARP = robot-assisted radical prostatectomy; RRP = retropubic radical prostatectomy. similar PSM rates and BCR-free survival estimates, regardless of the surgical approach. In both Europe and the United States, RARP has gained popularity because of the supposed superior perioperative and functional outcomes in the absence of major data on the oncologic outcomes of the procedure. Several recently published population-based studies compared all possible outcomes following RARP and standard RRP and suggest that the selective use of lymph node dissection [,97] or [(Fig._2)TD$FIG] adjuvant therapies [,11] might jeopardize patient outcome following RARP. Data on long-term cancer-specific survival following RARP are still lacking because of the limited time the procedure has been performed and the long natural history of clinically localized PCa. However, Menon et al. [6] and Suardi et al. [7] recently reported BCR-free survival rates similar to those reported in RRP series [2,3], and all the comparative studies evaluating PSA recurrence failed to demonstrate any significant difference among RARP, RRP, Fig. 2 Cumulative analyses of overall positive surgical margin rates following robot-assisted radical prostatectomy or retropubic radical prostatectomy. RARP = robot-assisted radical prostatectomy; RRP = retropubic radical prostatectomy.