Predictive Factors for Positive Surgical Margins and Their Locations After Robot-Assisted Laparoscopic Radical Prostatectomy

Transcription

1 EUROPEAN UROLOGY 57 (2010) available at journal homepage: Prostate Cancer Predictive Factors for Positive Surgical Margins and Their Locations After Robot-Assisted Laparoscopic Radical Prostatectomy Rafael F. Coelho a,b,c, Sanket Chauhan a, Marcelo A. Orvieto a, Kenneth J. Palmer a, Bernardo Rocco a,d, Vipul R. Patel a, * a Global Robotics Institute, Florida Hospital Celebration Health, Celebration, FL, USA b University of Central Florida School of Medicine, Orlando, FL, USA c Hospital das Clinicas da Faculdade de Medicina da Universidade de São Paulo, Divisão de Urologia, São Paulo, Brazil d Divisione di Urologia, Instituto Europeo di Oncologia, Milan, Italy Article info Article history: Accepted January 26, 2010 Published online ahead of print on February 5, 2010 Keywords: Prostate cancer Prostatectomy Risk factors Treatment outcome Surgical margins Robotic Abstract Background: Positive surgical margin (PSM) after radical prostatectomy (RP) has been shown to be an independent predictive factor for cancer recurrence. Several investigations have correlated clinical and histopathologic findings with surgical margin status after open RP. However, few studies have addressed the predictive factors for PSM after robot-assisted laparoscopic RP (RARP). Objective: We sought to identify predictive factors for PSMs and their locations after RARP. Design, setting, and participants: We prospectively analyzed 876 consecutive patients who underwent RARP from January 2008 to May Intervention: All patients underwent RARP performed by a single surgeon with previous experience of >1500 cases. Measurements: Stepwise logistic regression was used to identify potential predictive factors for PSM. Three logistic regression models were built: (1) one using preoperative variables only, (2) another using all variables (preoperative, intraoperative, and postoperative) combined, and (3) one created to identify potential predictive factors for PSM location. Preoperative variables entered into the models included age, body mass index (BMI), prostate-specific antigen, clinical stage, number of positive cores, percentage of positive cores, and American Urological Association symptom score. Intra- and postoperative variables analyzed were type of nerve sparing, presence of median lobe, percentage of tumor in the surgical specimen, gland size, histopathologic findings, pathologic stage, and pathologic Gleason grade. Results and limitations: In the multivariable analysis including preoperative variables, clinical stage was the only independent predictive factor for PSM, with a higher PSM rate for T3 versus T1c (odds ratio [OR]: 10.7; 95% confidence interval [CI], ) and for T2 versus T1c (OR: 2.9; 95% CI, ). Considering pre-, intra-, and postoperative variables combined, percentage of tumor, pathologic stage, and pathologic Gleason score were associated with increased risk of PSM in the univariable analysis ( p < for all variables). However, in the multivariable analysis, pathologic stage (pt2 vs pt1; OR: 2.9; 95% CI, ) and percentage of tumor in the surgical specimen (OR: 8.7; 95% CI, ; p = ) were the only independent predictive factors for PSM. Finally, BMI was shown to be an independentpredictive factor (OR: 1.1; 95% CI, ; p = ) for apicalpsms, with increasing BMI predicting higher incidence of apex location. Because most of our patients were referred from other centers, the biopsy technique and the number of cores were not standardized in our series. Conclusions: Clinical stage was the only preoperative variable independently associated with PSM after RARP. Pathologic stage and percentage of tumor in the surgical specimen were identified as independent predictive factors for PSMs when analyzing pre-, intra-, and postoperative variables combined. BMI was shown to be an independent predictive factor for apical PSMs. # 2010 European Association of Urology. Published by Elsevier B.V. All rights reserved. * Correspondingauthor. 410 Celebration Place, Suite 200, Celebration, FL34747, USA.Tel address: Vipul.patel.md@flhosp.org (V.R. Patel) /$ see back matter # 2010 European Association of Urology. Published by Elsevier B.V. All rights reserved. doi: /j.eururo

2 EUROPEAN UROLOGY 57 (2010) Introduction Positive surgical margin (PSM) after radical prostatectomy (RP) is an independent predictive factor of biochemical recurrence, local recurrence, and development of distant metastasis [1]. PSM also can cause significant psychologic distress to patients because men with PSMs remain more fearful in the long term following surgery compared with those with negative margins [2]. Therefore, the status of surgical margins is one of the most important outcomes to be evaluated in any innovative surgical treatment proposed for prostate cancer. In the last 2 decades, with the widespread diffusion of prostate-specific antigen (PSA) testing and improvements in surgical technique, the risk of PSMs has been reduced considerably [3]. Even so, PSMs remain a significant clinical problem and have been reported at % in large series of open RP (ORP) [4,5]. Therefore, identification of risk factors for PSMs may be helpful for the selection of candidates for a nerve-sparing procedure and for the introduction of technical refinements aiming to avoid this adverse oncologic outcome [6,7]. Several investigations have correlated clinical and histopathologic findings with margin status after ORP. Serum PSA, clinical stage, biopsy Gleason score, surgical technique, and individual surgeon s expertise have been consistently associated with PSMs after ORP [8 13]. However, few studies have addressed clinical and pathologic predictive factors for PSM after robot-assisted laparoscopic radical prostatectomy (RARP) [13,14]. The aim of this study was to identify predictive factors for PSM by evaluating pre-, intra-, and postoperative variables from 876 consecutive patients who underwent RARP. We also analyzed predictive factors associated with specific PSM locations following RARP Prostate biopsy and pathologic analysis of the surgical specimens Because most of the patients were referred from other centers, the biopsy technique and the number of cores were not standardized, although all of the slides were reviewed by a dedicated uropathologist at our institution. Primary and secondary Gleason patterns (grades) and total number and percentage of positive cores in the biopsy specimens were evaluated. Prostate weight was assessed in the surgical specimens. The apex and bladder-neck cones of the surgical specimen were amputated and sectioned in the sagittal plane. The remaining specimen was sectioned transversely at intervals of 4 mm formalin fixed and routinely processed for paraffin embedding. PSMs were defined as the presence of tumor tissue on the inked surface of the specimen and were categorized into four groups based on their locations: apex, bladder neck (BN), posteriolateral (PL), and multifocal (MF) Statistical analysis Stepwise logistic regression was used to identify potential predictive factors for PSM. Two logistic regression models were initially built, one using preoperative variables only (model 1) and another (model 2) using pre-, intra-, and postoperative variables combined. Preoperative variables entered into the models were age, body mass index (BMI), prostate-specific antigen (PSA) level, clinical stage, number of positive cores, percentage of positive cores, biopsy Gleason grade, and American Urological Association symptom score (AUA-SS). Intra- and postoperative variables entered into the models included type of nerve sparing (defined as non nerve sparing [NNS], unilateral nerve sparing [UNS], or bilateral nerve sparing [BNS]), presence of median lobe, percentage of tumor in the surgical specimen, gland size, histopathologic characteristics (perineural invasion, high-grade prostatic intraepithelial neoplasia, prostatitis, atrophy), pathologic stage, and pathologic Gleason grade. Finally, a third logistic regression model (model 3) was built to identify potential predictive factors of PSM location within the subgroup of patients with positive margins. Statistical models were developed individually for apex and PL locations; however, modeling was not feasible for BN and MF PSMs due to inadequate sample size. A two-sided 2. Patients and methods We analyzed 876 consecutive patients who underwent RARP at our institution from January 2008 to May All of the procedures were performed by a single surgeon (VRP) with previous experience of >1500 cases. After institutional review board approval, the data were prospectively collected in a customized database and retrospectively analyzed Surgical technique All procedures were performed using a transperitoneal, six-port technique described by the authors previously [15]. A nerve-sparing procedure was performed in those patients with ct1 ct2a prostate cancer, biopsy Gleason score 7, and preoperative Sexual Health Inventory for Men score >21. In selected patients with Gleason score 8 and small tumor volume, nerve sparing was also performed. The nervesparing technique was modified and performed athermally, with an early retrograde release of the neurovascular bundle (NVB) [16]. This technique involves releasing the NVB in a retrograde fashion, from the apex toward the base of the prostate, developing an interfascial plane between the neurovascular bundle laterally and the prostatic fascia medially. Table 1 Patient characteristics Characteristics N = 876 Age, yr, median (IQR) 61 (56 66) BMI, kg/m 2, median (IQR) 28 (26 31) PSA level before RARP, ng/ml, median (IQR) 4.9 ( ) Prostate weight, g, median (IQR) 48 (40 60) AUA-SS, median (IQR) 7 (3 12) Biopsy Gleason score, % Pathologic stage, % pt2 81 pt pt4 0.8 Pathologic Gleason score, % AUA-SS = American Urological Association symptom score; BMI = body mass index; IQR = interquartile range; PSA = prostate-specific antigen; RARP = robot-assisted laparoscopic radical prostatectomy.

3 1024 EUROPEAN UROLOGY 57 (2010) Table 2 Association between individual categorical and continuous variables with positive surgical margins: univariable analysis Potential predictive factor Margin status p value Negative Positive Preoperative Age Median (IQR) 61 (56 66) 62 (56 66) BMI Median (IQR) 28 ( ) 28 (26 30) PSA Median (IQR) 4.9 ( ) 5 ( ) Clinical stage, n (%) <0.001 T1c 632 (91.4) 59 (8.54) T2 139 (78.5) 38 (21.47) T3 4 (50) 4 (50) AUA-SS, n (%) Asymptomatic 30 (83.33) 6 (16.67) (88.05) 52 (11.95) (89.82) 34 (10.18) (87.32) 9 (12.68) Biopsy Gleason grade, n (%) (89) 56 (11) (87.92) 36 (12.08) 8 60 (86.96) 9 (13.04) No. of positive cores Median (IQR) 2 (1, 4) 3 (1, 5) Percentage of positive cores, % Median (IQR) 25 ( ) 33.3 ( ) Intraoperative Type of nerve sparing, n (%) pt2 NNS 43 (91.48) 4 (8.51) BNS 518 (91.84) 46 (8.15) UNS 107 (93.86) 7 (6.14) pt NNS 27 (69.23) 12 (30.76) BNS 65 (72.2) 25 (27.77) UNS 11 (73.33) 4 (26.66) Median lobe, n (%) Absent 643 (87.8) 89 (12.1) Present 132 (91.7) 12 (8.3) Postoperative PNI, n (%) Absent 285 (91.35) 27 (8.65) Present 490 (86.88) 74 (13.2) HGPIN, n (%) Absent 235 (87.04) 35 (12.96) Present 540 (89.11) 66 (10.89) Prostatitis, n (%) Absent 557 (88.98) 69 (11.02) Present 218 (87.2) 32 (12.8) Atrophy, n (%) Absent 724 (88.29) 96 (11.71) Present 51 (91.07) 5 (8.93) Pathologic stage, n (%) <0.001 pt2 668 (92.14) 50 (6.8) pt3 105 (71.92) 55 (34.02) pt4 1 (25) 3 (75) Pathologic Gleason grade, n (%) < (93.66) 21 (6.34) (86.42) 66 (13.58) 8 43 (75.44) 14 (24.56) Percentage of tumor, % <0.001 Median (IQR) 15 (10 20) 20 (15 30) Prostate weight, g Median (IQR) 48 (40 60) 47 ( ) AUA-SS = American Urological Association symptom score; BMI = body mass index; BNS = bilateral nerve sparing; HGPIN = high-grade prostatic intraepithelial neoplasia; IQR = interquartile range; NNS = non nerve sparing; PNI = perineural invasion; PSA = prostate-specific antigen; UNS = unilateral nerve sparing.

4 EUROPEAN UROLOGY 57 (2010) p < 0.05 was considered statistically significant. All statistical analyses were performed using SPSS v.16.0 (SPSS Inc, Chicago, IL, USA). 3. Results Patient demographics are shown in Table 1. The overall PSM rate was 11.5% (101 of 876). The distribution of these PSMs by location was 38.6% (39 of 101) in the apex, 34.6% (35 of 101) PL, 15.8% (16 of 101) MF, and 10.9% (11 of 101) in the BN Model 1: Preoperative factors only No association was observed in the univariable analysis (Table 2) between PSM and age ( p =0.882),BMI(p =0.746), PSA ( p = 0.300), number of positive cores ( p = 0.078), percentage of positive cores ( p = 0.102), biopsy Gleason grade ( p = 0.805), and AUA-SS ( p = 0.643). In the multivariable analysis including preoperative variables, clinical stage was the only significant predictive factor, with higher PSM rates for T3 versus T1c (odds ratio [OR]: 10.7; p < ) and for T2 versus T1c (OR: 2.9; p < ) (Table 3) Model 2: Preoperative, intraoperative, and postoperative factors combined Among patients with pt2 tumors, the PSM rates were similar regardless of the type of nerve-sparing procedure performed (8.15%, 6.14%, and 8.51% for BNS, UNS, and NNS, respectively; p = 0.755). Similarly, among patients with pt3 tumors, the PSM rates were also comparable for BNS, UNS, and NNS procedures (27.7%, 26.66%, and 30.76%, respectively; p = 0.929). Percentage of tumor, pathologic stage, and pathologic Gleason score were associated with increased risk of PSM in the univariable analysis ( p < for all variables; Table 2). In the multivariable analysis combining pre-, intra-, and postoperative variables, pathologic stage ( p < ) and percentage of tumor in the surgical specimen ( p = ) were the only independent predictive factors for PSM. Pathologic stage was highly correlated with clinical stage and pathologic Gleason grade, but pathologic stage alone sufficiently explained the proportion of variance associated with these variables. Based on this model, pathologic stage pt3 was associated with a risk of PSM 3.8 times higher compared to pt2 (95% confidence interval [CI]: ), and for each additional percentage point increase in tumor volume, the risk of PSM increased 8.7 times (95% CI: ) Model 3: Positive surgical margin location Table 4 summarizes clinical and pathologic features according to location of PSMs. BMI, analyzed as a continuous variable, was shown to be an independent predictive factor for apical PSMs, with increasing BMI predicting higher incidence of apex location (OR: 1.1; p =0.0119).Regarding PL PSM location, BMI ( p =0.0321)wasdeterminedtobethe only predictive factor, where increasing BMI predicts lower incidence of PL location (OR: 0.89; p =0.0321). 4. Discussion PSM after RP has been uniformly associated with an increased hazard of biochemical and local disease recurrence as well as the need for secondary treatment [5]. Refinements in surgical technique and a downward stage migration during the PSA era have collectively contributed to declining rates of PSMs in contemporary ORP series [3]. Despite this trend, the incidence of PSMs has remained relatively stable in the last few years, particularly in high-risk tumors [17]. Recent studies have suggested that RARP, with the advantages of enhanced vision and reduced blood loss, could be associated with reduction in PSM rates [18,19]. Nevertheless, data from recent comparative studies are controversial [4].In the current series, the overall PSM rate was 11.5%, a result at least comparable to those reported by high-volume ORP and laparoscopic RP (LRP) series [4,5]. Prostate cancer nomograms have been previously developed to predict risk of disease recurrence and adverse pathologic outcomes after RP based on preoperative variables [20,21]. These nomograms are generally based on preoperative Gleason grade, serum PSA, and clinical stage, which are the most commonly used factors to assess an individual patient s risk of having extraprostatic disease. In the present study, we initially built a regression model based only on preoperative variables, aiming to identify factors that can be helpful in the planning a nerve-sparing procedure. The only preoperative factor predictive of PSM in this model was clinical stage, with a higher PSM rate for T3 versus T1c and for T2 versus T1c. The other preoperative variables evaluated were not correlated with PSM in our Table 3 Prognostic role of positive surgical margins: multivariable analysis Predictive factor Comparison p value OR (95% CI) Preoperative variables Clinical stage T2 vs T1 < ( ) T3 vs T1 < ( ) Pre-, intra-, and postoperative variables combined Percentage of tumor Continuous ( ) Pathologic stage pt3 vs pt2 < ( ) pt4 vs pt ( ) CI = confidence interval; OR = odds ratio.

5 1026 EUROPEAN UROLOGY 57 (2010) Table 4 Clinical and pathologic features according to the positive surgical margin location Apex BN MF PL Preoperative variables Age Median (IQR) 62 ( ) 58 ( ) 63 ( ) 62 ( ) BMI Median (IQR) 29 (26 34) 28 ( ) 28 ( ) 27 ( ) PSA Median (IQR) 4.9 ( ) 6 ( ) 5.67 ( ) 5 ( ) Clinical stage, n (%) T1c 27 (45.76) 7 (11.86) 4 (6.78) 21 (35.59) T2 12 (31.58) 3 (7.89) 10 (26.32) 13 (34.21) T3 1 (25) 2 (50) 1 (25) AUA-SS, n (%) Asymptomatic 3 (50) 1 (16.67) 2 (33.33) (44.23) 4 (7.69) 6 (11.54) 19 (36.54) (41.18) 1 (2.94) 7 (20.59) 12 (35.29) (22.22) 3 (33.33) 2 (22.22) 2 (22.22) Biopsy Gleason grade, n (%) 6 24 (42.86) 7 (12.5) 4 (7.14) 21 (37.5) 7 11 (30.56) 4 (11.11) 9 (25) 12 (33.33) 8 4 (44.44) 3 (33.33) 2 (22.22) No. of positive cores Median (IQR) 3 (1.25 4) 4 (2 7) 5 (2 6) 3 (2 4.75) Percentage of positive cores, % Median (IQR) 25 ( ) 33.3 ( ) 43.3 ( ) 33.3 ( ) Intraoperative variables Type of nerve sparing, n (%) NNS 6 (30) 4 (20) 4 (20) 6 (30) UNS 1 (14.29) 1 (14.29) 3 (42.86) 2 (28.57) BNS 32 (43.24) 6 (8.11) 9 (12.16) 27 (36.49) Median lobe, n (%) Absent 33 (37.2) 11 (12.3) 14 (15.7) 31 (34.8) Present 6 (50) 2 (16.6) 4 (33.33) Postoperative variables PNI, n (%) Absent 13 (48.15) 5 (18.52) 1 (3.7) 8 (29.63) Present 26 (35.14) 6 (8.11) 15 (20.27) 27 (36.49) HGPIN, n (%) Absent 15 (42.86) 5 (14.29) 4 (11.43) 11 (31.43) Present 24 (36.36) 6 (9.09) 12 (18.18) 24 (36.36) Prostatitis, n (%) Absent 28 (40.58) 7 (10.14) 10 (14.49) 24 (34.78) Present 11 (34.38) 4 (12.5) 6 (18.75) 11 (34.38) Atrophy, n (%) Absent 38 (39.58) 10 (10.42) 15 (15.63) 33 (34.38) Present 1 (20) 1 (20) 1 (20) 2 (40) Pathologic stage, n (%) pt2 28 (49.12) 7 (12.28) 4 (7.02) 18 (31.58) pt3 10 (24.39) 4 (9.76) 10 (24.39) 17 (41.46) pt4 1 (33.33) 2 (66.67) Pathologic Gleason grade, n (%) 6 11 (52.38) 3 (14.29) 2 (9.52) 5 (23.81) 7 (37.88) 8 (12.12) 7 (10.61) 26 (39.39) 8 3 (21.43) 7 (50) 4 (28.57) Percentage of tumor, % Median (IQR) 15 (10 30) 30 ( ) 30 (20 40) 20 (15 27) Prostate weight Median (IQR) 50 ( ) 50 ( ) 46 ( ) 45 ( ) AUA-SS = American Urological Association symptom score; BMI = body mass index; BN = bladder neck; BNS = bilateral nerve sparing; HGPIN = high-grade prostatic intraepithelial neoplasia; IQR = interquartile range; MF = multifocal; NNS = non nerve sparing; PL = posteriolateral; PNI = perineural invasion; PSA = prostate-specific antigen; UNS = unilateral nerve sparing.

6 EUROPEAN UROLOGY 57 (2010) cohort of patients. Previous ORP, LRP, and RARP series have also correlated PSM with clinical stage [3,8,22]. Ramos et al. [23], for example, showed lower PSM rates in T1c tumors compared to T2b (20% vs 29%, p < 0.05) in 1620 ORPs performed by a single surgeon. Multivariable analysis also indicated a decreased risk of recurrence for the T1c group compared to the T2a and T2b groups. Likewise, Ficarra et al. [14] recently evaluated 322 RARPs and identified clinical stage as an independent predictive factor for PSM (hazard ratio [HR]: 2.217; p < 0.008). However, by contrast with these results, Liss et al. [13], evaluating 216 RARPs, showed that serum PSA level ( p = 0.012) and PSA density ( p = 0.005) were the only preoperative factors predictive of PSM. Clinical stage did not predict PSM in the univariable or multivariable analysis in their cohort of patients. Biopsy Gleason grade and number and percentage of positive cores have also been correlated with higher PSM rates in some previous studies [8 10,24,25]. Secin et al. [24] recently evaluated preoperative and intraoperative risk factors for side-specific PSMs in 407 consecutive LRPs. On multivariable analysis, higher preoperative PSA ( p = 0.02) and Gleason grade 7 compared with 6 (p = 0.001) were associated with a higher probability of PSM. Likewise, Freedland et al. [25], evaluating 1094 RPs (Shared Equal Access Regional Cancer Hospital [SEARCH] database), showed that serum PSA and percent of positive cores were significant predictors of PSMs. Percent of positive cores ( p < 0.001), serum PSA ( p = 0.008), and biopsy Gleason score ( p = 0.014) were also significant independent predictors of time to biochemical recurrence. In our study, however, biopsy Gleason grade, number, and percentage of positive cores were not correlated with PSMs in the multivariable analysis. Because most of our patients were referred from other centers, biopsy technique and the number of cores were not standardized, which could help explain this disparity in the results. Surgical technique and individual surgeon s expertise also play an important role in the incidence of PSM following RP [6,7,9,12,13]. Atug et al. [26] evaluated 140 consecutive patients who underwent RARP by the same surgical team. The patients were divided into three groups based on the time of surgery: Group 1 included the first 33 cases, group 2 included the second 33 cases, and group 3 comprised the last 34 cases. The PSM rates were 45.4%, 21.2%, and 11.7% for groups 1, 2, and 3, respectively. The difference among the groups was statistically significant ( p = ), showing lower PSM rates with increasing surgeon experience. Our current series included RARPs performed by a single surgeon after his first 1500 surgeries, minimizing the effect of the learning curve in the PSM rates. Lower PSM rates for high-volume surgeons suggest that experience and careful attention to surgical details, adjusting the nerve-sparing procedure for each patient, can decrease PSM rates and improve cancer control with RARP. Regarding the nerve-sparing technique adopted during RP, it is rational to believe that the wider the dissection around the prostate, the lower the risk of PSM [24]. However, the results reported in the literature conflict. In an ORP series, Villers and colleagues [27] decreased their PSM rates by performing extrafascial dissection of the NVB among patients with pt2 stage prostate cancer. The overall PSM rates decreased from 32% to 25%, and for patients with tumor volume <2cm 3, the PSM rates decreased from 21% to 5%. Similarly, Liss et al. [13] showed higher PSM rates in a RARP series of patients who underwent a nerve-sparing procedure compared with NNS. The PSM rate was 5.9% with pt2 tumors versus 3.3% for NNS for the nerve-sparing procedure and 39.5% for pt3 tumors versus 21.7% for NNS. In the multivariable analysis, nerve-sparing surgery was associated with a significant increase in PSMs ( p = 0.030; OR: 5.5; 95% CI, ) after adjusting for stage, age, and pathologic Gleason score. Likewise, Potdevin et al. [28], evaluating intrafascial versus interfascial nerve-sparing procedures during RARP, showed higher PSM rates in the intrafascial group for pt3 disease (41.8% vs 22.2%, p < 0.05). In contrast, Ward and colleagues [9] reported higher PSM rates for wide excision of the NVB compared to a nervesparing procedure in 7268 ORPs (42% vs 34%; p < 0.001). The authors concluded that a nerve-sparing procedure is not an independent adverse risk factor for PSMs or progression-free survival; tumor biology seems to predict the PSM rates independently of wide excision of the NVB or nerve-sparing techniques. Likewise, in our study, the PSM rates adopting BNS, UNS, or NNS procedures were similar. We believe that the surgeon s experience and adequate planning of the nerve-sparing procedure in each patient can explain the lack of difference in the PSM rates comparing nerve-sparing procedures in our series. We plan our nervesparing procedures individually for each case based on preoperative clinical and biopsy parameters. As expected, the most important predictive factors for PSMs in our study were related to the aggressiveness of the tumor. Pathologic stage and percentage of tumor in the surgical specimen were identified as the only independent predictive factors for PSM. These results also suggest that cancer biology and aggressiveness of the tumor appear to be more important in determining PSM rates after RARP thanthetypeofnerve-sparingprocedureperformedbyan experienced surgeon adopting a planned nerve-sparing protocol. Similar results were recently published by Ficarra et al. [14]. These authors evaluated 322 patients who underwent RALP performed by two surgeons. Among the preoperative variables analyzed, prostate volume on transrectal ultrasound (HR: 0.420; p < 0.002) and clinical stage (HR: 2.217; p < 0.008) were independent predictors of the presence of any PSM. Considering pathologic variables, only extraprostatic extension of the primary tumor was an independent predictor of any PSM (HR: ; p < 0.001). The authors concluded that factors related to biology of the tumor, such as pathologic extension of the primary tumor, represent the most relevant predictors of PSMs. Regarding PSM locations, BMI was identified as an independent predictive factor for PSM at the apex. Higher BMIs were associated with higher PSM rates at the apex and lower PL PSM rates. Similarly, results from the SEARCH database [29], including 1434 patients who underwent retropubic RPs, showed high risk of apical margins among

7 1028 EUROPEAN UROLOGY 57 (2010) the most obese men (BMI 35 kg/m 2 ) compared to healthyweight patients (35% vs 16%; p < 0.001). After adjusting for several preoperative clinical and pathologic characteristics, higher BMI was associated with an increased risk of PSMs overall and at all specific anatomic locations (all p 0.007). The authors suggest that additional care is needed in apical dissection to prevent iatrogenic PSMs when operating on obese men. Likewise, Castle et al. [30], comparing 91 nonobese and 49 obese patients who underwent RALP, showed higher overall PSM rates in the obese group (26.5% vs. 13.1%; p = ); however, PSM rates in specific locations were not analyzed. Our study has some limitations. Even thought the total number of patients analyzed was satisfactory, the overall incidence of events (PSMs) in our series was proportionally low; consequently, modeling was not feasible for BN and MF PSMs and some of our statistical analysis might be underpowered. In addition, most of our patients were referred from other urologic centers; therefore, the biopsy technique was not standardized, which could help to explain the lack of correlation of some of the preoperative factors with PSM. Furthermore, stepwise logistic regression, which was used to identify predictive factors for PSM, has some inherent limitations. The main weakness of this statistical approach is the bias in parameter estimation; this technique tends to overestimate the prognostic role of the covariates left in the model. Finally, we analyzed predictors for PSM after RARP performed by a very experienced surgeon; hence, our results might not be applicable to lowvolume RARP series. 5. Conclusions Factors correlated with aggressiveness of cancer, such as clinical stage, pathologic stage, and tumor volume, were the most important predictive factors for PSMs, whereas the nerve-sparing procedure was not associated with higher PSM rates after RARP. Clinical stage was the only preoperative variable independently associated with PSMs. Regarding PSM location, BMI was shown to be an independent predictive factor for apical PSMs. Author contributions: Vipul R. Patel 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: Coelho, Orvieto, Chauhan, Patel. Acquisition of data: Coelho, Chauhan. Analysis and interpretation of data: Coelho. Drafting of the manuscript: Coelho. Critical revision of the manuscript for important intellectual content: Patel, Palmer, Orvieto. Statistical analysis: Coelho. Obtaining funding: None. Administrative, technical, or material support: None. Supervision: Patel. Other (specify): None. 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] Pfitzenmaier J, Pahernik S, Tremmel T, Haferkamp A, Buse S, Hohenfellner M. Positive surgical margins after radical prostatectomy: do they have an impact on biochemical or clinical progression? BJU Int 2008;102: [2] Hong YM, Hu JC, Paciorek AT, Knight SJ, Carroll PR. Impact of radical prostatectomy positive surgical margins on fear of cancer recurrence: results from CaPSURE TM. Urol Oncol. In press. doi: /j.urolonc [3] Wieder JA, Soloway MS, Incidence. etiology, location, prevention and treatment of positive surgical margins after radical prostatectomy for prostate cancer. J Urol 1998;160: [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: [5] Yossepowitch O, Bjartell A, Eastham JA. Positive surgical margins in radical prostatectomy: outlining the problem and its long-term consequences. Eur Urol 2009;55: [6] Stephenson RA, Middleton RG, Abbott TM. Wide excision (nonnerve sparing) radical retropubic prostatectomy using an initial perirectal dissection. J Urol 1997;157: [7] Ahlering TE, Eichel L, Edwards RA, Lee DI, Skarecky DW. Robotic radical prostatectomy: a technique to reduce pt2 positive margins. Urology 2004;64: [8] Cheng L, Slezak J, Bergstralh EJ, Myers RP, Zincke H, Bostwick DG. Preoperative prediction of surgical margin status in patients with prostate cancer treated by radical prostatectomy. J Clin Oncol 2000; 18: [9] Ward JF, Zincke H, Bergstralh EJ, Slezak JM, Myers RP, Blute ML. The impact of surgical approach (nerve bundle preservation versus wide local excision) on surgical margins and biochemical recurrence following radical prostatectomy. J Urol 2004;172: [10] Sofer M, Hamilton-Nelson KL, Schlesselman JJ, Soloway MS. Risk of positive margins and biochemical recurrence in relation to nerve-sparing radical prostatectomy. J Clin Oncol 2002;20: [11] Ackerman DA, Barry JM, Wicklund RA, et al. Analysis of risk factors associated with prostate cancer extension to the surgical margin and pelvic node metastasis at radical prostatectomy. J Urol 1993; 150: [12] Vickers AJ, Bianco FJ, Serio AM, et al. The surgical learning curve for prostate cancer control after radical prostatectomy. Natl Cancer Inst 2007;99: [13] Liss M, Osann K, Ornstein D. Positive surgical margins during robotic radical prostatectomy: a contemporary analysis of risk factors. BJU Int 2008;102: [14] Ficarra V, Novara G, Secco S, et al. Predictors of positive surgical margins after laparoscopic robot assisted radical prostatectomy. J Urol 2009;182: [15] Patel VR, Tully AS, Holmes R, Lindsay J. Robotic radical prostatectomy in the community setting the learning curve and beyond: initial 200 cases. J Urol 2005;174: [16] Coughlin G, Dangle PP, Palmer KJ, Samavedi S, Patel VR. Athermal early retrograde release of the neurovascular bundle during nerve-sparing robotic-assisted laparoscopic radical prostatectomy. J Robotic Surg 2009;3:13 7.

8 EUROPEAN UROLOGY 57 (2010) [17] Han M, Partin AW, Chan DY, Walsh PC. An evaluation of the decreasing incidence of positive surgical margins in a large retropubic prostatectomy series. J Urol 2004;171:23 6. [18] Smith Jr JA, Chan RC, Chang SS, et al. A comparison of the incidence and location of positive surgical margins in robotic assisted laparoscopic radical prostatectomy and open retropubic radical prostatectomy. J Urol 2007;178: [19] White MA, De Haan AP, Stephens DD, Maatman TK, Maatman TJ. Comparative analysis of surgical margins between radical retropubic prostatectomy and RALP: are patients sacrificed during initiation of robotics program? Urology 2009;73: [20] Partin A, Yoo J, Carter H, et al. The use of prostate specific antigen, clinical stage and Gleason score to predict pathological stage in men with localized prostate cancer. J Urol 1993;150: [21] D Amico AV, Whittington R, Malkowicz SB, et al. Pretreatment nomogram for prostate-specific antigen recurrence after radical prostatectomy or external-beam radiation therapy for clinically localized prostate cancer. J Clin Oncol 1999;17: [22] Guillonneau B, el-fettouh H, Baumert H. Laparoscopic radical prostatectomy: oncological evaluation after 1,000 casesa Montsouris Institute. J Urol 2003;169: [23] Ramos CG, Carvalhal GF, Smith DS, Mager DE, Catalona WJ. Clinical and pathological characteristics, and recurrence rates of stage T1c versus T2a or T2b prostate cancer. J Urol 1999;161: [24] Secin FP, Serio A, Bianco Jr FJ, et al. Preoperative and intraoperative risk factors for side-specific positive surgical margins in laparoscopic radical prostatectomy for prostate cancer. Eur Urol 2007;51: [25] Freedland SJ, Aronson WJ, Terris MK, et al. SEARCH Database Study Group. Percent of prostate needle biopsy cores with cancer is significant independent predictor of prostate specific antigen recurrence following radical prostatectomy: results from SEARCH database. J Urol 2003;169: [26] Atug F, Castle EP, Srivastav SK, Burgess SV, Thomas R, Davis R. Positive surgical margins in robotic-assisted radical prostatectomy: impact of learning curve on oncologic outcomes. Eur Urol 2006; 49: [27] Villers A, Stamey TA, Yemoto C, Rischmann P, McNeal JE. Modified extrafascial radical retropubic prostatectomy technique decreases frequency of positive surgical margins in T2 cancers <2cm 3. Eur Urol 2000;38: [28] Potdevin L, Ercolani M, Jeong J, Kim IY. Functional and oncologic outcomes comparing interfascial and intrafascial nerve sparing in robot-assisted laparoscopic radical prostatectomies. J Endourol 2009; 23: [29] SEARCH Database Study Group, Jayachandran J, Aronson WJ, et al. Obesity and positive surgical margins by anatomic location after radical prostatectomy: results from the Shared Equal Access Regional Cancer Hospital database. BJU Int 2008; 102: [30] 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;2:91 5.