JOURNAL OF ENDOUROLOGY Volume 28, Number 7, July 2014 ª Mary Ann Liebert, Inc. Pp. 784–791 DOI: 10.1089/end.2013.0774
Is Robot-Assisted Radical Prostatectomy Safe in Men with High-Risk Prostate Cancer? Assessment of Perioperative Outcomes, Positive Surgical Margins, and Use of Additional Cancer Treatments Giorgio Gandaglia, MD,1,2 Firas Abdollah, MD,2 Jim Hu, MD,3 Simon Kim, MD,4 Alberto Briganti, MD,2 Jesse D. Sammon, MD,5 Andreas Becker, MD,1 Florian Roghmann, MD,1 Markus Graefen, MD,6 Francesco Montorsi, MD,2 Paul Perrotte, MD,7 Pierre I. Karakiewicz, MD,1,7 Quoc-Dien Trinh, MD,1,8 and Maxine Sun, BSc1
Abstract
Introduction: Despite a rapid dissemination of robot-assisted radical prostatectomy (RARP) over open radical prostatectomy (ORP), to date no study has compared perioperative outcomes between the two approaches in patients with high-risk prostate cancer (PCa). The aim of our study was to evaluate the safety and feasibility of RARP in this setting. Patients and Methods: Overall, 1,512 patients with high-risk PCa within the Surveillance, Epidemiology, and End Results (SEER) Medicare-linked database diagnosed between 2008 and 2009 were abstracted. Patients were treated with RARP or ORP. Postoperative complications, blood transfusions, prolonged length of stay (pLOS), positive surgical margins, and additional cancer therapy rates were compared. Propensity-score matched analyses and logistic regression models fitted with generalized estimating equations for clustering among hospitals were performed. Results: Overall, 706 (46.7%) and 806 (53.3%) patients underwent ORP and RARP, respectively. Following propensity-matched analyses, 706 patients remained. No differences were observed in complications (P = 0.6), positive surgical margins (P = 0.4), or additional therapy after surgery (P = 0.2) between patients treated with RARP and ORP; however, RARP was associated with lower rates of transfusions and shorter hospitalization (all P < 0.001). In multivariable analyses, patients undergoing RARP were less likely to receive a blood transfusion (P = 0.002) or to experience pLOS (P < 0.001) compared with men treated with ORP. Conclusions: RARP and ORP have comparable complications, positive surgical margins, and additional cancer therapy rates in high-risk PCa. RARP is associated with lower rates of blood transfusions and shorter hospital stays. These findings suggest that RARP is safe and feasible even in this clinical scenario. Introduction
D
espite the advent of prostate-specific antigen (PSA), between 20% and 30% of patients diagnosed with prostate cancer (PCa) present with high-risk characteristics, defined as clinical stage ‡ T2c, biopsy Gleason score 8
to 10, and/or serum PSA levels > 20 ng/mL.1,2 Historically, due to increased risk of biochemical recurrence (BCR), systemic progression, and worse oncologic outcomes,3,4 the consideration of surgical management was often discarded in such individuals.5 Since then, several studies have shown comparable oncologic outcomes between radical prostatectomy relative to
1
Cancer Prognostics and Health Outcomes Unit, University of Montreal Health Centre, Montreal, Canada. Department of Urology, Urological Research Institute, Vita Salute San Raffaele University, San Raffaele Scientific Institute, Milan, Italy. 3 Department of Urology, David Geffen School of Medicine at UCLA, Los Angeles, California. 4 Department of Urology, Yale University, New Haven, Connecticut. 5 Vattikuti Urology Institute, Henry Ford Health System, Detroit, Michigan. 6 Martiniclinic, University Medical Center Hamburg-Eppendorf, Hamburg, Germany. 7 Department of Urology, University of Montreal Health Centre, Montreal, Canada. 8 Department of Surgery, Division of Urology, Brigham and Women’s Hospital and Dana-Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts. 2
784
ROBOT-ASSISTED RADICAL PROSTATECTOMY IN HIGH-RISK PROSTATE CANCER
radiotherapy and/or androgen deprivation therapy (ADT) in the context of high-risk disease.6–8 As a result, the use of open radical prostatectomy (ORP) in patients with aggressive disease characteristics has significantly increased during the last decade,5,9 thereby confirming its role as a reasonable treatment option in selected patients with high-risk features.3,5 In a parallel setting, robot-assisted radical prostatectomy (RARP) has gained increasing popularity and, in the United States alone, use of RARP has overtaken ORP in recent years.10 That said, its use remains predominantly reserved for men with low- and intermediate-risk PCa,11,12 as many physicians remain apprehensive of potential adverse outcomes that could occur in the context of high-risk disease. Bulky disease and the risk of seminal vesicle invasion or bladder neck involvement in particular can increase the technical difficulties related to the adoption of RARP. This could possibly translate in higher risk of detrimental perioperative outcomes in these patients. Although previous studies reported superior short-term outcomes for patients undergoing minimally invasive surgery,13 no study has compared perioperative outcomes between ORP and RARP in patients with high-risk PCa. To address the lack of such data, we sought to evaluate and compare short-term outcomes between the two techniques among Medicare beneficiaries diagnosed with high-risk PCa. Our hypothesis states that even in the high-risk scenario, RARP represents a safe and feasible procedure. Patients and Methods Population source
The current study relied on the Surveillance, Epidemiology, and End Results (SEER) Medicare insurance program– linked database. This database is 98% complete for case ascertainment. The SEER registries covered approximately 28% of the U.S. population with Medicare administrative data. Medicare insurance includes approximately 97% of Americans aged 65 years or older. Study population
Overall 6,275 patients with nonmetastatic PCa (ICD-O site code 61.9, histologic code 8140) aged 65 years or older treated for radical prostatectomy from October 2008 to December 2009 were identified. Additional exclusions included individuals older than 80 years old (n = 88), those with unknown tumor stage (n = 130), and those with an unknown tumor grade (n = 111). For the purpose of our study, only patients with high-risk PCa (clinical stage ‡ T2c, biopsy Gleason score 8 to 10, or PSA levels > 20 ng/mL) were included.1 This resulted in 1,512 assessable patients. Subsequently, we distinguished RARP from ORP patients via a specific robot-assisted modifier code introduced as of October 1, 2008 (ICD-9 clinical modification procedure code: 17.42). Covariates
For each patient, age at diagnosis, race, population density, marital status, 2000 census tract percent with 4-year college education, 2000 census tract annual median income, region, biopsy Gleason score, preoperative PSA, clinical stage, pathologic stage, and cumulative number of risk factors were assigned.14 The Charlson comorbidity index (CCI) was de-
785
rived from the Medicare claims 1 year before PCa diagnosis using a commonly used and validated algorithm.15 We determined hospital volume by aggregating the number of procedures for all men included in our analyses, as previously described.13 Outcomes
Postoperative complications and blood transfusions that occur within 30 days after surgery were recorded. Similar to previous methodologies,13 seven schemes of postoperative complications were assessed using the ICD-9 and current procedural terminology codes: cardiac, respiratory, genitourinary, vascular, wound, and miscellaneous (medical and surgical). Prolonged length of stay (pLOS) was defined as a hospitalization beyond the median (‡ 3 days) after surgery. An additional end point consisted of the identification of men undergoing other postoperative cancer therapy (i.e., radiotherapy and ADT) during follow-up. Finally, we identified patients with positive surgical margins. This variable was available exclusively for patients with pT2 and pT3a disease (n = 1221). Statistical analyses
Means, medians, and interquartile ranges were reported for continuous variables. Frequencies and proportions were reported for categorical variables. The Mann-Whitney test and w2-test were used to compare the statistical significance of differences in medians and proportions, respectively. Due to inherent differences between patients undergoing ORP and RARP, adjustment was performed using 0.5 to 0.5 propensity score matching ratio. Matching is commonly used in observational studies to select control subjects who are matched with treated subjects on the controlled background covariates, which, if uncontrolled for, might lead to biased estimates of treatment effects.16 Propensity scores were computed by modeling a logistic regression with the dependent variable as the odds of undergoing RARP, and the independent variable as age, race, CCI, marital status, population density, income, education, clinical stage, biopsy Gleason score, and preoperative PSA levels. Subsequently, covariate balance between the matched groups were examined.17 Separate multivariable logistic regression models tested the effect of surgical approach on postoperative complications, blood transfusions, pLOS, positive surgical margins, and additional therapies. Covariates included age, race, marital status, population density, CCI, income, education, biopsy Gleason score, clinical stage, and preoperative PSA levels. Generalized estimating equations were fitted to account for clustering within hospitals.18 All statistical tests were performed using SPSS version 20 (SPSS, Chicago, IL). All tests were two-sided with a significance level set at P < 0.05. Results Baseline characteristics
Overall, 1,512 patients were identified (Table 1). Average age at diagnosis was 69.5 years. Respectively, 706 (46.7%) and 806 (53.3%) patients were treated with ORP and RARP. The rate of patients receiving pelvic lymph node dissection (PLND) was significantly higher among individuals treated
786
GANDAGLIA ET AL.
Table 1. Descriptive Statistics of 1,512 Patients Treated with Radical Prostatectomy for High-Risk PCa Within the SEER Medicare Database Between 2008 and 2009 Before and After Propensity Score Matching Before propensity score matching Total (n = 1512) Age at diagnosis Mean (Median) 69.5 (69) IQR 66–71 Race (%) White 1223 (80.9) African American 136 (9.0) Other 153 (10.1) Marital status (%) Married 1195 (79.0) Unmarried 317 (21.0) Population density (%) Metropolitan 1335 (88.3) Non-metropolitan 177 (11.7) Annual median income, USD (%) £$38,332 414 (27.4) $38,333–$51,348 380 (25.1) $51,349–$70,007 365 (24.1) ‡$70,008 353 (23.3) College education (% of persons) £14.7 409 (27.1) 14.8–26.1 385 (25.5) 26.2–42.6 377 (24.9) ‡42.7 341 (22.6) CCI (%) 0 1010 (66.8) 1 158 (10.4) 2 166 (11.0) ‡3 178 (11.8) Clinical stage (%) £T1c 542 (35.8) 365 (24.1) T2a/b T2c 447 (29.3) 158 (10.4) T3/4 Biopsy Gleason score (%) £6 150 (9.9) 7 454 (30.0) 8–10 908 (60.1) Preoperative PSA (%) £10 ng/mL 1014 (67,1) 10–20 ng/mL 212 (14.0) >20 ng/mL 286 (18.9) No. of risk factors (%) 1 1254 (82.9) 2 229 (15.1) 3 29 (1.9) Pelvic lymph node dissection (%) No 384 (25.4) Yes 1128 (74.6) Pathologic Gleason score (%) £6 180 (11.9) 7 387 (25.6) 8–10 769 (50.9) Unknown 176 (11.6)
ORP (n = 706)
RARP (n = 806)
69.6 (69) 67–71
69.3 (69) 67–71
574 (81.3) 64 (9.1) 68 (9.6)
After propensity score matching Total (n = 706)
ORP (n = 353)
RARP (n = 353)
0.001
69.7 (69) 67–71
69.8 (69) 67–71
69.6 (69) 67–71
0.5
649 (80.5) 72 (8.9) 85 (10.5)
0.8
571 (80.9) 69 (9.8) 66 (9.3)
283 (80.2) 39 (11.0) 31 (8.8)
288 (81.6) 30 (8.5) 35 (9.9)
0.5
573 (76.1) 169 (23.9)
658 (81.6) 148 (18.4)
0.01
531 (75.2) 175 (24.8)
264 (74.8) 89 (25.2)
267 (75.6) 86 (24.4)
0.8
601 (85.1) 105 (14.9)
734 (91.1) 72 (8.9)
Is Robot-Assisted Radical Prostatectomy Safe in Men with High-Risk Prostate Cancer? Assessment of Perioperative Outcomes, Positive Surgical Margins, and Use of Additional Cancer Treatments Giorgio Gandaglia, MD,1,2 Firas Abdollah, MD,2 Jim Hu, MD,3 Simon Kim, MD,4 Alberto Briganti, MD,2 Jesse D. Sammon, MD,5 Andreas Becker, MD,1 Florian Roghmann, MD,1 Markus Graefen, MD,6 Francesco Montorsi, MD,2 Paul Perrotte, MD,7 Pierre I. Karakiewicz, MD,1,7 Quoc-Dien Trinh, MD,1,8 and Maxine Sun, BSc1
Abstract
Introduction: Despite a rapid dissemination of robot-assisted radical prostatectomy (RARP) over open radical prostatectomy (ORP), to date no study has compared perioperative outcomes between the two approaches in patients with high-risk prostate cancer (PCa). The aim of our study was to evaluate the safety and feasibility of RARP in this setting. Patients and Methods: Overall, 1,512 patients with high-risk PCa within the Surveillance, Epidemiology, and End Results (SEER) Medicare-linked database diagnosed between 2008 and 2009 were abstracted. Patients were treated with RARP or ORP. Postoperative complications, blood transfusions, prolonged length of stay (pLOS), positive surgical margins, and additional cancer therapy rates were compared. Propensity-score matched analyses and logistic regression models fitted with generalized estimating equations for clustering among hospitals were performed. Results: Overall, 706 (46.7%) and 806 (53.3%) patients underwent ORP and RARP, respectively. Following propensity-matched analyses, 706 patients remained. No differences were observed in complications (P = 0.6), positive surgical margins (P = 0.4), or additional therapy after surgery (P = 0.2) between patients treated with RARP and ORP; however, RARP was associated with lower rates of transfusions and shorter hospitalization (all P < 0.001). In multivariable analyses, patients undergoing RARP were less likely to receive a blood transfusion (P = 0.002) or to experience pLOS (P < 0.001) compared with men treated with ORP. Conclusions: RARP and ORP have comparable complications, positive surgical margins, and additional cancer therapy rates in high-risk PCa. RARP is associated with lower rates of blood transfusions and shorter hospital stays. These findings suggest that RARP is safe and feasible even in this clinical scenario. Introduction
D
espite the advent of prostate-specific antigen (PSA), between 20% and 30% of patients diagnosed with prostate cancer (PCa) present with high-risk characteristics, defined as clinical stage ‡ T2c, biopsy Gleason score 8
to 10, and/or serum PSA levels > 20 ng/mL.1,2 Historically, due to increased risk of biochemical recurrence (BCR), systemic progression, and worse oncologic outcomes,3,4 the consideration of surgical management was often discarded in such individuals.5 Since then, several studies have shown comparable oncologic outcomes between radical prostatectomy relative to
1
Cancer Prognostics and Health Outcomes Unit, University of Montreal Health Centre, Montreal, Canada. Department of Urology, Urological Research Institute, Vita Salute San Raffaele University, San Raffaele Scientific Institute, Milan, Italy. 3 Department of Urology, David Geffen School of Medicine at UCLA, Los Angeles, California. 4 Department of Urology, Yale University, New Haven, Connecticut. 5 Vattikuti Urology Institute, Henry Ford Health System, Detroit, Michigan. 6 Martiniclinic, University Medical Center Hamburg-Eppendorf, Hamburg, Germany. 7 Department of Urology, University of Montreal Health Centre, Montreal, Canada. 8 Department of Surgery, Division of Urology, Brigham and Women’s Hospital and Dana-Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts. 2
784
ROBOT-ASSISTED RADICAL PROSTATECTOMY IN HIGH-RISK PROSTATE CANCER
radiotherapy and/or androgen deprivation therapy (ADT) in the context of high-risk disease.6–8 As a result, the use of open radical prostatectomy (ORP) in patients with aggressive disease characteristics has significantly increased during the last decade,5,9 thereby confirming its role as a reasonable treatment option in selected patients with high-risk features.3,5 In a parallel setting, robot-assisted radical prostatectomy (RARP) has gained increasing popularity and, in the United States alone, use of RARP has overtaken ORP in recent years.10 That said, its use remains predominantly reserved for men with low- and intermediate-risk PCa,11,12 as many physicians remain apprehensive of potential adverse outcomes that could occur in the context of high-risk disease. Bulky disease and the risk of seminal vesicle invasion or bladder neck involvement in particular can increase the technical difficulties related to the adoption of RARP. This could possibly translate in higher risk of detrimental perioperative outcomes in these patients. Although previous studies reported superior short-term outcomes for patients undergoing minimally invasive surgery,13 no study has compared perioperative outcomes between ORP and RARP in patients with high-risk PCa. To address the lack of such data, we sought to evaluate and compare short-term outcomes between the two techniques among Medicare beneficiaries diagnosed with high-risk PCa. Our hypothesis states that even in the high-risk scenario, RARP represents a safe and feasible procedure. Patients and Methods Population source
The current study relied on the Surveillance, Epidemiology, and End Results (SEER) Medicare insurance program– linked database. This database is 98% complete for case ascertainment. The SEER registries covered approximately 28% of the U.S. population with Medicare administrative data. Medicare insurance includes approximately 97% of Americans aged 65 years or older. Study population
Overall 6,275 patients with nonmetastatic PCa (ICD-O site code 61.9, histologic code 8140) aged 65 years or older treated for radical prostatectomy from October 2008 to December 2009 were identified. Additional exclusions included individuals older than 80 years old (n = 88), those with unknown tumor stage (n = 130), and those with an unknown tumor grade (n = 111). For the purpose of our study, only patients with high-risk PCa (clinical stage ‡ T2c, biopsy Gleason score 8 to 10, or PSA levels > 20 ng/mL) were included.1 This resulted in 1,512 assessable patients. Subsequently, we distinguished RARP from ORP patients via a specific robot-assisted modifier code introduced as of October 1, 2008 (ICD-9 clinical modification procedure code: 17.42). Covariates
For each patient, age at diagnosis, race, population density, marital status, 2000 census tract percent with 4-year college education, 2000 census tract annual median income, region, biopsy Gleason score, preoperative PSA, clinical stage, pathologic stage, and cumulative number of risk factors were assigned.14 The Charlson comorbidity index (CCI) was de-
785
rived from the Medicare claims 1 year before PCa diagnosis using a commonly used and validated algorithm.15 We determined hospital volume by aggregating the number of procedures for all men included in our analyses, as previously described.13 Outcomes
Postoperative complications and blood transfusions that occur within 30 days after surgery were recorded. Similar to previous methodologies,13 seven schemes of postoperative complications were assessed using the ICD-9 and current procedural terminology codes: cardiac, respiratory, genitourinary, vascular, wound, and miscellaneous (medical and surgical). Prolonged length of stay (pLOS) was defined as a hospitalization beyond the median (‡ 3 days) after surgery. An additional end point consisted of the identification of men undergoing other postoperative cancer therapy (i.e., radiotherapy and ADT) during follow-up. Finally, we identified patients with positive surgical margins. This variable was available exclusively for patients with pT2 and pT3a disease (n = 1221). Statistical analyses
Means, medians, and interquartile ranges were reported for continuous variables. Frequencies and proportions were reported for categorical variables. The Mann-Whitney test and w2-test were used to compare the statistical significance of differences in medians and proportions, respectively. Due to inherent differences between patients undergoing ORP and RARP, adjustment was performed using 0.5 to 0.5 propensity score matching ratio. Matching is commonly used in observational studies to select control subjects who are matched with treated subjects on the controlled background covariates, which, if uncontrolled for, might lead to biased estimates of treatment effects.16 Propensity scores were computed by modeling a logistic regression with the dependent variable as the odds of undergoing RARP, and the independent variable as age, race, CCI, marital status, population density, income, education, clinical stage, biopsy Gleason score, and preoperative PSA levels. Subsequently, covariate balance between the matched groups were examined.17 Separate multivariable logistic regression models tested the effect of surgical approach on postoperative complications, blood transfusions, pLOS, positive surgical margins, and additional therapies. Covariates included age, race, marital status, population density, CCI, income, education, biopsy Gleason score, clinical stage, and preoperative PSA levels. Generalized estimating equations were fitted to account for clustering within hospitals.18 All statistical tests were performed using SPSS version 20 (SPSS, Chicago, IL). All tests were two-sided with a significance level set at P < 0.05. Results Baseline characteristics
Overall, 1,512 patients were identified (Table 1). Average age at diagnosis was 69.5 years. Respectively, 706 (46.7%) and 806 (53.3%) patients were treated with ORP and RARP. The rate of patients receiving pelvic lymph node dissection (PLND) was significantly higher among individuals treated
786
GANDAGLIA ET AL.
Table 1. Descriptive Statistics of 1,512 Patients Treated with Radical Prostatectomy for High-Risk PCa Within the SEER Medicare Database Between 2008 and 2009 Before and After Propensity Score Matching Before propensity score matching Total (n = 1512) Age at diagnosis Mean (Median) 69.5 (69) IQR 66–71 Race (%) White 1223 (80.9) African American 136 (9.0) Other 153 (10.1) Marital status (%) Married 1195 (79.0) Unmarried 317 (21.0) Population density (%) Metropolitan 1335 (88.3) Non-metropolitan 177 (11.7) Annual median income, USD (%) £$38,332 414 (27.4) $38,333–$51,348 380 (25.1) $51,349–$70,007 365 (24.1) ‡$70,008 353 (23.3) College education (% of persons) £14.7 409 (27.1) 14.8–26.1 385 (25.5) 26.2–42.6 377 (24.9) ‡42.7 341 (22.6) CCI (%) 0 1010 (66.8) 1 158 (10.4) 2 166 (11.0) ‡3 178 (11.8) Clinical stage (%) £T1c 542 (35.8) 365 (24.1) T2a/b T2c 447 (29.3) 158 (10.4) T3/4 Biopsy Gleason score (%) £6 150 (9.9) 7 454 (30.0) 8–10 908 (60.1) Preoperative PSA (%) £10 ng/mL 1014 (67,1) 10–20 ng/mL 212 (14.0) >20 ng/mL 286 (18.9) No. of risk factors (%) 1 1254 (82.9) 2 229 (15.1) 3 29 (1.9) Pelvic lymph node dissection (%) No 384 (25.4) Yes 1128 (74.6) Pathologic Gleason score (%) £6 180 (11.9) 7 387 (25.6) 8–10 769 (50.9) Unknown 176 (11.6)
ORP (n = 706)
RARP (n = 806)
69.6 (69) 67–71
69.3 (69) 67–71
574 (81.3) 64 (9.1) 68 (9.6)
After propensity score matching Total (n = 706)
ORP (n = 353)
RARP (n = 353)
0.001
69.7 (69) 67–71
69.8 (69) 67–71
69.6 (69) 67–71
0.5
649 (80.5) 72 (8.9) 85 (10.5)
0.8
571 (80.9) 69 (9.8) 66 (9.3)
283 (80.2) 39 (11.0) 31 (8.8)
288 (81.6) 30 (8.5) 35 (9.9)
0.5
573 (76.1) 169 (23.9)
658 (81.6) 148 (18.4)
0.01
531 (75.2) 175 (24.8)
264 (74.8) 89 (25.2)
267 (75.6) 86 (24.4)
0.8
601 (85.1) 105 (14.9)
734 (91.1) 72 (8.9)
