Laparoscopy and Robotics Comparison of Perioperative Outcomes of Retroperitoneal and Transperitoneal Minimally Invasive Partial Nephrectomy After Adjusting for Tumor Complexity Greg E. Gin, Alexandra C. Maschino, Massimiliano Spaliviero, Emily A. Vertosick, Melanie L. Bernstein, and Jonathan A. Coleman OBJECTIVE MATERIALS AND METHODS
RESULTS
CONCLUSION
To evaluate and compare perioperative outcomes of transperitoneal and retroperitoneal (RP) laparoscopic and robotic partial nephrectomies (LPNs) while adjusting for tumor complexity. Retrospective review was conducted of 191 patients who underwent transperitoneal (n ¼ 116) or RP (n ¼ 75) LPN. To adjust for tumor complexity, individual components of the radius, exophytic or endophytic properties, nearness to the collecting system or sinus, anterior or posterior location, and location in reference to polar lines (R.E.N.A.L.) nephrometry score were used in multivariate linear and logistic regression models to compare perioperative outcomes between the 2 groups. A propensity approach was also used to adjust for multiple covariates. Investigated outcomes included estimated blood loss (EBL), ischemia and operative times, length of hospital stay, margin status, opioid use, postoperative estimated glomerular filtration rate, complications within 30 days, and readmission rates. Tumors resected by RPLPN were more likely to have lower complexity score by nephrometry (P ¼ .04). Four of the 5 components of the R.E.N.A.L. nephrometry score were significantly different between the groups. After adjustment for these factors, a lower EBL was noted in the RP group (b, 97; 95% confidence interval, 156 to 39; P ¼ .001). Risk of readmission for the RP group was significantly lower (odds ratio, 0.15; P ¼ .024) using propensity analysis. Using adjustment for tumor complexity, RPLPN was associated with lower EBL and readmission rates supporting the potential clinical advantage for this approach when feasible. UROLOGY 84: 1355e1360, 2014. 2014 Elsevier Inc.
R
enal cell carcinoma accounts for about 3% of adult cancers and 85%-90% of renal malignancies. Approximately 65,150 new cases of kidney cancers were diagnosed and 13,680 cancer-related deaths were projected in the United States in 2013.1 Detection of incidental renal masses has risen over the past decade with the increased utilization of computerized tomography. Concurrent with these trends has been the development and adaptation of less-invasive nephronsparing management strategies intended to optimally achieve cancer-specific outcomes while offering the putative physiologic advantages associated with renal function preservation. Surgical resection has been the Financial Disclosure: The authors declare that they have no relevant financial interests. From the Department of Urology, Mount Sinai Hospital, New York, NY; the Department of Epidemiology and Biostatistics, Memorial Sloan-Kettering Cancer Center, New York, NY; and the Department of Surgery, Sidney Kimmel Center for Prostate and Urologic Cancers, Memorial Sloan-Kettering Cancer Center, New York, NY Address correspondence to: Jonathan A. Coleman, M.D., Department of Surgery, Sidney Kimmel Center for Prostate and Urologic Cancers, Memorial Sloan-Kettering Cancer Center, 353 East 68th Street, New York, NY 10065. E-mail: [email protected] Submitted: February 11, 2014, accepted (with revisions): July 27, 2014
ª 2014 Elsevier Inc. All Rights Reserved
mainstay of therapeutic options for clinically localized renal tumors. Partial nephrectomy is considered the gold standard for treatment, demonstrating disease-specific outcomes equivalent to radical nephrectomy at experienced centers and confirmed in a prospective, randomized, multicenter trial.2-4 Advancements in technology and increased experience with laparoscopy have allowed urologic surgeons to broaden the indications of minimally invasive nephron-sparing surgery for the treatment of renal masses with results comparable with open procedures. Laparoscopic and robotic partial nephrectomy (LPN) can be accomplished through a transperitoneal (TP) or retroperitoneal (RP) approach, which is chosen based on the tumor location and surgeon experience.5,6 Because of limited experience, RPLPN is less commonly used but felt to offer advantages in regard to accessing vasculature and tumors based on the location, less manipulation of abdominal organs, and convalescence.7 To determine possible advantages related to the RP approach while attempting to control for many of the confounding features associated with retrospective http://dx.doi.org/10.1016/j.urology.2014.07.045 0090-4295/14
1355
multicenter studies, we analyzed outcomes in a series of consecutive TP and RP LPN cases performed with a similar technique managed under a standardized institutional care pathway using a multivariate and propensity statistical model.
MATERIALS AND METHODS Following institutional review board approval, a retrospective review of 252 LPNs by a single surgeon (J.A.C.) from November 2006 through March 2012 was performed. Patients having multiple renal tumors removed during the same procedure, insufficient imaging to assign a nephrometry score, or undergoing a concomitant procedure (eg, radical prostatectomy or laparoscopic hysterectomy) were excluded from the study. We identified 75 patients who underwent RPLPN and 116 patients who underwent TPLPN and met these criteria during this time. The approach was determined preoperatively by the discretion of the surgeon. A robot-assisted LPN was performed in 55 (73%) RP patients and 85 (73%) TP patients. All patients were managed under a standardized perioperative pathway for partial nephrectomy, which has been previously described.8 Briefly, the patients were started on a full liquid diet on postoperative day 0 and advanced as tolerated. Patients were offered patientcontrolled intravenous analgesia upon request and were transitioned to oral narcotics on day 1. Ketorolac is rarely provided. Patients were offered discharge when medically stable, capable of ambulating 1 mile without assistance, adequately pain controlled with oral medications, and tolerating a full liquid diet.
Surgical Technique Technical aspects of LPN performed by either approach have been previously described and underwent little modification over the study period.9 Briefly, a clear liquid diet was started 1 day before the surgery. Positioning was secured in a modified flank at 60 elevation without axillary roll. A Veress needle or balloon expansion was used to gain intraperitoneal or RP access, respectively, as previously described by others. Three to 4 ports were used for laparoscopic procedures and 4-5 ports for robotassisted cases. Tumors were evaluated with intraoperative ultrasound inspection. Vascular control, when used, was provided by bulldog clamps typically achieving complete renal ischemia. Excision was performed sharply through the cortex initially then with intrarenal anatomic dissection technique using diathermy to seal arterial vessels. Sutured closure of large venotomies and collecting system defects were performed separately with 3-0 coated Vicryl on RB needle (Ethicon Surgical, Somersville, NJ). Renorrhaphy was completed using interrupted 0 coated Vicryl sutures with CT-1 needle (Ethicon Surgical) over collagenthrombin matrix coagulant agents with oxidized cellulose material placed over the matrix agent before capsule apposition. After bulldog clamp removal, renal revascularization was confirmed with Doppler sonography. Intraoperative findings and surgical details were collected prospectively using a standardized reporting form maintained in the institutional surgical database system.
Statistical Methods Patient characteristics evaluated included gender, age, race, presentation, American Society of Anesthesiologists score, and preoperative estimated glomerular filtration rate (eGFR). 1356
Tumors were characterized using the radius, exophytic or endophytic properties, nearness to the collecting system or sinus, anterior or posterior location, and location in reference to polar lines (R.E.N.A.L.) nephrometry score.10 The subscale components of the scoring system were compared between groups. Differences were assessed using the Wilcoxon rank sum or the chi-square test. Perioperative outcomes compared between groups included estimated blood loss (EBL), ischemia and operative times, length of hospital stay (LOS), margin status, 24-hour opioid use, postoperative eGFR, Clavien complication grade (within 30 days of surgery), and readmission rates. Multivariate linear and logistic regression models were used. Because we expected that patients with larger and more complex tumors were more likely to have a TP approach, we adjusted for disease characteristics. Only variables significantly associated with surgical approach on univariate analysis were analyzed because of the relatively fewer numbers of RPLPNs. Given the relatively large number of covariates and the small number of occurring events, a propensity approach was performed to confirm the validity of the adjustment. A predicted probability of undergoing RP or TP LPN based on patient and disease characteristics was obtained. Such probability was then entered into a model predicting the outcome of interest to obtain an estimate of association between the surgical approach and each outcome. To ensure that the propensity score adequately controlled for disease characteristics, we reevaluated the association between surgical approach and each characteristic, adjusting for the propensity score. All analyses were conducted using Stata 12.0 (StataCorp, College Station, TX).
RESULTS Demographic and surgical characteristics of the 75 patients who underwent RPLPN and the 116 patients treated with TPLPN were similar (Table 1). Location of tumor was a significant predictor of surgical approach: 65% of the TPLPN group had anterior tumors, whereas 69% of the RPLPN group had posterior tumors (P .9 .2 .009
ASA, American Society of Anesthesiologists; CKD-EPI, Chronic Kidney Disease Epidemiology Collaboration Formula estimating glomerular filtration rate; R.E.N.A.L., radius, exophytic or endophytic properties, nearness to the collecting system or sinus, anterior vs posterior location, and location in reference to polar lines. Data are presented as frequency (percentage) or median (interquartile range). * P values obtained using the Wilcoxon rank sum or the chi-square test.
lower readmission rates (odds ratio [OR], 0.2; 95% CI, 0.04-1.0; P ¼ .05). When reevaluating the association between the surgical approach and each of the included disease characteristics, the propensity score appeared to adequately control for all included disease characteristics (all P >.8) and reaffirmed the findings of the multivariate approach (Table 4). Although the estimates for readmission rates were similar (OR, 0.2 in the multivariate vs 0.15 in the propensity approach), results were only found statistically significant under the propensity approach (P ¼ .024 vs P ¼ .05).
COMMENT The utilization of minimally invasive techniques for the management of small renal masses, including nephronsparing surgery, has increased in the past decade. LPN has been shown to have comparable oncologic outcomes to the open approach and allow for shorter hospitalizations.11,12 Comparisons among laparoscopic approaches have been less studied and often with confounding factors when comparing results between hospitals, surgeons, or at different eras of surgeon experience. In this series, such factors are likely to play less of a role although the choice of approach was influenced by tumor location and morphometry as indicated, reflecting the utilitarian nature of the different approaches under study. RPLPN is a UROLOGY 84 (6), 2014
less-commonly performed technique with some challenges because of a smaller working space and lessforgiving geometry; however, many surgeons prefer this approach for posterior and posteromedial tumors.13 Robotic technology with instrument articulation and capabilities in small confined spaces incorporating excellent optical stability and resolution has augmented the surgeon’s abilities for RPLPN with a potential for achieving results comparable with those of standard TPLPN as supported by our data and others.7,14,15 Advantages of the RP approach include avoidance of bowel mobilization or surgical fluid contamination (eg, blood and urine) within the peritoneal cavity. Purported benefits include less risk for bowel irritation and earlier return of bowel function after surgery.15 Moreover, the RP approach offers direct, rapid access to the hilum and posteriorly located renal tumors, allowing for less dissection and shorter surgical time in selected cases.13 Considerations must include nontraditional anatomic landmarks, smaller working space, and increased CO2 absorption relative to TP procedures. Prospective randomized trials comparing these approaches for radical nephrectomy have indicated little difference in typical metrics of surgical outcomes (eg, positive surgical margin rates, EBL, LOS, and analgesia).16 Prior reported studies comparing RP and TP LPNs have notable limitations creating difficulty in interpreting 1357
Table 2. Disease characteristics of patients treated with a retroperitoneal or transperitoneal partial nephrectomy for renal mass at the Memorial Sloan-Kettering Cancer Center (n ¼ 191) Disease Characteristics
Retroperitoneal (N ¼ 75)
Transperitoneal (N ¼ 116)
30 (40) 36 (48) 9 (12)
28 (24) 63 (54) 25 (22)
.04
51 (68) 24 (32) 0 (0)
79 (68) 22 (19) 15 (13)
.002
31 (41) 39 (52) 5 (7)
42 (36) 62 (53) 12 (10)
.6
21 (28) 15 (20) 39 (52)
28 (24) 6 (5) 82 (71)
.003
14 (19) 52 (69) 9 (12)
75 (65) 16 (14) 25 (22)
RESULTS
CONCLUSION
To evaluate and compare perioperative outcomes of transperitoneal and retroperitoneal (RP) laparoscopic and robotic partial nephrectomies (LPNs) while adjusting for tumor complexity. Retrospective review was conducted of 191 patients who underwent transperitoneal (n ¼ 116) or RP (n ¼ 75) LPN. To adjust for tumor complexity, individual components of the radius, exophytic or endophytic properties, nearness to the collecting system or sinus, anterior or posterior location, and location in reference to polar lines (R.E.N.A.L.) nephrometry score were used in multivariate linear and logistic regression models to compare perioperative outcomes between the 2 groups. A propensity approach was also used to adjust for multiple covariates. Investigated outcomes included estimated blood loss (EBL), ischemia and operative times, length of hospital stay, margin status, opioid use, postoperative estimated glomerular filtration rate, complications within 30 days, and readmission rates. Tumors resected by RPLPN were more likely to have lower complexity score by nephrometry (P ¼ .04). Four of the 5 components of the R.E.N.A.L. nephrometry score were significantly different between the groups. After adjustment for these factors, a lower EBL was noted in the RP group (b, 97; 95% confidence interval, 156 to 39; P ¼ .001). Risk of readmission for the RP group was significantly lower (odds ratio, 0.15; P ¼ .024) using propensity analysis. Using adjustment for tumor complexity, RPLPN was associated with lower EBL and readmission rates supporting the potential clinical advantage for this approach when feasible. UROLOGY 84: 1355e1360, 2014. 2014 Elsevier Inc.
R
enal cell carcinoma accounts for about 3% of adult cancers and 85%-90% of renal malignancies. Approximately 65,150 new cases of kidney cancers were diagnosed and 13,680 cancer-related deaths were projected in the United States in 2013.1 Detection of incidental renal masses has risen over the past decade with the increased utilization of computerized tomography. Concurrent with these trends has been the development and adaptation of less-invasive nephronsparing management strategies intended to optimally achieve cancer-specific outcomes while offering the putative physiologic advantages associated with renal function preservation. Surgical resection has been the Financial Disclosure: The authors declare that they have no relevant financial interests. From the Department of Urology, Mount Sinai Hospital, New York, NY; the Department of Epidemiology and Biostatistics, Memorial Sloan-Kettering Cancer Center, New York, NY; and the Department of Surgery, Sidney Kimmel Center for Prostate and Urologic Cancers, Memorial Sloan-Kettering Cancer Center, New York, NY Address correspondence to: Jonathan A. Coleman, M.D., Department of Surgery, Sidney Kimmel Center for Prostate and Urologic Cancers, Memorial Sloan-Kettering Cancer Center, 353 East 68th Street, New York, NY 10065. E-mail: [email protected] Submitted: February 11, 2014, accepted (with revisions): July 27, 2014
ª 2014 Elsevier Inc. All Rights Reserved
mainstay of therapeutic options for clinically localized renal tumors. Partial nephrectomy is considered the gold standard for treatment, demonstrating disease-specific outcomes equivalent to radical nephrectomy at experienced centers and confirmed in a prospective, randomized, multicenter trial.2-4 Advancements in technology and increased experience with laparoscopy have allowed urologic surgeons to broaden the indications of minimally invasive nephron-sparing surgery for the treatment of renal masses with results comparable with open procedures. Laparoscopic and robotic partial nephrectomy (LPN) can be accomplished through a transperitoneal (TP) or retroperitoneal (RP) approach, which is chosen based on the tumor location and surgeon experience.5,6 Because of limited experience, RPLPN is less commonly used but felt to offer advantages in regard to accessing vasculature and tumors based on the location, less manipulation of abdominal organs, and convalescence.7 To determine possible advantages related to the RP approach while attempting to control for many of the confounding features associated with retrospective http://dx.doi.org/10.1016/j.urology.2014.07.045 0090-4295/14
1355
multicenter studies, we analyzed outcomes in a series of consecutive TP and RP LPN cases performed with a similar technique managed under a standardized institutional care pathway using a multivariate and propensity statistical model.
MATERIALS AND METHODS Following institutional review board approval, a retrospective review of 252 LPNs by a single surgeon (J.A.C.) from November 2006 through March 2012 was performed. Patients having multiple renal tumors removed during the same procedure, insufficient imaging to assign a nephrometry score, or undergoing a concomitant procedure (eg, radical prostatectomy or laparoscopic hysterectomy) were excluded from the study. We identified 75 patients who underwent RPLPN and 116 patients who underwent TPLPN and met these criteria during this time. The approach was determined preoperatively by the discretion of the surgeon. A robot-assisted LPN was performed in 55 (73%) RP patients and 85 (73%) TP patients. All patients were managed under a standardized perioperative pathway for partial nephrectomy, which has been previously described.8 Briefly, the patients were started on a full liquid diet on postoperative day 0 and advanced as tolerated. Patients were offered patientcontrolled intravenous analgesia upon request and were transitioned to oral narcotics on day 1. Ketorolac is rarely provided. Patients were offered discharge when medically stable, capable of ambulating 1 mile without assistance, adequately pain controlled with oral medications, and tolerating a full liquid diet.
Surgical Technique Technical aspects of LPN performed by either approach have been previously described and underwent little modification over the study period.9 Briefly, a clear liquid diet was started 1 day before the surgery. Positioning was secured in a modified flank at 60 elevation without axillary roll. A Veress needle or balloon expansion was used to gain intraperitoneal or RP access, respectively, as previously described by others. Three to 4 ports were used for laparoscopic procedures and 4-5 ports for robotassisted cases. Tumors were evaluated with intraoperative ultrasound inspection. Vascular control, when used, was provided by bulldog clamps typically achieving complete renal ischemia. Excision was performed sharply through the cortex initially then with intrarenal anatomic dissection technique using diathermy to seal arterial vessels. Sutured closure of large venotomies and collecting system defects were performed separately with 3-0 coated Vicryl on RB needle (Ethicon Surgical, Somersville, NJ). Renorrhaphy was completed using interrupted 0 coated Vicryl sutures with CT-1 needle (Ethicon Surgical) over collagenthrombin matrix coagulant agents with oxidized cellulose material placed over the matrix agent before capsule apposition. After bulldog clamp removal, renal revascularization was confirmed with Doppler sonography. Intraoperative findings and surgical details were collected prospectively using a standardized reporting form maintained in the institutional surgical database system.
Statistical Methods Patient characteristics evaluated included gender, age, race, presentation, American Society of Anesthesiologists score, and preoperative estimated glomerular filtration rate (eGFR). 1356
Tumors were characterized using the radius, exophytic or endophytic properties, nearness to the collecting system or sinus, anterior or posterior location, and location in reference to polar lines (R.E.N.A.L.) nephrometry score.10 The subscale components of the scoring system were compared between groups. Differences were assessed using the Wilcoxon rank sum or the chi-square test. Perioperative outcomes compared between groups included estimated blood loss (EBL), ischemia and operative times, length of hospital stay (LOS), margin status, 24-hour opioid use, postoperative eGFR, Clavien complication grade (within 30 days of surgery), and readmission rates. Multivariate linear and logistic regression models were used. Because we expected that patients with larger and more complex tumors were more likely to have a TP approach, we adjusted for disease characteristics. Only variables significantly associated with surgical approach on univariate analysis were analyzed because of the relatively fewer numbers of RPLPNs. Given the relatively large number of covariates and the small number of occurring events, a propensity approach was performed to confirm the validity of the adjustment. A predicted probability of undergoing RP or TP LPN based on patient and disease characteristics was obtained. Such probability was then entered into a model predicting the outcome of interest to obtain an estimate of association between the surgical approach and each outcome. To ensure that the propensity score adequately controlled for disease characteristics, we reevaluated the association between surgical approach and each characteristic, adjusting for the propensity score. All analyses were conducted using Stata 12.0 (StataCorp, College Station, TX).
RESULTS Demographic and surgical characteristics of the 75 patients who underwent RPLPN and the 116 patients treated with TPLPN were similar (Table 1). Location of tumor was a significant predictor of surgical approach: 65% of the TPLPN group had anterior tumors, whereas 69% of the RPLPN group had posterior tumors (P .9 .2 .009
ASA, American Society of Anesthesiologists; CKD-EPI, Chronic Kidney Disease Epidemiology Collaboration Formula estimating glomerular filtration rate; R.E.N.A.L., radius, exophytic or endophytic properties, nearness to the collecting system or sinus, anterior vs posterior location, and location in reference to polar lines. Data are presented as frequency (percentage) or median (interquartile range). * P values obtained using the Wilcoxon rank sum or the chi-square test.
lower readmission rates (odds ratio [OR], 0.2; 95% CI, 0.04-1.0; P ¼ .05). When reevaluating the association between the surgical approach and each of the included disease characteristics, the propensity score appeared to adequately control for all included disease characteristics (all P >.8) and reaffirmed the findings of the multivariate approach (Table 4). Although the estimates for readmission rates were similar (OR, 0.2 in the multivariate vs 0.15 in the propensity approach), results were only found statistically significant under the propensity approach (P ¼ .024 vs P ¼ .05).
COMMENT The utilization of minimally invasive techniques for the management of small renal masses, including nephronsparing surgery, has increased in the past decade. LPN has been shown to have comparable oncologic outcomes to the open approach and allow for shorter hospitalizations.11,12 Comparisons among laparoscopic approaches have been less studied and often with confounding factors when comparing results between hospitals, surgeons, or at different eras of surgeon experience. In this series, such factors are likely to play less of a role although the choice of approach was influenced by tumor location and morphometry as indicated, reflecting the utilitarian nature of the different approaches under study. RPLPN is a UROLOGY 84 (6), 2014
less-commonly performed technique with some challenges because of a smaller working space and lessforgiving geometry; however, many surgeons prefer this approach for posterior and posteromedial tumors.13 Robotic technology with instrument articulation and capabilities in small confined spaces incorporating excellent optical stability and resolution has augmented the surgeon’s abilities for RPLPN with a potential for achieving results comparable with those of standard TPLPN as supported by our data and others.7,14,15 Advantages of the RP approach include avoidance of bowel mobilization or surgical fluid contamination (eg, blood and urine) within the peritoneal cavity. Purported benefits include less risk for bowel irritation and earlier return of bowel function after surgery.15 Moreover, the RP approach offers direct, rapid access to the hilum and posteriorly located renal tumors, allowing for less dissection and shorter surgical time in selected cases.13 Considerations must include nontraditional anatomic landmarks, smaller working space, and increased CO2 absorption relative to TP procedures. Prospective randomized trials comparing these approaches for radical nephrectomy have indicated little difference in typical metrics of surgical outcomes (eg, positive surgical margin rates, EBL, LOS, and analgesia).16 Prior reported studies comparing RP and TP LPNs have notable limitations creating difficulty in interpreting 1357
Table 2. Disease characteristics of patients treated with a retroperitoneal or transperitoneal partial nephrectomy for renal mass at the Memorial Sloan-Kettering Cancer Center (n ¼ 191) Disease Characteristics
Retroperitoneal (N ¼ 75)
Transperitoneal (N ¼ 116)
30 (40) 36 (48) 9 (12)
28 (24) 63 (54) 25 (22)
.04
51 (68) 24 (32) 0 (0)
79 (68) 22 (19) 15 (13)
.002
31 (41) 39 (52) 5 (7)
42 (36) 62 (53) 12 (10)
.6
21 (28) 15 (20) 39 (52)
28 (24) 6 (5) 82 (71)
.003
14 (19) 52 (69) 9 (12)
75 (65) 16 (14) 25 (22)
