World J Urol DOI 10.1007/s00345-015-1538-z
ORIGINAL ARTICLE
Cancer control of partial nephrectomy for high‑risk localized renal cell carcinoma: population‑based and single‑institutional analysis Rebecca L. O’Malley1,2 · Matthew H. Hayn1,3 · Katherine A. Brewer1 · Willie Underwood III1 · Nicholas J. Hellenthal4 · Hyung L. Kim1,5 · Igor Sorokin2 · Thomas Schwaab1
Received: 30 October 2014 / Accepted: 10 March 2015 © Springer-Verlag Berlin Heidelberg 2015
Abstract Purpose Cancer control of partial nephrectomy for highrisk localized renal cell carcinoma is unclear. To assess whether PN provides adequate cancer control in high-risk disease (HRD), survival outcomes were compared in both a population-based cohort and an institutional cohort. Methods Surveillance, Epidemiology, and End Results database and a prospectively maintained institutional database were queried for patients with RCC who underwent PN or RN for a localized tumor ≤7 cm and were found to have high-grade and/or high-stage disease (HRD). Cancerspecific (CSS) or recurrence-free survival (RFS) and overall survival (OS) were primary outcomes measured and were compared between those who underwent PN and RN using multivariable Cox proportional hazards and propensity analysis. Results The population cohort consisted of 12,757 (24.9 %) patients with HRD, 85.2 and 14.8 % of which underwent RN and PN, respectively. RN was not associated with CSS (HR 1.23, p = 0.08) but was independently
* Rebecca L. O’Malley [email protected] 1
Department of Urology, Roswell Park Cancer Institute, Buffalo, NY, USA
2
Department of Surgery, Albany Medical College, Urologic Institute of Northeastern New York, 23 Hackett Blvd., MC‑208, Albany, NY 12208, USA
3
Department of Urology, Maine Medical Center, Portland, ME, USA
4
Division of Urology, Department of Surgery, Bassett Healthcare, Cooperstown, NY, USA
5
Department of Urology, Cedars-Sinai Medical Center, Los Angeles, CA, USA
associated with poor OS (HR 1.16, p = 0.031). Propensity analysis showed that RN resulted in a 20 % increased risk of death from all causes (p = 0.008). In the institutional cohort, of 317 patients, 35.9 % had HRD, 56 and 52 of which underwent RN and PN, respectively. Adjusting for age-adjusted Charlson index, RN was a predictor of poor OS (OR 6.20, p = 0.041). Propensity analysis showed that RFS and OS were not related to nephrectomy type (RN HR 0.65, p = 0.627 and RN HR 1.70, p = 0.484). Conclusions In patients with pathologic high-risk RCC, partial excision is associated with similar cancer control as compared to radical excision. Keywords Renal cell carcinoma · Outcomes · Nephrectomy · Mortality · Complications
Introduction Estimates predict nearly 60,000 new kidney tumors in the USA in 2010, the majority of which will be incidentally discovered small renal masses [1]. The link between chronic kidney disease that results from radical surgery and cardiac morbidity and mortality is now well recognized, thus prompting recommendations for nephron-sparing approaches whenever feasible [2, 3]. Findings from the only randomized controlled trial comparing survival outcomes of radical (RN) versus partial nephrectomy (PN), however, were surprising. In the intention to treat analysis, RN was superior to PN for clinically localized masses ≤5 cm in size [4]. With removal of patients that pathologically did not meet inclusion criteria (size >5 cm, pathologic stage ≥T3), RN was no longer superior. This suggests that the superiority of RN could in part be attributable to patients with unsuspected high-risk disease and calls into
13
question the ability of PN to provide adequate cancer control in high-risk disease. This is in contrast to the preponderance of retrospective data which supports the equivalence of oncologic outcomes of PN and RN [3, 5–9]. The retrospective nature of these analyses introduces significant selection bias which is difficult to account for using only variables available in existing datasets. That being said, there is not likely to be another randomized trial due to substantial obstacles inherent in assigning patients to trial arms with respect to surgical treatments, as evidenced by the considerable difficulty in reaching accrual goals in the van Poppel et al. trial. Furthermore, because this trial was aimed at low-risk disease, it cannot be extrapolated to assess adequacy of cancer control of PN in high-risk RCC. For these reasons, it continues to be important to assess outcomes of treatment of high-risk disease in retrospective series. All variables that may affect survival outcomes are rarely available in large population cohorts. On the contrary, the large dataset necessary to obtain adequately powered analyses is rarely available in single-institution series, particularly if the analysis is limited to high-risk disease. To further investigate cancer control of PN as compared to RN, we examined cancer-specific and overall survival outcomes of patients who underwent nephrectomy for highrisk RCC both in a large population-based cohort and in a single institution series.
Methods Population‑based cohort The Surveillance, Epidemiology, and End Results (SEER) database, a national prospective cancer registry maintained by the National Cancer Institute which captures approximately 26 % of the population in the USA, was utilized. Patients from all regions were included. All patients with a diagnosis of RCC (International Classification of Diseases-9: 189.0) between 1988 and 2007 were identified (patients diagnosed prior to 1988 were excluded because accurate staging information for this time period was not available). Patients were included if they underwent PN or RN for a localized tumor ≤7 cm (because either type of nephrectomy is considered standard of care in this size group [3]) representing a cohort of 51,183 patients. All analyses used the American Joint Committee on Cancer (AJCC) pathologic TNM staging system relevant to the time of diagnosis. Because grade and stage remain the most important predictors of cancer outcome in RCC, only those with high-risk disease, defined as poorly or undifferentiated grade (equivalent to grade ≥3) and/or pathologic stage T3 were included leaving a cohort of 12,757 patients
13
World J Urol
[10]. For each case, demographic and pathologic data were abstracted, including age, gender, race (white, black or other), marital status (married vs. separated/widowed/ divorced/never married), year of diagnosis, tumor size and histologic subtype. Institutional cohort Following institutional review board approval, using a prospectively maintained, institutional renal tumor database, we identified patients clinically eligible for elective PN, who underwent PN or RN, from 1997 to 2010. All patients with a localized, solitary tumor ≤7 cm in size with a normal contralateral kidney and preoperative creatinine
ORIGINAL ARTICLE
Cancer control of partial nephrectomy for high‑risk localized renal cell carcinoma: population‑based and single‑institutional analysis Rebecca L. O’Malley1,2 · Matthew H. Hayn1,3 · Katherine A. Brewer1 · Willie Underwood III1 · Nicholas J. Hellenthal4 · Hyung L. Kim1,5 · Igor Sorokin2 · Thomas Schwaab1
Received: 30 October 2014 / Accepted: 10 March 2015 © Springer-Verlag Berlin Heidelberg 2015
Abstract Purpose Cancer control of partial nephrectomy for highrisk localized renal cell carcinoma is unclear. To assess whether PN provides adequate cancer control in high-risk disease (HRD), survival outcomes were compared in both a population-based cohort and an institutional cohort. Methods Surveillance, Epidemiology, and End Results database and a prospectively maintained institutional database were queried for patients with RCC who underwent PN or RN for a localized tumor ≤7 cm and were found to have high-grade and/or high-stage disease (HRD). Cancerspecific (CSS) or recurrence-free survival (RFS) and overall survival (OS) were primary outcomes measured and were compared between those who underwent PN and RN using multivariable Cox proportional hazards and propensity analysis. Results The population cohort consisted of 12,757 (24.9 %) patients with HRD, 85.2 and 14.8 % of which underwent RN and PN, respectively. RN was not associated with CSS (HR 1.23, p = 0.08) but was independently
* Rebecca L. O’Malley [email protected] 1
Department of Urology, Roswell Park Cancer Institute, Buffalo, NY, USA
2
Department of Surgery, Albany Medical College, Urologic Institute of Northeastern New York, 23 Hackett Blvd., MC‑208, Albany, NY 12208, USA
3
Department of Urology, Maine Medical Center, Portland, ME, USA
4
Division of Urology, Department of Surgery, Bassett Healthcare, Cooperstown, NY, USA
5
Department of Urology, Cedars-Sinai Medical Center, Los Angeles, CA, USA
associated with poor OS (HR 1.16, p = 0.031). Propensity analysis showed that RN resulted in a 20 % increased risk of death from all causes (p = 0.008). In the institutional cohort, of 317 patients, 35.9 % had HRD, 56 and 52 of which underwent RN and PN, respectively. Adjusting for age-adjusted Charlson index, RN was a predictor of poor OS (OR 6.20, p = 0.041). Propensity analysis showed that RFS and OS were not related to nephrectomy type (RN HR 0.65, p = 0.627 and RN HR 1.70, p = 0.484). Conclusions In patients with pathologic high-risk RCC, partial excision is associated with similar cancer control as compared to radical excision. Keywords Renal cell carcinoma · Outcomes · Nephrectomy · Mortality · Complications
Introduction Estimates predict nearly 60,000 new kidney tumors in the USA in 2010, the majority of which will be incidentally discovered small renal masses [1]. The link between chronic kidney disease that results from radical surgery and cardiac morbidity and mortality is now well recognized, thus prompting recommendations for nephron-sparing approaches whenever feasible [2, 3]. Findings from the only randomized controlled trial comparing survival outcomes of radical (RN) versus partial nephrectomy (PN), however, were surprising. In the intention to treat analysis, RN was superior to PN for clinically localized masses ≤5 cm in size [4]. With removal of patients that pathologically did not meet inclusion criteria (size >5 cm, pathologic stage ≥T3), RN was no longer superior. This suggests that the superiority of RN could in part be attributable to patients with unsuspected high-risk disease and calls into
13
question the ability of PN to provide adequate cancer control in high-risk disease. This is in contrast to the preponderance of retrospective data which supports the equivalence of oncologic outcomes of PN and RN [3, 5–9]. The retrospective nature of these analyses introduces significant selection bias which is difficult to account for using only variables available in existing datasets. That being said, there is not likely to be another randomized trial due to substantial obstacles inherent in assigning patients to trial arms with respect to surgical treatments, as evidenced by the considerable difficulty in reaching accrual goals in the van Poppel et al. trial. Furthermore, because this trial was aimed at low-risk disease, it cannot be extrapolated to assess adequacy of cancer control of PN in high-risk RCC. For these reasons, it continues to be important to assess outcomes of treatment of high-risk disease in retrospective series. All variables that may affect survival outcomes are rarely available in large population cohorts. On the contrary, the large dataset necessary to obtain adequately powered analyses is rarely available in single-institution series, particularly if the analysis is limited to high-risk disease. To further investigate cancer control of PN as compared to RN, we examined cancer-specific and overall survival outcomes of patients who underwent nephrectomy for highrisk RCC both in a large population-based cohort and in a single institution series.
Methods Population‑based cohort The Surveillance, Epidemiology, and End Results (SEER) database, a national prospective cancer registry maintained by the National Cancer Institute which captures approximately 26 % of the population in the USA, was utilized. Patients from all regions were included. All patients with a diagnosis of RCC (International Classification of Diseases-9: 189.0) between 1988 and 2007 were identified (patients diagnosed prior to 1988 were excluded because accurate staging information for this time period was not available). Patients were included if they underwent PN or RN for a localized tumor ≤7 cm (because either type of nephrectomy is considered standard of care in this size group [3]) representing a cohort of 51,183 patients. All analyses used the American Joint Committee on Cancer (AJCC) pathologic TNM staging system relevant to the time of diagnosis. Because grade and stage remain the most important predictors of cancer outcome in RCC, only those with high-risk disease, defined as poorly or undifferentiated grade (equivalent to grade ≥3) and/or pathologic stage T3 were included leaving a cohort of 12,757 patients
13
World J Urol
[10]. For each case, demographic and pathologic data were abstracted, including age, gender, race (white, black or other), marital status (married vs. separated/widowed/ divorced/never married), year of diagnosis, tumor size and histologic subtype. Institutional cohort Following institutional review board approval, using a prospectively maintained, institutional renal tumor database, we identified patients clinically eligible for elective PN, who underwent PN or RN, from 1997 to 2010. All patients with a localized, solitary tumor ≤7 cm in size with a normal contralateral kidney and preoperative creatinine
