- Email: [email protected]
A Population-based Comparison of Cancer-control Rates Between Radical and Partial Nephrectomy for T1A Renal Cell Carcinoma
Oncology A Population-based Comparison of Cancercontrol Rates Between Radical and Partial Nephrectomy for T1A Renal Cell Carcinoma Maxime Crépel, Claudio Jeldres, Maxine Sun, Giovanni Lughezzani, Hendrik Isbarn, Ahmed Alasker, Umberto Capitanio, Shahrokh F. Shariat, Philippe Arjane, Hugues Widmer, Markus Graefen, Francesco Montorsi, Paul Perrotte, and Pierre I. Karakiewicz OBJECTIVES
METHODS
RESULTS
CONCLUSIONS
To complement existing data with population-based cancer control outcomes that account for the effect of other-cause mortality (OCM). Cancer control rates are virtually equivalent between partial (PN) and radical nephrectomy (RN) for patients with T1aN0M0 renal cell carcinoma (RCC). To date, only 6 studies from centers of excellence examined cancer control rates after PN vs RN for T1aN0M0 RCC. OCM was unaccounted for in those studies, which may introduce a bias. We relied on the surveillance, epidemiology, and end results (SEER) database and assessed cancer-specific mortality (CSM) after either PN or RN for T1aN0M0 RCC, in competing-risks models. Between 1988 and 2004, the SEER-9 database identified 1622 PN (22.3%) and 5658 RN (77.7%) T1aN0M0 RCC. Competing-risks regression models, controlling for OCM and matched for age, year of surgery, tumor size, and Fuhrman grade, addressed the effect of nephrectomy type (PN vs RN) on CSM. At 5 years, in a PN and RN matched-population controlling for OCM, CSM after PN and RN was respectively 1.8% vs 2.5% (P ⫽ .5). The CSM rates in this cohort for patients aged ⱖ 70 years were respectively 1.0% and 3.4% (P ⫽ .7). This competing-risks population-based analysis confirmed the CSM equivalence between PN and RN for T1aN0M0 RCC and showed virtually perfect CSM-free rates (97.5% or better) even in older patients. UROLOGY 76: 883– 888, 2010. Crown Copyright © 2010 Published by Elsevier Inc.
S
ix reports available from centers of excellence in the United States and Europe indicate that partial nephrectomy (PN) represents a highly effective treatment modality for patients with T1aN0M0 renal cell carcinoma (RCC).1 At academic and/or tertiary care referral centers, an increasing proportion of patients are treated with PN.2-4 However, in previous popula-
Maxime Crépel and Claudio Jeldres contributed equally to the manuscript. Pierre I. Karakiewicz is partially supported by the University of Montreal Health Center Urology Associates, Fonds de la Recherche en Santé du Quebec, the University of Montreal Department Of Surgery, and the University of Montreal Health Center (CHUM) Foundation. Maxime Crépel is partially supported by the Association Française d’Urologie. From the Cancer Prognostics and Health Outcomes Unit, University of Montreal, Montreal, Quebec, Canada; Department of Urology, Rennes University Hospital, Rennes, France; Department of Urology, University of Montreal, Montreal, Canada; Department of Urology, Vita-Salute San Raffaele, Milan, Italy; and Martini-Clinic, Prostate Cancer Center Hamburg-Eppendorf, Hamburg, Germany Reprint requests: Pierre I. Karakiewicz, M.D., F.R.C.S.C., Cancer Prognostics and Health Outcomes Unit, University of Montreal Health Center (CHUM), 1058, rue St-Denis, Montréal, Québec, Canada, H2X 3J4. E-mail: pierre.karakiewicz@ umontreal.ca Submitted: July 3, 2009, accepted (with revisions): August 18, 2009
Crown Copyright © 2010 Published by Elsevier Inc. All Rights Reserved
tion-based analyses, most subjects underwent a radical nephrectomy (RN) even for T1a renal masses.5,6 For example, in the surveillance, epidemiology, and end results (SEER) database, only 26% of patients underwent a PN for T1a lesions ⬍ 2 cm and only 12% had a PN for T1a lesions 2-4 cm in size.7 In the Nationwide Inpatient Sample, the rate of PN for all T stages was as low as 10% at academic centers vs 5.2% at nonacademic centers.5 These population-based trends suggest PN underutilization and are worrisome. Radical nephrectomy instead of PN may predispose to surgically induced renal insufficiency,3,8-11 and may directly or indirectly increase its effect on other-cause mortality (OCM) rates. The population-based trends also prompted us to examine the rates of PN vs RN use over time and to compare cancerspecific mortality (CSM) rates according to nephrectomy type (PN vs RN). Other-cause mortality may undermine overall survival and may result in fewer individuals who remain at risk of CSM. This may in turn spuriously increase the CSM rate. To avoid such bias, we relied on 0090-4295/10/$34.00 doi:10.1016/j.urology.2009.08.028
883
competing-risks models that control for the confounding effect of OCM.12 Our hypothesis stated that the rate of PN use in the general population increased to a similar extent as in tertiary care centers. Moreover, we hypothesized that population-based cancer control outcomes of PN are equal to those after RN.4
MATERIAL AND METHODS Patient Population Patients diagnosed with primary invasive kidney cancer between 1988 and 2004 were identified within 9 SEER cancer registries.13 The registries include the Atlanta, Detroit, San Francisco–Oakland, and Seattle–Puget Sound Metropolitan areas, as well as the states of Connecticut, Hawaii, Iowa, New Mexico, and Utah. Characteristics of the SEER population are comparable to the general US population. Two kidney cancer diagnostic codes (international classification of diseases for oncology, second edition C64.9 code and ninth revision 189.0 code) were used as inclusion criteria. None of the cases included upper tract transitional cell carcinoma, ureteral cancer, or noncortical renal tumors (ie, melanomas, sarcomas, and lymphomas). Within this population, we focused on patients without nodal or distant metastases, who were aged ⱖ 18 years and who were treated within either PN or RN for tumors of size up to 4 cm. The cause of death was defined according to the SEER– specific cause of death (code 29020). Patients who did not die of RCC were considered to have died of other causes.
Statistical Analyses We relied on competing-risks regression models that adjust for the effect of OCM. The latter may spuriously reduce the number of individuals at risk of CSM, which may inflate the importance of cancer-related deaths.12 Because better treatment outcomes may be expected in more contemporary patients, matching by decade was also performed for the year of surgery. Caliper-matching, within 1 cm, was used for tumor size. Similarly, caliper-matching by decade was performed for patient age and for the year of surgery. Each patient treated with PN was matched with up to 4 RN patients to maximize the power of the comparisons. In addition to the main analyses, a subset analysis was performed in patients with available Fuhrman grade. Here, apart from matching for age, year of surgery, and tumor size, PN and RN cases were also matched for corresponding Fuhrman grades. The same analyses were repeated in patients aged ⱖ 70 years. Cumulative incidence plots were used to graphically explore the observed CSM rate differences according to nephrectomy type (PN vs RN), after controlling for OCM. All statistical tests were performed using S-PLUS Professional, version 1 (Mathsoft, Seattle, WA). Moreover, all tests were two-sided with a significance level set at ⬍ .05.
RESULTS The selection criteria resulted in the identification of 7280 patients with N0M0 RCC, aged ⱖ 18 years, who had malignant renal cortical tumors of size up to 4 cm (T1a). Of those patients, 1622 (22.3%) underwent a PN, and 5658 (77.7%), a RN. The temporal trends of PN use 884
are shown in Figure 1. Panel A shows the overall rate of PN in T1a patients, which increased from 7.1% to 35.9% during the study period (absolute increase: 28.8%; relative increase: 505.6%). After stratification according to tumor sizes ⱕ 2 vs 2.1-4.0 cm (Fig. 1B and C), the greatest increase in PN use was found for ⱕ 2-cm masses (from 12.7% to 53.5%; absolute increase: 40.8%; relative increase: 421.3%). Nonetheless, an important increase in PN rate was also recorded in patients with tumors of size 2.1-4.0 cm (from 6.1% to 29.2%; absolute increase: 23.1%; relative increase: 478.7%). Overall, 1564 of the 1622 PN patients (96.4%) were matched with 3955 (69.9%) of 5658 available RN patients (Table 1). Matching was done for age, tumor size, and year of surgery. When Fuhrman grade was added to the other 3 matching variables, 961 (59.2%) of 1622 PN cases could be matched with 2341 (41.3%) of 5658 RN cases. Age, gender, tumor size, histologic subtype, Fuhrman grade, follow-up length, and the year of surgery characteristics of the PN and RN matched-populations are shown in Table 1. On average, PN patients and RN patients were respectively 59 to 61 years old. Regardless of surgery type, the majority (57.9%-61.5%) were men and the mean and median tumor sizes were 2.5 and 3.0 cm, respectively. In the age, tumor size, and year of surgery-matched cohort (n ⫽ 5519), PN patients had smaller tumors (2.5 cm) than RN patients (2.8 cm). This difference persisted after matching was also made for Fuhrman grade (n ⫽ 3302) (2.5 vs 2.8 cm). Fuhrman grade II was recorded in most PN (55.9%) and RN (55.2%) cases. Most surgeries were performed between 2000 and 2004. The median follow-up of censored cases was 24-29 and 38-41 months for PN and RN patients, respectively.
Effect of PN vs RN on CSM In age, year of surgery, and tumor size–matched competingrisks regression models that focused on 5519 patients (1564 PN vs 3955 RN patients), the 5-year CSM rate recorded for PN [2.6% (95% confidence interval (CI): 1.5-3.8)] patients was not statistically significantly different (log-rank P ⫽ .3) than that of RN patients [2.4% (95% CI: 1.9%-3.0%)], after adjustment for OCM (Fig. 2A). In age, year of surgery, tumor size, and Fuhrman gradematched competing-risks regression models that focused on 3302 patients (961 PN vs 2341 RN patients), the 5-year CSM rate recorded for PN patients [1.8% (95% CI: 0.63.0)] was also not statistically significantly different (logrank P ⫽ .5) than that of RN patients [2.5% (95% CI: 1.7%-3.3%)], after adjustment for OCM (Fig. 2B). When data were restricted to patients aged ⱖ 70 years (n ⫽ 1495), after matching for age, tumor size, and year of surgery, the 5-year CSM rates of PN (n ⫽ 385) vs RN (n ⫽ 1110) were respectively 1.8% (95% CI: 0.1-3.4) and 3.9% (95% CI: 2.5-5.3) (log-rank P ⫽ .5). In the same individuals after matching for age, tumor size, year UROLOGY 76 (4), 2010
of surgery, and Fuhrman grade (n ⫽ 927), the 5-year CSM rates of PN (n ⫽ 234) vs RN (n ⫽ 693) were respectively 1.0% (95% CI: 0.0-2.4) and 3.4% (95% CI: 1.7-5.1) (log-rank P ⫽ .7).
COMMENT
Figure 1. Distribution of PN vs RN according to year of nephrectomy in 7280 patients (A) stratified according to tumor size ⱕ 2.0 cm (B) and 2.1-4.0 cm (C).
UROLOGY 76 (4), 2010
The advantages of PN relative to RN are manifold. First, the preservation of renal parenchyma results in lower rates of renal insufficiency after PN compared with RN.4,9,14-20 Second, OCM is on average 4.8% lower at 5 years after surgery if PN is performed instead of RN.18 Third, health-related quality of life is better after PN than after RN.21 These advantages should represent the main indications for the use of PN relative to RN. Nonetheless some disadvantages of PN also exist. These consist of acute complications, length of hospital stay, and intraoperative blood loss.22 These might represent some of the reasons for the low PN use in the community relative to tertiary care centers. We hypothesized that PN rate increased in the general population to the same extent, as in tertiary care centers.3,4 Moreover, we postulated that CSM of PN, after controlling for the confounding effect of OCM, is the same between PN and RN. With regard to PN use, our results demonstrated a sharp increase in PN rates over time. PN rates increased from 7.1% to 35.9% for all T1a tumors and from 12.7% to 53.5% and 6.1% to 29.2%, respectively, for tumors of size ⱕ 2.0 and 2.1-4.0 cm. The absolute PN rate increase in patients with T1aN0M0 was 28.8% during the study period (relative increase: 505.6%). For patients with tumor size ⱕ 2.0 cm, the absolute increase was highest at 40.8% (relative increase: 421.3%) and exceeded that of patients with tumor size 2.1-4.0 cm, in whom the absolute increase was 23.1% (relative increase: 478.7%). Although these increases are substantially higher than those indicated in previous population-based reports (SEER 1988-2001: 14%-42%; the Nationwide Inpatient Sample 1988-2002: 3.7%-12.3%), they are lower than in tertiary care centers. For example, at Memorial Sloan-Kettering Cancer Center 89% of patients with T1a RCC were treated with PN.4 Wider use of laparoscopic RN and concomitant more limited use of laparoscopic PN in the community may explain this observation.23 Alternatively, a high rate of acute complications after PN and greater technical complexity of PN might also contribute to PN underutilization. Nonetheless, the ultimate complication, namely, perioperative mortality, was not higher after PN vs RN for T1a lesions.24 Moreover, the long-term survival is better after PN vs RN.18 Although the PN rate observed in this study is not ideal (35.9%-53.5%), we demonstrated a steep increase in PN use during the most contemporary years relative to the more historic years. On the basis of these trends, one could speculate that this increasing rate will be maintained after the year 2004, which represents the most available year for SEER records. 885
Table 1. Clinical characteristics of the population matched for age, tumor size, and year of surgery (n ⫽ 5519) and of the population matched for age, tumor size, year of surgery, and Fuhrman grade (n ⫽ 3302), stratified according to type of surgery (PN vs RN) Population Matched for Age, Tumor Size, and Year of Surgery, Unmatched for Fuhrman Grade (n ⫽ 5519) PN (%) RN (%) No patients (%) Age (years) Mean Range Gender Male Tumor size (cm) Mean (median) Histology Clear cell Papillary Chromophobe Unclassified Fuhrman grade I II III IV Unknown Year of surgery 1988-1989 1990-1999 2000-2004 Median follow-up of censored patients (months)
1564 (28.3)
3955 (71.7)
961 (29.1)
2341 (70.9)
59.5 18-90
60.5 18-97
59.3 18-89
60.6 19.93
961 (61.4)
2361 (59.7)
591 (61.5)
1355 (57.9)
2.5 (2.5)
2.8 (2.8)
2.5 (2.5)
2.8 (3.0)
1282 (82.0) 192 (12.3) 58 (3.7) 32 (2.0)
3273 (87.8) 331 (8.4) 106 (2.7) 45 (1.1)
806 (83.9) 107 (11.1) 35 (3.6) 13 (1.4)
2058 (87.9) 210 (9.0) 53 (2.3) 20 (0.9)
286 (18.3) 556 (35.5) 133 (8.5) 17 (1.1) 572 (36.6)
628 (15.9) 1303 (32.9) 315 (8.0) 26 (0.7) 1683 (42.6)
276 (28.7) 537 (55.9) 131 (13.6) 17 (1.8) 0 (0.0)
662 (28.3) 1340 (57.2) 308 (13.2) 31 (1.3) 0 (0.0)
25 (1.6) 410 (26.2) 1129 (72.2) 29.0
93 (2.4) 1536 (38.8) 2326 (58.8) 41.0
0 (0.9) 229 (23.8) 723 (75.2) 24.0
34 (1.5) 831 (35.5) 1476 (63.0) 38.0
In the second part of the analysis, we examined CSM after either PN or RN. Because OCM may decrease the number of individuals at risk of CSM, it may artificially increase the CSM rate. Therefore, we relied on competingrisks regression models that adjust for the confounding effect of OCM.12 Our study is the first to rely on competing-risks regression to assess CSM in a large population-based sample of patients with T1a tumors who were treated with either PN or RN from North America. Another French group of investigators used competing-risks regression to assess OCM in other populations.9 Similarly, a multi-institutional collaborative group relied on competing-risks regression models to assess the efficacy of surveillance on OCM in patients with T1a tumors.18 Our study represents the first published report to compare population-based cancer control outcomes between PN and RN in T1a RCC. Since 1990, merely 6 studies from tertiary care centers have addressed this topic.25-29 Of those studies, the largest one relied on 379 PN that were compared with 1075 RN.26 In this study, most of the PN originated from European institutions. The second largest study originated from the Mayo Clinic Foundation and compared 358 PN with 290 RN.29 both studies confirmed the equivalence of CSM-free rates between PN and RN. Finally, investigators from Memorial Sloan-Kettering Cancer Center, University of California in Los Angeles, and from the Mayo Clinic addressed the same topic and respectively relied on 79, 146, and 185 PN cases in their compar886
Population Matched for Age, Tumor Size, Year of Surgery, and Furhman Grade (n ⫽ 3302) PN (%) RN (%)
isons, all of which confirmed the equivalence of cancer control efficacy of PN vs RN.25,27,28 The uniqueness of our findings stems from our patient population, sample size, and statistical methodology. Our patient population does not include any of the North American centers of excellence that previously reported their PN results. In consequence, our findings truly support the equivalence of cancer control between PN vs RN in the community. Moreover, the 5-year CSM rates indicate virtually perfect cancer control after either PN or RN. Only between 2.6% and 2.5% of patients died of RCC respectively after PN or RN. These findings indicate that the natural history of surgically treated small renal cortical tumors is favorable. This suggests that less invasive treatment modalities such as cryoablation, high intensity focused ultrasound, or even active surveillance could be safely applied to a large proportion of patients with T1a lesions.30 Specifically, elderly patients may represent perfect candidates for such treatment modalities as 5-year CSM rates in this subgroup were low (1.0%3.8%). Despite its strengths, limitations do apply to our study. For example, we could not account for a number of potentially important variables. These include institution type, insurance status, surgical volume, and comorbidities.23 These variables should ideally be considered in treatment selection. Moreover, we could not consider American Society of Anesthesiologists scores and baseUROLOGY 76 (4), 2010
Figure 2. Cumulative incidence plots depicting cancer-specific mortality after accounting for other-cause mortality in a population of 5519 patients matched for year of surgery, age, and tumor size (A) and in a population of 3302 patients matched for year of surgery, age, tumor size, and Fuhrman grade (B).
line glomerular filtration rate, which also represent important determinants of nephrectomy treatment.9 Moreover, we had no access to comorbidity information, which may also affect the frequency of PN use as well as CSM rates. Finally, PN rates may be affected by hospital and surgical volume. Despite the presence of some disadvantages related to the use of PN vs RN, the equivalence of cancer control outcomes, combined with renal funcUROLOGY 76 (4), 2010
tion preservation and quality of life benefits, make a strong case for PN for lesions that are technically amenable to such surgery.
CONCLUSIONS Our findings demonstrate the equivalence of PN relative to RN, when cancer control outcomes are considered in 887
patients with T1a RCC. This complements reports from tertiary care centers at which PN was shown to be equally effective to RN. In the light of important overuse of RN for small renal cortical tumors in North America, the current report should be perceived as a confirmation of PN efficacy and it should prompt wider use of PN in the community. References 1. Ljungberg B, Hanbury DC, Kuczyk MA, et al. Renal cell carcinoma guideline. Eur Urol. 2007;51:1502-1510. 2. Patard JJ, Tazi H, Bensalah K, et al. The changing evolution of renal tumours: a single center experience over a two-decade period. Eur Urol. 2004;45:490-493; discussion: 493-494. 3. Zini L, Patard JJ, Capitanio U, et al. The use of partial nephrectomy in European tertiary care centers. Eur J Surg Oncol. 2009;35:636642. 4. Thompson RH, Kaag M, Vickers A, et al. Contemporary use of partial nephrectomy at a tertiary care center in the United States. J Urol. 2009;181:993-997. 5. Hollenbeck BK, Taub DA, Miller DC, et al. National utilization trends of partial nephrectomy for renal cell carcinoma: a case of underutilization? Urology. 2006;67:254-259. 6. Miller DC, Saigal CS, Banerjee M, et al. Diffusion of surgical innovation among patients with kidney cancer. Cancer. 2008;112: 1708-1717. 7. Miller DC, Hollingsworth JM, Hafez KS, et al. Partial nephrectomy for small renal masses: an emerging quality of care concern? J Urol. 2006;175:853-857; discussion: 858. 8. James M, Ofer Y, Michael WK, et al. Partial nephrectomy for renal cortical tumors: pathologic findings and impact on outcome. Urology. 2002;60:1003-1009. 9. Zini L, Patard JJ, Capitanio U, et al. Cancer-specific and noncancer-related mortality rates in European patients with T1a and T1b renal cell carcinoma. BJU Int. 2009;103:894-898. 10. Russo P. Localized renal cell carcinoma. Curr Treat Options Oncol. 2001;2:447-455. 11. Go AS, Chertow GM, Fan D, et al. Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med. 2004;351:1296-1305. 12. Jason PF, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94:496-509. 13. Ries LAG, Melbert D, Krapcho M, et al. SEER Cancer Statistics Review, 1975-2004, National Cancer Institute. Bethesda, MD. Available at: http://seer.cancer.gov/csr/1975-2004/, based on November 2006 SEER data submission, posted to SEER website. Accessed January 2007. 14. Snow DC, Bhayani SB. Rapid communication: chronic renal insufficiency after laparoscopic partial nephrectomy and radical nephrectomy for pathologic t1a lesions. J Endourol. 2008;22:337-341.
888
15. Huang WC, Levey AS, Serio AM, et al. Chronic kidney disease after nephrectomy in patients with renal cortical tumours: a retrospective cohort study. Lancet Oncol. 2006;7:735-740. 16. Lau WK, Blute ML, Weaver AL, et al. Matched comparison of radical nephrectomy vs nephron-sparing surgery in patients with unilateral renal cell carcinoma and a normal contralateral kidney. Mayo Clin Proc. 2000;75:1236-1242. 17. McKiernan J, Yossepowitch O, Kattan MW, et al. Partial nephrectomy for renal cortical tumors: pathologic findings and impact on outcome. Urology. 2002;60:1003-1009. 18. Zini L, Perrotte P, Capitanio U, et al. Radical versus partial nephrectomy. Cancer. 2009;115:1465-1471. 19. O’Connor KM, Davis N, Lennon GM, et al. Can we avoid surgery in elderly patients with renal masses by using the Charlson comorbidity index? BJU Int. 2009;103:1492-1495. 20. Huang WC, Elkin EB, Levey AS, et al. Partial nephrectomy versus radical nephrectomy in patients with small renal tumors—is there a difference in mortality and cardiovascular outcomes? J Urol. 2009;181:55-61; discussion: 61-62. 21. Lesage K, Joniau S, Fransis K, et al. Comparison between open partial and radical nephrectomy for renal tumours: perioperative outcome and health-related quality of life. Eur Urol. 2007;51:614620. 22. Stephenson AJ, Hakimi AA, Snyder ME, et al. Complications of radical and partial nephrectomy in a large contemporary cohort. J Urol. 2004;171:130-134. 23. Porter MP, Lin DW. Trends in renal cancer surgery and patient provider characteristics associated with partial nephrectomy in the United States. Urol Oncol. 2007;25:298-302. 24. Cloutier V, Capitanio U, Zini L, et al. Thirty-day mortality after nephrectomy: clinical implications for informed consent. Eur Urol, in press. 25. Lerner SE, Hawkins CA, Blute ML, et al. Disease outcome in patients with low stage renal cell carcinoma treated with nephron sparing or radical surgery. J Urol. 1996;155:1868-1873. 26. Patard JJ, Shvarts O, Lam JS, et al. Safety and efficacy of partial nephrectomy for all T1 tumors based on an international multicenter experience. J Urol. 2004;171:2181-2185; quiz: 2435. 27. Lee CT, Katz J, Shi W, et al. Surgical management of renal tumors 4 cm. or less in a contemporary cohort. J Urol. 2000;163:730-736. 28. Belldegrun A, Tsui KH, deKernion JB, et al. Efficacy of nephronsparing surgery for renal cell carcinoma: analysis based on the new 1997 tumor-node-metastasis staging system. J Clin Oncol. 1999;17: 2868-2875. 29. Thompson RH, Boorjian SA, Lohse CM, et al. Radical nephrectomy for pT1a renal masses may be associated with decreased overall survival compared with partial nephrectomy. J Urol. 2008; 179:468-471; discussion: 472-473. 30. Mabjeesh NJ, Avidor Y, Matzkin H. Emerging nephron sparing treatments for kidney tumors: a continuum of modalities from energy ablation to laparoscopic partial nephrectomy. J Urol. 2004; 171:553-560.
UROLOGY 76 (4), 2010
METHODS
RESULTS
CONCLUSIONS
To complement existing data with population-based cancer control outcomes that account for the effect of other-cause mortality (OCM). Cancer control rates are virtually equivalent between partial (PN) and radical nephrectomy (RN) for patients with T1aN0M0 renal cell carcinoma (RCC). To date, only 6 studies from centers of excellence examined cancer control rates after PN vs RN for T1aN0M0 RCC. OCM was unaccounted for in those studies, which may introduce a bias. We relied on the surveillance, epidemiology, and end results (SEER) database and assessed cancer-specific mortality (CSM) after either PN or RN for T1aN0M0 RCC, in competing-risks models. Between 1988 and 2004, the SEER-9 database identified 1622 PN (22.3%) and 5658 RN (77.7%) T1aN0M0 RCC. Competing-risks regression models, controlling for OCM and matched for age, year of surgery, tumor size, and Fuhrman grade, addressed the effect of nephrectomy type (PN vs RN) on CSM. At 5 years, in a PN and RN matched-population controlling for OCM, CSM after PN and RN was respectively 1.8% vs 2.5% (P ⫽ .5). The CSM rates in this cohort for patients aged ⱖ 70 years were respectively 1.0% and 3.4% (P ⫽ .7). This competing-risks population-based analysis confirmed the CSM equivalence between PN and RN for T1aN0M0 RCC and showed virtually perfect CSM-free rates (97.5% or better) even in older patients. UROLOGY 76: 883– 888, 2010. Crown Copyright © 2010 Published by Elsevier Inc.
S
ix reports available from centers of excellence in the United States and Europe indicate that partial nephrectomy (PN) represents a highly effective treatment modality for patients with T1aN0M0 renal cell carcinoma (RCC).1 At academic and/or tertiary care referral centers, an increasing proportion of patients are treated with PN.2-4 However, in previous popula-
Maxime Crépel and Claudio Jeldres contributed equally to the manuscript. Pierre I. Karakiewicz is partially supported by the University of Montreal Health Center Urology Associates, Fonds de la Recherche en Santé du Quebec, the University of Montreal Department Of Surgery, and the University of Montreal Health Center (CHUM) Foundation. Maxime Crépel is partially supported by the Association Française d’Urologie. From the Cancer Prognostics and Health Outcomes Unit, University of Montreal, Montreal, Quebec, Canada; Department of Urology, Rennes University Hospital, Rennes, France; Department of Urology, University of Montreal, Montreal, Canada; Department of Urology, Vita-Salute San Raffaele, Milan, Italy; and Martini-Clinic, Prostate Cancer Center Hamburg-Eppendorf, Hamburg, Germany Reprint requests: Pierre I. Karakiewicz, M.D., F.R.C.S.C., Cancer Prognostics and Health Outcomes Unit, University of Montreal Health Center (CHUM), 1058, rue St-Denis, Montréal, Québec, Canada, H2X 3J4. E-mail: pierre.karakiewicz@ umontreal.ca Submitted: July 3, 2009, accepted (with revisions): August 18, 2009
Crown Copyright © 2010 Published by Elsevier Inc. All Rights Reserved
tion-based analyses, most subjects underwent a radical nephrectomy (RN) even for T1a renal masses.5,6 For example, in the surveillance, epidemiology, and end results (SEER) database, only 26% of patients underwent a PN for T1a lesions ⬍ 2 cm and only 12% had a PN for T1a lesions 2-4 cm in size.7 In the Nationwide Inpatient Sample, the rate of PN for all T stages was as low as 10% at academic centers vs 5.2% at nonacademic centers.5 These population-based trends suggest PN underutilization and are worrisome. Radical nephrectomy instead of PN may predispose to surgically induced renal insufficiency,3,8-11 and may directly or indirectly increase its effect on other-cause mortality (OCM) rates. The population-based trends also prompted us to examine the rates of PN vs RN use over time and to compare cancerspecific mortality (CSM) rates according to nephrectomy type (PN vs RN). Other-cause mortality may undermine overall survival and may result in fewer individuals who remain at risk of CSM. This may in turn spuriously increase the CSM rate. To avoid such bias, we relied on 0090-4295/10/$34.00 doi:10.1016/j.urology.2009.08.028
883
competing-risks models that control for the confounding effect of OCM.12 Our hypothesis stated that the rate of PN use in the general population increased to a similar extent as in tertiary care centers. Moreover, we hypothesized that population-based cancer control outcomes of PN are equal to those after RN.4
MATERIAL AND METHODS Patient Population Patients diagnosed with primary invasive kidney cancer between 1988 and 2004 were identified within 9 SEER cancer registries.13 The registries include the Atlanta, Detroit, San Francisco–Oakland, and Seattle–Puget Sound Metropolitan areas, as well as the states of Connecticut, Hawaii, Iowa, New Mexico, and Utah. Characteristics of the SEER population are comparable to the general US population. Two kidney cancer diagnostic codes (international classification of diseases for oncology, second edition C64.9 code and ninth revision 189.0 code) were used as inclusion criteria. None of the cases included upper tract transitional cell carcinoma, ureteral cancer, or noncortical renal tumors (ie, melanomas, sarcomas, and lymphomas). Within this population, we focused on patients without nodal or distant metastases, who were aged ⱖ 18 years and who were treated within either PN or RN for tumors of size up to 4 cm. The cause of death was defined according to the SEER– specific cause of death (code 29020). Patients who did not die of RCC were considered to have died of other causes.
Statistical Analyses We relied on competing-risks regression models that adjust for the effect of OCM. The latter may spuriously reduce the number of individuals at risk of CSM, which may inflate the importance of cancer-related deaths.12 Because better treatment outcomes may be expected in more contemporary patients, matching by decade was also performed for the year of surgery. Caliper-matching, within 1 cm, was used for tumor size. Similarly, caliper-matching by decade was performed for patient age and for the year of surgery. Each patient treated with PN was matched with up to 4 RN patients to maximize the power of the comparisons. In addition to the main analyses, a subset analysis was performed in patients with available Fuhrman grade. Here, apart from matching for age, year of surgery, and tumor size, PN and RN cases were also matched for corresponding Fuhrman grades. The same analyses were repeated in patients aged ⱖ 70 years. Cumulative incidence plots were used to graphically explore the observed CSM rate differences according to nephrectomy type (PN vs RN), after controlling for OCM. All statistical tests were performed using S-PLUS Professional, version 1 (Mathsoft, Seattle, WA). Moreover, all tests were two-sided with a significance level set at ⬍ .05.
RESULTS The selection criteria resulted in the identification of 7280 patients with N0M0 RCC, aged ⱖ 18 years, who had malignant renal cortical tumors of size up to 4 cm (T1a). Of those patients, 1622 (22.3%) underwent a PN, and 5658 (77.7%), a RN. The temporal trends of PN use 884
are shown in Figure 1. Panel A shows the overall rate of PN in T1a patients, which increased from 7.1% to 35.9% during the study period (absolute increase: 28.8%; relative increase: 505.6%). After stratification according to tumor sizes ⱕ 2 vs 2.1-4.0 cm (Fig. 1B and C), the greatest increase in PN use was found for ⱕ 2-cm masses (from 12.7% to 53.5%; absolute increase: 40.8%; relative increase: 421.3%). Nonetheless, an important increase in PN rate was also recorded in patients with tumors of size 2.1-4.0 cm (from 6.1% to 29.2%; absolute increase: 23.1%; relative increase: 478.7%). Overall, 1564 of the 1622 PN patients (96.4%) were matched with 3955 (69.9%) of 5658 available RN patients (Table 1). Matching was done for age, tumor size, and year of surgery. When Fuhrman grade was added to the other 3 matching variables, 961 (59.2%) of 1622 PN cases could be matched with 2341 (41.3%) of 5658 RN cases. Age, gender, tumor size, histologic subtype, Fuhrman grade, follow-up length, and the year of surgery characteristics of the PN and RN matched-populations are shown in Table 1. On average, PN patients and RN patients were respectively 59 to 61 years old. Regardless of surgery type, the majority (57.9%-61.5%) were men and the mean and median tumor sizes were 2.5 and 3.0 cm, respectively. In the age, tumor size, and year of surgery-matched cohort (n ⫽ 5519), PN patients had smaller tumors (2.5 cm) than RN patients (2.8 cm). This difference persisted after matching was also made for Fuhrman grade (n ⫽ 3302) (2.5 vs 2.8 cm). Fuhrman grade II was recorded in most PN (55.9%) and RN (55.2%) cases. Most surgeries were performed between 2000 and 2004. The median follow-up of censored cases was 24-29 and 38-41 months for PN and RN patients, respectively.
Effect of PN vs RN on CSM In age, year of surgery, and tumor size–matched competingrisks regression models that focused on 5519 patients (1564 PN vs 3955 RN patients), the 5-year CSM rate recorded for PN [2.6% (95% confidence interval (CI): 1.5-3.8)] patients was not statistically significantly different (log-rank P ⫽ .3) than that of RN patients [2.4% (95% CI: 1.9%-3.0%)], after adjustment for OCM (Fig. 2A). In age, year of surgery, tumor size, and Fuhrman gradematched competing-risks regression models that focused on 3302 patients (961 PN vs 2341 RN patients), the 5-year CSM rate recorded for PN patients [1.8% (95% CI: 0.63.0)] was also not statistically significantly different (logrank P ⫽ .5) than that of RN patients [2.5% (95% CI: 1.7%-3.3%)], after adjustment for OCM (Fig. 2B). When data were restricted to patients aged ⱖ 70 years (n ⫽ 1495), after matching for age, tumor size, and year of surgery, the 5-year CSM rates of PN (n ⫽ 385) vs RN (n ⫽ 1110) were respectively 1.8% (95% CI: 0.1-3.4) and 3.9% (95% CI: 2.5-5.3) (log-rank P ⫽ .5). In the same individuals after matching for age, tumor size, year UROLOGY 76 (4), 2010
of surgery, and Fuhrman grade (n ⫽ 927), the 5-year CSM rates of PN (n ⫽ 234) vs RN (n ⫽ 693) were respectively 1.0% (95% CI: 0.0-2.4) and 3.4% (95% CI: 1.7-5.1) (log-rank P ⫽ .7).
COMMENT
Figure 1. Distribution of PN vs RN according to year of nephrectomy in 7280 patients (A) stratified according to tumor size ⱕ 2.0 cm (B) and 2.1-4.0 cm (C).
UROLOGY 76 (4), 2010
The advantages of PN relative to RN are manifold. First, the preservation of renal parenchyma results in lower rates of renal insufficiency after PN compared with RN.4,9,14-20 Second, OCM is on average 4.8% lower at 5 years after surgery if PN is performed instead of RN.18 Third, health-related quality of life is better after PN than after RN.21 These advantages should represent the main indications for the use of PN relative to RN. Nonetheless some disadvantages of PN also exist. These consist of acute complications, length of hospital stay, and intraoperative blood loss.22 These might represent some of the reasons for the low PN use in the community relative to tertiary care centers. We hypothesized that PN rate increased in the general population to the same extent, as in tertiary care centers.3,4 Moreover, we postulated that CSM of PN, after controlling for the confounding effect of OCM, is the same between PN and RN. With regard to PN use, our results demonstrated a sharp increase in PN rates over time. PN rates increased from 7.1% to 35.9% for all T1a tumors and from 12.7% to 53.5% and 6.1% to 29.2%, respectively, for tumors of size ⱕ 2.0 and 2.1-4.0 cm. The absolute PN rate increase in patients with T1aN0M0 was 28.8% during the study period (relative increase: 505.6%). For patients with tumor size ⱕ 2.0 cm, the absolute increase was highest at 40.8% (relative increase: 421.3%) and exceeded that of patients with tumor size 2.1-4.0 cm, in whom the absolute increase was 23.1% (relative increase: 478.7%). Although these increases are substantially higher than those indicated in previous population-based reports (SEER 1988-2001: 14%-42%; the Nationwide Inpatient Sample 1988-2002: 3.7%-12.3%), they are lower than in tertiary care centers. For example, at Memorial Sloan-Kettering Cancer Center 89% of patients with T1a RCC were treated with PN.4 Wider use of laparoscopic RN and concomitant more limited use of laparoscopic PN in the community may explain this observation.23 Alternatively, a high rate of acute complications after PN and greater technical complexity of PN might also contribute to PN underutilization. Nonetheless, the ultimate complication, namely, perioperative mortality, was not higher after PN vs RN for T1a lesions.24 Moreover, the long-term survival is better after PN vs RN.18 Although the PN rate observed in this study is not ideal (35.9%-53.5%), we demonstrated a steep increase in PN use during the most contemporary years relative to the more historic years. On the basis of these trends, one could speculate that this increasing rate will be maintained after the year 2004, which represents the most available year for SEER records. 885
Table 1. Clinical characteristics of the population matched for age, tumor size, and year of surgery (n ⫽ 5519) and of the population matched for age, tumor size, year of surgery, and Fuhrman grade (n ⫽ 3302), stratified according to type of surgery (PN vs RN) Population Matched for Age, Tumor Size, and Year of Surgery, Unmatched for Fuhrman Grade (n ⫽ 5519) PN (%) RN (%) No patients (%) Age (years) Mean Range Gender Male Tumor size (cm) Mean (median) Histology Clear cell Papillary Chromophobe Unclassified Fuhrman grade I II III IV Unknown Year of surgery 1988-1989 1990-1999 2000-2004 Median follow-up of censored patients (months)
1564 (28.3)
3955 (71.7)
961 (29.1)
2341 (70.9)
59.5 18-90
60.5 18-97
59.3 18-89
60.6 19.93
961 (61.4)
2361 (59.7)
591 (61.5)
1355 (57.9)
2.5 (2.5)
2.8 (2.8)
2.5 (2.5)
2.8 (3.0)
1282 (82.0) 192 (12.3) 58 (3.7) 32 (2.0)
3273 (87.8) 331 (8.4) 106 (2.7) 45 (1.1)
806 (83.9) 107 (11.1) 35 (3.6) 13 (1.4)
2058 (87.9) 210 (9.0) 53 (2.3) 20 (0.9)
286 (18.3) 556 (35.5) 133 (8.5) 17 (1.1) 572 (36.6)
628 (15.9) 1303 (32.9) 315 (8.0) 26 (0.7) 1683 (42.6)
276 (28.7) 537 (55.9) 131 (13.6) 17 (1.8) 0 (0.0)
662 (28.3) 1340 (57.2) 308 (13.2) 31 (1.3) 0 (0.0)
25 (1.6) 410 (26.2) 1129 (72.2) 29.0
93 (2.4) 1536 (38.8) 2326 (58.8) 41.0
0 (0.9) 229 (23.8) 723 (75.2) 24.0
34 (1.5) 831 (35.5) 1476 (63.0) 38.0
In the second part of the analysis, we examined CSM after either PN or RN. Because OCM may decrease the number of individuals at risk of CSM, it may artificially increase the CSM rate. Therefore, we relied on competingrisks regression models that adjust for the confounding effect of OCM.12 Our study is the first to rely on competing-risks regression to assess CSM in a large population-based sample of patients with T1a tumors who were treated with either PN or RN from North America. Another French group of investigators used competing-risks regression to assess OCM in other populations.9 Similarly, a multi-institutional collaborative group relied on competing-risks regression models to assess the efficacy of surveillance on OCM in patients with T1a tumors.18 Our study represents the first published report to compare population-based cancer control outcomes between PN and RN in T1a RCC. Since 1990, merely 6 studies from tertiary care centers have addressed this topic.25-29 Of those studies, the largest one relied on 379 PN that were compared with 1075 RN.26 In this study, most of the PN originated from European institutions. The second largest study originated from the Mayo Clinic Foundation and compared 358 PN with 290 RN.29 both studies confirmed the equivalence of CSM-free rates between PN and RN. Finally, investigators from Memorial Sloan-Kettering Cancer Center, University of California in Los Angeles, and from the Mayo Clinic addressed the same topic and respectively relied on 79, 146, and 185 PN cases in their compar886
Population Matched for Age, Tumor Size, Year of Surgery, and Furhman Grade (n ⫽ 3302) PN (%) RN (%)
isons, all of which confirmed the equivalence of cancer control efficacy of PN vs RN.25,27,28 The uniqueness of our findings stems from our patient population, sample size, and statistical methodology. Our patient population does not include any of the North American centers of excellence that previously reported their PN results. In consequence, our findings truly support the equivalence of cancer control between PN vs RN in the community. Moreover, the 5-year CSM rates indicate virtually perfect cancer control after either PN or RN. Only between 2.6% and 2.5% of patients died of RCC respectively after PN or RN. These findings indicate that the natural history of surgically treated small renal cortical tumors is favorable. This suggests that less invasive treatment modalities such as cryoablation, high intensity focused ultrasound, or even active surveillance could be safely applied to a large proportion of patients with T1a lesions.30 Specifically, elderly patients may represent perfect candidates for such treatment modalities as 5-year CSM rates in this subgroup were low (1.0%3.8%). Despite its strengths, limitations do apply to our study. For example, we could not account for a number of potentially important variables. These include institution type, insurance status, surgical volume, and comorbidities.23 These variables should ideally be considered in treatment selection. Moreover, we could not consider American Society of Anesthesiologists scores and baseUROLOGY 76 (4), 2010
Figure 2. Cumulative incidence plots depicting cancer-specific mortality after accounting for other-cause mortality in a population of 5519 patients matched for year of surgery, age, and tumor size (A) and in a population of 3302 patients matched for year of surgery, age, tumor size, and Fuhrman grade (B).
line glomerular filtration rate, which also represent important determinants of nephrectomy treatment.9 Moreover, we had no access to comorbidity information, which may also affect the frequency of PN use as well as CSM rates. Finally, PN rates may be affected by hospital and surgical volume. Despite the presence of some disadvantages related to the use of PN vs RN, the equivalence of cancer control outcomes, combined with renal funcUROLOGY 76 (4), 2010
tion preservation and quality of life benefits, make a strong case for PN for lesions that are technically amenable to such surgery.
CONCLUSIONS Our findings demonstrate the equivalence of PN relative to RN, when cancer control outcomes are considered in 887
patients with T1a RCC. This complements reports from tertiary care centers at which PN was shown to be equally effective to RN. In the light of important overuse of RN for small renal cortical tumors in North America, the current report should be perceived as a confirmation of PN efficacy and it should prompt wider use of PN in the community. References 1. Ljungberg B, Hanbury DC, Kuczyk MA, et al. Renal cell carcinoma guideline. Eur Urol. 2007;51:1502-1510. 2. Patard JJ, Tazi H, Bensalah K, et al. The changing evolution of renal tumours: a single center experience over a two-decade period. Eur Urol. 2004;45:490-493; discussion: 493-494. 3. Zini L, Patard JJ, Capitanio U, et al. The use of partial nephrectomy in European tertiary care centers. Eur J Surg Oncol. 2009;35:636642. 4. Thompson RH, Kaag M, Vickers A, et al. Contemporary use of partial nephrectomy at a tertiary care center in the United States. J Urol. 2009;181:993-997. 5. Hollenbeck BK, Taub DA, Miller DC, et al. National utilization trends of partial nephrectomy for renal cell carcinoma: a case of underutilization? Urology. 2006;67:254-259. 6. Miller DC, Saigal CS, Banerjee M, et al. Diffusion of surgical innovation among patients with kidney cancer. Cancer. 2008;112: 1708-1717. 7. Miller DC, Hollingsworth JM, Hafez KS, et al. Partial nephrectomy for small renal masses: an emerging quality of care concern? J Urol. 2006;175:853-857; discussion: 858. 8. James M, Ofer Y, Michael WK, et al. Partial nephrectomy for renal cortical tumors: pathologic findings and impact on outcome. Urology. 2002;60:1003-1009. 9. Zini L, Patard JJ, Capitanio U, et al. Cancer-specific and noncancer-related mortality rates in European patients with T1a and T1b renal cell carcinoma. BJU Int. 2009;103:894-898. 10. Russo P. Localized renal cell carcinoma. Curr Treat Options Oncol. 2001;2:447-455. 11. Go AS, Chertow GM, Fan D, et al. Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med. 2004;351:1296-1305. 12. Jason PF, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94:496-509. 13. Ries LAG, Melbert D, Krapcho M, et al. SEER Cancer Statistics Review, 1975-2004, National Cancer Institute. Bethesda, MD. Available at: http://seer.cancer.gov/csr/1975-2004/, based on November 2006 SEER data submission, posted to SEER website. Accessed January 2007. 14. Snow DC, Bhayani SB. Rapid communication: chronic renal insufficiency after laparoscopic partial nephrectomy and radical nephrectomy for pathologic t1a lesions. J Endourol. 2008;22:337-341.
888
15. Huang WC, Levey AS, Serio AM, et al. Chronic kidney disease after nephrectomy in patients with renal cortical tumours: a retrospective cohort study. Lancet Oncol. 2006;7:735-740. 16. Lau WK, Blute ML, Weaver AL, et al. Matched comparison of radical nephrectomy vs nephron-sparing surgery in patients with unilateral renal cell carcinoma and a normal contralateral kidney. Mayo Clin Proc. 2000;75:1236-1242. 17. McKiernan J, Yossepowitch O, Kattan MW, et al. Partial nephrectomy for renal cortical tumors: pathologic findings and impact on outcome. Urology. 2002;60:1003-1009. 18. Zini L, Perrotte P, Capitanio U, et al. Radical versus partial nephrectomy. Cancer. 2009;115:1465-1471. 19. O’Connor KM, Davis N, Lennon GM, et al. Can we avoid surgery in elderly patients with renal masses by using the Charlson comorbidity index? BJU Int. 2009;103:1492-1495. 20. Huang WC, Elkin EB, Levey AS, et al. Partial nephrectomy versus radical nephrectomy in patients with small renal tumors—is there a difference in mortality and cardiovascular outcomes? J Urol. 2009;181:55-61; discussion: 61-62. 21. Lesage K, Joniau S, Fransis K, et al. Comparison between open partial and radical nephrectomy for renal tumours: perioperative outcome and health-related quality of life. Eur Urol. 2007;51:614620. 22. Stephenson AJ, Hakimi AA, Snyder ME, et al. Complications of radical and partial nephrectomy in a large contemporary cohort. J Urol. 2004;171:130-134. 23. Porter MP, Lin DW. Trends in renal cancer surgery and patient provider characteristics associated with partial nephrectomy in the United States. Urol Oncol. 2007;25:298-302. 24. Cloutier V, Capitanio U, Zini L, et al. Thirty-day mortality after nephrectomy: clinical implications for informed consent. Eur Urol, in press. 25. Lerner SE, Hawkins CA, Blute ML, et al. Disease outcome in patients with low stage renal cell carcinoma treated with nephron sparing or radical surgery. J Urol. 1996;155:1868-1873. 26. Patard JJ, Shvarts O, Lam JS, et al. Safety and efficacy of partial nephrectomy for all T1 tumors based on an international multicenter experience. J Urol. 2004;171:2181-2185; quiz: 2435. 27. Lee CT, Katz J, Shi W, et al. Surgical management of renal tumors 4 cm. or less in a contemporary cohort. J Urol. 2000;163:730-736. 28. Belldegrun A, Tsui KH, deKernion JB, et al. Efficacy of nephronsparing surgery for renal cell carcinoma: analysis based on the new 1997 tumor-node-metastasis staging system. J Clin Oncol. 1999;17: 2868-2875. 29. Thompson RH, Boorjian SA, Lohse CM, et al. Radical nephrectomy for pT1a renal masses may be associated with decreased overall survival compared with partial nephrectomy. J Urol. 2008; 179:468-471; discussion: 472-473. 30. Mabjeesh NJ, Avidor Y, Matzkin H. Emerging nephron sparing treatments for kidney tumors: a continuum of modalities from energy ablation to laparoscopic partial nephrectomy. J Urol. 2004; 171:553-560.
UROLOGY 76 (4), 2010