Fax: (507) 284-4951
Assessment of the pathologic inclusion criteria from contemporary adjuvant clinical trials for predicting disease progression after nephrectomy for renal cell carcinoma
Article first published online: 3 JAN 2012
Copyright © 2011 American Cancer Society
Volume 118, Issue 18, pages 4412–4420, 15 September 2012
How to Cite
Kim, S. P., Crispen, P. L., Thompson, R. H., Weight, C. J., Boorjian, S. A., Costello, B. A., Lohse, C. M. and Leibovich, B. C. (2012), Assessment of the pathologic inclusion criteria from contemporary adjuvant clinical trials for predicting disease progression after nephrectomy for renal cell carcinoma. Cancer, 118: 4412–4420. doi: 10.1002/cncr.26695
- Issue published online: 5 SEP 2012
- Article first published online: 3 JAN 2012
- Manuscript Accepted: 27 SEP 2011
- Manuscript Revised: 22 SEP 2011
- Manuscript Received: 23 JUN 2011
- renal cell carcinoma;
- disease progression;
- adjuvant trials
The objective of this study was to evaluate the accuracy of the pathologic inclusion criteria from all contemporary adjuvant trials in predicting disease progression (DP) for renal cell carcinoma (RCC).
A retrospective review was conducted on 1363 patients treated surgically for M0 RCC at the Mayo Clinic (Rochester, MN), from 1990 to 2001. Clinicopathologic features were reviewed to determine eligibility for the following trials: ARISER, ASSURE, EVEREST, PROTECT, SORCE, and S-TRAC. DP was defined as local recurrence or distant metastasis after surgery. The ability of each trial's inclusion criteria to accurately predict DP was evaluated by the c (concordance) index.
From the Mayo Clinic cohort, we determined that 41%, 45%, 45%, 33%, 47%, and 23% of the patients would have been eligible for the ARISER, ASSURE, EVEREST, PROTECT, SORCE, and S-TRAC clinical trials, respectively. Overall, 23% of all patients experienced DP (n = 317). Among eligible patients, 53%, 44%, 44%, 57%, 43%, and 59% developed DP during follow-up and 10%, 6%, 6%, 13%, 6%, and 18% went onto DP while not being eligible for the ARISER, ASSURE, EVEREST, PROTECT, SORCE, and S-TRAC trials, respectively. The c index of each trial to accurately predict DP from the pathologic inclusion criteria of ARISER, ASSURE, EVEREST, PROTECT, SORCE, and S-TRAC were 0.751, 0.751, 0.751, 0.742, 0.745, and 0.691, respectively.
Although the pathologic inclusion criteria of contemporary adjuvant trials have notable differences, all 6 adjuvant trials demonstrated high predictive accuracy of DP. Overall, 43% to 59% of patients included for the adjuvant trials would develop DP, whereas 6% to 18% of patients excluded from the trials would develop DP during follow-up. Cancer 2012. © 2012 American Cancer Society.
In 2010, it was estimated that there were 58,240 incident cases and 8210 deaths attributable to renal cell carcinoma (RCC) in the United States.1 Although surgical extirpation remains the gold standard for localized RCC and confers excellent cancer-specific survival at 5 years, a substantial proportion of patients remains at risk for disease progression (DP). With approximately 25% of patients having metastatic disease and another 25% with locally advanced disease at the time of diagnosis,2, 3 10% to 28% of patients will later experience local recurrence or distant metastasis following nephrectomy.4-6
As such, phase 3 randomized trials have been initiated to assess the clinical efficacy of adjuvant therapy for patients considered at intermediate or high risk for DP after partial or radical nephrectomy. Evidence to support the use of adjuvant therapy has been extrapolated from prior randomized trials for metastatic RCC.7-10 To date, 6 clinical trials have been initiated and are currently ongoing in order to evaluate the potential benefits of adjuvant therapy using targeted agents for intermediate-risk and high-risk M0 RCC: ARISER (Adjuvant Rencarex Immunotherapy Trial to Study Efficacy in Nonmetastasised Renal cell carcinoma; ClinicalTrials.gov number NCT00087022), ASSURE (Adjuvant Sorafenib or Sunitinib for Unfavorable Renal Carcinoma; ClinicalTrials.gov number NCT00326898), EVEREST (Everolimus for Renal Cancer Ensuing Surgical Therapy; ClinicalTrials.gov number NCT01120249), PROTECT (A Study to Evaluate Pazopanib as an Adjuvant Treatment for Localized Renal Cell Carcinoma; ClinicalTrials.gov number NCT01235962), SORCE (Sorafenib in Treating Patients at Risk of Relapse After Undergoing Surgery to Remove Kidney Cancer; ClinicalTrials.gov number NCT00492258), and S-TRAC (Sunitinib Treatment of Renal Adjuvant Cancer; ClinicalTrials.gov number NCT00375674).
In designing clinical trials, it is essential to designate inclusion criteria that accurately predict prognosis in order to appropriately identify individuals who are at high risk for DP and exclude those individuals with low-risk disease who may be unnecessarily treated.11 Appropriate designation of inclusion criteria for clinical trials should ideally be based on validated models for risk stratification. However, all 6 of the contemporary adjuvant RCC trials apply different pathologic inclusion criteria to determine eligibility, which may reflect the lack of consensus regarding a universally accepted definition of high-risk RCC.12-18 Furthermore, the accuracy of inclusion criteria for each adjuvant clinical trial in predicting DP has yet to be evaluated. In this context, our objective was to evaluate the accuracy of the pathologic inclusion criteria from all ongoing adjuvant trials in predicting DP following surgery for RCC.
MATERIAL AND METHODS
Upon approval from the Institutional Review Board, we identified 1363 patients treated with radical nephrectomy or nephron-sparing surgery for unilateral, sporadic, M0 RCC between 1990 and 2001 from the Mayo Clinic Nephrectomy Registry.
Clinical and Pathologic Features
Clinical features evaluated in our study included age, sex, and Eastern Cooperative Oncology Group performance status. Pathologic features studied included histologic subtype classified according to the Union Internationale Contre le Cancer, American Joint Committee on Cancer, and Heidelberg guidelines; tumor size; the 2002 primary tumor and regional lymph node classifications; nuclear grade; coagulative tumor necrosis; and sarcomatoid differentiation.19
Combinations of these pathologic features were used to identify patients from our cohort who would be eligible to participate based on the pathologic inclusion criteria from each of the following phase 3 randomized adjuvant trials: ARISER, ASSURE, EVEREST, PROTECT, SORCE, and S-TRAC (Table 1). In ARISER, which has completed accrual, this 2-arm trial allocated patients in a 1:1 randomization to receive intravenous girentuximib, a monoclonal antibody, or placebo. The inclusion trial selected only patients with clear cell RCC histologic subtype. In addition, patients were considered eligible and high-risk if the renal tumors had pT1b/pT2, grade 3 to 4 disease; pT3/pT4 disease; or pN1/pN2 disease. The primary outcomes for this study are disease-free survival (DFS) and overall survival (OS).
|Trial||Projected Accrual||Start Date–Estimated Primary Completion Date||Pathologic Inclusion Criteria||Treatment Arms||Outcome|
|ARISER||864||July 2004–September 2013||Clear cell RCC histology only||Intravenous girentuximab weekly for 24 wk||Primary: DFS, OS|
|pT1a-T2, Fuhrman grade ≥3|
|pT3-T4, Fuhrman grade any, N0/Nx||Placebo||Secondary: QOL, safety of therapy, and pharmacokinetics|
|ASSURE||1923||May 2006–April 2016||Clear cell or non–clear cell RCCa||Sunitinib daily for 4 wk followed by 2 wk of rest for a total of 9 courses||Primary: DFS|
|pT1b, Fuhrman grade 3 or 4, N0||Sorafenib daily for 4 wk followed by 2 wk of rest for a total of 9 courses||Secondary: OS, QOL toxicity of therapy, cardiac function, and fatigue|
|pT2, Fuhrman grade any, N0|
|pT3a, Fuhrman grade 1 or 2, N0||Placebo|
|pT3a, Fuhrman grade 3 or 4, N0|
|pT3b-c, Fuhrman grade any, N0|
|pT4, Fuhrman grade any, N0|
|EVEREST||1218||April 2011–August 2013||Clear cell or non–clear cell RCCa||Everolimus daily for 6 wk for a total 9 courses||Primary: Relapse-free survival|
|pT1b, Fuhrman grade 3 or 4, N0||Placebo||Secondary: OS, Toxicity|
|pT2, Fuhrman grade any, N0|
|pT3a, Fuhrman grade 1 or 2, N0|
|pT3a, Fuhrman grade 3 or 4, N0|
|pT3b-c, Fuhrman grade any, N0|
|pT4, Fuhrman grade any, N0|
|PROTECT||1500||November 2010–October 2015||Clear cell or predominant clear cell RCC||Pazopanib daily for 12 mo||Primary: DFS|
|pT2, Fuhrman grade 3 or 4, N0|
|pT3, Fuhrman grade any, N0||Placebo||Secondary: OS, safety, and QOL|
|pT4, Fuhrman grade any, N0|
|pTany, Fuhrman grade any, N1|
|SORCE||1656||June 2007–August 2012||All RCC variants||Sorafenib twice daily for 1 y followed by placebo for 2 y||Primary: DFS|
|Mayo Clinic progression score 3-11|
|Sorafenib twice daily for 3 y||Secondary: Metastasis-free survival, OS,|
|Placebo for 3 y||cost-effectiveness, toxicity|
|S-TRAC||720||July 2007–June 2017||Predominant clear cell RCC||Sunitinib daily for 4 wk followed by 2 wk of rest for 2 y||Primary: DFS|
|Modified UCLA Integrated Staging System high-risk|
|Placebo||Secondary: OS, safety, and tolerability|
The ASSURE study, which also has completed accrual, is the only 3-arm, 2-agent clinical trial that randomly allocated patients to sorafenib, sunitinib, or placebo. The inclusion criteria for the ASSURE trial defined eligibility for patients with a histologic diagnosis of clear cell, papillary, or chromophobe RCC; and those with pT1b, grade 3 or 4 disease; pT2/pT2/pT3/pT4, grade 1 through 4 disease; or those with pN1/pN2 disease. The primary outcome for the ASSURE trial is DFS after partial or radical nephrectomy. The EVEREST study was recently opened as a phase 3 clinical trial and is accruing patients who will be allocated in a 1:1 randomization to everolimus or placebo after surgery. The pathologic inclusion criteria and primary outcome for the EVEREST trial are identical to those in the ASSURE trial.
In the PROTECT study, this phase 3 clinical trial is currently accruing patients after nephrectomy for localized or locally advanced RCC. Patients are randomly assigned to receive either pazopanib or placebo in a 1:1 allocation. However, the pathologic criteria for PROTECT uses 2010 TNM classification and considers patients eligible to participate if there is clear cell RCC with pT2, grade 3/4, pNX/N0 disease; pT3/4, any grade, pNX/N0 disease; or pN1/pN2 disease. The primary outcome for PROTECT is DFS.
SORCE is an ongoing phase 3 clinical trial that is currently recruiting patients to be enrolled into 1 of 3 arms, where patients will be randomly assigned sorafenib for 3 years, sorafenib for 1 year, or placebo. SORCE is the only study to use a validated stratification tool to calculate the Mayo Clinic progression score to determine eligibility for patients who have undergone nephrectomy.13 Patients are considered eligible for enrollment with a Mayo Clinic progression score of 3 to 11, regardless of histologic subtype. This score was calculated as +2 if pT1b; +3 if pT2; +4 if pT3a, pT3b, pT3c, or pT4; +2 if pN1 or pN2; +1 if tumor size ≥10 cm; +1 if grade 3; +3 if grade 4; +1 if coagulative tumor necrosis was present; and 0 otherwise.
In S-TRAC, this phase 3 clinical trial is also currently accruing patients who have undergone nephrectomy for high-risk RCC who will be randomly assigned to sunitinib or placebo with 1:1 allocation. Patients who have clear cell RCC and pT3/pT4, or those with pN1/pN2 disease will be considered eligible. The primary outcome for S-TRAC is DFS with an estimated time frame to completion of 36 months.
To explore the effects of limiting the pathologic inclusion criteria to the strict definitions of high-risk RCC, we then used the University of California Los Angeles Integrated Staging System (UISS) and the Mayo Clinic progression score to define high-risk RCC as the pathologic inclusion criteria for clinical trial eligibility. As such, we used a cumulative UISS score of 3 or more to define high risk for eligibility.15, 20 According to the Mayo Clinic progression score, we used the previous definition based on a cumulative progression score of 6 or more to define high risk for only the clear cell RCC subtype.13
Disease status for patients in the Nephrectomy Registry is updated annually. If a patient has not been seen at our institution in the previous year, the patient is sent a disease status questionnaire. If there is evidence of DP (ie, local recurrence or distant metastases) in this questionnaire, the date, location, and treatment are verified in writing with the patient's local physician.
Progression-free survival (PFS), which was defined as survival free of local recurrence or distant metastases, was estimated using the Kaplan-Meier method. The ability of each trial's inclusion criteria to predict DP was evaluated using the c (for concordance) index for the Cox model proposed by Harrell et al.21 The c index corresponds to the proportion of all usable pairs of patients in which the observed and predicted survival times are concordant. The interpretation of the c index is identical to the interpretation of the area under a receiver operating characteristic curve. A c index of 1.0 indicates that the features in the model perfectly separate patients with different outcomes, whereas a value of 0.5 indicates that the features contain prognostic information equal to that obtained by chance alone.
The clinical and pathologic features of the 1363 patients with RCC in our analytic cohort are summarized in Table 2. Median age at surgery was 65 years (range, 21-89 years) and median tumor size was 5.0 cm (range, 0.2-24.0 cm). Overall, 317 (23.3%) patients in our analytic cohort experienced DP at a median of 1.6 years after nephrectomy (range, 0.1-16.2 years). Among those patients with DP, 260 patients experienced distant metastases only, 19 patients had local recurrence only, and 38 patients experienced both local recurrence and distant metastases. At last follow-up, 629 patients had died, including 259 patients who died from RCC at a median of 2.7 years after surgery (range, 0.0-16.5 years). Among the 734 patients who were still alive at last follow-up, the median duration of follow-up was 9.4 years (range, 0.1-18.4 years). Overall, PFS rates (95% confidence interval [CI]; number still at risk) at 1, 5, and 10 years after surgery were 91.1% (89.6-92.6; 1175), 81.2% (79.0-83.3; 914), and 74.4% (71.9-77.1; 400), respectively.
|Age at surgery, y|
|ECOG Performance Status|
|Primary tumor size, cm|
|5 to <7||274 (20.1)|
|7 to <10||249 (18.3)|
|2002 Primary tumor classification|
|2002 Regional lymph node classification|
|pNX and pN0||1330 (97.6)|
|pN1 and pN2||33 (2.4)|
|Clear cell||1036 (76.0)|
|Collecting duct||3 (0.2)|
|RCC, not otherwise specified||12 (0.9)|
|Coagulative tumor necrosis||368 (27.0)|
|Sarcomatoid differentiation||39 (2.9)|
Although 638 (46.8%) patients were eligible for the SORCE trial, 428 (31.4%) of these eligible patients had intermediate-risk RCC according to their Mayo Clinic progression scores. Among 1036 patients with clear cell RCC histologic subtype, 425 (41.0%), 338 (32.6%), and 239 (23.1%) patients were eligible for the ARISER, PROTECT, and S-TRAC trials, respectively. For the ASSURE and EVEREST trials, 603 (44.7%) patients would have been considered eligible from the 1348 patients with clear cell, papillary, or chromophobe RCC.
All patients eligible for enrollment into the ARISER, ASSURE, EVEREST, PROTECT, and S-TRAC trials were also eligible to participate in the SORCE trial (Table 3). However, only 8 patients with clear cell RCC and intermediate Mayo Clinic progression score were eligible to participate in the SORCE trial but no other trials. Of these patients with intermediate-risk RCC, 2 patients had pT1a, pNX/pN0, grade 4, necrotic clear cell RCC tumors whereas the remaining 6 patients had pT1b, pNX/pN0, grade 1/2, necrotic clear cell RCC tumors, yielding a progression score of 3 or 4. Seventy-three patients with clear cell RCC were eligible for the ASSURE, EVEREST, and SORCE trials, but not the ARISER, S-TRAC, and PROTECT trials. All of these patients had pT2, pNX/pN0, and grade 1/2 tumors.
DP varied according to the pathologic inclusion criteria eligibility for each clinical trial (Table 4). All trials exhibited high rates of DP among eligible patients treated with nephrectomy for localized RCC in our analytic cohort. For example, among eligible patients from each clinical trial, DP rates were 52.5%, 44.3%, 44.3%, 57.1%, 43.1%, and 59.0% for the ARISER, ASSURE, EVEREST, PROTECT, SORCE, and S-TRAC, respectively. As a result, the c index of each clinical trial in predicting DP was similar and relatively accurate for the ARISER (0.751; 95% CI = 0.730-0.772), ASSURE and EVEREST (0.751; 95% CI = 0.732-0.770), PROTECT (0.742; 95% CI = 0.718-0.766), SORCE (0.745; 95% CI = 0.727-0.764), and S-TRAC (0.691; 95% CI = 0.664-0.719) trials. However, the percentages of patients who met the inclusion criteria for the ARISER, ASSURE, EVEREST, PROTECT, SORCE, and S-TRAC trials but who did not experience DP varied at 47.5%, 55.7%, 55.7%, 42.9%, 56.9%, and 41.0%, respectively. Conversely, the percentages of patients who failed to meet the inclusion criteria for the ARISER, ASSURE, EVEREST, PROTECT, SORCE, and S-TRAC trials but did experience DP were 9.8%, 6.2%, 6.2%, 12.9%, 5.8%, and 17.8%, respectively.
|Clinical Trial||Progressed n/N (%)||Median Years to Progression||5-y PFS Rate (95% CI)|
|Eligible for ARISER (N = 1036)|
|No||60/611 (9.8)||5.9||95.9 (94.3-97.6)|
|Yes||223/425 (52.5)||1.2||52.3 (47.6-57.5)|
|Eligible for ASSURE (N = 1348)|
|No||46/745 (6.2)||6.1||97.5 (96.3-98.7)|
|Yes||267/603 (44.3)||1.3||61.2 (57.3-65.4)|
|Eligible for EVEREST (N = 1348)|
|No||46/745 (6.2)||6.1||97.5 (96.3-98.7)|
|Yes||267/603 (44.3)||1.3||61.2 (57.3-65.4)|
|Eligible for PROTECT (N = 1036)|
|No||90/698 (12.9)||5.0||93.0 (91.1-95.0)|
|Yes||193/338 (57.1)||1.1||46.9 (41.7-52.8)|
|Eligible for SORCE (N = 1363)|
|No||42/725 (5.8)||6.1||97.7 (96.6-98.9)|
|Yes||275/638 (43.1)||1.3||62.3 (58.5-66.3)|
|Eligible for S-TRAC (N = 1036)|
|No||142/797 (17.8)||3.5||88.3 (86.0-90.6)|
|Yes||141/239 (59.0)||1.0||43.0 (36.9-50.2)|
|High-risk Mayo Clinic PS ≥6 (N = 1036)|
|No||156/854 (18.3)||4.0||88.2 (86.0-90.5)|
|Yes||127/182 (69.8)||0.8||29.6 (23.4-37.6)|
|High-Risk UISS score ≥3 (N = 1363)|
|No||286/1321 (21.7)||2.1||82.9 (80.8-85.1)|
|Yes||31/42 (73.8)||0.4||24.2 (13.7-42.8)|
All 6 trials demonstrated similar PFS for eligible and ineligible patients, based on the pathologic inclusion criteria for each of the randomized adjuvant trials, with exception of the S-TRAC trial. In Figure 1, the 5-year PFS rate (95% CI; number still at risk) for the 425 patients eligible for the ARISER trial was 52.3% (47.6%-57.5%; 180). Of the 223 patients eligible for the ARISER trial who progressed, 190 progressed within 5 years of surgery (85.2%). In the ASSURE and EVEREST trials (Fig. 2), the 5-year PFS rate (95% CI; number still at risk) for the 603 patients eligible was 61.2% (57.3%-65.4%; 302). Of the 267 patients eligible for the ASSURE and EVEREST trials who progressed, 219 progressed within 5 years of surgery (82.0%). For the PROTECT trial (Fig. 3), the 5-year PFS rate (95% CI; number still at risk) for the 338 patients eligible was 46.9% (41.7%-52.8%; 130). Of the 193 patients eligible for the PROTECT trial who progressed, 168 patients progressed within 5 years of surgery (87.1%). In the SORCE trial (Fig. 4), the 5-year PFS rate (95% CI; number still at risk) for the 638 patients eligible was 62.3% (58.5%-66.3%; 326). Of the 275 patients eligible for the SORCE trial who progressed, 225 progressed within 5 years of surgery (81.8%). For the S-TRAC trial (Fig. 5), the 5-year PFS rate (95% CI; number still at risk) for the 239 patients eligible was moderately lower at 43.0% (36.9%-50.2%; 83). Of the 141 patients eligible for the S-TRAC trial who progressed, 126 patients progressed within 5 years of surgery (89.4%).
To ascertain variations in DP if a clinical trial adopted more stringent inclusion criteria for strictly defining high risk to determine eligibility, a UISS score ≥ 3 and a progression score ≥ 6 yielded higher rates of eligible individuals who experienced DP with a shorter median time to DP. For example, in the UISS high-risk group for all histologic subtypes, 42 patients (3.1%) would be eligible from the RCC cohort with a median time to DP of 0.4 years (Table 4). Moreover, adoption of a strict definition of the high-risk UISS classification increased DP to 73.8% (n = 31) among all eligible patients. In clear cell RCC with a progression score ≥ 6, 182 (17.6%) patients would be eligible with a median time to progression of 0.8 years. Use of the standardized definition of high risk from the Mayo Clinic progression score increased the DP to 69.8% (n = 127) among all eligible patients. The percentages of patients who met the inclusion criteria for the high-risk stratification by the UISS or the Mayo Clinic progression score, but who did not experience DP, also decreased to 26.2% and 30.2%, respectively. However, the percentages of patients who did not meet the inclusion criteria for high-risk stratification by UISS or the Mayo Clinic progression score, but who experienced DP, rose to 21.7% and 18.3%, respectively.
In this study, we report the effects of differing pathologic inclusion criteria from contemporary adjuvant clinical trials in accurately predicting DP. Our study found that the differing pathologic inclusion criteria from the ongoing adjuvant trials had similar and relatively strong predictive accuracy of DP. Our results also suggest that the inclusion of patients with intermediate-risk stratification by the UISS score or the Mayo Clinic progression score may possibly lead to overtreatment. For example, approximately half of the patients enrolled would have not progressed and may be unnecessarily exposed to adverse events associated with targeted therapy in our large retrospective cohort. Overtreatment of such individuals eligible for enrollment by current criteria, but without DP, could be decreased to 26% to 30% if the pathologic inclusion criteria became more stringent by selecting individuals with only high-risk stratification, as defined by the Mayo Clinic progression score or UISS classification. Moreover, median time to progression for those eligible with high-risk stratification by UISS or the Mayo Clinic progression score would be decreased to less than 1 year in both cases, resulting in shorter duration of follow-up needed to detect differences in progression and survival. Hence, use of a strict pathologic inclusion criteria from UISS and the Mayo Clinic progression score, thereby excluding intermediate-risk RCC, would take a shorter time period to accumulate the necessary events of disease recurrence and, as a result, report the results of each trial sooner in order to possibly change clinical practice or inform the designs of future clinical trials.
Designation of inclusion criteria for randomized clinical trials ought to require a transparent process that uses the best existing predictive models to appropriately risk-stratify individuals who will most likely benefit from treatment, while minimizing the enrollment of individuals with low risk of DP to avoid overtreatment. By doing so, the inclusion criteria can be standardized to also allow for evaluation of the comparative effectiveness of efficacy and side effects for each targeted agent. Only the SORCE trial used a previously validated stratification tool specifically designed for prospective clinical trials to better predict distant metastasis from clear cell RCC after radical nephrectomy.13 For the ARISER, ASSURE, and EVEREST trials, the inclusion criteria were similar by using a modified UISS.15, 20
Given these differences in determining patient eligibility for each trial, it may be also difficult to assess the comparative effectiveness from the clinical trials in order to determine a superior targeted agent in reducing the risk of disease relapse for several reasons. Each of the clinical trials is designed to initially determine whether adjuvant therapy with targeted agents for intermediate- or high-risk RCC after nephrectomy reduces the risk of recurrence, by randomly assigning patients to either the targeted agent or placebo. At present, most of the ongoing phase 3 adjuvant trials are not designed to evaluate which agent is the most effective in reducing the risk of disease relapse after surgery, because these trials are not performing head-to-head clinical trials or using an adaptive design to determine superiority, noninferiority, or futility.22, 23 Only the results from the ASSURE trial will allow for assessing the relative effectiveness of sunitinib versus sorafenib in reducing the risk of DP. Moreover, until all the trials are complete, it is possible the clinical characteristics of the patients enrolled in each adjuvant trial may be vastly different, such that one would be unable to assess the relative effectiveness of the different adjuvant agents from the clinical trials in this study. Taken together, unless all trials either use a uniform pathologic inclusion criteria or an adaptive clinical trial design, or perform a head-to-head trial, which is premature until efficacy is initially demonstrated, it will be difficult for patients and providers to make fully informed decisions about which agent and the dose is most effective in reducing the risk of recurrence and minimizes treatment-related adverse events. Another important consideration is that patients who are eligible but without DP may be unnecessarily exposed to adjuvant therapy with limited clinical efficacy, but potentially significant adverse events.
There are several limitations about the inferences from our study, regarding the accuracy of each clinical trial's pathologic inclusion criteria. First, our study is subject to the inherent biases from a large, single institution and retrospective nephrectomy registry. Moreover, characteristics of the patient population from our institution may ultimately differ from those individuals enrolled in the multicenter phase 3 randomized trials. Second, our study also did not evaluate the possible benefits derived from each adjuvant agent in our cohort eligible for enrollment in each phase 3 trial. Third, adoption of more stringent pathologic inclusion criteria also resulted in a moderately increased rate of patients who are not eligible that will experience DP, if a strict definition of high risk, as originally described, was adopted from either the UISS or the Mayo Clinic classification systems. Fourth, we also acknowledge that each of the randomized trials' existing pathologic inclusion criteria had relatively high DP rates (>40%) following nephrectomy, such that adjuvant therapy would be considered an acceptable aspect of the oncologic treatment. Furthermore, an implicit message of our study assumes that patients found to have high-risk RCC will achieve greater efficacy from adjuvant agents than those with intermediate-risk disease. Indeed, a subgroup analysis by risk stratification from the pathologic inclusion criteria from these adjuvant clinical trials may in fact demonstrate that intermediate-risk patients will achieve equal efficacy compared to high-risk patients, and that the competing risk of disease recurrence may be greater in the placebo arm of intermediate-risk patients with RCC.
In summary, our study demonstrates that each of the ongoing adjuvant trial's pathologic inclusion criteria had similar and accurate abilities in predicting DP following surgery. However, there may be a substantial number of eligible patients who are not destined for DP. Moreover, inclusion of intermediate-risk patients may require longer follow-up time to detect differences in DP. Further research is needed to implement a more uniform risk stratification of high-risk RCC in the development of inclusion criteria for randomized trials for adjuvant therapy for patients who have undergone nephrectomy.
Funding provided by the Department of Urology, Mayo Clinic, and Healthcare Delivery Research Scholars Program, Mayo Clinic.
CONFLICT OF INTEREST DISCLOSURE
The authors made no disclosure.
- 2The medical treatment of metastatic renal cell cancer. EAU Update Ser. 2003; 1: 237-246..