Sunitinib has replaced interferon (IFN) as a first-line standard of care in the treatment of metastatic renal cell carcinoma (RCC). This study aimed to determine overall survival and to confirm effectiveness in a population that includes poor prognosis patients.
Data were collected on all patients identified by the BC Cancer Registry with metastatic RCC who were treated with IFN or sunitinib. The IFN group consisted of patients who received IFN between January 2000 and October 2005, and the sunitinib group included patients treated with first-line sunitinib from October 2005 to September 2007.
There were 131 and 69 patients in the IFN and sunitinib groups, respectively. The median follow-up of those still alive was 12.6 months. The median age (62 vs 63 years; P = .41), Memorial Sloan Kettering Cancer Center (MSKCC) prognostic criteria (poor in 19% vs 30%; P = .41), and proportion with >1 metastasis (53% vs 62%; P = .21) were similar between the IFN and sunitinib groups, respectively. The median survival of the IFN and sunitinib groups was 8.7 and 17.3 months, respectively (log-rank P = .004). The median survival of patients with favorable, intermediate, and poor MSKCC prognostic profiles in the IFN group was 22.9, 8.7, and 4.1 months, respectively (P < .001), whereas in the sunitinib group it was not reached, 16.8, and 10.7 months, respectively (P = .006). The hazard ratio of death after adjusting for MSKCC criteria was 0.49 (95% confidence interval, 0.31-0.76; P = .001).
Metastatic renal cell carcinoma (RCC) portends a poor prognosis and was estimated to cause 12,890 deaths in the United States in 2007.1 Previously, immunotherapy agents such as interleukin-2 and interferon (IFN) alpha were the only treatments available and demonstrated low response rates of approximately 15% with minimal to no impact on overall survival.2-6 Because the biology underlying RCC has been elucidated, agents targeting relevant biologic pathways have been investigated.7 Antiangiogenic agents are rapidly changing the therapeutic landscape in metastatic RCC. Sunitinib, sorafenib, bevacizumab in combination with IFN, temsirolimus, and everolimus have thus far demonstrated significant activity within randomized phase 3 trials.8-10, 18, 21
Sunitinib is an oral tyrosine kinase inhibitor with activity against vascular endothelial growth factor (VEGF) receptor types 1, 2, 3, platelet-derived growth factor receptors alpha and beta, c-kit, and FLT-3. In the landmark phase 3 trial examining treatment-naive patients with metastatic RCC, there was a statistically significant difference in progression-free survival (PFS) in patients treated with sunitinib versus IFN (11 vs 5 months), with a hazard ratio of 0.42 (P < .001).8 These results served as the basis for introducing sunitinib as a reference first-line standard of care.
When a significant PFS benefit in an interim analysis of the pivotal phase 3 sunitinib trial8 was detected, patients randomized to IFN were allowed to cross over to sunitinib at progression or be treated with other anti-VEGF therapy. Overall survival data for sunitinib was presented recently showing an impressive difference in survival compared with IFN (26.4 vs 21.8 months) but with a log-rank P value of .051 (Wilcoxon P value .0128).13 A post hoc analysis indicated that the true overall survival difference between sunitinib- and IFN-treated groups was diluted because a substantial number of patients in the standard IFN treatment arm received either sunitinib or another anti-VEGF targeted therapy after progression. Other phase 3 trials comparing targeted therapies such as sorafenib and bevacizumab with IFN have not yet demonstrated an overall survival benefit for potentially the same reason, despite impressive differences in PFS.9, 10
It is important to note that only 6% to 7% of patients enrolled in the randomized phase 3 sunitinib trial had poor Memorial Sloan Kettering Cancer Center (MSKCC) prognostic profiles. This reflects the eligibility criteria and selection bias in randomized controlled trials, in that patients enrolled in prospective trials are relatively healthier and may have better outcomes than patients in the general population.14
The question of generalizability of the initially reported results with a PFS benefit only8 and the high drug costs led to extensive discussions between experts and funding institutions about the true survival benefit of these drugs and the value of sunitinib in a general patient population including poor prognosis patients. This has become a major obstacle for obtaining funding for these expensive drugs from healthcare insurers or funding institutions in various countries around the world.
This population-based retrospective analysis aims to provide important information about the impact of a new treatment strategy on the outcomes of RCC in British Columbia. It examines the effect of sunitinib on overall survival when used in the general, unselected population, reflecting “real world” clinical practices. All patients in the province of British Columbia's population of 4.3 million people are reflected in this analysis, including patients treated at academic and community oncology centers.
MATERIALS AND METHODS
The BC Cancer Agency coordinates all cancer care for the entire province of British Columbia, Canada. This has been previously described in other population-based studies performed at this institution.15 Most patients receive care from a medical oncologist at 1 of the 4 major cancer treatment centers: the Vancouver Cancer Center, Vancouver Island Cancer Center, Fraser Valley Cancer Center, and Cancer Center for the Southern Interior. The remaining patients receive care from local community oncologists throughout the province. The BC Cancer Registry records all new cases of cancer diagnosed in the province of British Columbia. Pathologists are legally required to report all new cancer cases to the registry. This registry allows independent verification of all cancers seen in the province, including RCC. The BC Cancer Agency Provincial Pharmacy Database maintains electronic records of all drugs dispensed. The Registry and Pharmacy database can be linked electronically to match patient diagnosis, treatment, and outcome data.
This is a population-based retrospective analysis examining outcomes for all patients with metastatic RCC in British Columbia, Canada during a 7-year interval: January 2000 to September 2007. This represents a 5-year period before (presunitinib) and 2-year period after (postsunitinib) the institution of a new standard treatment for patients with RCC. A provincial policy switched standard therapy for metastatic RCC from IFN to sunitinib in October 2005, when sunitinib became available for patients in the province of British Columbia. This study was approved by the British Columbia Cancer Agency Research Ethics Board.
The BC Cancer Registry was cross-referenced with the Provincial Pharmacy Database to identify all patients with metastatic RCC who were treated with IFN and/or sunitinib in the province of British Columbia between January 2000 and September 2007. The IFN group consisted of all patients who received IFN alone between January 2000 and October 2005. The sunitinib group included all patients treated with first-line sunitinib from October 2005 to September 2007, when sunitinib became available as standard therapy.
Inclusion and Exclusion Criteria
Patients included in this study had biopsy-proven renal cell cancer with evidence of metastases. Patients who were too frail or unsuitable for treatment were not included in this analysis. Patients who received first-line IFN followed by second-line sunitinib were excluded to preserve the validity of strictly comparing the IFN and sunitinib groups.
Before October 2005, patients with metastatic RCC were treated with IFN 9,000,000 international units (IU) 3 times per week on nonconsecutive days. After October 2005, patients were treated with sunitinib at a starting dose of 50 mg daily for 4 weeks followed by a 2-week break each cycle. Doses were adjusted according to toxicity, with dose reductions to 37.5 mg and then 25 mg if needed. These treatment regimens are virtually the same regimens for IFN and sunitinib as used in the pivotal randomized phase 3 trial.8
Treatment of metastatic RCC has been uniform across British Columbia. All physicians are expected to follow the BC Cancer Agency treatment guidelines, which are explicit documents outlining recommended therapy with drug dosages, management of treatment complications, and follow-up recommendations. The aim of these guidelines is to promote a uniform standard of care according to current evidence-based medicine. Changes in treatment policy are decided by a multidisciplinary tumor group and are communicated directly to all physicians that treat patients with kidney cancer via immediate e-mail notification and by updating the web-based cancer management guidelines of the BC Cancer Agency (http://www.bccancer.bc.ca/HPI/CancerManagementGuidelines/default.htm).
Data Collection and Statistical Analysis
Baseline characteristics including those that have previously been shown to be prognostic (Karnofsky performance status [KPS], number of sites of metastases, diagnosis to treatment interval >1 year, baseline lactate dehydrogenase [LDH], hemoglobin, corrected calcium, neutrophil count, platelet count)16, 17 were collected. A subgroup analysis according to MSKCC prognostic profiles (1 point each for diagnosis to treatment interval <1 year, KPS <80, low hemoglobin, elevated serum calcium, elevated LDH 1.5 times the upper limit of normal; 0 points constituting the favorable prognosis group, 1 or 2 points the intermediate prognosis group, and 3 or more points the poor prognosis group17) was performed. Overall survival was calculated from date of first-line treatment initiation (either IFN or sunitinib) until the date of death or last follow-up.
Statistical analysis was performed to compare clinical characteristics and differences in overall survival in first-line IFN versus first-line sunitinib patients. Two-sample t test for continuous variables and chi-square tests for categoric variables were used. For the known prognostic factors, P values were calculated to determine whether the 2 treatment groups were equally balanced to better elucidate the treatment effect. A multivariate analysis was performed using a Cox proportional hazards model to assess the independent effect of treatment on overall survival after controlling for relevant prognostic covariates. Data were analyzed using SAS 9.1 statistical software (Cary, NC).
A total of 200 eligible patients were identified from the BC Cancer Registry and the Provincial Pharmacy Database. A total of 131 patients were treated with IFN first-line, and 69 patients were treated with sunitinib first-line. The number of patients dead at the time of analysis was 144 (72%). The median follow-up for all patients treated with IFN was 8.4 months (range, 0.9-88.6), and 9.9 months (range, 1.9-33.4) for those treated with sunitinib. The median follow-up for patients still alive at the time of analysis was 12.6 months.
The median age (62 vs 63 years; P = .41) and proportion of patients with >1 metastasis (53% vs 62%; P = .21) were similar between the IFN and sunitinib groups, respectively (Table 1). The MSKCC prognostic profiles distributed between the IFN and sunitinib groups were not statistically different (P = .41), although there was a greater percentage of patients with poor MSKCC prognostic profiles in the sunitinib-treated group. The vast majority of patients had clear cell histology.
Table 1. Patient Characteristics
Group A: IFN Alone, n=131
Group B: Sunitinib, n=69
IFN indicates interferon; Tx, treatment; LDH, lactate dehydrogenase; ULN, upper limit of normal; Dx, diagnosis; MSKCC, Memorial Sloan Kettering Cancer Center.
The median survival of patients treated with IFN versus sunitinib was 8.7 and 17.3 months, respectively (log-rank P = .004) (Fig. 1). The 1-year overall survival rate for the IFN- and sunitinib-treated groups was 38.7% and 59.9%, respectively. When adjusted for MSKCC prognosis profile, the hazard ratio of death for those patients treated with sunitinib versus IFN was 0.49 (95% confidence interval, 0.31-0.76; P = .001).
When examining patients with favorable, intermediate, and poor MSKCC prognostic categories, the median survival of patients receiving IFN was 22.9, 8.7, and 4.1 months, respectively (P < .001), whereas in patients receiving sunitinib it was not reached (NR), 16.8, and 10.7 months, respectively (P = .006).
In subgroup analyses of patients in different MSKCC prognostic profiles, there was a statistically significant difference in overall survival between the IFN- and sunitinib-treated groups in patients with either intermediate (median 8.7 vs 16.8 months; log-rank P = .0487) or poor (median 4.1 vs 10.7 months; log-rank P = .0329) prognostic profiles. Because of small patient numbers and extended survival, no difference in overall survival in the favorable prognostic subgroup was detected (median 22.9 months vs NR; log-rank P = .0589). A significant difference may be detected in this subgroup with further follow-up.
The treatment of metastatic RCC has been revolutionized by VEGF-targeted therapy. Figlin et al13 recently reported the final results of the pivotal randomized phase 3 study comparing IFN with sunitinib. With a median survival of more than 2 years in this trial, sunitinib has now been widely accepted as a standard of care for first-line therapy for patients with metastatic RCC.
Our current study is a population-based analysis designed to evaluate the impact of the introduction of this combination therapy across the general population of patients with metastatic RCC. As many of these costly new therapies emerge, it will become increasingly important to demonstrate their effectiveness and generalizability in the general population to justify societal costs. These issues are major obstacles for many government bodies in Canada and some European countries to provide funding for these drugs.18, 19 It is therefore of utmost importance to clarify the impact of sunitinib on overall survival in an unselected, “real world” population of RCC patients.
Our population-based study confirms the profound impact of the introduction of sunitinib on overall survival when administered to an unselected population of patients with metastatic RCC in routine clinical practice. Patients treated in the sunitinib era had a median survival of 17.3 months compared with 8.7 months in the IFN era (log-rank P = .004). This is an important verification that the randomized controlled trial data can be extended to a general population of patients, including many who did not participate in clinical trials. Very similar results have recently been reported by another Canadian province.20, 21
Importantly, this survival benefit was extended to the subset of patients with poor MSKCC prognostic profiles17 (4.1 vs 10.7 months; P = .0329). In comparison, in the phase 3 trial of sunitinib versus IFN, the subgroup with poor MSKCC prognostic profiles had a trend toward benefit in PFS, but the confidence intervals crossed, likely because of limited patient numbers.8 The only phase 3 clinical trial to study patients with poor prognostic criteria compared IFN, temsirolimus, and both combined.11 The median overall survival of patients treated with IFN versus temsirolimus was 7.3 and 10.9 months, respectively (hazard ratio 0.73; P = .008). These overall survival times are very similar to those demonstrated in our poor prognosis subset of patients treated with sunitinib versus IFN. In the expanded access study of sunitinib,22, 23 9.8% of patients had MSKCC poor risk criteria, and 13.2% had an Eastern Cooperative Oncology Group score >1. First-line sunitinib patients had a median PFS of 8.9 months, which is 2 months less than the PFS of the aforementioned phase 3 trial. Again, it is demonstrated that patient survival outside of a clinical trial could potentially be decreased by the inclusion of real-world patients who may have poorer risk profiles. Overall survival results are pending for the expanded access study, which would then enable us to directly compare our data.
The strength of these data are that they are population-based and reflect “real-world” practice. The population was derived from databases that capture all patients with a diagnosis of RCC and their treatment in a province with uniform treatment guidelines that are strictly adhered to for reimbursement purposes. This allows data collection to be consistent in the electronic medical records and ensures a level of standardized care between sunitinib- and IFN-treated groups.
The analysis of clinical factors demonstrated no statistically significant differences between the 2 patient groups, although there were slightly more patients in the sunitinib-treated group with poor MSKCC prognostic profiles, >1 site of metastasis, anemia, and hypercalcemia. All of these factors may have slightly biased the results against sunitinib in terms of overall survival.
The IFN regimen used in British Columbia was considered a standard regimen at the time and was virtually the same regimen used in the randomized phase 3 study. It is important to note that the overall survival observed in both our IFN and sunitinib groups was somewhat inferior to those in previously reported clinical trials, ranging from 13 to 14 months2-5 and from 23.9 to 26.4 months,13, 24 respectively. The patients in the randomized trials are generally well selected, with good performance status compared with the patients in this study. Compared with the patients in the pivotal phase 3 trial,8 the patients of our study were slightly older (62 vs 60 years), had substantially fewer nephrectomies (68% vs 90%), and had more poor risk patients (23% vs 6% poor risk patients), all of which may lead to poorer outcome. Compared with the MSKCC database of immunotherapy patients composed of patients enrolled in clinical trials,17 the patients of our population-based study were slightly older (62 vs 59 years) with worse performance status (median KPS 70 vs 90), which again may lead to poorer outcome.
In addition, almost 60% of patients in the phase 3 trial13 had active second-line treatment, which would likely increase their overall survival compared with patients in our study. Our IFN group received no active second-line therapy (4 patients on non-VEGF phase 1 trials, 1 vinblastine, and 1 capecitabine). IFN patients who received second-line targeted therapy were excluded from the primary analysis and are discussed below. In our sunitinib group, <10% of patients had active second-line therapy (3 sorafenib, 2 temsirolimus). Although data collection was very complete through pharmacy registries, we cannot entirely exclude the possibility of missing data.
The major limitations of this study include the retrospective nature, the inability to perform a central pathology review, and the relatively modest cohort of patients collected over an extended period of time. However, the population of British Columbia was relatively stable over this period, which allowed the reliable tracking of clinical outcomes. In addition, data collection was very complete, with data available in >99% of patients. Of the possible 7000 data points collected, only 62 were missing (0.88%).
Technological advances in imaging or changes in surgical or supportive care over the years may have affected the results of this analysis. However, there were no obvious dramatic changes in the treatment of metastatic RCC in British Columbia in the last decade other than the introduction of targeted therapy. Furthermore, patients in the sunitinib group did not receive treatment earlier than patients in the IFN group, which would otherwise have artificially created an overall survival difference similar to the concept of lead-time bias in screening studies. Specifically, the median time from diagnosis of metastatic disease to treatment was 1.9 months in the IFN group and 2.6 months in the sunitinib group (log-rank P = .66).
When targeted therapy was first introduced, there was a transition period where 30 patients were treated with first-line IFN followed by second-line sunitinib and were subsequently excluded in this analysis as documented in the Materials and Methods section. This was because these patients would contaminate the direct comparison between IFN and sunitinib and may unfairly bias the data toward the IFN group. Of note, if these 30 patients were combined with the 131 IFN-only patients for a pure intention-to-treat analysis on first-line therapy, median overall survival between first-line IFN and first-line sunitinib treated groups would be 10.4 versus 17.3 months, respectively (IFN n = 161, sunitinib n = 69; log-rank P = .03), which is still clinically and statistically significant.
Finally, known prognostic factor profiles have been developed for patients with metastatic RCC including the commonly used, validated MSKCC criteria.17, 25 This study demonstrates that the MSKCC criteria were also able to discriminate between the 3 prognostic categories in both patients treated with IFN and those treated with sunitinib.
The introduction of sunitinib as standard first-line treatment for metastatic RCC has resulted in a substantial improvement in overall survival compared with patients treated with IFN alone in this retrospective analysis. This survival benefit extended into the subgroup of patients with poor MSKCC prognostic profiles. This study confirms the effectiveness of sunitinib in the general, unselected population of patients with metastatic RCC.
Conflict of Interest Disclosures
This study was funded in part by the Canadian Institutes of Health Research.
Dr. Garcia has received honoraria and advisory board fees from Wyeth, Pfizer, Bayer, and Genentech.
Dr. Rini has received honoraria and advisory board fees from Bayer, Wyeth, Genentech, Pfizer, and Onyx.
Dr. Bukowski has received honoraria and advisory board fees from Bayer, Pfizer, Novartis, Genentech, and Wyeth.