Bernard Escudier, Tim Eisen, Walter M. Stadler, Cezary Szczylik, Stephane Oudard, Michael Staehler,´ Sylvie Negrier, Christine Chevreau, Apurva A. Desai, Frederic Rolland, Tomasz Demkow,´´ Thomas E. Hutson, Martin Gore, Sibyl Anderson, Gloria Hoﬁlena, Minghua Shan, Carol Pena, Chetan Lathia, and Ronald M. Bukowski
Mature survival data and evaluation of vascular endothelial growth factor (VEGF) as a prognostic biomarker from the Treatment Approaches in Renal Cancer Global Evaluation Trial (TARGET) study in patients with renal cell carcinoma (RCC) are reported.
Patients and Methods
Nine hundred three previously treated patients were randomly assigned to receive sorafenib versus placebo. On demonstration of progression-free survival (PFS) beneﬁt with sorafenib, patients assigned to placebo were offered sorafenib. Overall survival (OS) was determined at two planned interim analyses and one ﬁnal analysis, with a secondary OS analysis conducted by censoring placebo patients who crossed over to sorafenib. The relationships between baseline VEGF level and prognosis and efﬁcacy were evaluated. Results The ﬁnal OS of patients receiving sorafenib was comparable with that of patients receiving placebo (17.8 v 15.2 months, respectively; hazard ratio [HR]=0.88; P=.146); however, when post– cross-over placebo survival data were censored, the difference became signiﬁcant (17.8 v 14.3 months, respectively; HR =0.78; P =.029). Adverse events at 16 months after cross over were similar to those previously reported. Baseline VEGF levels correlated with Eastern Cooper- ative Oncology Group performance status (P＜.0001), Memorial Sloan-Kettering Cancer Center score (P .0001), and PFS and OS in univariate (PFS, P =.0013; OS, P =.0009) and multivariate (PFS, P =.0231; OS, P =.0416) analyses of placebo patients and with short OS by multivariate analysis of patients receiving sorafenib (P=.0145). Both high-VEGF (P＜.01) and low-VEGF (P ＜.01) groups beneﬁted from sorafenib.
Although an OS beneﬁt was not seen on a primary intent-to-treat analysis, results of a secondary OS analysis censoring placebo patients demonstrated a survival advantage for those receiving sorafenib, suggesting an important cross-over effect. VEGF levels are prognostic for PFS and OS in RCC. The results of TARGET establish the efﬁcacy and safety of sorafenib in advanced RCC.
J Clin Oncol 27:3312-3318. © 2009 by American Society of Clinical Oncology
Renal cell carcinoma (RCC) comprises 5% of epi- thelial cancers diagnosed in the United States each year, with the majority being of clear cell histology.1,2 Approximately 20% to 30% of patients with RCC have metastases at diagnosis, and 20% to 40% of those with localized disease who undergo nephrec- tomy subsequently develop metastases.3 Until re- cently, therapeutic options for unresectable and/or metastatic RCC were limited. RCC is generally resis- tant to conventional chemotherapy, and only a small percentage of patients with RCC beneﬁt from cytokine treatment.2,4,5
Clear cell RCC is characterized by inactiva- tion of the Von Hippel-Lindau (VHL) pathway with somatic mutations or methylation of the VHL gene in the majority of patients. The result- ant increased production of vascular endothelial growth factor (VEGF)/platelet-derived growth fac- tor (PDGF) is considered to be fundamental to the highly angiogenic nature of RCC and critical to on- cogenesis.6,7 Several inhibitors of VEGF and its cog- nate receptor VEGF receptor (VEGFR) 2 have dramatic antitumor activity in RCC.8 Sorafenib tosylate (Nexavar; Bayer HealthCare Pharmaceuticals Corporation, Montville, NJ; Onyx Pharmaceuticals, Emeryville, CA) is an orally active multikinase in- hibitor that blocks VEGFR-2, VEGFR-3, and PDGF receptor (PDGFR- ), as well as RAF-1, Flt-3, and c-KIT.9,10 In a phase II randomized discontinuation trial, sorafenib improved progression- free survival (PFS) of RCC patients.11
These ﬁndings prompted initiation of the phase III multicenter Treatment Approaches in Renal Cancer Global Evaluation Trial (TARGET), a randomized, double-blind, placebo-controlled study of treatment with sorafenib in clear cell RCC patients who experienced treatment failure with one prior systemic therapy.12 In January 2005, a preplanned interim analysis conducted via independent assessment reported that sorafenib-treated patients had PFS that was signiﬁcantly superior to that of patients randomly assigned to placebo (5.5 v 2.8 months, respectively; hazard ratio [HR] =0.44; 95% CI, 0.35 to 0.55; P ＜.000001). On the basis of this analysis, the data safety monitoring board halted the trial, leading to cross over of patients still on placebo to sorafenib and approval of sorafenib for the treatment of advanced RCC by the US Food and Drug Administration and other regulatory authorities. The initial preliminary survival data on cross over did not meet the original planned statistical criteria (P .0094) for success on interim analysis (HR =0.71, P =.015). More mature survival data from an analysis 16 months after cross over (September 2006), along with data on VEGF as a predictive/prognostic molecular biomarker, are now presented.
PATIENTS AND METHODS
Patients and Study Design
Study design and patient inclusion criteria for TARGET have been pre- viously described.12 Patients were enrolled from November 2003 to March 2005. Brieﬂy, this was a phase III, multicenter, randomized, double-blind, placebo-controlled trial for patients with unresectable and/or metastatic RCC who had undergone one prior systemic therapy. Other inclusion criteria in- cluded low- or intermediate-risk Memorial Sloan-Kettering Cancer Center (MSKCC) score5 and adequate organ function. Patients were stratiﬁed by MSKCC score and country of enrollment and then randomly assigned to receive either continuous treatment with sorafenib 400 mg twice a day (n =451) or placebo (n =452). Treatment interruptions and two dose modiﬁcations (ﬁrst to 400 mg every day, then 400 mg every 2 days) were permitted for therapy-related toxicities. Patients remained on study drug until disease progression or discontinuation as a result of intolerable toxicity or death, but those in the sorafenib group could continue open-label treatment beyond the end point of radiologic progression at investigator discretion.
A planned analysis of PFS in January 2005 demonstrated an advantage in the sorafenib group (HR =0.44; 95% CI, 0.35 to 0.55; P ＜.001), leading to a data monitoring committee recommendation to close the study and cross patients on placebo over to sorafenib. Overall, 48% of the patients from the placebo group crossed over to sorafenib.
Study End Points
Efﬁcacy and safety end points. The original primary end point of TARGET was overall survival (OS). Secondary end points included PFS (de- ﬁned as time from random assignment to disease progression based on radio- logic or clinical assessment), response rate, and patient-reported outcomes. All randomly assigned patients were included in the intent-to-treat (ITT) popu- lation for the efﬁcacy analyses, and at the time of cross over, three preplanned efﬁcacy analyses were speciﬁed: an initial OS analysis corresponding to start of cross over (pre– cross-over analysis); a second OS analysis with a data cutoff date of 6 months after cross over; and a ﬁnal primary OS analysis with a data cutoff date originally planned to be when 540 events (deaths) had occurred, which took place 16 months after cross over (Appendix Fig A1A and Appendix Table A1, online only). Antitumor activity, safety, and the initial pre– cross- over survival have been previously described.12 Patient visits for safety moni- toring were conducted every 3 weeks during the ﬁrst 24 weeks of treatment and every 4 weeks thereafter. A toxicity assessment was conducted 30 days after discontinuation of study drug. After the conclusion of treatment with the study drug, patients were contacted to determine survival status approxi- mately every 3 months until death or loss to follow-up was documented.
VEGF analysis. Plasma samples were collected from patients at the screening visit. Blood samples (10 mL) were drawn into a vacutainer contain- ing potassium EDTA. Within 10 to 15 minutes after collecting, samples were centrifuged at 4°C to separate plasma. In the absence of a refrigerated centri- fuge, samples were chilled on ice for 5 to 15 minutes and subsequently centri- fuged at room temperature. Each plasma sample was frozen at –70°C within 20 minutes of centrifugation. VEGF was determined by sandwich enzyme- linked immunosorbent assay assay (code no. PDVE00; R&D Systems, Minne- apolis, MN) according to the manufacturer’s protocol. Results are expressed as the average of duplicate assays.
Original sample size was calculated to detect a 0.77 HR in OS, assuming an exponential survival distribution, a Lan-DeMets spending function for constructing group sequential boundaries, an overall two-sided α=.04, and power of 90%. For one interim and one ﬁnal analysis to be performed, 540 events (deaths) were required.13 The original planned sample size was also sufﬁcient to detect a 0.67 HR in PFS using a single analysis, with an overall two-sidedα=.01 and a power of 90%. For the ﬁnal OS analysis based on the 561 events (16 months after cross over, September 2006), a two-sided α≤.037 was used. Survival analyses were stratiﬁed by country and MSKCC prognostic risk category and conducted using the Kaplan-Meier method. Additional Cox proportional hazard analyses using the prespeciﬁed stratiﬁcation variables, biomarker data (see next section), and treatment arm were also planned and conducted. To account for a possible survival beneﬁt after cross over, a pre- speciﬁed ITT survival analysis uniformly censoring patients originally ran- domly assigned to placebo at the time of cross over was also conducted.
The relationship between VEGF, treatment arm, and PFS or OS was examined using Cox proportional hazards regression models and Kaplan- Meier analyses. Univariate analyses of baseline VEGF and outcome were performed on placebo patients only. Multivariate models included baseline VEGF as a continuous variable, Eastern Oncology Cooperative Group (ECOG) performance status (PS; 0 v 1 or 2), and MSKCC score (low v intermediate). To maximize the number of patients with noncensored data for correlative analyses, tumor assessment–related data (progression) were evalu- ated and determined by an investigator at the patient’s site as opposed to an independent review committee. The PFS data were from the interim analysis cutoff date (May 2005, before cross over), and the OS data were from the 16-month post– cross-over cutoff data (September 2006), with placebo data censored to June 2005 (before cross over; Appendix Figs A1A and A1B). The signiﬁcance of the relationships between demographic variables and VEGF was evaluated using an analysis of variance F test.
Patient Characteristics and Disposition
A total of 903 patients were enrolled onto TARGET; 451 and 452 patients were randomly assigned to receive sorafenib or pla- cebo, respectively (Appendix Fig A1B). Demographic and baseline characteristics of the ITT population have previously been de- scribed.12 A total of 216 patients randomly assigned to placebo (48%) crossed over to treatment with sorafenib; 93 patients were receiving double-blind treatment at the time of cross over, one was receiving open-label treatment, 121 were in long-term post-treatment follow- up, and one patient had no status information (Appendix Fig A1B). The majority of patients were male (75% to 76%), had a median age of 59 years, and had received prior interleukin-2/interferon alfa therapy (81% to 82%). The only signiﬁcant differences between patients ran- domly assigned to placebo (n =452) and who crossed over (n =216) and those who did not cross over (n =236) were that a smaller proportion of the former had an ECOG PS ≥1 (37% v 67%, respec- tively) and a greater proportion of those who crossed over had a low MSKCC risk (62% v 36%, respectively; Table 1).
Of the 903 randomly assigned patients, 700 (78%) entered post- treatment follow-up at any time during the trial (337 [75%] from the sorafenib group and 363 [80%] from the placebo group). The most frequent reasons for discontinuation in the sorafenib and placebo groups were death (229 v 248 patients, respectively), loss to follow-up (eight v 13 patients, respectively), and withdrawal of consent (six v ﬁve patients, respectively). At the time of ﬁnal survival analysis (16 months after cross over, September 2006), 170 patients (38%) in the sorafenib group were ongoing in the study (86 in post-treatment follow-up and 84 in treatment), and 155 patients (34%) in the placebo group were ongoing in the study (88 in post-treatment follow-up and 67 in treat- ment; Appendix Figs A1A and A1B).
Exposure to Study Medication: ITT
Patients who were randomly assigned to placebo received treat- ment for a median of 12.0 weeks before cross over and 40.2 weeks after cross over. In the ﬁnal OS analysis (16 months after cross over), the median administered dose of sorafenib was similar between sorafenib patients and placebo-assigned patients who had crossed over (790 mg; range, 213 to 1,707 mg v 778 mg; range, 229 to 1,444 mg, respectively). Fifty percent (n 225) of sorafenib-assigned and 40% (n 182) of placebo-assigned patients received 90% of planned dosing. Dose reductions or interruptions were infrequent in both groups, with only 28% (n 126) and 22% (n 100) of patients in the groups originally randomly assigned to sorafenib and placebo having two or more of such events, respectively. Eighty-ﬁve percent of patients (n 385) in the sorafenib-assigned group and 90% of patients (n 407) in the placebo-assigned group had no dose delays. It is of note that 285 patients (63%) randomly assigned to sorafenib and 104 cross-over patients (48%) continued sorafenib therapy after Response Evalua- tion Criteria in Solid Tumors (RECIST) – deﬁned disease progression per investigator-assessed clinical beneﬁt for a mean of 24.9 weeks (median, 14.9 weeks) and 16.8 weeks (median, 10.6 weeks), respec- tively. Sorafenib therapy accounted for 61% of the treatment received throughout the entire study for the placebo-assigned group.
PFS and OS
In January 2005, a preplanned interim PFS analysis, conducted via independent assessment, reported that patients treated with sor- afenib had PFS that was signiﬁcantly superior to that of patients randomly assigned to placebo (median PFS, 5.5 v 2.8 months, respec- tively; HR =0.44; 95% CI, 0.35 to 0.55; P ＜.000001).12 On the basis of these data, the protocol for TARGET was amended to permit patients randomly assigned to placebo to cross over to sorafenib treatment. The ﬁrst preplanned survival analysis for TARGET was performed in May 2005, when 48% of patients initially randomly assigned to pla- cebo crossed over to treatment with sorafenib. At that time, 220 deaths had occurred, including 97 in the sorafenib group (22%) and 123 in the placebo arm (27%). Although treatment with sorafenib was asso- ciated with an improved OS (not available v 14.7 months for placebo; HR=0.71; 95% CI, 0.54 to 0.94; P=.015; for interim analy- sis =.0094), these data did not reach the O’Brien-Fleming boundary for statistical signiﬁcance (Table 2 and Appendix Table A1). A second survival analysis was performed 6 months after cross over (November 2005) when 367 deaths had occurred (sorafenib =171; pla- cebo =196). In this analysis, patients initially randomly assigned to sorafenib had superior OS versus those initially assigned to placebo (19.3 v 15.9 months, respectively; HR =0.77; 95% CI, 0.63 to 0.95; P =.015), although once again, the ﬁndings did not meet the prespeci- ﬁed boundary for statistical signiﬁcance. A total of 561 deaths (sor- afenib =278; placebo =283) were reported at the ﬁnal data cutoff point 16 months after cross over (September 2006). In the ﬁnal anal- ysis of the ITT population, survival in the sorafenib group was not superior to placebo (17.8 v 15.2 months, respectively; HR= 0.88; 95% CI, 0.74 to 1.04; P =.146; Fig 1A). To determine whether cross over had an impact on OS, an analysis with censoring of placebo-assigned
patients who crossed over to sorafenib at the start of cross over was conducted. In this case, treatment with sorafenib was associated with an improved survival compared with placebo (17.8 v 14.3 months, respectively; HR =0.78; 95% CI, 0.62 to 0.97; P =.0287; Fig 1B).
Overall, sorafenib was well tolerated, and most events were grade 1 or 2, reversible, and clinically manageable. Treatment-related ad- verse events (AEs) for all patients randomly assigned to sorafenib or placebo and for patients who crossed over from placebo to sorafenib are listed in Table 3. Grade 3 and 4 treatment-related AEs were un- common in both the sorafenib-assigned and cross-over groups; diar- rhea, fatigue, hypertension, hand-foot skin reaction, and rash/ desquamation were reported in more than 2% of patients.
Fig 1. Final overall survival (OS) and progression-free survival (PFS) data. (A) Kaplan-Meier analysis of ﬁnal OS for the intent-to-treat (ITT) population. (B) Kaplan-Meier analysis of ﬁnal OS for the ITT population with placebo patients censored. (C) Kaplan-Meier analysis of PFS in placebo patients only by baseline vascular endothelial growth factor (VEGF) levels. (D and E) Kaplan-Meier analyses of PFS by baseline VEGF levels. HR, hazard ratio.
Twenty-two patients (4.9%) randomly assigned to sorafenib reported cardiac ischemic/infarct AEs, with six events reported as related to study drug. Dose interruption was required in six pa- tients receiving sorafenib, and one participant reported dose re- duction. One cardiac ischemic event in the sorafenib group led to permanent discontinuation of study drug. Of note, the sorafenib arm had a longer follow-up time and inclusion of postprogression patients. Three cross-over patients (1.4%) and two patients on placebo (0.4%) reported cardiac ischemia/infarct events. CNS ischemia was reported by seven sorafenib patients (1.5%), com- pared with three patients (0.7%) in the placebo group and none in the cross-over group.
VEGF As a Prognostic Factor for PFS and OS in RCC
Previous studies have suggested that VEGF levels may be a prognostic factor in RCC.14 These ﬁndings, in combination with the inhibition of VEGFR-2 by sorafenib, led to the hypothesis that patients with elevated baseline VEGF levels may derive enhanced beneﬁt from sorafenib treatment. Therefore, we examined VEGF levels in the present study. Assessable baseline VEGF data were available for 712 patients. In univariate analyses of VEGF (as a continuous variable) versus outcome in placebo patients, VEGF levels correlated inversely with PFS (P =.0013; Fig 1C) and OS (P =.0009; data not shown).
Patients with a higher MSKCC score (poorer prognosis) or higher ECOG PS (poorer performance) had signiﬁcantly higher base- line VEGF levels than those with low MSKCC (P ＜.0001) or ECOG scores (P ＜.0001; Table 4), lending further support for the association of high VEGF levels with poor prognosis in RCC patients. Sex (male v female; P =.7595), age (＜ v ≥65 years; P =.2904), and stage at study entry (stage III v IV; P =.3879) did not have signiﬁcant association with baseline VEGF levels (Table 4).
To determine whether VEGF offers additional prognostic infor- mation independently of ECOG PS and MSKCC score, multivariate analyses were performed (Table 5). Results of these analyses indicate that both MSKCC score and baseline VEGF are independent prognos- tic factors for PFS when analyzed for patients on placebo only (MSKCC, P =.0012; VEGF, P= .0231), but not in patients treated with sorafenib (MSKCC, P=.0657; VEGF, P=.6252). For the analysis of OS, MSKCC score, ECOG PS, and baseline VEGF were all independently prognostic in both placebo patients (MSKCC, P ＜.0001; ECOG, P =.0018; VEGF, P =.0416) and sorafenib-treated patients (MSKCC, P ＜.0001; ECOG, P =.0248; VEGF, P =.0145; Table 5).
Initial analysis of baseline VEGF and sorafenib antitumor activity (in terms of PFS) using the median value (131 pg/mL) to deﬁne high- versus low-VEGF groups suggested that both groups beneﬁt from sorafenib treatment (Figs 1D and 1E). However, the high-VEGF group trended toward deriving more beneﬁt from sor- afenib (HR =0.48; 95% CI, 0.38 to 0.62) than the low-VEGF group (HR =0.64; 95% CI, 0.49 to 0.83; P for interaction between VEGF and treatment arm=.096). In an attempt to optimize this difference, further exploratory analyses using the 25th and 75th percentiles to deﬁne low versus high VEGF were performed using half of the data set (for hypothesis generation, chosen randomly). The difference in ben- eﬁt between the high- and low-VEGF groups was most pronounced using the 75th percentile (254 pg/mL; HR =0.27; 95% CI, 0.15 to 0.460 for high VEGF v 0.58; 95% CI, 0.43 to 0.78 for low VEGF; P for interaction =.020), with the high-VEGF group deriving more beneﬁt from sorafenib (Appendix Table A2, online only). This ﬁnding was supported by analyzing the second half of the data (conﬁrmatory data set; P =.023; Appendix Table A2), again suggesting that, although both high- and low-VEGF groups beneﬁt from sorafenib (in terms of PFS), the high-VEGF group may beneﬁt more.
The TARGET trial demonstrated that treatment with sorafenib doubled PFS versus placebo in patients with previously treated advanced RCC. The ﬁnal survival analyses were confounded by the crossing over of study participants from placebo to sorafenib. This likely contributed to the lack of observed survival advantage in the ﬁnal analysis. In a preplanned secondary analysis in which patients crossing over from placebo to sorafenib were censored, treatment with sorafenib led to a survival advantage over placebo, further suggesting that the primary end point was confounded by the cross over.
Sorafenib was well tolerated, and most AEs were grade 1 or 2, easily managed, and consistent with prior reports. The observed car- diovascular events are more notable than in the original report and are similar to what has been reported with other VEGF pathway– directed agents. Although these events were confounded by the longer treat- ment time on sorafenib than placebo, they illustrate the potential vascular toxicity of these agents. Whether more aggressive blood pres- sure management, as has been suggested with increasing experience with VEGF pathway– directed agents, will ameliorate this toxicity re- mains to be determined.
Increased tumor VEGF activity is frequently observed in patients with RCC.7 Previous studies have shown that serum VEGF is corre- lated with tumor grade, stage, and recurrence after deﬁnitive local resection.14 Serum VEGF has also been associated with poorer out- come with sunitinib therapy.15 The ﬁndings of the current study support a prognostic role for VEGF in the largest RCC population tested to date. The prognostic value of VEGF is preserved in multivar- iate analyses including MSKCC score and ECOG PS, suggesting that VEGF reﬂects an aggressive tumor biology not captured by these scoring scales. In addition, although patients with either high or low baseline VEGF beneﬁt from sorafenib (in terms of PFS), the data presented here suggest that patients with high VEGF levels, who are at an initial disadvantage with poorer prognosis, may beneﬁt more.
Updated results from TARGET, including data from pre– cross- over, 6-month post– cross-over, and 16-month post– cross-over cut- off dates, demonstrated that sorafenib provided signiﬁcant clinical beneﬁt for patients with RCC. Although ﬁnal OS analyses were con- founded by participants who had crossed over from placebo to sor- afenib, censoring these patients at cross over revealed a signiﬁcant improvement in OS for patients randomly assigned to sorafenib. These data conﬁrm the ﬁndings of the planned interim analysis of TARGET; sorafenib is efﬁcacious for the treatment of advanced RCC. These data are also in accordance with data recently reported with sunitinib, suggesting that tyrosine kinase inhibitors improve OS in RCC, although cross over and sequential treatments decrease the magnitude of the difference.
AUTHORS’ DISCLOSURES OF POTENTIAL CONFILCTS OF INTEREST
Although all authors completed the disclosure declaration, the following author(s) indicated a ﬁnancial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO’s conﬂict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conﬂicts of Interest section in Information for Contributors.
Employment or Leadership Position: Sibyl Anderson, Bayer HealthCare (C); Gloria Hoﬁlena, Bayer HealthCare (C); Minghua Shan, Bayer HealthCare (C); Carol Pena, Bayer HealthCare (C); Chetan Lathia, Bayer HealthCare (C) Consultant or Advisory Role: Bernard Escudier, Bayer HealthCare (C), Pﬁzer (C), Roche (C); Tim Eisen, Bayer HealthCare (C), Pﬁzer (C), Wyeth (C); Walter M. Stadler, Onyx (C), Bayer HealthCare (C), Pﬁzer (C), Genentech (C), Novartis (C), Wyeth (C); Cezary Szczylik, Bayer HealthCare (C); Michael Staehler, Bayer HealthCare (C); Thomas E. Hutson, Bayer HealthCare (C), Pﬁzer (C), Wyeth (C); Martin Gore, Bayer HealthCare (C); Carol Pena, Bayer HealthCare (C); Ronald M. Bukowski, Pﬁzer (C), Bayer HealthCare (C), Wyeth (C), Novartis (C), Genentech (C), Antigenics (C), Regeneron (C) Stock Ownership: Walter M. Stadler, Abbott; Sibyl Anderson, Bayer HealthCare; Gloria Hoﬁlena, Bayer HealthCare; Minghua Shan, Bayer HealthCare; Carol Pena, Bayer Healthcare; Chetan Lathia, Bayer HealthCare Honoraria: Bernard Escudier, Bayer HealthCare, Roche, Pﬁzer, Genentech, Novartis; Tim Eisen, Bayer HealthCare, Pﬁzer, Wyeth; Cezary Szczylik, Bayer HealthCare; Michael Staehler, Bayer HealthCare; Sylvie Negrier, Bayer HealthCare, Wyeth, Pﬁzer; Thomas E. Hutson, Bayer HealthCare, Pﬁzer, Wyeth; Martin Gore, Bayer HealthCare; Ronald M. Bukowski, Pﬁzer, Bayer HealthCare, Onyx, Genentech, Wyeth, Novartis Research Funding: Tim Eisen, Bayer HealthCare, Pﬁzer; Walter M. Stadler, Bayer HealthCare, Pﬁzer, Genentech, Novartis, Exilixis, Amgen, Bristol-Meyers Squibb, AstraZeneca, Imclone Systems; Michael Staehler, Bayer HealthCare; Sylvie Negrier, Wyeth; Thomas E. Hutson, Bayer HealthCare, Pﬁzer, Wyeth; Martin Gore, Bayer HealthCare; Ronald M. Bukowski, Pﬁzer, Bayer HealthCare, Novartis, Wyeth Expert Testimony: Walter M. Stadler, Novartis (C); Cezary Szczylik, Bayer HealthCare (C) Other Remuneration: Tim Eisen, Bayer HealthCare; Cezary Szczylik, Bayer HealthCare
Conception and design: Bernard Escudier, Tim Eisen, Walter M. Stadler, Sylvie Negrier, Christine Chevreau, Martin Gore, Gloria Hoﬁlena, Minghua Shan, Chetan Lathia, Ronald M. Bukowski Provision of study materials or patients: Walter M. Stadler, Cezary Szczylik, Stephane Oudard, Michael Staehler, Sylvie Negrier, Christine´ Chevreau, Apurva A. Desai, Frederic Rolland, Tomasz Demkow, Thomas´´ E. Hutson, Martin Gore, Ronald M. Bukowski Collection and assembly of data: Stephane Oudard, Frederic Rolland,´´´ Martin Gore, Sibyl Anderson, Gloria Hoﬁlena, Carol Pena, Chetan Lathia, Ronald M. Bukowski Data analysis and interpretation: Walter M. Stadler, Cezary Szczylik, Apurva A. Desai, Thomas E. Hutson, Sibyl Anderson, Gloria Hoﬁlena, Minghua Shan, Carol Pena, Chetan Lathia, Ronald M. Bukowski Manuscript writing: Walter M. Stadler, Thomas E. Hutson, Sibyl Anderson, Carol Pena, Ronald M. Bukowski Final approval of manuscript: Walter M. Stadler, Cezary Szczylik, Stephane Oudard, Sylvie Negrier, Christine Chevreau, Apurva A. Desai,´ Frederic Rolland, Thomas E. Hutson, Martin Gore, Gloria Hoﬁlena,´´ Minghua Shan, Carol Pena, Chetan Lathia, Ronald M. Bukowski
1. Jemal A, Siegel R, Ward E, et al: Cancer statistics, 2007. CA Cancer J Clin 57:43-66, 2007
2. Costa LJ, Drabkin HA: Renal cell carcinoma: New developments in molecular biology and potential for targeted therapies. Oncologist 12:1404-1415, 2007
3. Janzen NK, Kim HL, Figlin RA, et al: Surveillance after radical or partial nephrectomy for localized renal cell carcinoma and management of recurrent disease. Urol Clin North Am 30:843-852, 2003
4. Gitlitz BJ, Figlin RA: Cytokine-based therapy for metastatic renal cell cancer. Urol Clin North Am 30:589-600, 2003
5. Motzer RJ, Mazumdar M, Bacik J, et al: Survival and prognostic stratiﬁcation of 670 patients with advanced renal cell carcinoma. J Clin Oncol 17:2530-2540, 1999
6. van Houwelingen KP, van Dijk BA, Hulsbergen-van de Kaa CA, et al: Prevalence of von Hippel-Lindau gene mutations in sporadic renal cell carcinoma: Results from The Netherlands cohort study. BMC Cancer 5:57, 2005
7. Na X, Wu G, Ryan CK, et al: Overproduction of vascular endothelial growth factor related to von Hippel-Lindau tumor suppressor gene mutations and hypoxia-inducible factor-1 alpha expression in renal cell carcinomas. J Urol 170:588-592, 2003
8. Herbst RS: Therapeutic options to target an- giogenesis in human malignancies. Expert Opin Emerg Drugs 11:635-650, 2006
9. Wilhelm SM, Carter C, Tang L, et al: BAY 43-9006 exhibits broad spectrum oral antitumor activity and targets the RAF/MEK/ERK pathway and receptor tyrosine kinases involved in tumor progres- sion and angiogenesis. Cancer Res 64:7099-7109, 2004
10. Bayer Pharmaceuticals: Nexavar (sorafenib) prescribing information. http://berlex.bayerhealthcare .com/html/products/pi/Nexavar_PI.pdf
11. Ratain MJ, Eisen T, Stadler WM, et al: Phase II placebo-controlled randomized discontin- uation trial of sorafenib in patients with metastatic renal cell carcinoma. J Clin Oncol 24:2505-2512, 2006
12. Escudier B, Eisen T, Stadler WM, et al: Sor- afenib in advanced clear-cell renal-cell carcinoma. N Engl J Med 356:125-134, 2007
13. DeMets L: Discrete sequential boundaries for clinical trial. Biometrika 70:659-663, 1983
14. Jacobsen J, Rasmuson T, Grankvist K, et al: Vascular endothelial growth factor as prognostic factor in renal cell carcinoma. J Urol 163:343-347, 2000
15. George DJ, Michaelson MD, Rosenberg JE, et al: Phase II trial of sunitinib in bevacizumab- refractory metastatic renal cell carcinoma (mRCC): Updated results and analysis of circulating biomar- kers. J Clin Oncol 25:243s, 2007 (suppl; abstr 5035)