Targeted therapy for advanced renal cell cancer (RCC): a Cochrane systematic review of published randomised trials


Christian Kollmannsberger, BC Cancer Agency, 600 W. 12th Avenue, Vancouver, BC V5Z 4E6, Canada. e-mail:


What's known on the subject? and What does the study add?

Targeted agents have greatly changed the therapeutic landscape in RCC. A substantial number of trials have been published in recent years.

The current review summarises and analyses the available data to date.


  • • To estimate the effects of drugs with molecular targets on patients with advanced renal cell cancer (RCC).


  • • MEDLINE, EMBASE, and the Cochrane Collaboration Library were systematically searched on-line through to June 2011 to identify eligible randomised trials. We also searched abstract reports from major oncology and urology meetings.
  • • We included randomised trials that tested a targeted agent and reported at least one outcome by allocation on an intent-to-treat basis. Completeness of ascertainment and risk of bias were assessed.
  • • Our primary outcome was progression-free survival (PFS).


  • • In all, 28 studies met our inclusion criteria and 10 were placebo-controlled. Two studies were too small to assess, and five early studies used nonspecific anti-angiogenic agents with poor activity. In all, 15 studies, in 5587 patients, tested anti-vascular epithelial growth factor (VEGF) agents: bevacizumab (BEV), sorafenib, sunitinib, pazopanib, tivozanib, or axitinib. Three studies, in 1147 patients, tested the mammalian target of rapamycin (mTOR) inhibitors, temsirolimus or everolimus. Two studies included epidermal growth factor receptor (EGFR) inhibitors, and one tested the combination of temsirolimus plus BEV.
  • • In treatment-naive patients with mostly good–moderate prognostic risk, in separate trials oral sunitinib (one trial) and intravenous BEV plus subcutaneousinterferon-α (two trials) improved PFS compared with the previous standard of care interferon-α within randomised phase III trials. Sorafenib did not improve PFS over interferon-α in the first-line setting and the addition of cytokines did not improve sorafenib efficacy. In poor-risk patients, the mTOR inhibitor temsirolimus improved PFS and overall survival (OS). The studies of other VEGF inhibitors have used placebo controls no longer appropriate in this setting, although pazopanib is an approved option.
  • • Several trials examined agents in the second-line setting. After cytokine therapy, sorafenib (one study) and pazopanib (one study) prolonged PFS over placebo. A preliminary report of the investigational VEGF receptorinhibitor axitinib gave superior PFS to sorafenib after either prior cytokine or prior sunitinib treatment. After cancer progression ≤6 months of sunitinib and/or sorafenib therapy, everolimusprolonged PFS.
  • • OS was marginally improved in several studies. A more substantial effect on OS may have been diluted by crossover from control therapy to the investigational arm and/or by other anti-angiogenic agents after trial closure. Patient-reported outcomes were considered unreliable in trials without ‘blinding’.
  • • A clear cell RCC (ccRCC) component was required for most trials, and information for non-ccRCCs is consequently limited


  • • Agents targeting VEGF and mTOR pathways improve PFS in both first-line and second-line settings. These treatments rarely yield complete responses and thus are not curative.
  • • No placebo-controlled trial has reported a health-related quality of life benefit.



clear cell RCC


vascular epithelial growth factor (receptor)


progression-free survival


overall survival


objective response rate


Response Evaluation Criteria in Solid Tumors (criteria)


mammalian target of rapamycin




hazard ratio


health-related quality of life


epidermal growth factor receptor


Systemic therapy of advanced RCC has been unsatisfactory. The former standard cytokine therapy with interferon-α (IFNα) had substantial subjective toxicities and poor effectiveness [1]. Targeted agents represent a new class of drugs that have much more specific sites of cellular action than chemotherapy or immunotherapy, with potential for improved effectiveness with fewer harmful effects. Subtypes of RCC based on molecular pathology are now well recognised [2]. In particular, the common clear cell subtype of RCC (ccRCC) usually has inactivation of both copies of the von Hippel–Lindau gene with constitutive activation of the hypoxia-inducible pathway that provides multiple therapeutic targets including vascular epithelial growth factor (VEGF). We recently updated a Cochrane systematic review of published randomised trials of targeted agents for RCC [3]. We summarise our approach and findings, further updated to June 2011.


We included studies if they were randomised controlled trials in advanced RCC published in the English language through to June 2011 that included a targeted agent in at least one study arm, and reported at least one efficacy outcome by allocation. Targeted agents could have specific known sites of action (single or multiple targets), or less specific anti-angiogenic action. Studies were identified from duplicate electronic searches of MEDLINE, EMBASE, and the Cochrane Library, supplemented by searches of the published abstracts of major meetings of the American Society of Clinical Oncology, European Cancer Organization, European Society of Medical Oncology, and the AUA. All identified publications were assembled for each study [3], but we list only the most recent key publication. Standard Cochrane Collaboration methodology was used for data extraction, analysis, and risk of bias assessment. Our primary outcome was progression-free survival (PFS), the most consistently reported endpoint. Key secondary effectiveness outcomes included overall survival (OS), objective response rate (ORR) by RECIST (Response Evaluation Criteria in Solid Tumors) criteria, and patient-reported outcomes. We also assessed treatment related harms.

RESULTS (Tables 1[3–29] and 2[10–14,16,19,21–23,25,26])

Table 1.  Summary of included randomised studies. Trial names are as in the Cochrane review [3] and primary outcome is indicated in bold; cc, clear cell component required; pp, poor prognosis required; ab, meeting report only; all studies are open label unless ‘placebo’ is indicated under comparator
ReferenceTrial nameStart datePatients, nPrior agent(s)Agent(s)ComparatorReported outcomes
  1. CAI, carboxyaminoimidazole; IL2, interleukin-2.

Nonspecific anti-angiogenesis inhibitors
 Escudier et al. (1) [6]NCT000059952000305CytokineAE-941PlaceboOS
 Gordon et al.[7]ECOG E29982000342NilThalidomideIFNαOS, PFS, ORR
 Lee et al.ab[8] 200060CytokineThalidomideHormoneOS, PFS, ORR
 Stadler et al.[4]CALGB 699012000368Nil, IFNαCAIPlaceboPFS
 Ebbinghaus et al.[9]NCT000731252003103NilABT-510100 vs 10 mgOS, PFS, ORR
VEGF or VEGFR inhibitors
 ccYang et al.[10]NCT000195391998116IL2BEV 10 mgPlaceboOS, PFS, ORR
 ccRini et al. (1) [11]CALGB 902062003732NilBEV + IFNαIFNαOS, PFS, ORR
 ccEscudier et al. (3) [12]AVOREN2004649NilBEV + IFNαPlacebo + IFNαOS, PFS, ORR
 Ratain et al.[5]NCT00079612200265cytokineSorafenibPlaceboPFS
 ccEscudier et al. (2) [13]TARGET2003903cytokineSorafenibPlaceboOS, PFS, ORR, HRQL
 ccEscudier et al. (4) [14]NCT001176372004189NilSorafenibIFNαPFS, HRQL
 ccJonasch et al.[15]NCT00126594200580NilSorafenib + IFNαSorafenibOS, PFS, ORR
 ccBracarda et al.ab[16]RAPSODY2007101NilSorafenib + IFNα x 5Sorafenib + IFNαPFS, ORR
 Procopio et al.[17]ROSORC2006128NilSorafenib + IL2SorafenibPFS, ORR
 ccRini et al. (3)ab[18]NCT00467025 152NilSorafenib + AMG386SorafenibPFS, ORR
 ccMotzer et al. (1) [19]NCT000838892004750NilSunitinib 4/2IFNαOS, PFS, ORR, HRQL
 ccMotzer et al. (3)ab[20]EFFECT2007292NilSunitinib dailySunitinib 4/2OS, PFS, ORR
 ccSternberg et al.[21]VEG 101922006435Nil, ifnαPazopanibPlaceboOS, PFS, ORR, HRQL
 Nosov et al.ab[22]NCT005023072007272Nil, ifnαTivozanib 3/1PlaceboOS, PFS, PFS@12 weeks
 ccRini et al. (2)ab[23]AXIS 723Any oneAxitinibSorafenibPFS, ORR
mTOR inhibitors       
 Atkins et al.[24] 2000111CytokineTemsirolimus3 dose levelsOS, PFS, ORR
 ppHudes et al.[25]Global ARCC2003626NilTemsirolimusIFNαOS, PFS, ORR, HRQL
 ccMotzer et al. (2) [26]RECORD-12006410VEGFR inhibitorEverolimusPlaceboOS, PFS, ORR, HRQL
 Ravaud et al.[27]EGF 200012002416CytokineLapatinibHormoneOS, PFS, ORR
 ccBukowski et al.[28]NCT000816142004104NilErlotinib + BEVPlacebo + BEVOS, PFS, ORR, HRQL
 Escudier et al. (5)ab[29]TORAVA2008171NilTemsirolimus + BEVBEV + IFNα, or sunitinibORR, PFS@48 weeks
Table 2.  Outcomes for trials reporting improved PFS
ReferencePhaseTrial nameAgent(s)ComparatorPFS, monthsHRPORR, %POS, monthsHRP
  • *

    ‘blinded’ independent imaging assessment; nr, not reported; RDT, randomized discontinuation trial; HR for progression or death; ns, not statistically significant; RII, randomised phase II trial; III, phase III trial. All data is as reported by investigators. P values are stratified where applicable. Boxes define major outcomes.

Yang et al.[10]RIINCT 19539BEV 10 mgplacebo4.8 vs 2.50. 39<0.00110 vs 0nsnr  
Rini et al. (1) [11]IIICALGB 90206BEV + IFNαIFNα8.5 vs 5.20.71<0.00126 vs 16<0.00118.3 vs 17.40.860.069
Escudier et al. (3) [12]IIIAVORENBEV + IFNαPlacebo + IFNα10.2 vs 5.40.61<0.00131 vs 13<0.00123.2 vs 21.30.860.13
Ratain et al.[14]RDTNCT 79612SorafenibPlacebo5.5 vs 1.4nr0.009   
Escudier et al. (2) [13]IIITARGETSorafenibPlacebo5.5 vs 2.80.44<0.0012 vs 0ns17.8 vs 15.20.880.15
Bracarda et al.[16]RIIRAPSODYSorafenib + IFNα x 5Sorafenib + IFNα8.6 vs 7.9nr0.04935 vs 18nsnr  
Motzer et al. (1) [19]IIINCT 83889Sunitinib 4/2IFNα*11 vs 50.54<0.00147 vs 12<0.00126.4 vs 21.80.820.049
Sternberg et al.[21]IIIVEG 10192PazopanibPlacebo*9.2 vs 4.20.46<0.001*30 vs 3<0.00122.9 vs 20.50.910.22
Nosov et al.[22]RDTNCT 502307Tivozanib 3/1Placebo10.3 vs 3.3nr0.01   
Rini et al. (2) [23]IIIAXISAxitinibSorafenib*6.7 vs 4.70.67<0.001*19 vs 9<0.001nr  
Hudes et al.[25]IIIGlobal ARCCTemsirolimusIFNα3.8 vs 1.9nr<0.059 vs 5ns10.9 vs 7.30.730.008
Motzer et al. (2) [26]IIIRECORD-1EverolimusPlacebo*4.9 vs 1.90.33<0.001*1 vs 0ns14.8 vs 14.40.870.18


In all, 28eligible trials were identified, of which eight were identified only as meeting reports. Two studies were too small for further consideration [30,31]. Targeted agents might be growth stabilizing rather than cytoreductive, and three studies randomly discontinued the active agent to placebo in patients with stable disease during an initial ‘run-in’ period (randomised discontinuation trials) [4,5,32]. The most frequent stratification was by performance status, nephrectomy status, and prognostic risk category. Five trials of four agents with non-specific anti-angiogenic activity were tested in the earlier years of this series and none showed clinical benefits [4,6–9]. We here divide the studies of target-specific agents into three groups based on the molecular target: specific VEGF pathway inhibitors (15 studies), mammalian target of rapamycin (mTOR) inhibitors (three studies), and other (three studies). Most studies required RCC with a clear cell component; consequently the following results are for clear cell RCC (ccRCC) unless stated otherwise. Presented statistics of time-dependent outcomes are based on the authors' reports of the two-sided log-rank test on an intent-to-treat basis. Quantitative subset analysis is considered for stratified variables only.


Bevacizumab (BEV)

BEV is a monoclonal antibody against VEGF, blocking access to its receptor and inhibiting downstream angiogenesis that is amplified in ccRCC. In a randomised phase II study, BEV at 3 mg/kg and 10 mg/kg i.v. biweekly was compared with placebo in 116 patients previously treated with, or unsuitable for, high-dose interleukin-2 [10]. The ORR was <10% in any arm but PFS was prolonged with BEV 10 mg/kg (median 4.8 vs 2.5 months, hazard ratio [HR] 0.39, P < 0.001). This dose and schedule of BEV was reported to be tolerable when added to IFNα (9 million IU, s.c., three-times weekly). Two studies compared BEV + IFNα vs IFNα alone with similar results [11,12]. Unlike CALGB 90206, the AVOREN trial also included a placebo control and conducted independent ‘blinded’ imaging assessment for response and progression endpoints. The ORR was substantially improved (28.4 vs 12.9% for BEV + IFNα vs IFNα or IFNα+ placebo, combined data for the two studies), as well as PFS, the HRs for progression were 0.71 and 0.61, respectively. The primary endpoint for both studies was OS (median survivals for CALGB 90206 were 18.3 vs 17.4 months, and for AVOREN 23.2 vs 21.3 months for BEV + IFNα vs IFNα or IFNα+ placebo, respectively). Survival differences were not significant in either study alone. However, the stratified HRs for survival indicated a 14% risk reduction for death in favour of the BEV + IFNα combination. On both studies, more than half of all patients received post-protocol therapies such as VEGFR inhibitors that may have diluted any survival benefit of the first-line treatment. The relative contributions of BEV and IFNα to these benefits is unknown, but each agent has shown efficacy over inactive therapy [1,10].


The intracellular kinase terminal of the VEGF receptor family can be blocked by orally available small molecule inhibitors developed over the past decade, beginning with sorafenib. A small randomised discontinuation trial of sorafenib showed prolonged PFS in patients stable during the run-in phase [5]. The TARGET phase III study evaluated sorafenib vs placebo in the second-line setting after IFNα progression, finding few RECIST-defined responses but prolonged PFS (median 5.5 vs 2.8 months, HR 0.44, P < 0.001) [13]. Improvement in the primary endpoint of OS did not reach significance in the intent-to-treat analysis but nearly half of the placebo-assigned patients crossed over to sorafenib at study closure. Overall health-related quality of life (HRQL) scores were not significantly improved [33]. Sorafenib has been the best-evidenced second-line option after cytokine failure, until the AXIS study discussed below. A randomised phase II study of sorafenib in the first-line setting reported no significant advantage in remissions or PFS for sorafenib over IFNα at the standard sorafenib dose of 400 mg twice daily [14]. Subsequent emphasis has been on attempted enhancement of activity by combining sorafenib with other agents, including low dose IFNα[15,16], low dose interleukin-2 [17], and AMG 386 (an investigational fusion protein that sequesters angiopoietin-1 and -2) [18]. Thus far, no clinically useful advance with these combinations has been identified.


The oral VEGFR inhibitor sunitinib was introduced to phase I/II testing using a discontinuous 6-week cycle of 4 weeks ‘on’ and 2 weeks ‘off’ therapy (4/2 regimen). Sunitinib went directly into a phase III study in systemically untreated patients [19]. At entry, patients were required to have pretreatment progression and measurable disease. The high rates of objective response in non-randomised phase II studies after cytokine failure were confirmed in the first-line phase III trial (independently assessed ORR 46.9 vs 12.3%, sunitinib vs IFNα). The primary endpoint of PFS was evaluated by independent imaging review and was substantially prolonged (median 11 vs 5 months, HR 0.54, P < 0.001), and was consistent across patient subsets, although only 7% were poor prognostic risk. Importantly, there was an improvement in patient-reported outcomes [34], although the validity of this evaluation is difficult to assess in the absence of a double-dummy placebo control. The secondary endpoint of OS was improved with borderline statistical significance (median 26.4 vs 21.8 months, HR 0.82, stratified P= 0.049). The survival benefit may have been diluted by crossover of ≈60% of placebo-assigned patients to sunitinib and/or other VEGFR inhibitor therapy. The phase II EFFECT trial in measurable treatment-naïve advanced disease compared the standard 4 weeks on/2 weeks off (4/2) 50 mg daily regimen with 37.5 mg daily taken continuously; compliance was better with the discontinuous regimen, patients appreciate the break in toxicities, and the primary outcome of PFS showed a trend in favour of the 4/2 regimen (median PFS 8.5 vs 7.0 months, HR 0.77, P= 0.07) [20].


Pazopanib is the third oral VEGFR inhibitor to achieve regulatory approval status in the USA and Europe. The study of pazopanib 800 mg daily vs placebo (2:1 randomisation) [21] included 202 patients after prior cytokine and was expanded soon after opening to include 233 treatment-naïve patients because of the emerging evidence of VEGFR inhibitor efficacy in the untreated setting. The intent-to-treat ORR was 30 vs 3% (pazopanib vs placebo, P < 0.001). The primary outcome of PFS was significantly improved in both cytokine-pretreated and treatment-naïve patients (all patients median PFS 9.2 vs 4.2 months, HR 0.46; cytokine-pretreated HR 0.54, and treatment-naïve patients HR 0.40; all P < 0.001, overlapping 95%CIs). OS was similar for pazopanib-assigned and placebo-assigned patients (median OS 22.9 vs 20.5 months, HR 0.91, P= 0.22) but 54% of the latter received pazopanib after progression. HRQL was neither better nor worse than placebo [35]. The results of the COMPARZ phase III study comparing pazopanib with sunitinib are awaited with interest (NCT00720941).


Tivozanib 1.5 mg daily (orally), 3 weeks ‘on’ then 1 week ‘off’, has shown activity over placebo from a randomised discontinuation trial in treatment-naïve or cytokine-pretreated patients with stable disease after the 16-week run-in phase [22,32]. The primary endpoint of freedom from progression after a further 12 weeks was 49% for tivozanib vs 21% for placebo (P= 0.001). This agent is currently in phase III vs sorafenib (NCT01030783).


Axitinib is the most recently reported active oral VEGFR inhibitor from the second-line phase III AXIS study of 723 patients who had received one prior agent reflecting the range of current choices for first-line therapy (sunitinib 54%, cytokine 35%, other 11%) [23]. Sorafenib 400 mg twice daily was compared with axitinib 5 mg twice daily dose-titrated up or down to tolerance. The primary outcome of PFS was significantly better for axitinib than sorafenib regardless of prior treatment (median PFS for all patients was 6.7 vs 4.7 months, HR 0.67, P < 0.001; HR 0.74 after prior sunitinib, HR 0.46 after cytokine). Partial responses were seen more often after axitinib than sorafenib (19.4 vs 9.4%, P < 0.001). Improved PFS in the AXIS study was confirmed by evaluation of patient-reported outcomes that was the same for the two agents during therapy but was better for axitinib in the post-treatment evaluation [36].

Adverse events with VEGFR inhibitors

The practical supervision and detailed monitoring of targeted agents in the clinic is central to the safety of the individual patient but beyond the scope of this article. Currently, the only direct comparison of two targeted agents is of the VEGFR inhibitors axitinib vs sorafenib in the AXIS trial, as yet only available as a meeting report [23,36]. Axitinib had more all-grade hypertension (40 vs 29%) and hypothyroidism (19 vs 8%) than sorafenib, considered on-target effects and consistent with axitinib having more selective action against VEGFRs. Axitinib also had more fatigue and dysphonia but less hand-foot syndrome, rash, and alopecia. Fewer axitinib patients discontinued therapy for adverse events (3.9 vs 8.2%) despite dose reductions or interruptions being allowed in either arm. Other toxicities seen with VEGFR inhibitors include cardiotoxicity, bleeding, impaired wound healing, sarcopenia, hepatotoxicity, stomatitis and diarrhoea.

Comparative efficacy of VEGFR inhibitors

These agents differ in their range of inhibition of the spectrum of VEGFRs and other similar receptor families, and at this time there is limited data to support the hypothesis that a narrower activity spectrum gives a superior therapeutic index of efficacy over toxicity. The AXIS trial is the first to provide comparative data and more trials are in progress.



mTOR inhibitors are semi-synthetic derivatives of the antifungal agent sirolimus, a product of a Streptomyces species discovered in a soil sample collected on Easter Island (now Rapa Nui, hence rapamycins or rapalogs). mTOR is a highly conserved kinase at a key regulatory locus on the PI3K-Akt-mTOR pathway with complex linkages to other pathways affecting cell cycle progression and angiogenesis. Temsirolimus, a prodrug of sirolimus, given by weekly i.v. infusion was examined in advanced RCC for dose-response [24], with the lowest dose of 25 mg projected from the in vitro inhibition of mTOR, the unique target for rapalogs. There was no advantage to the higher dose arms so the low dose was taken forward. An international phase III study of temsirolimus vs IFNα was conducted in 626 systemically untreated patients with at least three of six adverse prognostic factors, a minority group in which the then available treatment with cytokines was considered minimally effective [25]. Despite a low ORR by RECIST criteria, other outcomes were superior for temsirolimus, including PFS, HRQL, and the primary endpoint of OS (median OS 10.9 vs 7.3 months, HR 0.73, P= 0.008). Lactate dehydrogenase was both prognostic and predictive for survival benefit [37].


Everolimus is the first oral mTOR inhibitor to be evaluated in RCC, and has a different active form from temsirolimus. RECORD-1 compared everolimus 10 mg daily with placebo in 410 patients with progressive disease ≤6 months of sunitinib and/or sorafenibtreatment [26]. The primary endpoint of PFS by independent central review was improved (median PFS 4.9 vs 1.9 months, HR 0.33, P < 0.001). RECIST-defined responses were infrequent. Overall HRQL was neither impaired nor improved, but there was a delay in performance status decline [38]. OS was the same in both arms, although everolimus was used in 76% of placebo-assigned patients after disease progression.

Adverse events with mTOR inhibitors

No direct toxicity comparisons are available for mTOR inhibitors. The parent drug sirolimus is in wide use for immunosuppression after organ transplantation, and lymphopenia and atypical infections are class effects, as is non-infectious pneumonitis. Fortunately these effects have been manageable and reversible in most patients. Other common toxicities include elevation of serum cholesterol, glucose, and triglycerides.


Epidermal growth factor receptor (EGFR) inhibitors

Lapatinib is an inhibitor of the EGFR and Her-2. As second-line treatment after cytokine failure, lapatinib was not superior to hormone therapy, considered a placebo equivalent [27]. Interestingly, a pre-planned subset analysis of patients with tumours that strongly overexpressed EGFR showed a trend to improved OS (P < 0.012). We are not aware of current studies of lapatinib for RCC. Another phase II study of BEV with or without the EGFR inhibitor, erlotinib, reported no difference in response or progression for the combination [28].


Sorafenib combinations are discussed under VEGFR inhibitors above. Non-randomised studies are finding high rates of adverse events with targeted agent combinations. TORAVA, a three-arm phase II study, compared BEV plus temsirolimus to standard sunitinib and to BEV + IFNα and encountered troublesome toxicity in the investigational arm that prevented useful treatment delivery [29].

Nephrectomy status

Three published studies with an IFNα control provide prospective data for ‘on-study’ nephrectomy status, that is, patients were stratified for this variable and outcomes are separately reported. Temsirolimus showed a survival benefit over IFNα for 138 poor-risk patients who had not undergone nephrectomy (HR 0.6, P < 0.05, values estimated from graph) [25]. BEV + IFNα also improved survival in 112 patients that were not nephrectomised (HR 0.65, P= 0.04) [11]. The same trend was seen for sunitinib (HR 0.79, P < 0.05) [19]. Therefore prior nephrectomy does not appear to be essential for benefit from targeted therapy with either VEGFR or mTOR inhibitor therapies, although patients who did not undergo a nephrectomy could have different important characteristics from the group that had the nephrectomy. Randomised trials are currently examining the role and timing of nephrectomy for sunitinib-treated patients (NCT00930033, NCT01099423).



The completeness of ascertainment of identified eligible trials was checked against a trials compilation ( Only three as yet unpublished additional trials were identified that appeared to have completed accrual for >1year (NCT00423332, NCT00244764, and NCT00606866) and selective publication of positive trials is considered unlikely. A limitation of the published studies is that, with the exception of BEV + IFNα, trials that establish standard of care have not been replicated. Industry involvement in designing, sponsoring, supervising, and co-authoring many studies, limits confidence that interpretation of results is always free of bias. The quantitative benefits of an intervention cannot safely be extrapolated beyond the patient group evaluated, although improved outcomes were usually consistent across subsets within pivotal studies.


Earlier studies of specific VEGF [12,13,33] and mTOR [25] inhibitors used OS as the primary endpoint, and this was achievable when there were no good alternatives, especially for poor-prognosis patients and not yet evidence of efficacy in the investigational arm to encourage or ethically require crossover. Within-trial crossover and earlier deployment of multiple lines of systemic therapy has made accurately assessing the impact of these therapies on survival problematic. Methods have been used to adjust OS for crossover effects [13,26] but we defer discussion until such time as there is consensus on this. PFS is a surrogate outcome of undetermined value as a predictor of OS or HRQL but has become accepted by regulatory authorities in the absence of a better alternative. However, data suggesting that PFS is a surrogate for OS [39] has not been universally confirmed. Absence of differences in time-to-disease progression and PFS can predict an absence of significant survival differences but existing differences in surrogate parameters may not predict existing survival differences. Tumour shrinkage is a measure of biological activity, and is presumably useful for palliation of cancer-related symptoms. No placebo-controlled trial has reported a HRQL benefit in this series.


Sunitinib [19] or BEV + IFNα[11,12] have been validated against prior standard interferon in good–intermediate-risk patients for multiple outcomes including OS. The TORAVA trial included these as control arms, and the preliminary report suggests a higher rate of grade 3/4 toxicities for BEV + IFNα than sunitinib (27 vs 14%) [29]. Pazopanib gives similar PFS to sunitinib but was compared with placebo rather than IFNα. For poor-prognosis patients, temsirolimus improves median survival by 3.6 months.


After first-line cytokine: based on preliminary data, axitinib [23] may be superior to sorafenib [13]. Pazopanib is an available alternative [21]. After first-line sunitinib: everolimus [26]; axitinib [23] appears to have a higher response rate. After first-line sorafenib: everolimus yields prolonged PFS, but few objective responses, and unchanged overall HRQL. After first-line temsirolimus or BEV+IFNα: no trials available.


No randomised trial has been completed for this group. Some trials included patients with non-ccRCC but none prospectively stratified for histological subtype. In an exploratory analysis of the non-ccRCC subset of poor-prognosis patients, temsirolimus was superior to IFNα for PFS (HR 0.38, P < 0.05) and OS (HR 0.49 P < 0.05) [40].


Agents targeting VEGF and mTOR pathways improve PFS in both first-line and second-line settings. These treatments rarely yield complete responses and are not curative. No placebo-controlled trial has reported a HRQL benefit.


None declared.