Several drugs targeting VEGF or mTOR pathways have been approved for treatment of advanced renal-cell carcinoma because of improvements noted in progression-free survival (PFS) in phase 3 trials.1 Validation of prognostic models showed that treatment with such drugs can lead to a median overall survival of around 43 months for patients in favourable risk categories and 23 months for patients in intermediate risk categories.2 With few exceptions, patients on first-line therapy progress and proceed to need one or more subsequent lines of targeted therapy. In a population-based study,3 patients in a favourable risk group had progression on first-line VEGF-targeted therapy after a median of 16·6 months (compared with 15 months for patients in an intermediate risk group) and progression after 6·2 months on second-line targeted therapy (5·5 months for intermediate risk). Two phase 3 trials4, 5 assessed outcomes after failure of a previous VEGF-targeted therapy to establish evidence for the mTOR-inhibitor everolimus and the selective inhibitor of VEGF receptors 1—3, axitinib. The AXIS trial5 is the only study that directly compared two active compounds (axitinib vs sorafenib) after failure of an approved first-line regimen. In AXIS, more than a third of patients had received cytokines and over half had received sunitinib as first-line therapy. Axitinib led to an improvement in median PFS compared with sorafenib in the intention-to-treat analysis. However, the difference in PFS for patients after sunitinib treatment based on investigator and independent review committee assessments was only slight. Data for overall survival, a secondary endpoint, were immature before the first report was published in 2011. Because guidelines and clinical practice favour targeted therapy in preference to cytokines as first-line treatment,1 axitinib is regarded as a treatment option for second-line therapy of advanced renal-cell carcinoma.5

In The Lancet Oncology, Robert Motzer and colleagues6 now report mature overall survival data for the AXIS trial. Such an analysis is important because crossover between the two study arms was not allowed. No significant differences in overall survival were noted between patients in both treatment arms who received the same first-line regimen (median overall survival 20·1 months [95% CI 16·7—23·4] with axitinib vs 19·2 months [17·5—22·3] with sorafenib; hazard ratio 0·969, 95% CI 0·800—1·174, p=0·3744). More than half the patients in each arm continued with a third-line treatment after progression on study drug and treatment after progression was allowed. This design confounded overall survival results and raises questions as to whether PFS is meaningful in this setting.7 Third-line therapy partly explains the long time interval noted between progression on second-line treatment and overall survival. However, inclusion of patients with a less aggressive tumour biology might have contributed to this outcome. Only a third of patients in the AXIS trial were Memorial Sloan Kettering Cancer Center (MSKCC) poor risk at entry,5 suggesting that individuals with rapid deterioration of performance or accelerated progression during first-line therapy are less likely to enter trials than are patients with more favourable risk profiles.

For patients previously treated with sunitinib in Motzer and colleagues’ study,6 median time on first-line therapy was about 10 months, with a median overall survival for all risk groups of 15·2 months (95% CI 12·8—18·3) for axitinib and 16·5 months (13·7—19·2) for sorafenib. Patients who received cytokines had first-line therapy for about 6 months and a median overall survival of 29·4 months (24·5—not assessable) for axitinib and 27·8 months (23·1—34·5) for sorafenib. After correction for the different length of first-line therapies, overall survival seemed to be increased by about 7—9 months in patients who had cytokines before VEGF-targeted therapy. Resistance to previous VEGF-targeted therapy, which might not be apparent in patients previously untreated with such an approach, cannot fully explain this difference. Motzer and colleagues noted a putative association of overall survival with length of previous sunitinib treatment for both axitinib and sorafenib, although there was substantial overlap in the 95% CIs.6 A retrospective Database Consortium analysis of 464 patients who had received two lines of VEGF-targeted therapies reported no correlation between first-line PFS and second-line PFS.8 Rather, a significant difference in multivariate analysis of baseline prognostic factors in favour of cytokine versus sunitinib pretreatment (HR 0·503, 95% CI 0·395—0·641; p<0·0001) suggested that patients with less advanced disease were most likely to start treatment with cytokines.6 However, information about the distribution of prognostic factors between patients who were pretreated with cytokines and sunitinib, which could have important implications for treatment sequences, is not provided in Motzer and colleagues’ study.

Data from trials and population-based analyses suggest that a ceiling has almost been reached in terms of outcome with present targeted therapies and prognosis that relies on models based on clinical factors.2 The mature AXIS data add axitinib to the choices for second-line treatment with similar outcome and different toxicity profiles.4—6 Despite prognostic factors assessed in the updated analysis and a correlation of development of hypertension during axitinib and sorafenib treatment with overall survival, the choice for a second-line drug or even treatment beyond progression at failure of first-line treatment remains an educated guess. The outcome of this study proves once again that renal-cell carcinoma is a heterogenous cancer9 that needs further research into predictive biomarkers to tailor treatment choices.

Source: Lancet