Encouraging successes have ushered in the era of targeted therapy for advanced cancers. Starting with great bench-to-bedside advances such as rituximab for lymphoid malignancies and imatinib for chronic myelogenous leukemia, these transformative changes have recently extended to the historically barren therapeutic landscapes of melanoma and lung cancer with ipilimumab, erlotinib, crizotinib, and others. Nowhere has the paradigm shifted toward targeted therapy as convincingly and as broadly as it has in renal cell carcinoma (RCC).
We have learned, to date, of 2 pathways with undeniable therapeutic importance in RCC: the vascular endothelial growth factor (VEGF) and mammalian target of rapamycin (mTOR) pathways. Each comes with compelling preclinical evidence that has been robustly translated into clinical evidence, and has over the past 5 years resulted in new approaches to the treatment of all patients with advanced RCC. Six individual drugs have been approved by the US Food and Drug Administration for the treatment of patients with metastatic RCC, including 4 that target the VEGF pathway and 2 that target the mTOR pathway.1-6 More are undoubtedly coming in the near future. Individually, each of these drugs has proven to delay progression of disease and in some cases to prolong survival, while maintaining an acceptable safety profile. Collectively, they have proven the principle of targeted therapy, over and again, and have resulted in true benefit that we can observe in the clinic on a daily basis.
Implied within the framework of targeted therapy is the notion that “collateral damage” is absent, or minimized. The specificity of targeted therapy is responsible for much of its appeal, and is the sine qua non that differentiates it from cytotoxic chemotherapy. The lofty goals of targeted therapy are to ascend the twin peaks of efficacy and safety by inhibiting the target of interest, and only the target of interest. And yet, the results in this area have been somewhat disappointing. First, even by inhibiting only the target of interest, we encounter inevitable toxicity, exemplified by indisputably VEGF-related adverse effects with bevacizumab such as hypertension, proteinuria, thrombosis, bleeding, and others. Second, many targeted therapies have been recast as “multitargeted” therapies, primarily by virtue of the multiple toxicity pathways. The cutaneous difficulties engendered by sorafenib, sunitinib, and other multitargeted inhibitors are an example. Off-target effects frequently lead to dose reduction, treatment interruption or discontinuation, and eventually disease progression.
The disappointment related to toxicity is magnified as we turn now to combination targeted therapy. Given the importance of VEGF and mTOR pathway inhibition, there is obvious and compelling rationale to combine the two. Ideally, we would create a solid tumor analog to cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) chemotherapy for lymphoma. We would treat patients with drugs that may be mildly effective individually but synergistic when coadministered, to the point of durable (complete) responses. Moreover, because these therapies are targeted, we would hope that the minimal toxicities of each would not overlap in the combination. Unfortunately, the reality has been unexpected and serious difficulty in combining any of these pathway inhibitors. This is demonstrated again by the report from Molina et al in this issue describing the combination of the VEGF receptor inhibitor, sunitinib, and the mTOR inhibitor, everolimus.7
The current report notes substantial dose-limiting acute and chronic toxicities in patients treated with sunitinib and everolimus, at reduced or modified doses of each of these drugs. It is therefore fairly clear that this is not a successful combination regimen. Similar challenges have been encountered with other combinations (Table 1).7-15 Sunitinib and temsirolimus resulted in 2 dose-limiting toxicities among the first 3 patients at the initial dose level and the study went no further.8 Although in some instances “tolerable” dosing regimens have been identified in phase 1 trials, these regimens as a rule have mandated dose reductions, usually of each drug, and still have resulted in excessive toxicity relative to monotherapy. In the TORAVA trial reported in 2010, a full 40% of patients receiving the combination of bevacizumab and temsirolimus discontinued treatment because of toxicity. That study produced a progression-free curve for the experimental arm that appeared to be inferior to each of the 2 control arms, sunitinib and bevacizumab/interferon.16
|VEGFR Inhibitor||VEGF Inhibitor||mTOR Inhibitor||Dose (mg)a||Study|
|Sorafenib||—||Temsirolimus||S: 200, T: 25||Patnaik 20079|
|Bevacizumab||Temsirolimus||B: 10, T: 25||Merchan 200710|
|Sunitinib||—||Everolimus||Current study||Molina 20117|
|Sorafenib||—||Everolimus||S: 400, E: 5||Harzstark 201111|
|—||Bevacizumab||Everolimus||B: 10, E: 10||Hainsworth 201012|
|Sorafenib||Bevacizumab||—||S: 200, B: 5||Sosman 200814|
Why have we observed this degree of unexpected toxicities across multiple combinations of targeted therapies? Molina et al have addressed and ruled out the possibility of pharmacokinetic interaction in their study.7 The same has been true in other combination studies. Amplification of overlapping toxicities does not appear to fully explain the findings either. Instead, the more common observation is that a typically mild adverse effect of one drug is exaggerated by the addition of the second drug. Although some of the adverse events in the study by Molina et al,7 such as thrombocytopenia and mucositis, do fit into the category of overlapping toxicities, many other adverse events do not. Notable examples include endocarditis, gastrointestinal bleeding, neutropenic fever, pulmonary embolus, and others. Other clinical trials have also demonstrated this phenomenon of magnification of the toxicity of drug A through the addition of drug B, through unexplained biological mechanisms rather than through pharmacokinetic interactions.
The notion of on-target versus off-target toxicity remains a source of hope for some. If combination approaches were restricted to agents that are more specifically targeted to their pathways of interest, perhaps the twin peaks of efficacy and safety could be conquered. Although multitargeted inhibitors may pose too great a toxicity challenge, combination therapy may be more feasible with agents that are highly specific for VEGF, including bevacizumab or others currently in development such as aflibercept,17 tivozinib,18 or axitinib.19 Among mTOR inhibitors, continued research may reveal more precisely targeted drugs that affect different proteins in the pathway, including phosphoinositide 3- kinase, Akt, or others, without causing such adverse effects as mucositis, pneumonitis, and hyperlipidemia. Combination trials could then assess these improved agents (more specific drugs and/or more specific targets) and move the field forward.
In the current report by Molina et al,7 a notable finding was that 3 of 6 patients with non-clear cell RCC achieved partial responses, along with disease control for more than 1 year. This may suggest that everolimus has particular activity in non-clear cell RCC, or that the combination has synergistic efficacy in this population. Other hints of improved efficacy have been observed with combination treatments in other studies, such as the promising activity of temsirolimus and bevacizumab in a phase 1/2 trial.20 Another combination, this time with VEGF ligand and VEGF receptor inhibition using bevacizumab and sunitinib, demonstrated responses in multiple cancers but was plagued by the development of microangiopathic hemolytic anemia.13 An important lesson learned has been to resist overinterpretation of early results. For example, the initial presentation of the combination of temsirolimus and bevacizumab noted 8 responses among 12 patients, and updated results 2 years later revealed a modest response rate of 16%.20 A similar phenomenon was observed with the combination of everolimus and bevacizumab, with the median progression-free survival going from an initial 11 months to a final 7.1 months.12 In each case, large follow-up trials were undertaken, perhaps prematurely, after the initial results were presented.
A final issue to consider is more rational selection of patients. Conceivably, a subset of patients with RCC could benefit from combination therapy whereas others may be better served with monotherapy. There are new classification schemes in the era of targeted therapy, which are comparable but distinct from past models.21 These schemes may help to select patients at good, intermediate, and poor risk with regard to prognosis, and may start translating into recruitment criteria for clinical trials. Research into predictive factors of response is ongoing but has not yet resulted in clear therapeutic algorithms. For the most part, the large ongoing trials are recruiting unselected populations of patients with RCC.
Currently, monotherapy remains the standard of care in patients with RCC. Combination therapy has been characterized by excessive toxicity without demonstrating durable complete responses or impressive clinical benefit. Incremental improvements in surrogate endpoints such as response rate or even progression-free survival for combination versus monotherapy should not prompt a change in the standard of care. Rather, we should require proof that combination therapy improves patient outcome in a clinically meaningful way compared with the less toxic and better tolerated approach of sequential monotherapy. Clinical trials, directed by deeper biologic insight into RCC and involving rational patient selection, remain crucial in helping to advance the treatment of patients with RCC.