To Treat or Not to Treat Adjuvantly in RCC: That Is the Question

To Treat or Not to Treat Adjuvantly in RCC: That Is the Question

Dr. Lauren C. Harshman

Dr. Rana R. McKay

Dr. Toni K. Choueiri, credit Dana Farber Cancer Institute

By Lauren C. Harshman, MD; Rana R. McKay, MD; and Toni K. Choueiri, MD

Article Highlights

  • The S-TRAC trial results revealed that 1 year of adjuvant therapy with sunitinib may increase disease-free survival in high-risk, non-metastatic clear cell RCC. However, there is controversy over why this study differed from the negative findings of the ASSURE study.
  • At least four other randomized controlled adjuvant studies evaluating VEGF and mTOR pathway blockade have completed accrual (SORCE, PROTECT, ATLAS, and EVEREST) and will provide further insight into the current controversy.
  • Many questions abound regarding the optimal perioperative strategy including a pure adjuvant versus neoadjuvant approach, the need for both pre- and post-resection systemic therapy, predictive selection criteria, and whether combinations of drugs would be more effective than monotherapy or sequential administration.

After decades of negative trials and countless patient data and clinical resources that have gone toward finding an effective adjuvant treatment for non-metastatic renal cell carcinoma (RCC), a beacon of hope has surfaced with the results of the S-TRAC trial. These results showed that 1 year of adjuvant therapy with VEGF blockade in the form of sunitinib may increase disease-free survival (DFS) in non-metastatic, high-risk, clear cell RCC.1 However, the positive result is somewhat tempered by controversy over why this industry-sponsored study differed from the negative findings of the first-in-class, larger, National Clinical Trials Network ASSURE study (E2805, NCT00326898).2 The conflicting results leave us with many questions: Does adjuvant sunitinib therapy really eliminate micrometastatic disease, is it for everyone, and if so, is the benefit in delaying recurrence large enough to balance the quality-of-life costs that may accompany this therapy?

Following multiple negative adjuvant studies of cytokine-based immunotherapeutics such as interleukin-2, interferon-alpha, and various vaccine strategies, and aiming to capitalize on the advent and broad-based clinical effectiveness of the targeted therapies in advanced RCC, the field moved its focus to the adjuvant administration of VEGF- and mTOR-targeted therapies.

Comparing ASSURE and S-TRAC

The ASSURE trial, which pitted 1 year of sunitinib plus placebo against 1 year of sorafenib plus placebo or 1 year of placebo/placebo in patients who had undergone partial or radical nephrectomy for clear cell or non–clear cell RCC, was the first of the adjuvant VEGF-targeted therapy trials to release results.2 After a median follow-up of close to 6 years, median DFS among the three arms was not significantly different at 5.8 years, 6.1 years, and 6.6 years, respectively. Haas et al.2 highlighted the significant toxicity induced with the adjuvant use of VEGFR tyrosine-kinase inhibitors (TKIs). Reduced dosing still incurred substantial toxicity, and a significant proportion of patients discontinued TKI therapy because of “excessive toxicity.”

S-TRAC, the second of the trials to report results, was an industry-sponsored study that tested 1 year of adjuvant sunitinib versus placebo after nephrectomy in patients with high-risk, clear cell disease. It revealed a close to 1-year benefit in DFS with adjuvant sunitinib administration (6.8 years vs. 5.6 years; HR 0.76, 95% CI [0.59, 0.98]).1

Whereas the negative ASSURE trial encompassed a broader swath of patients including those with lower-stage disease (approximately 37% pT1b and pT2 disease) and approximately 20% non–clear cell histologies, S-TRAC restricted its patient population to clear cell histology and higher-risk T3 and T4 disease. The obvious question is whether the benefit in adjuvant VEGFR blockade is limited to the higher-risk clear cell subsets. However, even upon subgroup evaluation of patients in the ASSURE trial, the lack of benefit persisted for the higher-risk pT3 and pT4 (HR 1.04, 97.5% CI [0.83, 1.31]) and clear cell (HR 1.01, 97.5% CI [0.82, 1.25]) subsets.2

An additional consideration is that dose may be critical to achieving efficacy. Given the degree of toxicity and treatment discontinuations initially observed, the ASSURE study was amended to decrease the starting dose of sunitinib to 37.5 mg (with the option to dose escalate after ensuring tolerability), whereas 100% of patients in the S-TRAC trial received the full starting dose. Indeed, in a subgroup analysis of the patients in the ASSURE trial, starting at the full dose showed a trend to benefit from adjuvant treatment.2 However, it is notable that in either study, less than 60% of patients completed a full year of adjuvant sunitinib, in large part because of toxicity.

The question of whether DFS is the optimal primary endpoint also looms. In the ASSURE study, overall survival (OS) was not significantly improved with adjuvant sunitinib or sorafenib. Despite a median follow-up of 5.4 years and a patient death rate of approximately 20% in each arm, the S-TRAC authors reported that the data were immature for the OS secondary endpoint, as median OS had not been reached. However, they did estimate that the hazard ratio for OS between the arms was 1.01 (95% CI [0.72, 1.44]).1  Interestingly, median follow-up on the ASSURE study was quite similar at 5.8 years when the Data Safety Monitoring Committee concluded that further blinded follow-up was highly unlikely to alter the results and recommended that the study findings be released.

Ultimately, if OS is positive in S-TRAC, it would provide further support for this approach despite the quality-of-life disadvantage of adjuvant sunitinib. However, if it is negative, it suggests that delayed administration of systemic therapy only in patients in whom disease recurs may be equally effective and preserve upfront quality of life post-nephrectomy. The latter approach requires that close long-term surveillance be consistently instituted.

The long durations required to prove a survival benefit in the adjuvant setting highlight the critical need to identify intermediate endpoints that will shorten the duration of our adjuvant investigations. For example, if we knew today that 5-year DFS is a good surrogate for OS, perhaps we would be more willing to immediately incorporate adjuvant sunitinib into clinical practice. In a meta-analysis of 10 adjuvant studies spanning the last 50 years, which had adequate published data on median DFS and OS, DFS had only a moderate correlation with OS at 5 years. (Harshman et al., personal communication, GU Cancers Symposium Abstract 459). Earlier endpoints such as 3-year DFS correlating with OS would be even more ideal.

Choosing the Optimal Agent, Perioperative Strategy

As we tease out the nuances of the results from the first two VEGFR inhibitor adjuvant studies, we are comforted by the fact that there are at least four other randomized controlled adjuvant studies evaluating VEGF and mTOR pathway blockade that have completed accrual and that will provide further insight. These include SORCE (1 year vs. 3 years of sorafenib; NCT00492258), PROTECT (1 year of adjuvant pazopanib; NCT01235962), ATLAS (1 year of adjuvant axitinib; NCT01599754), and the only mTOR-targeted study, EVEREST (1 year of adjuvant everolimus; NCT01120249). Future meta-analyses of these studies are likely to be the most germane way to end the current controversy.

Supposing that these subsequent studies confirm the positivity of S-TRAC, the next step would be to identify what the optimal agent is in terms of efficacy, duration, and tolerability. One could imagine embarking on the next set of studies for the adjuvant setting akin to COMPARZ and RECORD-3, pitting one agent against the other in large studies that take countless years.3,4

A more parsimonious approach would be to assume the amount of benefit is similar across the VEGF-targeted agents and that clinical practice restrictions (e.g., “pathways”) will be limited to one agent of choice. In that case, we could extrapolate from the past comparative phase III studies in the metastatic setting that demonstrated that sunitinib was more effective than everolimus and that pazopanib and sunitinib are comparable in efficacy.

Detractors of this approach could reasonably counter that the biology of metastatic disease does not mirror the earlier, perhaps less-complex disease being targeted in the adjuvant setting, and that there could be different mechanisms of innate or acquired resistance at play in the earlier disease states.  These questions highlight the need to learn from the numerous and costly biomarker correlative studies built into each of these large adjuvant studies. We should amass and interrogate these precious data to identify predictive signatures of recurrence and/or response to the various agents studied. In an ideal world, we would merge the tissue and sera data to perform large meta-analyses that would accomplish these objectives.

Even with its positive outcome for DFS,  S-TRAC illustrates that we continue to have much room for improvement: Despite the 1 year of adjuvant sunitinib therapy, approximately 37% of patients developed recurrence, developed a second cancer, or died.1 The oncology field, including RCC, has shifted a large proportion of its focus to the latest immuno-oncology agents. PD-1 pathway inhibitors are the frontrunners, and nivolumab, an antibody against PD-1, has received U.S. Food and Drug Administration approval in metastatic RCC that is refractory to VEGF-targeted therapy.5 In addition, the high response rates achieved by combining PD-1 pathway blockade with anti–CTLA-4 or VEGF-targeted therapies are also promising and will be in even greater demand if the ongoing and maturing frontline metastatic trials bear out. These studies include investigations of ipilimumab plus nivolumab versus sunitinib (NCT02231749), atezolizumab in combination with bevacizumab versus sunitinib (NCT02420821), and avelumab in combination with axitinib versus sunitinib (NCT02684006).

The choice of optimal immuno-oncologic agent aside, many questions abound regarding the optimal perioperative strategy. Should it be a pure adjuvant or neoadjuvant approach, or do we need both pre- and post-resection systemic therapy? Are combinations of drugs more effective than sequential administration? Finally, should we select patients based on a tumor biomarker such as PD-L1, mutational load, or genetic or immunologic signatures?

Prediction of optimal candidates for adjuvant therapy based on risk and markers of response is an active area of investigation that is in its nascency. There are no validated biomarkers or tumoral or immunologic signatures that are ready for primetime. Despite substantial promise in terms of predicting response in the phase I and II studies that tested the use of nivolumab in metastatic RCC,5-8 the lead candidate, tumoral PD-L1, was not found to be predictive of survival in the phase III study of nivolumab for the advanced treatment of refractory metastatic RCC.5

In terms of feasibility, the pure adjuvant approach is the standard paradigm. Clinicians are comfortable with this strategy, and it does not alter workflow. However, if we think about the mechanism of PD-1/PD-L1 blockade and factor in some intriguing data from mouse models, it makes sense to consider the need for priming the immune system with PD-1 blockade before surgical resection, in other words, the neoadjuvant approach.  Work from the Drake Lab using murine models has shown that PD-1 blockade prior to resection enhances the proliferation of antitumor CD-8 T cells in the primary tumor, tumor-draining lymph nodes, and tumor microenvironment.9,10 Campbell and colleagues have observed a sharp decline in PD-1–positive effector T cells after nephrectomy,11 suggesting a loss of the potential T-cell army. Most recently, using a murine model, Liu and colleagues showed a significant benefit in survival with neoadjuvant PD-1 blockade administration compared to an adjuvant approach.12 These human and murine studies suggest that the immune system is less primed for response to adjuvant stimulation after the primary tumor has been resected. 

Neoadjuvant and Pure Adjuvant Approaches

The National Clinical Trials Network Team spanning investigators from ECOG-ACRIN, the Alliance, SWOG, the National Cancer Institute of Canada, and Radiation Therapy Oncology Group, have united to study the three-pronged approach of presurgical priming, resection, and consolidation with adjuvant PD-1 blockade in the PROSPER RCC study (Fig. 1). Armed with the aforementioned preclinical data and knowledge of PD-1 pathway mechanism of action that is reliant on tumor antigen, the PROSPER investigators believe that priming of the immune system with PD-1 blockade with the tumor in situ is rational and necessary for effective PD-1 pathway blockade in the M0 setting.

Ongoing studies are providing preliminary evidence of the safety and feasibility of neoadjuvant nivolumab in non-metastatic RCC with three to four doses prior to nephrectomy (personal communications from M. Allaf October 25, 2016, [NCT02575222] and presented during the 2016 International Kidney Cancer Association Symposium meeting), and personal communications from M. Voss October 27, 2016 [NCT02595918]). Further, this neoadjuvant approach affords the greatest opportunity to improve patient outcomes with correlative science that will study the effects of PD-1 priming on the immune milieu of the tumor and sera and how these changes—or, more simply, the patient’s baseline tumor and immune environment—correlate with clinical outcomes. We believe that this three-dimensional approach to high-risk RCC has the greatest potential to identify predictive genetic and immune signatures, which will aid in our selection of the ideal patient for perioperative immunotherapy administration and move us beyond our current rudimentary guides of stage, histology, and performance status.

Other industry-sponsored studies are in development that will undertake the pure adjuvant approach, with the belief that it is the most feasible approach and that dosing after primary tumor removal will be sufficient. In the interim, ongoing and planned presurgical studies including some in the cytoreductive nephrectomy setting will serve as proof of concept that the neoadjuvant administration of immunotherapeutics is safe and feasible, and that the agents are working through the mechanisms we intend. These presurgical studies may also uncover potential resistance pathways and identify potential selective biomarkers. Some of the studies that are furthest along in this space include presurgical blockade with combinations of anti–PD-1 and anti–CTLA-4 therapies such as durvalumab with or without tremelimumab (NCT02762006) and nivolumab with or without ipilimumab or bevacizumab (NCT02210117).

The future is bright, and 2017 promises to be an exciting year, with at least two large randomized controlled trials launching that will test the perioperative administration of promising immuno-oncotherapeutics in RCC. The wealth of pre- and postsurgical studies of immune-stimulating agents with different designs and scientific underpinnings must be complementary and enhance the outcomes of our patients. 

About the Authors: Dr. Harshman is an assistant professor of medicine at Harvard Medical School and the co-director of the Kidney Cancer Center at Dana-Farber/Brigham and Women’s Cancer Center. She is the principal investigator of the PROSPER RCC study. Dr. McKay is a medical oncologist who has transitioned from the Dana-Farber Cancer Institute to the University of California, San Diego, where she is an assistant professor of medicine focused on genitourinary cancers. Dr. Choueiri is the director of the Lank Center for Genitourinary Oncology at the Dana-Farber Cancer Institute and the Jerome and Nancy Kohlberg Chair in Medicine at Harvard Medical School.