Naomi B. Haas, MD
Conflicting results from large adjuvant studies for renal cell carcinoma (RCC) may have been caused by differences between the clinical trial designs, patient populations, and assessment criteria, according to Naomi B. Haas, MD, who discussed the data from these trials at the 2016 International Kidney Cancer Symposium.
In order to fully understand the viability of adjuvant therapy, the nuances that likely impacted the outcomes of each trial need to be fully explored, including differing risk groups, populations, dose ranges, starting dose, duration, endpoints, and statistical definitions of activity, noted Haas, director of the Prostate and Kidney Cancer Program at the Hospital of the University of Pennsylvania.
To date, 3 trials have published results: ARISER, ASSURE, and S-TRAC. Additionally, there are currently several ongoing adjuvant small-molecule inhibitor trials ongoing (SORCE, EVEREST, PROTECT, ATLAS, and E2810), most of which are nearly completely accrued, ensuring the release of a great deal of information within the next year or two.Confounding Adjuvant Trials
The ASSURE trial enrolled 1,943 patients with both clear cell RCC (ccRCC) and non-ccRCC, though the study was powered for ccRCC. Patients received either 50 mg/day of sunitinib, 400 mg twice daily of sorafenib, or placebo for 1 year. Following the accrual of 1,300 patients, however, investigators noticed a high drop-out rate from intolerance and side effects, and responded by amending the trial to start patients at a lower dose of sunitinib to 37.5 mg/day.
Overall, the study showed no difference in disease free survival (DFS), which was the primary endpoint, or overall survival (OS) between sorafenib or sunitinib and placebo. The hazard ratio for both treatments was near 1.0 across all populations enrolled.1
The S-TRAC study was a 1-year trial that enrolled only patients with high-risk ccRCC. Patients were stratified by the UCLA international staging criteria, and only patients with pT3 disease or higher were enrolled. Patients received 50 mg/day of sunitinib, with the option to reduce the dose to 37.5 mg/day if needed.
For this study, the primary endpoint was DFS by blinded independent central review, and sunitinib had a statistically significant advantage over placebo. Remarkably, the curves for sunitinib and placebo stay well separated both 3 and 5 years out.2
The third completed clinical trial, ARISER,3
was a 6-month study of carbonic anhydrase antibody girentuximab versus placebo. Patients were divided into three high-risk groups: pT3/pT4Nx/N0M0, pTanyN+M0, or pT1b/pT2Nx/N0M0. The investigators evaluated carbonic anhydrase membrane staining; however, concerns surrounding validation prohibited the use of membrane staining as an endpoint. Instead, the staining was used only for staging.
Overall, patients treated with girentuximab had no statistically significant DFS compared with placebo. However, patients with low carbonic anhydrase-IX (CAIX) scores (0-100) were more likely to recur and die than those with high CAIX scores (201-300).4Examining Patient Characteristics
These contrasting results, specifically for the studies exploring VEGF-directed therapies, raised several questions regarding adjuvant therapy, with answers that may be found within the trials themselves. “We might look at different parts of the trials and how we might learn something,” proposed Haas.
Patient populations in these studies were selected based on histopathology and risk nomograms. While the ASSURE study included both ccRCC and non-ccRCC, and patients with lower stages of disease, in the S-TRAC study patients all had ccRCC and ST-3 disease or higher.
There were differences between the 2 trials in terms of what measures are used as risk nomograms. Further, American Joint Committee on Cancer (AJCC) criteria for staging has been refined and imaging technology has improved since some of the recurrence data was collected, raising the possibility that micrometastatic disease may not have been detected in earlier studies. In spite of these advances, it remains difficult to assess risk of recurrence in patients with pT2 and pT3 disease, where risk of recurrence could be 20% to 60%. Also confounding a direct comparison was the fact that the S-TRAC trial used central radiology review while the ASSURE trial did not.
One weakness of DFS as a sole endpoint is that the frequency of scanning is not uniform among studies. Though more frequent scanning may lead to the end of the study more quickly, it may limit the ability to compare DFS between studies. For adjuvant therapy, cure is a laudable goal, and future studies should be powered for OS, a step that requires regulatory support.