Despite positive clinical trial results, the role of immunotherapy in renal cell carcinoma remains unclear.
David McDermott, MD
Is immunotherapy the future of first-line treatment for advanced renal cell carcinoma (RCC)? The answer, according to David F. McDermott, MD, leader of the kidney cancer program at DanaFarber/Harvard Cancer Center, is “maybe.”
“Immunotherapy is already a standard of care for those who fail on a VEGF TKI [tyrosine kinase inhibitor]. With CheckMate-214, it almost certainly will be FDA approved in the frontline setting. Whether that indication will be for the intermediate- and poor-risk groups or for the all-comers, we must wait and see—but it will be standard of care in 2018, no doubt. Even if it is standard of care, it still won’t benefit every patient, and there are still some patients who may get more benefit from VEGF strategies,” he said.
McDermott outlined the potential for immunotherapy in RCC at the 2nd Annual International Congress on Immunotherapies in Cancer™, hosted by Physicians’ Education Resource®, LLC, in New York, New York, in December 2017. Despite positive clinical trial results, the role of immunotherapy in RCC remains unclear, he said. Cost of treatment is a huge concern and physicians must develop biomarkers to identify the patients most likely to benefit from particular therapies. Further, he said, physicians need to learn if combination therapy can overcome innate and acquired resistance and provide durable benefits.
To show where the field might be headed, McDermott reviewed results from the phase III CheckMate-214 trial1 evaluating PD-1/CTLA-4 blockade in kidney cancer and the 3-arm phase II IMmotion 150 trial.2
In CheckMate-214, treatment-naïve patients with advanced or metastatic clear-cell RCC were randomly assigned to 3 mg/kg nivolumab (Opdivo) and 1 mg/kg ipilimumab (Yervoy) every 3 weeks for 4 doses, followed by 3 mg/ kg nivolumab every 2 weeks (n = 425) or 50 mg daily oral sunitinib (Sutent) for 4 weeks in 6-week cycles. Patients were stratified by risk group and region, and treatment was continued until disease progression or unacceptable toxicity.
Across the full intent-to-treat population (ITT), the median PFS was not improved (12.4 vs 12.3 months; HR, 0.98; 99.1% CI, 0.79-1.23; P = .8498), results that were announced at the 2017 ESMO Congress. PFS in the intermediate- and poor-risk group was 11.6 months with the combination versus 8.4 months with sunitinib (HR, 0.82; 99.1% CI, 0.64-1.05; P = .0331). A P value of .009 was required for significance (TABLE).
Favorable risk patients, in contrast, had a significantly higher confirmed objective response rate (ORR) with sunitinib versus the combination arm (52% vs 29%, P = .0002), as well as a significantly longer PFS (25.1 vs 15.3 months; P <.0001).
CheckMate-214 was among the first clinical trials to show improved OS in kidney cancer. Median OS for intermediate/poor risk patients was 26 months in the sunitinib arm and not reached in the combination arm (HR, 0.68; 99.8% CI, 0.44-0.89; P <.0001).
“Why do these endpoints start to merge when you include the good-risk patients? If you look at the good-risk patients alone [n = 249], the response rate strongly favors sunitinib [52% vs 29%; P = .0002]. This is some of the best sunitinib data we’ve ever seen,” McDermott said. Median PFS for good risk patients similarly favored sunitinib at 25.1 months versus 15.3 months for the combination (HR, 2.18; 99.1% CI, 1.29-3.68; P <.0001).
“There are some patients, based on their biology, who might be better off getting a VEGF-TKI first and others who might benefit more from immune combination strategies,” he said.
He noted that the combination produced fewer grade 3/4 toxicities, which he suggested could be due to CheckMate-214 participants receiving more nivolumab and less ipilimumab compared with patients in melanoma trials. McDermott said the safety outcomes could have implications for other solid tumor types such as lung and bladder cancers.
However, he said, it is premature to suggest the nivolumab-ipilimumab combination will become the standard-of-care in first-line treatment. “I would put a caveat on this data before we get too excited. When you look at the ITT analysis of all patients—good-, intermediate-, and poor-risk— the benefits are not quite as dramatic,” he said. “The difference in response still favors nivo-ipi, but it’s not significant. You see PFS is identical when you add the good-risk patients. OS is still improved in the ITT population, which is obviously the most important endpoint.”Turning to IMmotion 150, a study in which he was a coauthor, McDermott said the study produced “some of the best data we have in kidney cancer.” Trial results showed that single-agent PD-L1 blockade is “very” active in untreated metastatic RCC (mRCC), he said, and VEGF blockade may be beneficial with immune-oncology therapy, findings that could be important in other cancers.
Treatment-naïve patients with mRCC were randomly assigned to 1200 mg of atezolizumab (Tecentriq) and 15 mg/kg of bevacizumab (Avastin) every 3 weeks on days 1 and 22 of each 6-week cycle until disease progression.
Median investigator-assessed PFS by Response Evaluation Criteria in Solid Tumors (RECIST) was 7.8 months with sunitinib, 5.5 months with atezolizumab, and 11 months with atezolizumab and bevacizumab combined (HR, 0.88; P = .44). By immune-modified RECIST (imRECIST) criteria, median PFS was 9.9 months with sunitinib, 8.5 months with atezolizumab alone, and 17.3 months with atezolizumab and bevacizumab (HR, 0.78; P = .16).2
The most important lesson, McDermott said, came from the biomarker strategy used in IMmotion 150. The biomarker strategy for atezolizumab trials, he said, should include RNA sequencing, whole exome sequencing, and immunohistochemistry. The research team combined these assays into a biomarker model that incorporated mutational load, CD8 density, PD-L1 expression, histology, interferon-gamma signature, and RNA sequencing to identify patients who would be most likely to benefit.
“If we look at PD-L1 expression, for the first time it was shown that, for single agents and combinations, the more PD-L1 in the tumor microenvironment, the more likely you were to benefit from immune therapy,” he said. “Also, maybe not surprisingly, if you had a high T-effector signature by RNA sequencing, you were more likely to benefit from immune therapy.”
He added that patients with a high angiogenic signature in the tumor—suggesting high levels of VEGF signaling—had better outcomes with sunitinib. Like the CheckMate-214 results, that suggests that physicians can identify patients who should get VEGF first and others who should get immune therapy first.
McDermott said that results from IMmotion 150 show that combining an RNA signature with a genetic mutation could further identify patients who would benefit from VEGF blockade. “If you put all this data together, you can imagine a world in which patients with an angiogenic signature get sunitinib,” he said. “Those with a T-effector signature who had myeloid-low expression—meaning low expression of genes associated with myeloid inflammation—[are] more likely to respond to atezo.”
Adding bevacizumab to atezolizumab seemed to improve both response and PFS in T-effector high, myeloid-high patients, suggesting that for them, bevacizumab may be acting as immune therapy and possibly overcoming resistance, he said. McDermott said that bevacizumab might have implications in other tumor types in which myeloid immunosuppression plays a key role.