Based on results from the phase 2 OMNIVORE study, the response-based strategy achieved no complete responses and a low partial response/complete response conversion in advanced renal cell carcinoma.
Although upfront dual checkpoint blockade is a suggested treatment for eligible patients with advanced renal cell carcinoma (RCC), administration of nivolumab (Opdivo) followed by 2 doses of ipilimumab (Yervoy) on a response-based strategy is not currently recommended due to the limited number of responses induced by this approach. Based on results from the phase 2 OMNIVORE study, the response-based strategy achieved no complete responses (CRs) and a low partial response (PR)/CR conversion.
The number of patients who were evaluated for nivolumab discontinuation was low, making the value of this approach not assessable, according to the study authors, led by Rana R. McKay, MD, associate professor of Medicine and medical oncologist at the University of California, San Diego.
The trial aimed to determine the possibility of escalating or discontinuing immune checkpoint inhibitor therapy as treatment of patients with advanced RCC. Immune-related toxicity is common and can be life-threatening with immune checkpoint inhibitors. Prior studies have demonstrated improved responses with the combination of nivolumab with ipilimumab. However, immune-related toxicity is common and can be life-threatening with the immune checkpoint inhibitors. Investigators aimed to maintain the benefit of this therapy with a less intensive treatment approach.
The primary end points of the response-adaptive study included the proportion of patients with durable CRs/PRs at 1 year after nivolumab discontinuation, which was the population that made up arm A, and the proportion of stable disease (SD)/progressive disease (PD) receiving nivolumab who converted to a PR/CR after the addition of ipilimumab, which made up the arm B patient population. Secondary end points included overall survival (OS), treatment-free interval, duration of disease control, progression-free survival (PFS), and toxicity.
Nivolumab was administered 240 intravenously (IV) every 2 weeks in patients on the study until the protocol was amended, then converted to 480 mg IV every 4 weeks. Patients with a PR/CR within 6 months of treatment then discontinued nivolumab and were observed in arm A, and if the patients experienced PD, they reinitiated nivolumab therapy while 2 doses of ipilimumab were added to nivolumab in those who had persistent PD or recurred.
All patients remained in arm A for analyses, while patients with SD or PD after a minimum of 2 to 6 months of nivolumab with the additional 2 doses of ipilimumab were evaluated as arm B. After allocation to either arm A or B, patients were assessed every 8 weeks in the first and after the first 12 weeks followed by every 8 thereafter in the latter arm.
A total of 83 patients were accrued. The majority of patients were male (82%), had an ECOG performance status of 0 (66.3%) or 1 (32.5%), and had intermediate risk (54.2%). Fourteen patients did not undergo allocation due to either PD (n = 7) or toxicity (n = 7). Among the patients who had discontinued treatment due to toxicity, the median treatment-free interval was 5.2 months (range, 1.2-20.4). Ten patients achieving PRs, 1 achieving an unconfirmed PR, and 1 with SD were allocated to arm A, and 28 patients with SD and 29 with PD were assigned to arm B. Three patients in arm A discontinued due to toxicity (n = 2) and PD (n = 1), while 54 patients from arm B discontinued the study due to toxicity (n = 11), PD (n = 37), other reasons (n = 5), and death unrelated to treatment (n = 1).
The median number of nivolumab induction cycles overall was 4 (range, 1-8), while the medians in arms A and B were 4 (range, 2-8) and 4 (range, 2-6), respectively. Fifty patients in arm B received 2 doses of ipilimumab, while 6 received 1 dose, and 1 received none. The median duration of treatment was 3.7 months (range, 1-24) in arm B.
The objective response rate was 14%, which included PRs and unconfirmed PRS within 6 months of nivolumab induction (90% CI, 9%-22%), while the rate was 17% among the treatment-naïve patients and 12% among those previously treated. The response rates appeared similar among those with favorable- (15%) and intermediate/poor-risk disease (14%).
Among patients in arm A, 5 (42%) remained off nivolumab therapy at 1-year following discontinuation (90% CI, 18%-68%), while 4 patients restarted nivolumab within 6 months after discontinuation, which was due to PD in 3 patients and ongoing PR in 1. Three patients remain in active follow-up who have not reached the 1-year mark post-treatment discontinuation.
In arm B, 2 patients (4%) converted to a PR (90% CI, 1%-11%), and both of these patients had been previously treated, did not have sarcomatoid/rhabdoid differentiation, and had PD as best response to nivolumab induction therapy. The duration of response among these 2 patients was 9.2 months and 10.9 months, and ultimately both discontinued treatment due to PD. The rate of SD as best response was 46%, while the rate of PD as best response was 40%.
The median duration of disease control in arm B was 10.4 months (95% CI, 5.5-13.0), and the median PFS was 4.7 months, which included 45 PFS events (95% CI, 2.7-8.3).
Overall, the median follow-up for OS was 19.5 months (range, 2.5-28.1), and 19 deaths were observed in study, 12 of which were from arm B, 1 in arm A, and 6 who withdrew before allocation. The median OS had not been reached, but the 18-month OS rate was 79% (95% CI, 67%-87%).
Any grade treatment-related adverse events (TRAEs) occurred in 80% of the patients during nivolumab induction, and grade 3 TRAEs were observed in 7%. Prednisone ≥ 40 mg or the equivalent was required in 8% of patients, another 8% discontinued treatment due to toxicity, and 14% had a dose delay due to toxicity.
Any grade TRAEs occurred in 81% of patients in arm B, while grade 3/4 TRAEs were observed in 25% of the population. Eighteen percent of patients required the prednisone or equivalent therapy, 19% discontinued due to toxicity, and 32% had dose delay because of toxicity.
No treatment-related deaths occurred on the study.
Overall, this multicenter study evaluated important clinical questions regarding the use of nivolumab and checkpoint inhibition in the advanced RCC patient population and demonstrated a lack of CRs with 2 cycles of ipilimumab added to nivolumab induction on a response-based strategy, and the PR/CR conversion rate was low. Durable responses, however, were observed among patients who discontinued nivolumab early in the absence of toxicity.
“Based on these findings, we cannot recommend a response-based adaptive strategy for nivolumab and ipilimumab, and recommend upfront dual checkpoint blockade in patients eligible to receive this treatment,” wrote McKay et al. “We are currently investigating novel biomarkers of response and resistance to therapy to better optimize treatment strategies for patients.”
McKay RR, McGregor BA, Xie W, et al. Optimized management of nivolumab and opilimumab in advanced renal cell carcinoma: A response-based phase II study (OMNIVORE). J Clin Oncol. Published online October 27, 2020. doi: 10.1200/JCO.20.02295