Background

Cancer of the kidney and renal pelvis was estimated to be diagnosed in 76,080 people in the United States in 2021, with 13,780 projected to die of the disease [1]. The most common histologic subtype is clear cell renal cell carcinoma (RCC) [2, 3]. For patients with relapsed or de novo stage IV disease, the preferred first-line treatment options include immune checkpoint inhibitor–based therapies such as axitinib plus pembrolizumab, axitinib plus avelumab, lenvatinib plus pembrolizumab, cabozantinib plus nivolumab or ipilimumab plus nivolumab [2, 4,5,6,7,8]. Specifically, in patients with disease considered intermediate or poor risk per the International Metastatic RCC Database Consortium (IMDC), options include all 5 immunotherapy-based doublets and cabozantinib monotherapy, whereas for all IMDC risk categories, axitinib plus pembrolizumab, cabozantinib plus nivolumab, or pembrolizumab plus lenvatinib have emerged as preferred standard regimens [2, 4,5,6,7,8]. In selected patients not eligible for immunotherapy, treatment with a vascular endothelial growth factor receptor (VEGFR) tyrosine kinase inhibitor (eg, sunitinib, pazopanib, cabozantinib) is often offered [2].

Because of toxicities and disease progression with currently available therapies, there is an unmet medical need for safer and more effective combination treatments for metastatic RCC (mRCC). Both indoleamine 2,3-deoxygenase 1 (IDO1) and programmed death 1 (PD-1) mediated pathways suppress T-cell–mediated antitumor immunity, and IDO1 and the PD-1 ligand (PD-L1) are co-expressed in multiple cancer types and correlate with poor prognosis [9,10,11,12,13,14,15]. Anti–PD-1 agents block immunosuppressive receptor PD-1 on T cells, thus enhancing immune responses against tumors [16]. IDO is an enzyme that catalyzes the rate-limiting step in the conversion of tryptophan to kynurenine, which results in apoptosis of effector T cells and activation of regulatory T cells, thereby promoting an immunosuppressive environment and tumor growth [17]. Further, high expression of IDO1 may be associated with resistance to PD-1 inhibition in non–small cell lung cancer, melanoma, and RCC [18,19,20,21]. Combining an anti–PD-1 agent with an IDO1 inhibitor may therefore enhance antitumor immunity. In preclinical models, anti–PD-1 treatment combined with an IDO1 inhibitor showed synergistic antitumor activity in models of melanoma and glioblastoma [22, 23].

Pembrolizumab is a potent and highly selective humanized monoclonal antibody directed against PD-1 that has shown promising antitumor activity as monotherapy in both advanced clear cell RCC and non-clear cell mRCC and is approved, in combination with axitinib [4, 24, 25] or with lenvatinib [26, 27], for the first-line treatment of advanced mRCC. Epacadostat (formerly INCB024360) represents a novel, potent, and selective inhibitor of IDO1 in human tumor and dendritic cells [20, 28, 29]. In a phase 1 study of patients with advanced solid tumors, the combination of pembrolizumab plus epacadostat has shown promising antitumor activity, including objective responses in patients with treatment-naive and previously treated melanoma, non–small cell lung cancer, urothelial cancer, and mRCC (25 of 62 patients responded, including 2 of 11 with RCC) [30]. These responses were durable across tumor types, with 68% of responses (17 of 25) ongoing at data cutoff.

The aim of the present study was to compare the antitumor activity and tolerability of pembrolizumab plus epacadostat with standard of care treatment (at the time of study conduct) with either sunitinib or pazopanib as first-line therapy in patients with mRCC.

Methods

This was a randomized, open-label, parallel-group, multicenter, phase 3 trial (KEYNOTE-679/ECHO-302; NCT03260894; first registration: 8/24/2017). The study was conducted at 74 centers in 14 countries in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines, and independent institutional review boards or ethics committees reviewed and approved the protocol and applicable amendments for each institution (Supplemental Table 1). All patients provided written informed consent.

Patients

Eligible patients had histologically confirmed locally advanced or mRCC (stage IV per American Joint Committee on Cancer) with a clear cell component, with or without sarcomatoid features, no prior systemic therapy for mRCC, measurable disease per RECIST 1.1, and Karnofsky performance status (KPS) ≥ 70. Patients were excluded if they had previously received therapy with an anti–PD-1, anti–PD-L1, or anti–PD-L2 agent, or with epacadostat or any anti-IDO1 agent, or with an agent directed to another stimulatory or co-inhibitory T-cell receptor. Patients were ineligible if they had previously received therapy with VEGF/VEGFR or with mammalian target of rapamycin–targeting agents for locally advanced or metastatic cancer or if they received systemic anti-cancer therapy, including investigational agents, within 4 weeks before randomization.

Treatments

Patients were randomly assigned 1:1 to receive pembrolizumab 200 mg IV every 3 weeks plus epacadostat 100 mg orally twice daily continuously or sunitinib or pazopanib (sunitinib 50 mg orally once daily [6-week cycles; 4 weeks on treatment followed by 2 weeks off treatment] or pazopanib 800 mg orally once daily continuously). Dose delays or dose reductions were performed per defined criteria in the protocol. If pembrolizumab dosing was held, dosing for epacadostat was also to be held. Patients who required dose reduction of epacadostat because of adverse events (AEs) remained at the lower dose; re-escalation was not permitted.

Randomization was performed centrally using an interactive voice response system/integrated web response system and stratified according to IMDC risk category (favorable vs intermediate vs poor), physician’s intended choice of comparator drug (sunitinib vs pazopanib), and geographical region (United States, Canada, and Western Europe vs rest of world).

Study conduct

The study was initiated on December 7, 2017, and on May 2, 2018, the sponsor made a strategic decision to stop enrollment permanently because a phase 3 study in patients with unresectable or metastatic melanoma did not meet the prespecified primary end point of improvement in progression-free survival (PFS) for pembrolizumab plus epacadostat compared with pembrolizumab plus placebo [31]. In that study, no new safety concerns had arisen with pembrolizumab plus epacadostat compared with pembrolizumab monotherapy. In the present study, patients who experienced ongoing clinical benefit could continue study treatment at the discretion of the investigator and would continue to be monitored.

End points

The original dual primary end points were comparison of PFS (as assessed by blinded, independent central review) and overall survival (OS) between treatment arms. The protocol was amended when enrollment was stopped, and the primary end point changed to estimation of objective response rate (ORR; as assessed per RECIST v1.1 by the investigator) for each treatment arm after the first on-study scan (week 12); PFS and OS were removed as end points.

ORR was defined as the proportion of patients in the analysis population whose best response was complete response (CR) or partial response (PR). Protocol-specified efficacy imaging was stopped at week 12 with the protocol amendment. Subsequent disease monitoring was performed according to local standard of care. For some patients imaging was completed beyond week 12 at the time enrollment was stopped; thus, ORR was based on all available imaging assessments.

Safety and tolerability were assessed as a secondary end point and included the number of patients with an AE and the number of patients who discontinued because of an AE.

Pharmacodynamic activity, an exploratory end point, was assessed as change from baseline to week 3 in serum kynurenine in both treatment arms. Serum kynurenine levels were determined using a proprietary, validated liquid chromatography–mass spectrometry assay using calibrated standards (Worldwide Clinical Trials, Morrisville, NC). Kynurenine serum levels at cycle 1, day 1 (C1D1) and cycle 2, day 1 (C2D1) were compared using paired t-tests within each treatment arm.

Statistical analysis

The original target enrollment was 630 patients; a subsequent amendment (June 15, 2018) resulted in enrollment being stopped early after 129 patients had been randomly assigned. In the original statistical analysis plan, PFS and OS, dual original primary end points, were to be evaluated using a stratified log-rank test, and hazard ratio was to be estimated using a stratified Cox regression model. Event rates over time were to be estimated within each treatment group using the Kaplan–Meier method. In the amended protocol, ORR was estimated within each treatment arm with 95% confidence intervals (CI) using Clopper-Pearson exact method based on binomial distribution [32] and was summarized by study treatment group. Counts and percentages of patients with AEs were provided by treatment group. AEs were assessed using Common Terminology Criteria for Adverse Events version 4.0.

The intention-to-treat population was used for the efficacy analysis and included all randomly assigned patients. The all-patients-as-treated population was used for the safety analysis and included all randomly assigned patients who received at least 1 dose of study treatment. Safety and tolerability parameters were summarized by descriptive statistics based on treatment group.

Results

Patients

The study was initiated on December 7, 2017, and enrollment was stopped on May 2, 2018. Median (range) follow-up, defined as time from randomization to data cutoff, was 10.3 months (2.2–14.3) in the pembrolizumab plus epacadostat arm and 10.3 months (2.7–13.8) in the sunitinib/pazopanib arm. A total of 129 patients were randomly assigned to treatment (pembrolizumab plus epacadostat, n = 64; sunitinib/pazopanib, n = 65); 10 patients discontinued in the pembrolizumab plus epacadostat arm and 19 patients discontinued in the sunitinib/pazopanib arm (11 primarily because of withdrawal) (Fig. 1). The primary reason for discontinuation in the pembrolizumab plus epacadostat arm was death (n = 6); none of these deaths were deemed treatment related. Treatment arms were balanced with regard to most patient characteristics (Table 1). Overall, 42.6% were aged ≥ 65 years, and most patients were male (72.9%), white (84.5%), had Eastern Cooperative Oncology Group performance status 0 (58.9%) or 1 (39.5%). Metastatic staging at initial diagnosis was M0 (44.2%) or M1 (55.0%). IMDC risk category was favorable in 29.5% of patients, intermediate in 55.0%, and poor in 15.5% at time of study entry.

Fig. 1
figure 1

Patient disposition for trial. aSunitinib 50 mg orally once daily (6-week cycles; 4 weeks on treatment followed by 2 weeks off treatment) or pazopanib 800 mg orally once daily continuously. bStatus was not reported as of the data cutoff date. Patients could be ongoing with study or treatment

Full size image
Table 1 Patient demographics
Full size table

Response rates

Response rates were similar in both treatment arms (Table 2). Objective responses were achieved by 20 patients (31.3% [95% CI 20.2–44.1]) receiving pembrolizumab plus epacadostat and by 19 patients (29.2% [95% CI 18.6–41.8]) receiving sunitinib/pazopanib. One patient, enrolled in the pembrolizumab plus epacadostat arm, achieved CR. The disease control rate (CR + PR + stable disease) was similar in both treatment arms (45 patients [70.3%] in the pembrolizumab plus epacadostat arm vs 46 [70.8%] in the sunitinib/pazopanib arm).

Table 2 Objective response rate per RECIST v1.1. by investigator assessmenta
Full size table

Change in tumor burden also indicated a similar profile of tumor response in both treatment arms; 62% of patients receiving pembrolizumab plus epacadostat and 78% of those receiving sunitinib/pazopanib therapy experienced a reduction in the sum of the diameter of target lesions (Fig. 2).

Fig. 2
figure 2

Best tumor change from baseline in patients with measurable disease at baseline and at least 1 postbaseline measurement in the (a) pembrolizumab plus epacadostat arm or the (b) sunitinib or pazopanib arm. aBased on investigator assessment per RECIST v1.1 in the ITT population which included all subjects with measurable disease at baseline and ≥ 1 post-baseline measurement by cutoff date. Abbreviations: ITT intention-to-treat, RECIST Response Evaluation Criteria in Solid Tumors

Full size image

In patients with favorable IMDC risk category, the ORR was 21.1% (95% CI 6.1–45.6) with pembrolizumab plus epacadostat versus 31.6% (95% CI 12.6–56.6) with sunitinib/pazopanib. In patients with intermediate or poor IMDC risk category, the ORR was 35.6% (95% CI 21.9–51.2) with pembrolizumab plus epacadostat versus 28.3% (95% CI 16.0–43.5) with sunitinib/pazopanib.

Safety

At least 1 AE occurred in all patients in both the pembrolizumab plus epacadostat arm and the sunitinib/pazopanib arm (Table 3). Overall, 81.3% of patients in the pembrolizumab plus epacadostat arm and 93.7% of patients in the sunitinib/pazopanib arm had AEs that were considered by the investigator to be related to study drug. Grade 1/2 treatment-related AEs occurred in 65.6% versus 57.1% of patients, and grade 3–5 treatment-related AEs occurred in 34.4% versus 42.9% of patients in the pembrolizumab plus epacadostat arm versus the sunitinib/pazopanib arm. There were no treatment-related deaths in either treatment arm; 1 patient in the sunitinib/pazopanib arm died of septic shock (not treatment related (per AMA)). All-cause AEs occurring in ≥ 10% of patients in the pembrolizumab plus epacadostat arm and the sunitinib/pazopanib arm were nausea (34.4% vs 31.7%), pruritus (25.0% vs 1.6%), diarrhea (20.3% vs 47.6%), and fatigue (18.8% vs 23.8%) (Table 4). Eighteen patients (28.1%) in the pembrolizumab plus epacadostat arm and 15 patients (23.8%) in the sunitinib/pazopanib arm had serious AEs.

Table 3 Summary of adverse eventsa
Full size table
Table 4 Incidence of any-grade nonserious adverse events (incidence ≥ 10% in either treatment group) regardless of attribution
Full size table

Eight patients (12.5%) in the pembrolizumab plus epacadostat arm discontinued because of a treatment-related AE; 5 patients (7.8%) discontinued pembrolizumab and 8 patients (12.5%) discontinued epacadostat. In the sunitinib/pazopanib arm, 5 patients (7.9%) discontinued because of a treatment-related AE (4 [6.3%] sunitinib and 1 [1.6%] pazopanib).

Pharmacodynamic endpoint

Samples were available for pharmacodynamic analysis in 61 and 52 patients from the pembrolizumab plus epacadostat and the sunitinib/pazopanib arms, respectively. Serum kynurenine levels from baseline to week 3 decreased by 10.3% with pembrolizumab plus epacadostat (from 2.9 µM at baseline to 2.6 µM at week 3; absolute change, –0.3 µM) but not to levels previously reported in healthy volunteers (1.5 µM) [28] (Supplemental Fig. 1). In contrast, serum kynurenine increased by 3.2% from baseline to week 3 with sunitinib/pazopanib (from 3.1 µM to 3.2 µM; absolute change, + 3.2%).

Discussion

Despite advances in the treatment of patients with clear cell mRCC, many patients experience disease relapse or progression. Results from the ECHO-202/KEYNOTE-037 study indicated that pembrolizumab combined with epacadostat was well tolerated and showed evidence of durable objective responses in 2 patients with mRCC, which warranted further investigation and informed the design of the current study [30]. However, the ECHO-301/KEYNOTE-252 study found that epacadostat in combination with pembrolizumab did not improve clinical benefit compared with pembrolizumab monotherapy for unresectable or metastatic melanoma [31]. Based on those findings, a decision was made to stop the current study and to change the primary end point to investigator-assessed ORR per RECIST v1.1.

In the current study, response rates were similar in patients with locally advanced or mRCC receiving first-line pembrolizumab plus epacadostat compared with response rates in those receiving treatment with sunitinib or pazopanib. Sunitinib or pazopanib were used as comparators because these were the standard treatments at the time the study was conducted. Response rates reported in this study were similar to those observed in KEYNOTE-427 with single-agent pembrolizumab in 110 patients who had previously untreated advanced clear cell RCC [24]. This suggests that the addition of epacadostat to pembrolizumab has little effect on response rates in RCC. Although interpretation of the findings of this study are limited by its relatively small sample size, response rates were numerically higher in patients with an intermediate or a poor IMDC risk category receiving pembrolizumab plus epacadostat versus sunitinib/pazopanib; conversely, response rates were lower in patients with a favorable IMDC risk category receiving pembrolizumab plus epacadostat versus sunitinib/pazopanib. Despite the small subgroups of patients, this observation is consistent with earlier observations and suggest greater benefit of VEGF-targeted therapies in favorable risk [4, 5].

Immunotherapy combinations are effective in the treatment of RCC. Pembrolizumab in combination with axitinib [4, 33] or lenvatinib [26, 27] are approved first-line treatment options for patients with locally advanced RCC or mRCC. Pembrolizumab in combination with lenvatinib also showed improved progression-free survival, overall survival, and objective response rate over sunitinib in the phase 3 CLEAR study of patients with advanced RCC [7]. Further, IDO1 expression may be a valuable biomarker for response to immune therapy in patients with mRCC. High IDO-1 expression in tumor endothelial cells in patients with mRCC is associated with better therapeutic response to nivolumab [10].

Epacadostat and other IDO-1 inhibitors, including indoximod and navoximod, are also in clinical development for other tumor types [17]. Several phase 3 studies with linrodostat (BMS-986205), a selective IDO-1 inhibitor, in combination with nivolumab for the treatment of muscle-invasive bladder cancer (NCT03661320) [34] and previously untreated metastatic or unresectable melanoma (NCT03329846) are ongoing.

Consistent with the results of the ECHO-301/KEYNOTE-252 study, pembrolizumab in combination with epacadostat was generally well tolerated [31]. AEs and AE rates were generally similar between the pembrolizumab plus epacadostat and sunitinib/pazopanib treatment arms and were consistent with data from previous studies evaluating this combination in other tumor types [31]. The pembrolizumab plus epacadostat combination was manageable, and no new safety concerns were identified. Study treatment discontinuations because of treatment-related AEs were similar in both arms. Moreover, the pharmacodynamic data showed decreases in serum kynurenine that did not reach those of healthy subjects [28, 35], suggesting that higher doses of epacadostat may be warranted. The observation that serum kynurenine levels decreased in the pembrolizumab plus epacadostat arm and slightly increased in the sunitinib/pazopanib arm is consistent with an inhibitory effect of epacadostat on IDO1 activity. Because IDO1 is induced by interferon-γ [36], it is possible that the antitumor response creates an environment in which IDO1 expression surpasses the ability of epacadostat 100 mg twice daily to sufficiently inhibit kynurenine production. In fact, a retrospective pooled analysis of epacadostat clinical studies determined that when combined with an anti-PD-1 agent, doses < 600 mg twice daily were not able to maintain inhibition of plasma kynurenine, suggesting that doses ≥ 600 mg twice daily warrant additional studies in combination with checkpoint inhibitors to determine any clinical benefit to normalizing kynurenine levels [35].

Limitations

Given the early termination of this study, the sample size was relatively small and the duration of follow-up was short, which has to be considered when interpreting the results.

Conclusion

Pembrolizumab plus epacadostat demonstrated similar response rates compared with sunitinib or pazopanib in patients with locally advanced RCC or mRCC who had not previously received systemic therapy. AE rates were generally similar between treatment arms and no new safety concerns were identified for either drug. When taking into account the initial promising activity of single-agent pembrolizumab in advanced RCC (KEYNOTE-427) [24, 25] and the lack of additional benefit with pembrolizumab plus epacadostat over pembrolizumab monotherapy in the melanoma setting (ECHO-301/KEYNOTE-252) [31], the findings of the current study provide further evidence that the antitumor responses observed in patients with RCC receiving pembrolizumab plus epacadostat 100 mg twice daily may be driven primarily by pembrolizumab.