INCB024360

A randomised, open-label, phase 2 study of the IDO1 inhibitor epacadostat (INCB024360) versus tamoxifen as therapy for biochemically recurrent (CA-125 relapse)–only epithelial ovarian cancer, primary peritoneal carcinoma, or fallopian tube cancer

H I G H L I G H T S

• Epacadostat monotherapy was evaluated in biochemically recurrent ovarian cancer.
• Immune-related adverse events were manageable and predominantly low grade.
• No significant efficacy difference between epacadostat and tamoxifen was observed.
• IDO1/PD-L1 coexpression supports evaluation of IDO1 plus checkpoint inhibitors.

Abstract

Objective. Indoleamine 2,3-dioxygenase-1 (IDO1) is a key regulator of immune tolerance in ovarian cancer. This study investigated efficacy and safety of the IDO1 enzyme inhibitor epacadostat versus tamoxifen in patients with biochemical-only recurrence (CA-125 elevation) following complete remission after first-line chemothera- py for advanced epithelial ovarian, primary peritoneal, or fallopian tube cancer.

Methods. In this open-label, phase 2 study (NCT01685255), patients were randomised 1:1 to epacadostat 600 mg or tamoxifen 20 mg twice daily for successive 28-day cycles and stratified by time since completion of first-line chemotherapy to first CA-125 elevation (3 to b 12 or ≥12 months). The primary endpoint was investi- gator-assessed progression-free survival (PFS; RECIST v1.1). Secondary endpoints included CA-125 response (Gynecologic Cancer InterGroup criteria), overall survival, safety, and tolerability.

Results. The study was terminated primarily due to slow accrual and lack of evidence of superiority. Median PFS was 3.75 months for epacadostat (n = 22) versus 5.56 months for tamoxifen (n = 20; HR, 1.34 [95% CI, 0.58– 3.14]; P = 0.54). Of evaluable patients, 1 (5.0%) epacadostat and 3 (15.8%) tamoxifen patients had confirmed CA- 125 responses. The most common treatment-emergent adverse event was fatigue (epacadostat, 36.4%; tamoxi- fen, 40.0%). Immune-related adverse events, observed with epacadostat only, were primarily rash (18.2%) and pruritus (9.1%). Epacadostat pharmacokinetics/pharmacodynamics were consistent with its known mechanism of action. IDO1 expression was observed in 94% of archival tumour samples.

Conclusions. This first report of immunotherapy evaluation in biochemical-only relapse ovarian cancer and of IDO1 inhibitor monotherapy in ovarian cancer found no significant difference in efficacy between epacadostat and tamoxifen. Epacadostat was generally well tolerated.

1. Introduction

Ovarian cancer is the leading cause of gynaecologic cancer-related deaths worldwide and has poor long-term survival [1–3]. For patients who relapse ≥ 6 months after responding to first-line treatment (typical- ly cytoreductive surgery and systemic platinum-based chemotherapy [2,4]), retreatment with platinum-based chemotherapy has encourag- ing response rates [5]; however, the majority of patients experiencing relapse are considered incurable [2,4,6]. There remains a substantial unmet clinical need for better strategies to improve disease-free surviv- al and cure in early treatment of ovarian cancer [5,6].

The development of symptoms is one indicator of disease relapse, prompting biochemical testing with the tumour marker CA-125 and imag- ing to confirm disease recurrence [5,7]. However, patients are frequently asymptomatic at the time of small-volume recurrence, with suspicion of relapse based solely on rising CA-125 levels [5,7]. In such patients, a watch-and-wait policy is justifiable. Second-line chemotherapy is initiated according to symptoms, extent of disease and CA-125 level, among other considerations [5]. When patients present with a biochemical relapse with- out clinical evidence of disease, there may be an opportunity to improve outcomes by extending the time that the cancer remains under control, po- tentially delaying progression and the need for further cytotoxic therapy.

Ovarian cancer is an immunogenic malignancy [8,9], supporting the rationale for immunomodulatory agents (eg, checkpoint inhibitors) as po- tentially effective therapeutic agents. Recruitment of regulatory T cells in ovarian cancer leads to immunosuppression [10], which has been associ- ated with decreased survival, paclitaxel resistance, and increased levels of vascular endothelial growth factor [8,10]. In patients with stage III/IV ovarian cancer, survival is also strongly correlated with the presence of tu- mour-infiltrating lymphocytes (TILs) [11], with a 5-year survival of 38% when TILs are present versus 4.5% when they are absent [12].

The intracellular indoleamine 2,3-dioxygenase-1 (IDO1) enzyme is a key regulator of the immunosuppression responsible for tumour escape from immune surveillance [15–17] and is predominantly expressed by tumour epithelial cells, antigen-presenting cells in primary tumours and tumour-draining lymph nodes in a variety of cancers [13,14]. IDO1 catalyses the degradation of tryptophan via oxidation to kynurenine (Kyn), which results in strong inhibitory effects on T-cell– mediated responses, including blocking T-cell activation and inducing T-cell apoptosis [18]. High intratumoural IDO1 expression in ovarian cancer has been found to correlate with a reduced number of TILs [19], advanced disease stage, paclitaxel resistance, and decreased sur- vival [15–17,19]. Taken together, these findings strongly support IDO1 as a rational target to reactivate the antitumour immunity in patients with ovarian cancer. Epacadostat (INCB024360), a selective IDO1 en- zyme inhibitor, has been developed and is currently under clinical in- vestigation in various tumour types [20–23].

Ovarian cancer treatment guidelines suggest that patients with bio- chemical relapse (serially increasing CA-125 levels and no clinical evi- dence of disease) have several options: (1) delay therapy until clinical relapse; (2) enrol in a clinical trial; or (3) undergo treatment with a sec- ond-line therapy that has an acceptable side-effect profile, such as bio- logic therapies (eg, tamoxifen) over cytotoxic therapies [2]. We hypothesised that these patients would be good candidates for im- mune-targeted therapies and investigated the effects of treatment with epacadostat in patients with a low cancer burden. Thus, the objec- tive of this study was to determine the efficacy of epacadostat compared with tamoxifen in biochemical-recurrent–only epithelial ovarian, pri- mary peritoneal, or fallopian tube cancer.

2. Methods

2.1. Study design and treatment

This international, multicentre, randomised, open-label phase 2 study conducted in 6 countries (United States, United Kingdom, Russia,
Ukraine, Australia, and Canada) evaluated epacadostat versus tamoxi- fen for efficacy, safety, and tolerability in women with ovarian cancer and CA-125 elevation following complete remission with first-line che- motherapy. At study initiation, the intention was to enrol 110 patients randomised 1:1 to receive epacadostat or tamoxifen and stratified based on the number of months since prior first-line chemotherapy to the time of their first CA-125 elevation (3 to b 12 months or ≥ 12 months). The study (ClinicalTrials.gov: NCT01685255) was conducted in accordance with the ethical principles of Good Clinical Prac- tice, according to the International Conference on Harmonisation guidelines, and was approved by the institutional review board or ethics committee at each participating institution. All patients provided writ- ten informed consent before initiation of treatment.

Study treatment was administered orally as continuous 28-day cycles of either epacadostat 600 mg twice daily (BID) or tamoxifen 20 mg BID. Dose reductions, interruptions, or discontinuations were allowed at any time for safety reasons (Supplement Table 1). However, only 2 dose re- ductions of epacadostat were allowed (400 mg BID and 300 mg BID). The study comprised a screening phase, treatment phase, and safety fol- low-up phase. During the treatment phase, patients received study drug in successive 28-day cycles until they met any criterion for withdrawal. Patients were monitored for 60 days after the last dose of epacadostat or tamoxifen during the safety follow-up. After this, patients were moni- tored for survival at approximately 12-week intervals.

2.2. Study population

Eligible patients were women aged 18 years or older with Eastern Cooperative Oncology Group performance status 0 or 1; histologically confirmed Federation of International Gynecologists and Obstetricians (FIGO) [24] stage IC, II, III, or IV epithelial ovarian cancer, primary peri- toneal carcinoma, or fallopian tube cancer at diagnosis; biochemical re- currence; and no other objective evidence of disease recurrence as defined by Response Evaluation Criteria in Solid Tumors (RECIST v1.1). Biochemical recurrence of disease (Gynecologic Cancer InterGroup [GCIG] criteria) was defined as 2 consecutive measurements of CA-125 above the upper limit of normal (ULN) that were ≥ 2 weeks apart, with the second measurement showing a further increase from the first measurement. If the first CA-125 measurement is ≥ 2 × ULN, the confirmatory CA-125 measurement only needs to be ≥ 1 week later. In the United Kingdom (UK-only requirement), biochemical recurrence of disease was defined as elevated CA-125 levels ≥ 2 × ULN on 2 occa- sions that were ≥ 1 week apart without evidence of disease as defined by RECIST 1.1. Before entering the study, patients must have had a com- plete response to chemotherapy and must have received a first-line platinum-containing chemotherapy regimen with documentation of CA-125 elevation at first diagnosis and at least 1 normal CA-125 level during or after first-line therapy.

Key exclusion criteria included protocol-specified active or inactive autoimmune processes (except vitiligo, thyroiditis, or eczema) and un- stable cardiovascular disease ≤ 6 months before starting study treat- ment. Patients were also excluded if they had received prior antitumour systemic therapy besides first-line chemotherapy; prior ra- diotherapy within 3 months of randomisation with unresolved toxic- ities; prior investigational drug or immunologically based treatment for any reason, including chronic use of systemic steroid ≥ 7.5 mg/d prednisone equivalents (except completed adjuvant therapy or use of inhaled or topical steroids); potent cytochrome P450 3A4 inducers or inhibitors; monoamine oxidase inhibitors within the 21 days before screening; prior serotonin syndrome after receiving ≥ 1 serotonergic drug; and contraindication to tamoxifen therapy.

2.3. Endpoints and assessments

The primary endpoint was efficacy by investigator-assessed progression-free survival (PFS; RECIST v1.1). Per RECIST v1.1,progressive disease was defined by the appearance of any new le- sion, whether target or non-target. Disease and tumour assess- ments were conducted every 8 weeks for the first 12 months, then every 12 weeks thereafter, and at end of treatment or early termination.Secondary endpoints included evaluation of CA-125 response or non-response (GCIG criteria; CA 125 already progressing at study entry), overall survival, and evaluation of the safety and tolerability of epacadostat. A CA-125 response (GCIG criteria) was defined as ≥ 50% re- duction in CA-125 levels from a pretreatment sample that was con- firmed and maintained for ≥ 28 days. Safety and tolerability assessments included treatment-emergent adverse events (TEAEs),treatment-related adverse events (TRAEs), and immune-related adverse events (irAEs), vital signs, electrocardiograms (ECGs), physical ex- amination, and clinical laboratory tests. CA-125 and adverse events were assessed at baseline then every 4 weeks thereafter, at end of treat- ment or early termination, and 30 and 60 days after end of treatment or conducted.

3. Results

3.1. Study population

Forty-two women were enrolled in the study between March 7, 2013, and October 23, 2014 (epacadostat, n = 22; tamoxifen, n = 20). All 42 patients were evaluated for efficacy and safety. All 42 patients discontinued from study drug treatment. The most common reason for discontinuation was disease progression (epacadostat, n = 10.

The modified intent-to-treat population, defined as all randomised patients who received ≥ 1 dose of study drug, was used for efficacy analyses. Safety analyses included all enrolled patients who received ≥ 1 dose of study drug. The PK/PD evaluable population included patients who received ≥ 1 dose of epacadostat or tamoxifen and provided ≥ 1 postdose plasma sample for PK/PD measurement.
SAS® software, version 9.2 (SAS Institute, Cary, NC) was used to generate all tables, graphs, and statistical analyses. Descriptive sta- tistics were used to present summaries of continuous and categorical variables. Safety, PK, PD, and immunologic marker data were analysed using summary statistics (eg, means and frequencies).Standard noncompartmental PK methods were used to analyse epacadostat plasma concentration data using Phoenix WinNonlin® ver- sion 6.0 (Pharsight Corporation, Mountain View, CA).

Fig. 1. PFS Kaplan-Meier curves by investigator assessment using RECIST 1.1 criteria (mITT population). mITT, modified intent-to-treat; PFS, progression-free survival; RECIST, Response Evaluation Criteria in Solid Tumours.

The majority of patients were white (epacadostat, 100%; tamoxifen, 90.0%), and the median age overall was 59.0 years. Baseline demo- graphics and disease characteristics were generally well balanced across the 2 treatment groups, including primary cancer site, number of months from completion of prior first-line chemotherapy to the first CA-125 elevation, prior surgery, prior systemic therapy, and breast can- cer susceptibility gene (BRCA) status. The most common FIGO stage at diagnosis was stage IIIC for both treatment groups (64.3% overall; Table 1).

3.2. Efficacy

At early study termination, epacadostat was not associated with su- periority over tamoxifen as measured by investigator-assessed PFS. Me- dian PFS was 3.75 months in the epacadostat group versus 5.56 months in the tamoxifen group. The hazard ratio (HR) for death or disease pro- gression was 1.34 (95% CI, 0.58–3.14; P = 0.54; Fig. 1). For patients who had their first CA-125 elevation 3 to b 12 months after completion of first-line chemotherapy (baseline stratification factors), median PFS was 2.24 months for epacadostat versus 5.48 months for tamoxifen (HR, 1.63 [95% CI, 0.48–5.50]; P = 0.41). For patients with their first CA-125 elevation at ≥ 12 months, median PFS was 3.98 and 5.56 months in epacadostat and tamoxifen arms (HR, 0.85 [95% CI, 0.26–2.80]; P = 0.78), respectively.

Of evaluable patients, 1 (5.0%) in the epacadostat group and 3 (15.8%) in the tamoxifen group had confirmed CA-125 responses. Two additional patients (10.0%) in the epacadostat group and 2 (10.5%) in the tamoxifen group had CA-125 responses but were unconfirmed (ie, ≥50% reduction of CA-125 at only 1 time point; Supplement Table 3. The mean best per- centage change from baseline in CA-125 was 36.6% for epacadostat-treat- ed patients and 59.9% for tamoxifen-treated patients (Fig. 2).

Overall survival was assessed over the treatment and follow-up phases, during which 21 patients (95.5%) receiving epacadostat and all 20 patients (100%) receiving tamoxifen discontinued from the study and were censored from the overall survival analysis. The remaining 1 patient (4.5%) in the epacadostat group completed study drug treat- ment and the 60-day safety follow-up but subsequently died because of disease progression during the survival follow-up, before the sponsor’s decision to terminate the study. No tamoxifen-treated pa- tients died at time of follow-up before study termination.

3.3. Safety and tolerability

Epacadostat was generally well tolerated. Twenty-two patients re- ceived ≥ 1 dose of epacadostat, with a median exposure of 56.0 days and a median total daily dose of 1200 mg. Twenty patients received ≥ 1 dose of tamoxifen, with a median exposure of 61.0 days and a median total daily dose of 40 mg.

Fig. 2. Best percentage change from baseline in CA-125 (modified intent-to-treat population).

Seventeen patients (77.3%) in the epacadostat group and 15 patients (75.0%) in the tamoxifen group experienced TEAEs (Table 2A). The most frequently reported all-grade TEAE in both groups was fatigue (epacadostat, 36.4%; tamoxifen, 40.0%). Six patients (27.3%) in the epacadostat group discontinued study drug because of an adverse event; the most frequently reported reason was rash (n = 3; all TRAEs). No patient in the tamoxifen group discontinued because of an adverse event.

Seven patients (31.8%) in the epacadostat group and 2 patients (10%) in the tamoxifen group had TEAEs of grade ≥ 3. The most fre- quently reported grade ≥ 3 TEAE in the epacadostat group was maculopapular rash (n = 2 [9.1%]). No individual TEAE grade ≥ 3 was re- ported in N 1 patient in the tamoxifen group.

As expected based on its mechanism of action, no irAEs were report- ed in patients receiving tamoxifen. In the epacadostat group, the most frequently reported irAEs were skin related: 4 patients (18.2%) had rash and 2 patients (9.1%) had pruritus (Table 2B). Two patients (9.1%) receiving epacadostat had a grade ≥ 3 irAE of maculopapular rash. The majority of patients had normal haematology and clinical chem- istry laboratory assessments at baseline, and the values remained nor- mal throughout the study. Overall, no clinically meaningful changes or trends in vital signs or ECG findings were observed. There were no reports of serotonin syndrome at any time during the study.

3.4. Pharmacokinetics

After oral administration of epacadostat 600 mg BID (n = 15), mean (SD) epacadostat maximum observed plasma concentration (Cmax), minimum observed plasma concentration (Cmin), and area under the plasma concentration-time curve (AUC0-τ) were 6.20 (2.68) μM, 0.868 (0.516) μM, and 31.7 (11.7) μM·h, respectively, on day 15 of cycle 1. Me- dian time to Cmax (tmax) was 1.9 h postdose (Fig. 3).

3.5. Pharmacodynamics

Samples from 29 patients were available for IDO1 inhibition analysis. In the epacadostat group (n = 16), the average Kyn level was reduced by 39% from 2010 nM at cycle 1 day 1 to 1227 nM at cycle 2 day 1, sug- gesting that treatment with epacadostat reduced plasma Kyn levels to within the observed range in healthy volunteers (median Kyn, 1499 nM) [25]. In contrast, the average Kyn level was reduced by 13% in the tamoxifen group (n = 13) during the same period (from 2192 to 1897 nM).

In the ex vivo PD analysis that evaluated IDO1 inhibition in IFN-γ– and LPS- stimulated whole blood samples, N 90% Kyn reduction was ob- served for at least the first 6 h after epacadostat administration on day 1 of cycle 1 and at all time points after epacadostat administration on day 15; average inhibition over the 6 h ranged from 95% to 98% (Fig. 4A, B). Meanwhile, reductions in Kyn levels in the tamoxifen group appeared to be more modest and inconstant, evidenced by an average inhibition of 0% to 40% over the 6 h (data not shown).
There were no significant differences in inflammatory markers be- tween treatment groups at baseline; however, these baseline markers, including C-reactive protein (CRP) and interleukin-6, were elevated in both treatment groups compared with healthy volunteers (data not shown). This was not unexpected because patients with cancer general- ly have been shown to have elevated CRP and evidence of a chronic sys- temic inflammatory response [26,27]. Although minor changes were observed in some analytes during treatment, they were not statistically significant.

A large number of archival biopsy tissue samples (30 of 32 samples; 94%) were positive for IDO1 expression in tumour cells, although rela- tively low levels of expression were detected in the majority of IDO1- positive patients. PD-L1 expression was observed in only 11 of 31 evaluable samples (35%); all PD-L1–positive samples were also IDO1- positive. Only 2 samples (6%) were negative for expression of both IDO1 and PD-L1. Eighteen samples (58%) were positive for IDO1 but negative for PD-L1 (Supplement Table 4).

4. Discussion

This was an international, multicentre, randomised, open-label phase 2 study of the efficacy, safety, and tolerability of epacadostat ver- sus tamoxifen in women with histologically confirmed epithelial ovari- an cancer, primary peritoneal carcinoma, or fallopian tube cancer who had biochemical recurrence of disease (CA-125 elevation) and no other objective evidence of disease recurrence after complete response with first-line chemotherapy. This is the first report, to our knowledge, of the use of immunotherapy in patients with early relapse of ovarian cancer and the first study of a small-molecule IDO1 enzyme inhibitor as monotherapy in ovarian cancer. Study enrolment was stopped at 42 patients out of the planned 110 upon early termination by the sponsor based on slow accrual and lack of evidence of superiority for epacadostat. The majority of patients who discontinued did so because of disease progression.

At the time of the interim analysis at study termination (with 14 progression events on epacadostat, 12 progression events on tamoxifen, and 8 censored events in each arm), there was no significant difference in efficacy between epacadostat and tamoxifen as measured by investi- gator-assessed PFS, and PFS was shorter in both treatment groups for patients who had early (3 to b 12 months) versus later (≥ 12 months) CA-125 relapse following previous complete responses. PFS data were similar to those reported in a comparable patient sample from a phase 3 trial evaluating tamoxifen 20 mg BID versus thalidomide 220 mg daily, suggesting that the efficacy of tamoxifen in this study was as expected in this patient population [28].

Fig. 3. Mean (SE) epacadostat plasma concentrations on Day 15 of Cycle 1 in patients receiving epacadostat 600 mg twice daily.

Epacadostat was generally well tolerated. The majority of patients in the epacadostat and tamoxifen groups experienced ≥ 1 TEAE during the study, with fatigue being the most frequently reported TEAE in both groups. As expected, given the mechanism of action of epacadostat, irAEs were observed in the epacadostat group and were primarily skin-related events. There were no TEAEs leading to death. Epacadostat PK parameters in this study were comparable to those observed in the 600-mg BID dose group in a dose-escalation study (NCT01195311) in which epacadostat monotherapy (dose range, 50 mg once daily to 700 mg BID) was evaluated in patients with advanced malignancies [23].

Despite epacadostat not showing single-agent activity, it is impor- tant to consider the timing of this study in the context of immunother- apy development in advanced ovarian cancer. Not only is this the first reported study investigating a small-molecule IDO1 enzyme inhibitor in ovarian cancer, the protocol development predated other clinical in- vestigations with checkpoint inhibitors in this cancer setting. In addi- tion, all immune responses are accompanied by (and limited by) the generation of negative feedback mechanisms that may suppress immu- nity; IDO1 is one of many such negative feedback mechanisms [29]. IDO1 expression, which is inducible by interferon, is part of adaptive im- mune resistance mechanisms to limit physiologic inflammation [30,31] and may have a broad role in combination immunotherapies for human malignancies. The concept of evaluating the predictive role of quantita- tive or qualitative IDO1 expression in tumour tissue and immune cells in the tumour microenvironment at baseline or during therapy has not been defined but is a goal of ongoing studies. Importantly, this study demonstrated a clinically manageable safety profile with epacadostat (including irAEs), effective IDO1 inhibition activity at 600 mg BID, and frequent IDO1 expression and coexpression with PD-L1 in tumour sam- ples, which are all significant considerations given the potential future use of epacadostat in this cancer setting – likely as part of immune- based combination therapy. Preclinical evidence suggests that IDO1 in- hibition may dramatically increase the efficacy of various chemothera- peutic agents without increased toxicity, including platinum-based compounds and taxanes [32], both of which are recommended for ovar- ian cancers [2]. Although the mechanisms responsible for this potentia- tion are not fully understood, these effects were not observed in T-cell– deficient animals, suggesting that the effects may be due to the disabling of immunosuppressive mechanisms within the tumour microenviron- ment [32]. In addition, the combination of high-dose PD-L1 inhibition and cisplatin was associated with tumour burden reduction in preclini- cal models of ovarian cancers [33]. Ongoing clinical studies are evaluat- ing anti–PD-1/PD-L1 agents, including pembrolizumab, atezolizumab, and avelumab, in combination with chemotherapy in previously un- treated patients with ovarian cancers or patients who have recurrent disease (NCT02608684, NCT02440425, NCT02659384, and NCT02718417) [34]. In the current study, coexpression of IDO1 and PD-L1 was apparent in tumour biopsy samples, suggesting that combi- nation treatment of a small-molecule IDO1 enzyme inhibitor, such as epacadostat, with an immunomodulatory checkpoint inhibitor may be an important therapeutic strategy for cancer treatment beyond this clin- ical setting. Early data from the phase 1/2 dose-escalation study of epacadostat plus ipilimumab in patients with unresectable or metasta- tic melanoma (NCT01604889) are promising [35] and suggest that com- bination therapy of IDO1 inhibition and cytotoxic T-lymphocyte antigen-4 blockade may be considered in advanced ovarian cancer. Other studies are currently being conducted to evaluate epacadostat in combination with various immunomodulatory agents, including pembrolizumab (in select advanced cancers; NCT02178722), nivolumab (in select advanced cancers; NCT02327078), durvalumab (in select advanced cancers; NCT02318277), and atezolizumab (in non–small-cell lung cancer; NCT02298153). A phase 3 trial of epacadostat combined with pembrolizumab in patients with unresectable or metastatic melanoma was initiated in 2016 (NCT02752074).

In conclusion, this is the first study of the IDO1 enzyme inhibitor epacadostat in ovarian cancer and also the first report of immunothera- py use in early-relapse ovarian cancer. Epacadostat was generally well tolerated, with manageable irAEs and other adverse events. Although epacadostat monotherapy did not exhibit activity at the time of interim analysis, additional studies are in progress to assess the activity of epacadostat in combination with other immunomodulatory agents.

Fig. 4. Mean kynurenine inhibition in whole blood on (A) Day 1 and (B) Day 15 in the epacadostat group.