Phase Ib Study of Binimetinib with Paclitaxel in Patients with Platinum-Resistant Ovarian Cancer: Final Results, Potential Biomarkers, and Extreme Responders

Abstract

Purpose

Epithelial ovarian cancer (EOC) is a molecularly diverse disease. Mitogen-activated protein kinase kinase (MEK) inhibition targets tumors harboring mitogen-activated protein kinase (MAPK) pathway alterations and enhances paclitaxel-induced apoptosis in EOC. This phase Ib study evaluated the MEK inhibitor binimetinib combined with paclitaxel in patients with platinum-resistant EOC.

Experimental Design

Patients received intravenous (IV) weekly paclitaxel with oral binimetinib in three different administration schedules. Outcomes were assessed by RECIST and CGIC CA-125 response criteria. Tumor samples were analyzed using next-generation sequencing.

Results

Thirty-four patients received ≥1 binimetinib dose. A 30-mg twice daily (BID) continuous or 45-mg BID intermittent binimetinib dose were deemed the recommended phase 2 doses (RP2Ds) in combination with 80 mg/m² IV weekly paclitaxel. Rate of grade 3/4 adverse events was 65%. The best overall response rate was 18%—1 complete (CR) and 4 partial responses (PRs)—among 28 patients with RECIST-measurable disease. Eleven patients achieved stable disease (SD), yielding a clinical benefit rate (CR+PR+SD) of 57%. Response rates, per both RECIST and CA-125 criteria, were highest in the 45-mg BID continuous cohort and lowest in the 45-mg BID intermittent cohort. All 4 evaluable patients with MAPK pathway-altered tumors experienced clinical benefit.

Conclusions

The combination of binimetinib and IV weekly paclitaxel was tolerable in this patient population. The RP2D of binimetinib in combination with paclitaxel was 30 mg BID as a continuous or 45 mg BID as an intermittent dose. Although response rates were modest, a higher clinical benefit rate was seen in patients harboring alterations affecting the MAPK pathway.

Introduction

Ovarian cancer is the most lethal gynecologic malignancy (1). Greater than 70% of patients who achieve remission following first-line therapy will ultimately relapse, and virtually all of these patients will eventually develop platinum-resistant disease.

The U.S. Food and Drug Administration has approved several chemotherapy regimens for patients with platinum-resistant disease, including paclitaxel, topotecan, and liposomal doxorubicin, either alone or in combination with bevacizumab, with response rates of 7% to 30% (2-7). Taxane-mediated apoptosis involves signaling through the proapoptotic Bcl-2-like protein 11 (BIM) (8,9). BIM can be regulated via mitogen-activated protein kinase kinase/extracellular signal-regulated kinase (MEK/ERK)-mediated phosphorylation, resulting in lowered BIM levels. MEK inhibition previously has been shown to enhance paclitaxel-induced tumor apoptosis (10). Additionally, aberrant signaling through the RAS/RAF/MEK/ERK pathway is a characteristic feature of many cancers and has been shown to lead to unconstrained cell growth and cellular transformation (11,12). Among patients with epithelial ovarian cancer (EOC), RAS and/or RAF pathway gene alterations are most commonly found in low-grade serous (35-57%), endometrioid (21%), and mucinous ovarian cancers (44-85%), histologies that are relatively resistant to standard cytotoxic chemotherapy (13-19). However, RAS and RAF alterations are rare in high-grade serous ovarian cancer. Therefore, paclitaxel coupled with MEK inhibition is hypothesized to be an active combination across the spectrum of recurrent EOC histologies, both through enhancement of taxane-mediated apoptosis and selected inhibition in those histologies with sensitizing alterations within the mitogen-activated protein kinase (MAPK) pathway.

Binimetinib (MEK162, ARRY-438162) is an oral adenosine triphosphate non-competitive, highly selective inhibitor of MEK1/2. Single-agent studies of binimetinib have shown it to have modest accumulation on repeat dosing, with a terminal half-life of <10 hours and a maximal concentration observed 1-2 hours post dose. A phase I dose-escalation study in patients with KRAS- or BRAF-mutant colorectal cancer determined the maximum tolerated dose (MTD) of single-agent binimetinib as 60 mg by mouth twice daily. The goal of the current study was to determine the recommended phase 2 dose (RP2D) of oral binimetinib in combination with intravenous (IV) paclitaxel in patients with platinum-resistant EOC and to estimate the efficacy and possible biomarkers of response within this heterogeneous population.

Patients and Methods

Patients

Eligible patients were women with EOC, including fallopian tube or primary peritoneal cancer, who had platinum-resistant or platinum-refractory disease or disease no longer amenable to further platinum therapy (e.g., those with platinum hypersensitivity) as determined by the treating physician. Patients had to have measurable or evaluable (nonmeasurable) disease per Response Evaluation Criteria in Solid Tumors version 1.1 (RECIST) criteria and an Eastern Cooperative Oncology Group (ECOG) performance status of ≤2. Furthermore, patients had to have adequate hematopoietic, hepatic, and renal function at baseline, as well as normal cardiac ejection fraction per institutional standards. The research was conducted in compliance with the Declaration of Helsinki. Each participating site for NCT01649336 obtained approval by their institutional review board to conduct this research. Written informed consent was obtained from each patient or her guardian.

A tumor specimen (primary or metastatic) from archival material or fresh biopsy was required. Prohibited medical history conditions included retinal vein occlusion, acute coronary syndromes, Class II–IV heart failure, QT value corrected for heart rate using Fridericia’s formula (QTcF) ≥470 msec, uncontrolled or symptomatic brain metastases, concurrent malignancies with <5-year disease-free survival, gastrointestinal abnormalities limiting ability to take oral medications, Gilbert’s syndrome, known positivity for the human immunodeficiency virus, active hepatitis B and/or active hepatitis C, and any other disease or condition that may have impaired study participation. Restrictions were implemented for the previous administration and/or washout of ritonavir, cyclical chemotherapies, biologic therapies, small-molecule therapeutics, radiotherapy, and other investigational agents.

Study design

This open-label, multicenter, phase Ib dose-escalation study was designed to determine the RP2D of twice daily (BID) oral binimetinib administered on continuous and intermittent schedules in combination with weekly paclitaxel. Secondary objectives included assessing the potential plasma pharmacokinetic interactions between binimetinib, its metabolite AR00426032, and paclitaxel; estimating the efficacy of binimetinib in combination with paclitaxel; and assessing possible predictive biomarkers of clinical activity for the combination of binimetinib and paclitaxel based on next-generation sequencing results indicating possible activation of the MAPK pathway.

Patients were screened for eligibility within 28 days of dosing and sequentially enrolled on binimetinib dosing cohorts. Dose escalation on the binimetinib continuous dosing schedule (Cohorts 1 and 2) proceeded using a modified 3+3 design. Patients were evaluated for dose-limiting toxicities (DLTs) during cycle 1 of treatment. The dose-escalation schema called for 3 to 6 evaluable patients to be enrolled on each of the continuous binimetinib dosing cohorts (30 to 60 mg BID) until the MTD or RP2D was determined, with at least 6 evaluable patients initially treated at the dose determined to be the MTD/RP2D and up to 6 additional patients enrolled after determination of the RP2D (for a total of 12 evaluable patients).

Only one dose level was planned for the intermittent dosing cohort, with expansion to up to 12 patients if 2 or fewer DLTs were seen during cycle 1 of treatment in the first 6 patients. If any DLT was observed during cycle 1 of treatment in 2 or more of the 6 patients, the dose would have been considered not tolerable. All patients received concurrent paclitaxel at a dose of 80 mg/m² administered weekly for 3 of 4 weeks as an IV infusion over 1 hour (±10 minutes). Paclitaxel premedications (e.g., corticosteroids, diphenhydramine, H2-receptor antagonists) were administered in accordance with institutional guidelines. Study treatment was administered in 28-day cycles. Patients were monitored for protocol-specified DLTs for the first cycle of study treatment to inform dose-escalation decisions. A DLT was defined as the inability to administer at least 75% of planned binimetinib doses and all three doses of paclitaxel in cycle 1 due to related adverse events (AEs), a decrease in left ventricular ejection fraction (LVEF) of >10% from baseline and below the institution’s lower limit of normal, a grade 4 alkaline phosphatase elevation, or any of the following AEs of grade 3 or higher despite conservative management: persistent hypertension, fatigue (if ≥ grade 2 increase from baseline), rash, diarrhea, nausea/vomiting, pancreatitis, aspartate transaminase/alanine transaminase (AST/ALT) or bilirubin elevation, thrombocytopenia with signs of bleeding or grade 4 thrombocytopenia, creatinine elevation, electrocardiogram QTc prolongation, eye disorders, creatine kinase (CK) elevation with an associated ≥ grade 2 increase in creatinine, or any other clinically significant grade 3 or higher AE other than lymphopenia.

Patients self-administered oral binimetinib with water, without regard to food intake, with 12±2 hours between the morning and evening doses. For the continuous dosing schedule, binimetinib was administered at a starting dose of 30 mg BID on days 1 through 28 of every 28-day cycle, except for cycle 1, for which binimetinib dosing began on day 2. For the intermittent dosing schedule, binimetinib was administered at a dose of 45 mg BID on days 1 through 5, days 8 through 12, and days 15 through 19 of every 28-day cycle.

Patients were permitted to remain on study treatment until progression of disease (PD) or intolerable toxicity. Patients who permanently discontinued paclitaxel but did not meet any of the other treatment discontinuation criteria were permitted to continue to receive single-agent binimetinib at their current dose level and dosing schedule at the investigator’s discretion.

Safety and efficacy assessments

AEs and clinical laboratory test results were graded using the National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.03. Echocardiogram (ECHO) and electrocardiogram were performed at baseline and prespecified timepoints throughout the study. A clinical ophthalmic exam included slit lamp examination, visual acuity testing, a color vision test, tonometry, optical coherence tomography (OCT), funduscopy, and color fundus photography for retinal abnormalities, and occurred at baseline, day 8 of cycle 1, day 1 of each cycle, and at treatment discontinuation. Radiographic responses were assessed every 2 months using RECIST [Version 1.1]. The safety cohort included all patients who received at least one dose of binimetinib and paclitaxel. The RECIST response evaluable cohort included all patients in the safety cohort who had measurable disease at baseline tumor assessment. The CA-125 response evaluable cohort included all patients in the safety cohort who had a baseline CA-125 that was at least twice the upper limit of normal (ULN) within 2 weeks prior to study treatment initiation. CA-125 response was evaluated according to the Gynecologic Cancer InterGroup (GCIG) CA-125 response criteria.

The objective response rate (ORR), defined as the number of patients with a best overall response of complete response (CR) or partial response (PR) divided by the total number of patients in the RECIST response evaluable cohort, was reported by treatment group and overall using exact binomial 95% confidence intervals (CIs). The clinical benefit rate was determined based on the number of patients in the RECIST response evaluable cohort experiencing radiographic response or stable disease (CR+PR+SD).

Pharmacokinetic analyses

Venous blood samples were collected at specified time points (Supplementary Table S1) to determine the plasma concentration of binimetinib, AR00426032, and paclitaxel.

Biomarker analyses

Biomarker analyses were performed on archived formalin-fixed, paraffin-embedded (FFPE) tumor tissue collected at screening for all patients in order to assess mutations and copy number variations in genes related to cancer and chemotherapy metabolism by next-generation sequencing of a panel of >150 genes. Matched tumor and blood samples from selected patients were also analyzed using the Memorial Sloan Kettering-Integrated Mutation Profiling of Actionable Cancer Targets (MSK-IMPACT) assay, and sequence read alignment, processing, as well as single nucleotide variant and copy number alteration detection were performed, as previously described (20).

Results

Study population

Thirty-six patients across four sites were enrolled in the study between August 9, 2012, and September 22, 2014, including 2 patients who withdrew before receiving study treatment (one due to grade 3 hyperglycemia and the other due to a lack of health insurance). The 34 patients who received at least one dose of study drugs (safety cohort) were enrolled on three sequential dosing cohorts (Table 1). Twenty-eight patients had RECIST-measurable disease at baseline, 22 had a CA-125 level that was at least twice the ULN at baseline and were included in the CA-125 response evaluable cohort, and 28 had FFPE tissue available for inclusion in the biomarker cohort.

Table 1.

Dosing cohorts

Cohort	Dose and schedule	Number of patients
Cohort 1	30 mg PO BID binimetinib (continuous) + 80 mg/m2 IV paclitaxel (3 weeks on/1 week off)	14
Cohort 2	45 mg PO BID binimetinib (continuous) + 80 mg/m2 IV paclitaxel (3 weeks on/1 week off)	8
Cohort 3	45 mg PO BID binimetinib (intermittent, days 1–5 of each week, 3 weeks on/1 week off) + 80 mg/m2 IV paclitaxel (3 weeks on/1 week off)	12

PO, by mouth; BID, twice daily; IV, intravenous

All patients have discontinued study treatment—17 patients (50%) due to radiographic PD, 7 (21%) due to clinical PD, 8 (24%) due to AEs or inability to tolerate study treatment, 1 (3%) due to investigator discretion based on the patient’s noncompliance with oral drug administration, and 1 (3%) due to the termination of the study by the sponsor following 24 months on study treatment (this patient continues on treatment through a single patient compassionate Investigational New Drug use exemption).

Patient characteristics per dose level are listed in Table 2. Most patients had disease of high-grade serous (65%) or low-grade serous (18%) histology. The median number of prior therapies was 4 (range, 1–14). All patients had undergone prior surgery and received prior taxane therapy, with 74% of patients having a best response of SD or progressive disease on their most recent prior taxane regimen. The progression-free interval on the most recent therapy prior to study entry was generally short (median, 1 month).

Table 2.

Baseline disease characteristics (safety cohort)

Characteristics	Binimetinib + paclitaxel dose
	30 mg BID continuous +80 mg/m2 (n = 14)	45 mg BID continuous +80 mg/m2 (n = 8)	45 mg BID intermittent +80 mg/m2 (n = 12)	Total (N = 34)
Age, y
Mean ± standard deviation	62.1 ± 10.44	62.9 ± 8.97	59.8 ± 9.20	61.5 ± 9.48
Median (range)	65 (42–77)	64 (43–72)	58 (42–74)	63 (42–77)
Race, N (%)
Asian	2 (14)	0	0	2 (6)
Black/African American	1 (7)	0	0	1 (3)
White	11 (79)	8 (100)	12 (100)	31 (91)
Tumor type, N (%)
Epithelial ovarian	13 (93)	7 (88)	8 (67)	28 (82)
Fallopian tube	1 (7)	0	0	1 (3)
Primary peritoneal	0	1 (13)	4 (33)	5 (15)
Histology, N (%)
Clear cell	0	0	1 (8)	1 (3)
Endometrioid	0	0	1 (8)	1 (3)
High-grade serous	11 (79)	4 (50)	7 (58)	22 (65)
Mixed epithelial histology	2 (14)	1 (13)	1 (8)	4 (12)
Low-grade serous	1 (7)	3 (38)	2 (17)	6 (18)
Stage, N (%)
IIA	1 (7)	0	0	1 (3)
IIC	0	0	1 (8)	1 (3)
III	0	0	1 (8)	1 (3)
IIIC	5 (36)	3 (38)	2 (17)	10 (29)
IV	7 (50)	2 (25)	8 (67)	17 (50)
Unknown	1 (7)	3 (38)	0	4 (12)
ECOG Performance Status, N (%)
0	6 (43)	3 (38)	8 (67)	17 (50)
1	8 (57)	5 (63)	3 (25)	16 (47)
2	0	0	1 (8)	1 (3)
Prior therapy, N (%)
Prior surgery	14 (100)	8 (100)	12 (100)	34 (100)
Prior taxane therapy	14 (100)	8 (100)	12 (100)	34 (100)
Prior biologic therapy	9 (64)	3 (38)	8 (67)	20 (59)
Prior hormonal therapy	3 (21)	1 (13)	1 (8)	5 (15)
Prior radiation	0	0	3 (25)	3 (9)
Number of prior lines of therapy
Mean ± standard deviation	5.1 ± 3.43	3.1 ± 1.64	4.5 ± 2.35	4.4 ± 2.76
Median (range)	5 (1–14)	3 (1–6)	5 (1–8)	4 (1–14)
Best response to most recent prior taxane, N (%)
CR	2 (14)	1 (13)	4 (33)	7 (21)
PR	1 (7)	0	1 (8)	2 (6)
SD	4 (29)	3 (38)	3 (25)	10 (29)
PD	5 (36)	2 (25)	3 (25)	10 (29)
Unknown	2 (14)	2 (25)	1 (8)	5 (15)

BID, twice daily; ECOG, Eastern Cooperative Oncology Group; CR, complete response; PR, partial response; SD, stable disease; PD, progression of disease

Twenty-six of the 28 patients in the biomarker cohort had gene alterations identified through next-generation sequencing. The most common mutations and copy number alterations identified are listed in Table 3; a complete list of identified alterations by patient is listed in Supplementary Table S2. An additional 4 patients had MSK-IMPACT analyses performed on tumor and matched blood.

Table 3.

Somatic gene alterations observed by next-generation sequencing in more than 1 patient (biomarker cohort^a)

Variant type gene	Binimetinib + paclitaxel dose, n (%)
	30 mg BID continuous +80 mg/m2 (n = 12)	45 mg BID continuous +80 mg/m2 (n = 6)	45 mg BID intermittent +80 mg/m2 (n = 10)	Total (N = 28)
Total patients with any variant in gene status	12 (100)	5 (83)	9 (90)	26 (93)
Short variant	10 (83)	4 (67)	9 (90)	23 (82)
TP53	9 (75)	3 (50)	6 (60)	18 (64)
KRAS	1 (8)	1 (17)	1 (10)	3 (11)
ARID1A	0	0	2 (20)	2 (7)
BRCA1	0	0	2 (20)	2 (7)
BRCA2	1 (8)	0	1 (10)	2 (7)
BRIP1	1 (8)	1 (17)	0	2 (7)
PIK3CA	0	0	2 (20)	2 (7)
TSC2	1 (8)	0	1 (10)	2 (7)
Copy number alteration	6 (50)	2 (33)	5 (50)	13 (46)
CCNE1	3 (25)	0	1 (10)	4 (14)
KRAS	1 (8)	1 (17)	1 (10)	3 (11)
MCL1	1 (8)	1 (17)	1 (10)	3 (11)
CDKN2A	2 (17)	0	0	2 (7)
CDKN2B	2 (17)	0	0	2 (7)
MYC	2 (17)	0	0	2 (7)
NFKBIA	0	1 (17)	1 (10)	2 (7)
PIK3CA	1 (8)	0	1 (10)	2 (7)
RB1	1 (8)	1 (17)	0	2 (7)
ZNF217	0	0	2 (20)	2 (7)
Rearrangement	2 (17)	1 (17)	0	3 (11)

BID, twice daily

^aThe biomarker cohortcontains 28 patients with adequate FFPE tissue available for analyses

Determination of MTD and DLTs

Thirteen of the 14 patients treated in cohort 1 were evaluable for DLT. Two (15%) of 13 evaluable patients treated in cohort 1 (30 mg of binimetinib BID continuously) developed a DLT (grade 2 nausea and neutropenia leading to inability to receive at least 75% of planned binimetinib doses during cycle 1 in 1 patient and grade 3 fatigue in the other patient). At the escalated dose of 45 mg BID continuously (cohort 2), 8 patients were enrolled and 7 were evaluable for DLT. Three (43%) of 7 evaluable patients developed a DLT (grade 2 maculopapular rash in 1 patient and grade 2 fatigue in another patient led to inability to receive at least 75% of planned doses of binimetinib during cycle 1, and grade 3 stomatitis occurred in a third patient). As such, 30 mg BID was determined to be the RP2D of binimetinib administered continuously in combination with weekly paclitaxel.

In cohort 3 (45 mg BID intermittent binimetinib dosing), the binimetinib dose of 45 mg BID administered on days 1 to 5 weekly for 3 of every 4 weeks was confirmed as the RP2D in combination with weekly paclitaxel. In this cohort, 1 (8%) of 12 evaluable patients developed a DLT (grade 3 asthenia).

Safety

The most commonly reported all-cause and all grade AEs (those reported in ≥30% of patients) were as follows: fatigue (22 patients, 65%); diarrhea, nausea, and vomiting (each in 19 patients, 56%); abdominal pain, maculopapular rash, and peripheral edema (each in 16 patients, 47%); anemia (14 patients, 41%); stomatitis (13 patients, 38%); and constipation (11 patients, 32%).

AEs reported in ≥10% of patients and those reported at a level of grade 3 to 5 in >1 patient are listed in Supplementary Tables S3 and S4. The most common AEs related to binimetinib or paclitaxel, as reported by investigators, are listed by cohort in Table 4. Grade 3 to 5 AEs reported for >2 patients were anemia, fatigue, nausea, neutropenia, pulmonary embolism, small intestinal obstruction, stomatitis, and vomiting. The overall incidence of grade 3 to 5 AEs was highest in cohort 2 (88%) and lowest in cohort 1 (64%). All grade 5 AEs were considered to be secondary to PD.

Table 4.

Adverse events assessed as related to binimetinib or paclitaxel reported in ≥10% of patients (safety cohort^a)

Preferred term	Binimetinib + paclitaxel dose, n (%)
	30 mg BID continuous +80 mg/m2 (n = 14)	45 mg BID continuous +80 mg/m2 (n = 8)	45 mg BID intermittent +80 mg/m2 (n = 12)	Total (N = 34)
Patients with any AE	14 (100)	8 (100)	12 (100)	34 (100)
Grade 1	0 (0)	0 (0)	4 (33)	4 (12)
Grade 2	9 (64)	2 (25)	3 (25)	14 (41)
Grade 3	5 (36)	6 (75)	5 (42)	16 (47)
Grade 4	0 (0)	0 (0)	0 (0)	0 (0)
Grade 5	0 (0)	0 (0)	0 (0)	0 (0)
Fatigue	8 (57)	6 (75)	5 (42)	19 (56)
Diarrhea	6 (43)	5 (63)	6 (50)	17 (50)
Rash maculopapular	5 (36)	3 (38)	6 (50)	14 (41)
Edema peripheral	6 (43)	4 (50)	3 (25)	13 (38)
Anemia	5 (36)	4 (50)	3 (25)	12 (35)
Nausea	4 (29)	5 (63)	3 (25)	12 (35)
Stomatitis	4 (29)	3 (38)	4 (33)	11 (32)
Alopecia	4 (29)	4 (50)	2 (17)	10 (29)
Vomiting	5 (36)	3 (38)	2 (17)	10 (29)
Neuropathy peripheral	3 (21)	3 (38)	2 (17)	8 (24)
Blood creatine phosphokinase increased	2 (14)	4 (50)	0	6 (18)
Dry skin	4 (29)	1 (13)	1 (8)	6 (18)
Decreased appetite	4 (29)	0	1 (8)	5 (15)
Dermatitis acneiform	4 (29)	1 (13)	0	5 (15)
Dry mouth	3 (21)	2 (25)	0	5 (15)
Dysgeusia	3 (21)	1 (13)	1 (8)	5 (15)
Epistaxis	1 (7)	3 (38)	1 (8)	5 (15)
Hypokalemia	2 (14)	3 (38)	0	5 (15)
Myalgia	2 (14)	3 (38)	0	5 (15)
Retinal edema	1 (7)	0	4 (33)	5 (15)
Gastroesophageal reflux disease	1 (7)	2 (25)	1 (8)	4 (12)
Hypomagnesemia	1 (7)	2 (25)	1 (8)	4 (12)
Neutropenia	3 (21)	0	1 (8)	4 (12)

BID, twice daily; AE, adverse event

^aThe safety cohort includes patients who received at least 1 dose of binimetinib and paclitaxel.

AEs resulted in dose reduction of binimetinib for 7 (88%) of 8 patients in cohort 2 and 1 (8%) of 12 patients in cohort 3; none of the patients in cohort 1 underwent a dose reduction. Eight patients (24%) experienced at least one AE that led to a dose reduction of paclitaxel, with no correlation observed with binimetinib dose level.

Ocular toxicities are a class effect of MEK inhibitors; therefore, all enrolled patients were carefully monitored with serial ophthalmic examinations, as detailed above. Newly occurring retinal AEs were considered clinically significant in 2 patients (6%), both of whom were treated in cohort 2. One patient developed a serous retinal detachment on cycle 1 day 8; her binimetinib dose was reduced from 45 mg BID to 30 mg BID, and her serous detachment resolved by cycle 2 day 1. The other patient developed new subretinal fluid collection on cycle 1 day 8, with an associated change in visual acuity from 20/20 to 20/30 in one eye. She continued the same dose of binimetinib (45 mg BID), and the fluid and visual acuity changes resolved by cycle 2 day1.

No patients died while receiving treatment. Four patients died during follow-up within 30 days of their last study treatment (9-22 days after their last study treatment). Three of these patients were admitted to the hospital for symptoms deemed related to PD and then ultimately discharged to hospice care. A fourth patient was admitted with a small bowel obstruction deemed related to PD and died on comfort care. An additional 2 patients died during the prespecified follow-up period, but more than 30 days after completion of study treatment (one 60 days after completion of study treatment due to PD with symptomatic brain metastasis, and the other 80 days after completion of study treatment due to PD while on hospice care). The cause of death for all 6 patients was determined to be related to PD.

Pharmacokinetic analyses

The cycle 1 geometric mean plasma concentration-time profiles of binimetinib and its metabolite AR00426032 were similar for days 7, 8, and 15 for the two binimetinib doses (30 mg BID and 45 mg BID) evaluated under the continuous dosing schedule and for days 1, 8, and 15 for the 45-mg BID binimetinib dose evaluated under the intermittent dosing schedule (Supplementary Figures S1–S3). Concentrations of binimetinib and AR00426032 were typically highest at 2 hours after administration. During the intermittent dosing schedule, plasma concentrations for binimetinib were similar on days 1, 8, and 15, with overlapping error bars, indicating that accumulation with repeat intermittent dosing was not significant. For the continuous dosing schedule, plasma concentrations of both binimetinib and AR00426032 were similar on day 7 (no paclitaxel) and days 8 and 15 (paclitaxel administered), with overlapping error bars. This suggests that paclitaxel did not have a clinically relevant effect on exposure for binimetinib.

Efficacy

Within the RECIST evaluable cohort (N = 28), 1 patient (4%) achieved a confirmed CR and 4 patients (14%) achieved a PR, for an ORR of 18% (95% CI, 6.1–36.9). An additional 11 patients (39%) had a best response of SD, yielding a clinical benefit rate (CR+PR+SD) of 57%. The ORR was highest in cohort 2 (4/7, 57%) and lowest in cohort 3 (0/11, 0%); there was 1 response in 10 patients (10%) in cohort 1. Figure 1 displays a waterfall plot of best response in patients with measurable disease. Of note, all patients exhibiting an alteration affecting the MAPK pathway experienced clinical benefit: the patient with a CR had a BRAF fusion, 1 patient with a PR had a KRAS G12D mutation, and 2 patients with SD had either a KRAS G12D or a KRAS G12V mutation.

Figure 1. Waterfall plot displaying responses among patients with RECIST-measurable disease.

The dosing cohort is displayed above each bar. Alterations of known significance identified through next-generation sequencing are displayed below each patient. #, next-generation sequencing not performed; 30, binimetinib 30 mg PO BID in combination with IV paclitaxel; 45, binimetinib 45 mg PO BID in combination with IV paclitaxel; 45i, binimetinib 45 mg PO BID intermittent dosing in combination with IV paclitaxel; RECIST, Response Evaluation Criteria in Solid Tumors; HGS, high-grade serous; LGS, low-grade serous; PO, by mouth.

Twenty of the 22 patients in the CA-125 response evaluable cohort had measurable disease. In the CA-125 response evaluable cohort (N = 22), the ORR was 27%; 3 patients (14%) achieved a CR, 3 (14%) achieved a PR, and the remaining 16 (73%) had SD per GCIG CA-125 response criteria. The ORR was highest in cohort 2 (67%) and lowest in cohort 3 (10%). Best overall response in patients who had either measurable disease (n = 28) or were evaluable for CA-125 (n = 22) was 23% (N = 30). Rules for determining best response in this cohort are described in Table S5. Results by cohort are presented in Table 5. The duration of binimetinib treatment was longest in the 45 mg BID binimetinib (continuous) cohort (mean 37.5 ± SD 14.9 weeks) and shortest in the 30 mg BID binimetinib (continuous) cohort (mean 16.8 ± SD 14.9 weeks). Eight patients (24%) received at least 6 months of binimetinib treatment.

Table 5.

Best response (combined RECIST and CA-125 response evaluable cohort)

Response	Binimetinib + paclitaxel dose, n (%)
	30 mg BID continuous +80 mg/m2 (n = 12)	45 mg BID continuous +80 mg/m2 (n = 7)	45 mg BID intermittent +80 mg/m2 (n = 11)	Total (N = 30)
Best response, N (%)
CR	0	1 (14)	0	1 (3)
PR	2 (17)	3 (43)	1 (9)	6 (20)
SD	4 (33)	2 (29)	3 (27)	9 (30)
PD	6 (50)	0	3 (27)	9 (30)
Not evaluable	0	1 (14)	4 (36)	5 (17)
Response rate, %	16.7	57.1	9.1	23.3
95% CI	2.1–48.4	18.4–90.1	0.2–41.3	9.9–42.3

RECIST, Response Evaluation Criteria in Solid Tumors; BID, twice daily; CR, complete response; PR, partial response; SD, stable disease; PD, progression of disease; CI, confidence interval.

Rules for determining best response in this cohort are described in Table S5.

Two patients had exceptional responses to therapy, with periods of prolonged remission following treatment discontinuation. One patient was initially diagnosed with high-grade serous ovarian cancer in 2010 and was treated with an optimal debulking and carboplatin/paclitaxel chemotherapy at that time. She developed recurrent disease 9 months after completion of initial chemotherapy and was subsequently treated with five lines of chemotherapy for recurrent disease, with progression on each line. The patient entered the study with nonmeasurable disease in October 2012 and was treated with continuous binimetinib 30 mg BID in combination with weekly paclitaxel until May 2013. At that time, paclitaxel was stopped due to grade 2 nail toxicity. The patient subsequently continued on binimetinib alone until July 2013, at which time she was removed from protocol due to her unwillingness to comply with the continued oral administration of the study drug, as directed. She achieved a CR per CA-125 criteria. She had small-volume calcific disease at the time of study treatment discontinuation and remained on observation alone for 4 years without evidence of progression. An analysis of her tumor by next-generation sequencing identified a TP53 exon 6 mutation and amplification of ERBB2 (fold change, 22.16), CCNE1 (fold change, 2.55), and PIK3CA (fold change, 2.06). Sequencing of her tumor and matched blood by the MSK-IMPACT assay also identified SMARCA4 amplification (fold change, 3.9).

Another patient was initially diagnosed with stage IIIC low-grade serous ovarian cancer in 2007. She underwent an optimal debulking followed by 6 cycles of carboplatin/paclitaxel chemotherapy. She was found to have biopsy-proven recurrent disease in March 2013 and entered the study in April 2013. On study, she was treated with continuous binimetinib 45 mg BID in combination with weekly paclitaxel. In July 2013, her binimetinib dose was decreased to 30 mg BID due to edema, and in August 2013, her paclitaxel dose was reduced to 60 mg weekly due to neuropathy. She achieved a radiographic CR with concurrent normalization of her CA-125 level and remained on treatment for 7 months, at which time she discontinued therapy due to the development of pneumonitis (Figure S4). Following her removal from the study, she remained on observation without evidence of disease for an additional 25 months until November 2015, when she developed PD. She subsequently received exemestane and progressed following 3 months of treatment. Next-generation sequencing revealed she had a paracentric inversion within the long arm of chromosome 7, leading to an in-frame fusion between the BRAF kinase domain and the Cullin 1 protein. Expression of this fusion was confirmed at the transcript level by whole transcriptome sequencing (14).

Discussion

This study evaluated the use of IV paclitaxel, administered 3 weeks on and 1 week off, in combination with the oral MEK inhibitor binimetinib in genomically unselected patients with platinum-resistant ovarian cancer. Next-generation sequencing was performed on archival tissue from all patients when adequate tissue could be obtained (28/36, 78%).

The RP2D of binimetinib in combination with IV weekly paclitaxel was found to be 30 mg BID administered orally in a continuous dosing regimen and 45 mg BID administered orally in an intermittent dosing regimen. The combination of oral binimetinib and IV paclitaxel was found to be tolerable. The most common AEs assessed as related to binimetinib or paclitaxel were fatigue, diarrhea, rash, and peripheral edema. Among those on the continuous binimetinib regimen, most patients (88%) treated at the 45-mg BID dose level were reduced to the 30-mg BID dose level due to AEs. A total of 12 patients (4 in the 30-mg continuous cohort, 5 in the 45-mg continuous cohort, and 3 in the 45-mg intermittent cohort) discontinued paclitaxel due to an AE. All objective responses (CR+PR) occurred in patients receiving continuous dosing, suggesting that this may be a preferable approach.

Evaluation of efficacy was limited, as this was a heavily pretreated patient population. The median number of prior therapies was 4 (range, 1–14), and all patients had received prior taxane therapy. Seventy-four percent of patients had a best response of SD or progression when last treated with taxane therapy. Among those with measurable disease, the ORR was 18%, with a clinical benefit rate of 57%. In the AURELIA study, patients with platinum-resistant ovarian cancer with 1 or 2 prior lines of systemic therapy (platinum refractory excluded) were randomized to physician’s choice of chemotherapy with weekly paclitaxel, topotecan or pegylated liposomal doxorubicin given as a single agent or in combination with bevacizumab; the ORR to single-agent chemotherapy was 12.6%, and 30.9% in combination with bevacizumab (21,22). A subsequent study examined weekly paclitaxel alone versus weekly paclitaxel in combination with linsitinib, an oral dual inhibitor of insulin-like growth factor-1 receptor and insulin receptor; eligibility was restricted to patients with platinum-resistant or -refractory ovarian cancer with a maximum of 2 prior lines of chemotherapy and no prior exposure to weekly paclitaxel. There was no reported improvement in ORR with the addition of linsitinib dosed either continuously or on an intermittent schedule versus paclitaxel alone. The ORR to weekly paclitaxel in that population was 34%, with an ORR of 17.7% for patients treated with weekly paclitaxel and intermittent linsitinib and 21.6% for patients treated with weekly paclitaxel and continuous linsitinib (23). It is difficult to compare response rates across these studies given that the current study enrolled a heavily pretreated patient population, most of whom had stable or progressive disease when most recently treated with taxane. It is notable, however, that among those with RECIST evaluable disease, all 4 patients with alterations affecting the MAPK pathway detected by next-generation sequencing achieved a clinical benefit (2 with SD [1 for 5.5 months and the other for 7.4 months], 1 with a PR, and 1 with a CR). Two of these patients had low-grade serous ovarian cancer, 1 had endometrioid, and 1 had a mixture of high-grade serous and clear cell histology. In other studies, KRAS mutations have been found in 12% to 29% of endometrioid ovarian cancer cases, 12% to 13% of clear cell ovarian cancers, and 13% to 60% of mucinous ovarian cancers (24-28). Using immunohistochemistry, expression of active MAPK has been detected in up to 80% of low-grade serous ovarian carcinomas, with KRAS mutations identified in 19% to 55% and BRAF mutations in 0% to 33% (29,30). These rarer histologies of ovarian cancer, which together account for approximately 20% of EOC cases, are frequently refractory to chemotherapy and extremely challenging to treat in the recurrent setting. This underscores the relevance of a predictive biomarker of response within these patient populations. Activation of the MAPK pathway within these rare histologies of ovarian cancer (low-grade serous, endometrioid, clear cell, and mucinous) may potentially sensitize to treatment with MEK inhibition.

Outside of BRCA-altered disease, no reliable biomarkers predictive of treatment response have been identified in ovarian cancer. Patients with platinum-resistant disease are frequently cycled through multiple lines of cytotoxic chemotherapy, with cumulative side effects but limited benefit. The results of this study suggest that alterations affecting the MAPK pathway may serve as predictive biomarkers across EOC histologies for response to paclitaxel in combination with MEK inhibition. Based on these results, a phase II study of IV paclitaxel in combination with oral binimetinib restricted to patients with alterations within the MAPK pathway is warranted.

Translational Relevance.

The majority of patients diagnosed with epithelial ovarian cancer (EOC) will eventually develop platinum-resistant disease with associated limited responses to chemotherapy. Preclinical evidence suggests that concurrent treatment with mitogen-activated protein kinase kinase (MEK) inhibition and taxanes may promote taxane-induced cell death. Many of the more chemoresistant histologies of EOC, including low-grade serous, mucinous, and endometrioid carcinomas, harbor MAPK pathway alterations, potentially conferring sensitivity to MEK inhibitor therapy. This phase Ib study evaluated the combination of intravenous weekly paclitaxel plus the oral MEK inhibitor binimetinib in heavily pretreated patients with platinum-resistant EOC. The results establish the recommended phase 2 dose of binimetinib in combination with paclitaxel. Across histologies, all evaluable patients with alterations affecting the mitogen-activated protein kinase (MAPK) pathway displayed clinical benefit, suggesting that MAPK pathway alterations may represent useful biomarkers to identify patients with EOC who are most likely to respond to treatment with paclitaxel in combination with binimetinib.