Phase I Study of the Combination of Vemurafenib, Carboplatin, and Paclitaxel in Patients with BRAF-Mutated Melanoma and Other Advanced Malignancies

Abstract

Background:

BRAF inhibitors are effective against selected BRAF^V600-mutated tumors. Preclinical data suggest that BRAF inhibition in conjunction with chemotherapy has increased therapeutic activity.

Methods:

Patients with advanced cancers and BRAF mutations were enrolled to the dose-escalation study (3+3 design) to determine the maximum tolerated dose (MTD), dose limiting toxicities (DLT).

Results:

Nineteen patients with advanced cancers and BRAF mutations were enrolled and received vemurafenib (480–720 mg orally twice a day [BID]), carboplatin (AUC [area under the curve] 5–6 intravenously every 3 weeks), and paclitaxel (100–135 mg/m² intravenously every 3 weeks). The MTD was not reached and vemurafenib 720mg BID, carboplatin AUC5 and paclitaxel 135mg/m2 was the last safe dose level. DLTs included persistent grade (G) 2 creatinine elevation (n=1), G3 transaminitis (n=1), and G4 thrombocytopenia (n=1). Non-DLTs ≥G3 that occurred in >2 patients included G3/4 neutropenia (n=5), G3/4 thrombocytopenia (n=5), G3 fatigue (n=4), and G3 anemia (n=3). Of the 19 patients, 5 (26%; all with melanoma) had a partial response (PR; n=4) or complete response (CR; n=1); these responses were mostly durable, lasting 3.1–54.1 months. Of 13 patients previously treated with BRAF and/or MEK inhibitors, 4 (31%) had CR (n=1) or PR (n=3). Patients not treated with prior platinum therapy had higher response rate than those who did (45% vs. 0%, P=0.045).

Conclusions:

The combination of vemurafenib, carboplatin, and paclitaxel is well tolerated and demonstrates encouraging activity, predominantly in patients with advanced melanoma and BRAF^V600 mutations irrespective of prior treatment with BRAF and/or MEK inhibitors.

Precis:

The combination of vemurafenib, carboplatin, and paclitaxel demonstrated an acceptable safety profile in patients with advanced cancers and BRAFV600 mutations. In addition, promising signals of anticancer efficacy were observed in advanced melanoma with BRAFV600 mutations previously treated with BRAF and/or MEK inhibitors.

INTRODUCTION

Mutations in the BRAF oncogene, which encodes a serine/threonine kinase leading to activation of the MAPK pathway, have been reported in 7–9% of advanced cancers, with the highest incidence in melanoma (50%) followed by papillary thyroid cancer (45%), low-grade serous ovarian cancer (35%), colorectal cancer (11–12%), and non–small cell lung cancer (NSCLC; 3–5%)^1–7. The most common activating somatic point mutation is BRAF^V600E, a substitution of valine to glutamic acid in the glycine-rich loop⁴. BRAF inhibitors such as vemurafenib and dabrafenib have demonstrated preclinical and clinical activity against several advanced cancers with BRAF^V600 mutations and have been approved by the U.S. Food and Drug Administration for the treatment of advanced melanoma, Erdheim-Chester disease, and NSCLC (in combination with trametinib) with BRAF^V600E or other BRAF^V600 mutations^8–11.

Carboplatin and paclitaxel have demonstrated antitumor activity in chemotherapy-naïve patients with metastatic melanoma, with response rates of 11–18% and a median progression-free survival (PFS) duration of 4.2 months^{12, 13}. Carboplatin and paclitaxel are the standard of care for other malignancies known to have BRAF mutations, including advanced ovarian cancer and NSCLC^{14, 15}. Preclinical studies demonstrated that BRAF inhibition plus chemotherapy could inhibit cell growth in in vitro and in vivo xenograft models. This led to the development of combined sorafenib, carboplatin, and paclitaxel treatment for metastatic melanoma, which despite encouraging early data failed in randomized trials^{13, 16–18}. This failure could have been due to sorafenib’s weak inhibitory activity on the BRAF kinase and its lack of molecular selection for tumors harboring BRAF mutations. We hypothesized that unlike sorafenib, vemurafenib is a potent selective inhibitor of the mutated BRAF kinase that can have synergistic activity with carboplatin and paclitaxel. Therefore, we designed a phase I clinical study to determine the maximum tolerated dose (MTD), safety, and early signals of clinical activity of the combination of vemurafenib, carboplatin, and paclitaxel in patients with BRAF-mutated advanced cancers.

PATIENTS AND METHODS

Study design and objectives

This was an open-label, non-randomized, 3+3 dose-escalation phase I study (NCT01636622) of the combination of vemurafenib, carboplatin, and paclitaxel in patients with BRAF-mutated advanced cancers. The primary endpoints were to establish the regimen’s MTD or recommended phase II dose and to evaluate its dose-limiting toxicities (DLTs) and safety. Secondary endpoints were to evaluate early signals of efficacy, such as response rate. The study was performed at The University of Texas MD Anderson Cancer Center. The study protocol was approved by MD Anderson’s Institutional Review Board (IRB). All patients provided written informed consent before starting study-related procedures.

Eligible patients had a histologically or cytologically confirmed diagnosis of a refractory advanced solid tumor harboring a BRAF mutation detected in a Clinical Laboratory Improvement Amendments–approved laboratory; had measurable disease according to Response Evaluation Criteria in Solid Tumors, version 1.1 (RECIST 1.1)¹⁹; and had been off prior systemic cancer therapies for at least 3 weeks (in the case of chemotherapy) or 5 half-lives (in the case of biologic treatments except BRAF inhibitors, which required no washout). Other inclusion criteria included a QTc interval of <500 ms; an Eastern Cooperative Oncology Group performance status of 0–2; and adequate bone marrow, liver, and renal function (total bilirubin level ≤2 × upper limit of normal [ULN]; aspartate aminotransferase [AST] and alanine aminotransferase [ALT] levels ≤2.5 × ULN or, if liver metastasis was present, ≤5 × ULN; serum creatinine level ≤2 × ULN; platelet count >75,000/ml; absolute neutrophil count >1000/ml; and hemoglobin level >8.0 g/dL).

Treatment dose levels are given in Table 1. Vemurafenib (480–720 mg) was given orally twice daily for 21 days starting the evening of the day after paclitaxel (100–135 mg/m²) and carboplatin (area under the curve [AUC] 5–6) administration. Paclitaxel and carboplatin were given intravenously on day 1 of the 21-day cycle until disease progression or unacceptable toxicity. Safety was assessed using the NCI Common Terminology Criteria for Adverse Events v4, and DLTs were evaluated during the first 21 days of therapy.

Table 1.

Treatment dose levels and dose-limiting toxicities.

Dose level	Vemurafenib, mg PO BID	Paclitaxel, mg/m2 IV every 3 weeks	Carboplatin, AUC IV every 3 weeks	No. of patients (no. evaluable for DLTs)	DLT event
1	480	100	5	7 (7)	G2 creatinine elevation >7 days
2	720	100	5	4 (4)
3	720	135	5	7 (6)	G3 AST elevation, G4 thrombocytopenia
4	720	135	6	1 (1)

Abbreviations: PO, per oral; BID, twice a day; IV, intravenous; AUC, area under the curve; DLTs, dose-limiting toxicities; G, grade; AST, aspartate aminotransferase.

DLTs were treatment-related grade 3 or higher hemolytic anemia; grade 4 thrombocytopenia; grade 3 or higher febrile neutropenia; grade 4 neutropenia lasting >7 consecutive days; grade 3 creatinine elevation; grade 3 elevation of bilirubin, ALT, or AST; any other grade 3 non-hematologic toxicity leading to treatment interruption for >7 consecutive days; any grade 4 non-hematologic toxicity; any intolerable grade 2 or 3 non-hematologic toxicity requiring a dose reduction for the second cycle of treatment; and any treatment-related adverse events resulting in a treatment delay of >7 consecutive days during cycle 1. Therapy response was assessed according to RECIST 1.1 with computed tomography (CT) and magnetic resonance imaging (MRI) done at baseline and then every 2 weeks¹⁹.

Plasma collection and cell-free DNA BRAF^V600 mutation testing

Whole blood was collected in ethylenediaminetetraacetic acid–containing tubes and centrifuged and spun twice within 2 hours to yield plasma. The QIAamp Circulating Nucleic Acid kit (Qiagen, Valencia, CA) was used to isolate cell-free DNA (cfDNA) according to the manufacturer’s instructions. At least 8 ng (if available) of unamplified cfDNA was tested with a droplet digital polymerase chain reaction (ddPCR) BRAF ^V600E mutation–specific assay and/or multiplex BRAF ^V600 Screening Kit (Bio-Rad, Pleasanton, CA) to distinguish the wild-type allele from the 3 most common mutations (BRAF ^V600E, BRAF ^V600K, and BRAF ^V600R); with a multiplex ddPCR KRAS ^G12/G13 Screening Kit (Bio-Rad) to distinguish the wild-type allele from the 7 most common mutations (KRAS ^G12A, KRAS ^G12C, KRAS ^G12D, KRAS ^G12R, KRAS ^G12S, KRAS ^G12V, and KRAS ^G13D); with a multiplex ddPCR NRAS ^G12/G13 Screening Kit (Bio-Rad) to distinguish the wild-type allele from the 8 most common mutations (NRAS^G12A, NRAS ^G12C, NRAS ^G12D, NRAS ^G12S, NRAS ^G12V, NRAS ^G13D, NRAS ^G13R, and NRAS^G13V); and with a multiplex ddPCR NRAS ^Q61 Screening Kit (Bio-Rad) to distinguish the wild-type allele from the 5 most common mutations (NRAS^Q61K, NRAS^Q61L, NRAS^Q61R, NRAS^{Q61H 183A>T}, and NRAS^{Q61H 183A>C}) using the QX200 Droplet Digital PCR System (Bio-Rad) according to the manufacturer’s standard protocol. The lower limit of detection was approximately 0.2% mutant allele frequency (MAF) for the multiplexed screening assay and <0.1% MAF per single well for the mutation-specific assays.

Statistical analysis

PFS duration was defined as the time from the initiation of systemic therapy to the date of disease progression or death from any cause. Overall survival (OS) duration was defined as the time from the initiation of systemic therapy to the date of death or last follow-up. The Kaplan–Meier method was used to estimate PFS and OS, and a log-rank test was used to compare PFS or OS among patient subgroups. All tests were 2-sided, and P values <0.05 were considered statistically significant. Statistical analyses were performed using the SPSS 21 (SPSS, Chicago, IL) software program.

RESULTS

Patient characteristics

Of 23 patients screened from October 2012 to April 2014, 19 met the eligibility criteria and were enrolled in the study (Supplementary Figure 1). The clinical characteristics of these 19 patients are described in Table 2. The median age was 53 years (range, 33–75 years), and most patients were men (n=11, 58%) and white (n=15, 79%). Melanoma was the most common tumor type (n=13, 68%) followed by papillary thyroid cancer (n=1, 5%), anal squamous cell carcinoma (n=1, 5%), cholangiocarcinoma (n=1, 5%), spindle cell sarcoma (n=1, 5%), pancreatic carcinoma (n=1, 5%), and adenocarcinoma of unknown primary (n=1, 5%). Patients had received a median of 3 lines of prior therapies (range, 1–7 therapies). In addition, 13 patients (68%) received ≥1 line of prior treatment with BRAF and/or MEK inhibitors (1 line, 9 patients; 2 lines, 2 patients; >2 lines, 2 patients).

Table 2.

Patients’ characteristics.

Characteristic	No. of patients (%), n=19
Median age (range), y	53 (33–75)
Gender
Male	11 (58)
Female	8 (42)
Ethnicity
White	15 (79)
Hispanic	2 (11)
Asian	1 (5)
Unknown	1 (5)
ECOG performance status
0	2 (11)
1	17 (89)
Prior therapies
Median no. of prior therapies (range)	3 (1–7)
≤3 therapies	11 (57.9)
>3 therapies	8 (42.1)
Prior BRAF/MEK inhibitor	13 (68)
1 line prior BRAF/MEK inhibitor	9 (50)
2 lines prior BRAF/MEK inhibitor	2 (11)
>2 lines prior BRAF/MEK inhibitor	2 (11)
Prior vemurafenib	10 (53)
Prior dabrafenib	1 (5)
Prior dabrafenib with trametinib	1 (5)
Prior encorafenib with binimetinib	1 (5)
Prior trametinib	2 (11)
Prior platinum-based therapy	8 (42)
Prior taxane-based therapy	3 (16)
Prior ipilimumab	7(37)
Prior pembrolizumab	1 (5)
Diagnosis
Melanoma	13 (68)
Othera	6 (32)
Type of BRAF mutation
BRAFV600E	15 (79)
BRAFV600K	1 (5)
Non-BRAFV600 b	3 (16)

Note: All data are no. of patients (%) unless otherwise indicated.

Abbreviation: ECOG, Eastern Cooperative Oncology Group.

^aPapillary thyroid carcinoma, anal squamous cell carcinoma, cholangiocarcinoma, spindle cell sarcoma, pancreatic carcinoma, adenocarcinoma of unknown primary.

^bBRAF^R671Q, BRAF^N486_P490del, BRAF^D594G

Toxicity

Of the 19 patients enrolled at 4 dose levels, 18 (95%) were evaluable for toxicity. (One patient withdrew consent before completing cycle 1). DLTs included grade 2 creatinine elevation for >7 consecutive days in 1 patient at dose level 1 and grade 3 transaminitis with grade 4 thrombocytopenia in 1 patient at dose level 3 (Table 1). The MTD was not reached by the end of the study, and dose level 3 (vemurafenib 720 mg orally twice per day, carboplatin AUC 5 intravenously every 3 weeks, paclitaxel 135 mg/m² intravenously every 3 weeks) was the last dose level proven to be safe. Only 1 patient was treated at dose level 4 before the IRB halted the study because of slow enrollment. DLTs aside, the most frequent treatment-related grade 3 or higher toxicities included grade 3 or 4 neutropenia in 5 patients, grade 3 or 4 thrombocytopenia in 5 patients, grade 3 fatigue in 4 patients, grade 3 anemia in 3 patients, grade 3 febrile neutropenia in 2 patients, and grade 3 hyponatremia in 2 patients (Table 3). Other grade 3 or 4 treatment-related toxicities included anorexia, altered mental status, bacteremia, dehydration, hand-foot syndrome, hypokalemia, hypophosphatemia, mucositis, nausea, vomiting, transaminitis, and weight loss. One patient had a fatal (grade 5) intracranial hemorrhage that was possibly related to the study drug. Seven patients (37%) required dose reductions of carboplatin and/or paclitaxel (n=5), vemurafenib (n=1), or carboplatin and vemurafenib (n=1) owing to hematological toxicity, fatigue, hand-foot syndrome, or peripheral neuropathy. In addition, 5 patients (26%) received granulocyte-colony stimulating factor to treat or prevent neutropenia, 6 (32%) required blood transfusions, and 4 (21%) received platelet transfusions. Treatment-related squamous cell carcinoma of the skin and actinic keratosis were each reported in 1 (5%) patient. No patients had grade 3 or higher or clinically significant QTc prolongation.

Table 3.

Treatment-related grade 3 or higher toxicities.

Toxicity	CTCAE grade	No.
Fatigue	3	4
Leukopenia	3	2
	4	3
Neutropenia	3	1
	4	4
Thrombocytopenia	3	2
	4	3
Anemia	3	3
Febrile neutropenia	3	2
Hyponatremia	3	2
Anorexia	3	1
Altered mental status	4	1
Bacteremia	3	1
Dehydration	3	1
Hand-foot syndrome	3	1
Hypokalemia	3	1
Hypophosphatemia	3	1
Mucositis	3	1
Nausea	3	1
Transaminitis	3	1
Weight loss	3	1
Vomiting	3	1
Intracranial hemorrhage	5	1

Abbreviations: CTCAE, Common Terminology Criteria for Adverse Events.

Efficacy

Of the 19 patients enrolled in the study, 16 (84%) were evaluable for response and 5 (26% of all patients and 31% of evaluable patients) had a complete response (CR) or partial response (PR). Three patients did not have restaging CT or MRI because they discontinued treatment owing to clinical progression, consent withdrawal, or treatment-related intracranial hemorrhage resulting in death, respectively. One patient with melanoma and a BRAF^V600E mutation, who received three lines of prior therapies for metastatic disease including 8 months of therapy with the MEK inhibitor trametinib, had a CR (Table 4). This patient was taken off the study after 26 cycles of treatment owing to a continuing CR and remained progression-free at 54.1 months (Figures 1A and 1B; Table 4). Four patients with melanoma, of whom three received prior BRAF inhibitors, attained PRs (−51%, −50%, −31%, −30%, respectively; Figures 1A, 1C and 1D; Table 4). In addition, 8 patients had stable disease (SD; −24% to 0%), which in 4 of them lasted longer than 4 months (4.8 months, 4.9 months, 8.6 months, and >25.1 months; Figure 1A).

Table 4.

BRAF mutation types, prior therapies and efficacy of vemurafenib, carboplatin and paclitaxel in 13 patients with melanoma.

ID	BRAF Mutation	Prior BRAF and/or MEK inhibitors	Prior platinum	Prior taxane	Dose Level	Response (%)	PFS in months
1	V600E	- trametinib	no	no	1	CR (−100%)	54.1a
2	V600E	- vemurafenib	cisplatin	no	1	SDb	4.5
3	V600E	- vemurafenib	no	no	1	SD (−23%)	13.7
4	V600K	- vemurafenib	no	paclitaxel	1	PR (−51%)	3.5
5	V600E	- vemurafenib	no	no	1	PR (−50%)	15.4
6	V600E	- vemurafenib	carboplatin	paclitaxel	2	SD (−12%)	4.8
7	V600E	no	no	no	2	PR (−31%)	7.0
8	V600E	- dabrafenib - vemurafenib/sorafenib	cisplatin	no	2	PD (+14%c)	1.5
9	V600E	- trametinib - vemurafenib - vemurafenib/sorafenib	no	no	3	PD (+15%c)	1.4
10	V600E	- dabrafenib/trametinib	no	no	3	not doned	0.7a
11	V600E	- vemurafenib - vemurafenib/ipilimumab - vemurafenib/crizotinib - sorafenib/vemurafenib	no	no	3	PR (−30%)	14.0
12	V600E	- encorafenib/binimetinib	no	no	3	SD (−6%)	4.9
13	V600E	- vemurafenib/sorafenib	cisplatin	No	4	SD (−15%)	2.8

Abbreviation: PFS, progression-free survival

^aPatient did not have disease progression at the time of analysis

^bPatient did not have measurable disease

^cClinical progression

^dPatient withdrew consent

Figure 1.

A. Waterfall plot depicting maximum percentage change in target lesions from baseline using RECIST 1.1. Baseline and week 6 computed tomography scans in a patient with melanoma and BRAF^V600E mutation previously treated with a MEK inhibitor, who attained a complete response, demonstrate complete resolution of the abdominal implant (B.); in patient with melanoma and BRAF^V600K mutation previously treated with a BRAF inhibitor, who attained a partial response (−51%), demonstrate improvement in bilateral pleural effusion and mediastinal lymphadenopathy (C.); in patient with melanoma and BRAF^V600E mutation, who attained a partial response (−31%), demonstrate improvement in left breast mass (D.). Of two remaining patients, who attained partial response (−30% and −50%, respectively), a patient with melanoma and BRAF^V600E mutation previously treated with a BRAF inhibitor is depicted with pertinent images and plasma cell-free DNA in Figure 3A and one patient with melanoma and BRAF^V600E mutation previously treated with a BRAF inhibitor is not depicted because of having computed tomography without intravenous contrast.

The CR/PR rates of 6 patients not previously treated with BRAF or MEK inhibitors and 13 patients previously treated with BRAF and/or MEK inhibitors did not differ significantly (1/6, 17% vs. 4/13, 31%; P=1.00); nor did those of 16 patients not previously treated with taxanes and 3 patients previously treated with taxanes (4/16, 25% vs. 1/3, 33%; P=1.00). However, the CR/PR rate of 11 patients not previously treated with platinum agents was significantly higher than that of 8 patients previously treated with platinum agents (5/11, 45% vs. 0/8, 0%; P=0.045).

The median PFS duration for all 19 patients was 3.7 months (95% confidence interval [CI], 2.5–4.9 months). The median PFS durations for 6 patients not previously treated with BRAF or MEK inhibitors and 13 patients previously treated with BRAF and/or MEK inhibitors did not differ significantly (1.9 months; 95% CI, 0.3–3.5 months vs. 3.7 months; 95% CI, 0.8–6.6 months; P=0.41; Figure 2A). In addition, the median PFS durations for 16 patients not previously treated with taxanes and 3 patients previously treated with taxanes did not differ significantly (3.7 months; 95% CI, 1.1–6.3 months vs. 3.1 months; 95% CI, 0.0–6.9 months; P=0.25; Figure 2B). In contrast, the median PFS duration for 11 patients not previously treated with platinum agents was significantly longer than that for 8 patients previously treated with platinum agents (8.6 months; 95% CI, 3.0–14.2 months vs. 2.0 months; 95% CI, 0.8–3.2 months; P=0.005; Figure 2C).

Figure 2.

A. The median PFS duration of 13 patients previously treated with BRAF and/or MEK inhibitors (3.7 months; 95% CI, 0.8–6.6 months; red) and that of 6 patients not previously treated with BRAF or MEK inhibitors (1.9 months; 95% CI, 0.3–3.5 months; blue) did not differ significantly (P=0.41). B. The median PFS duration of 3 patients previously treated with taxanes (3.1 months; 95% CI, 0.0–6.9 months; red) and that of 16 patients not previously treated with taxanes (3.7 months; 95% CI, 1.1–6.3 months; blue) did not differ significantly (P=0.25). C. The median PFS duration of 8 patients previously treated with platinum therapy (2.0 months; 95% CI, 0.8–3.2 months; red) was significantly shorter than that of 11 patients not previously treated with platinum therapy (8.6 months; 95% CI, 3.0–14.2 months; blue; P=0.005).

The median OS duration for all 19 patients was 7.8 months (95% CI, 3.8–11.8 months). The median OS durations for 6 patients not previously treated with BRAF or MEK inhibitors and 13 patients previously treated with BRAF and/or MEK inhibitors did not differ significantly (2.0 months; 95% CI, 1.1–2.8 months vs. 9.1 months; 95% CI, 3.7–14.5 months; P=0.50; Supplementary Figure 2A). In addition, there was a trend towards a longer median OS duration for 16 patients not previously treated with taxanes compared with 3 patients previously treated with taxanes (9.1 months; 95% CI, 0.1–18.1 months vs. 3.5 months; 95% CI, 0.5–6.5 months; P=0.06; Supplementary Figure 2B). The median OS duration for 11 patients not previously treated with platinum agents was significantly longer than that of 8 patients previously treated with platinum agents (13.7 months; 95% CI, 0.0–29.5 months vs. 4.5 months; 95% CI, 0.0–10.5 months; P=0.033; Supplementary Figure 2C).

Plasma cfDNA analysis

Of the 19 patients, 11 (58%; all with BRAF^V600 mutations) had blood collection before therapy for molecular analysis of plasma-derived cfDNA. Of these 11 patients, 8 were receiving a BRAF inhibitor and 1 was receiving an MEK inhibitor immediately before study enrollment. Of these 11 patients, 8 had additional serial blood collection of ≥1 sample during therapy. BRAF^V600 mutations in plasma cfDNA were detected in blood samples collected before therapy in 5 patients (45%) and in blood samples collected at any time in 7 patients (64%). The median numbers of BRAF^V600 mutant copies in plasma cfDNA collected before therapy (24.5; range, 0–1672), during therapy (16.5; range, 0–608), and at disease progression (275.5; range, 13–710) differed significantly (P=0.03; Supplementary Figure 3). Of the 8 patients who had serial blood collection, 6 had detectable BRAF^V600E-mutant cfDNA. Dynamic changes in BRAF^V600E-mutant cfDNA corresponded with the clinical disease course and often indicated upcoming disease progression (Figure 3). We also performed testing for KRAS mutations in exon 2 and NRAS mutations in exons 2 and 3, since these mutations could be implicated in adaptive resistance to BRAF inhibitors; however, none of the plasma cfDNA samples demonstrated any of these mutations²⁰.

Figure 3.

A–F. Dynamic tracking of the number of BRAF^V600E-mutant copies in cfDNA (red) isolated from serially collected plasma samples and corresponding images of selected targeted lesions in a patient with melanoma and a PR (A), a patient with melanoma and stable disease (B), a patient with melanoma and stable disease (C), a patient with papillary thyroid cancer and stable disease (D), a patient with melanoma and stable disease (E), and a patient with intrahepatic cholangiocarcinoma and unconfirmed stable disease (F).

DISCUSSION

The combination of vemurafenib, carboplatin, and paclitaxel demonstrated an acceptable safety profile and promising signals of anticancer activity in patients with advanced cancers and BRAF^V600 mutations. Most of the observed grade 3 or higher toxicities were hematological such as thrombocytopenia requiring platelet transfusion in 21% of patients and were deemed to be related to the chemotherapy (i.e., carboplatin and paclitaxel) portion of the combination. Although the study did not meet its primary endpoint to determine MTD, it demonstrated that vemurafenib, carboplatin, and paclitaxel has noteworthy clinical activity. The IRB requested to stop the enrollment because of the slower-than-anticipated enrollment. The accrual had been impacted during the second half of the study by the boom of immunotherapy clinical trials, which resulted in the opening of a variety of competing protocols. The study enrolled patients to 4 of 6 planned dose levels, and since only 1 patient was treated at dose level 4, dose level 3 (vemurafenib 720 mg, carboplatin AUC 6, paclitaxel 135 mg/m²) was established as the last safe dose level. Nevertheless, despite these limitations, our trial suggests that the combination of vemurafenib, carboplatin, and paclitaxel has noteworthy clinical activity across all dose levels, with 3 of 5 PRs occurring at dose level 1.

The PR/CR rates in the present study were 26% (5/19) for all patients and 31% (5/16) for patients evaluable for response; and responses were independent of prior BRAF and/or MEK inhibitor use; however, it needs to be noted that 3 of 6 patients in our study, who were naïve to BRAF and MEK inhibitors, had BRAF^non-V600 mutations. The enrollment of patients with BRAF^nonV600 mutations was then discouraged taking into account available preclinical data and lack of activity in this population observed in our own presented patient population²¹. In patients previously treated with BRAF and/or MEK inhibitors, the PR/CR rate was 31% (4/13), and the responses were mostly durable, lasting 3.1, 14.0, >15.4, and >54.1 months. This compares favorably to previously published response rates of 11–18% in chemotherapy-naïve patients with metastatic melanoma treated with carboplatin and paclitaxel and to some published results with targeted treatment strategies for patients with BRAF-mutated melanoma previously treated with BRAF inhibitors^{12, 13, 22}. For instance, the first-in-class ERK inhibitor ulixertinib demonstrated a PR/CR rate of 15% (3/19) in patients with advanced BRAF^V600-mutated melanoma previously treated with BRAF and/or MEK inhibitors²³. However, the combination of the BRAF inhibitor dabrafenib and the MEK inhibitor trametinib in patients with advanced BRAF^V600-mutated melanoma previously treated until progression with a BRAF inhibitor without or with a MEK inhibitor followed by a treatment break demonstrated CR/PR rates of 32%−37%^{24, 25}. This relatively high CR/PR rate can be explained by a suppression of resistance clones after discontinuation of a BRAF inhibitor. In our study, it remains unclear if encouraging activity observed in BRAF^V600-mutated melanoma previously treated with BRAF and/or MEK inhibitors reflects the ability of the studied combination to combat resistance or regained sensitivity. In our study, an exploratory analysis demonstrated no significant difference in the PR/CR rates (P=1.00) or median PFS durations (P=0.41) of patients with or without prior treatment with BRAF and/or MEK inhibitors. However, patients not previously treated with platinum-based chemotherapy had a significantly higher PR/CR rate (P=0.045) and median PFS duration (P=0.005) than patients with a history of prior platinum-based therapy did. Therefore, it is plausible that platinum-based chemotherapy can play a role in overcoming resistance induced by BRAF inhibitors, MEK inhibitors or other therapies.

We did not identify any BRAF^V600 mutations in plasma cfDNA samples collected before therapy in 6 of 11 patients, which resulted in a sensitivity of only 45% compared to the archival tumor tissue testing. This is noticeably less than the sensitivity of 73–81% reported by us and others from other studies using various digital PCRs^26–28. This can be explained by the fact that, of the 11 patients whose plasma cfDNA was collected, 9 had been receiving BRAF or MEK inhibitors before enrolling in the study, which could have led to the elimination of BRAF^V600-mutated cfDNA from plasma. In agreement with previously published literature, the number of BRAF^V600-mutant cfDNA copies differed among samples collected before therapy, during therapy, and at disease progression (P=0.03)²⁹. Dynamic changes in BRAF^V600E cfDNA in plasma overall corresponded with clinical course; however, the emergence of RAS mutations indicating therapeutic resistance was not detected.

Our study had several potential limitations. First, despite the clinical activity was noteworthy the study did not meet its primary endpoint of identifying the MTD. Second, the number of patients enrolled was relatively small, and all PRs and CRs were observed in patients with BRAF^V600-mutated melanoma; therefore, whether the drug combination has potential for further development in other tumor types remains unclear. Third, the study was designed in 2012, shortly before the results of the first phase I trial of the anti–programmed cell death protein 1 (PD1) antibody nivolumab were published³⁰. In our study, only 5 patients with melanoma received prior checkpoint inhibitors, and only 1 had received an anti-PD1 antibody; thus, how the combination tested in the present study should be positioned in current therapeutic strategies in advanced melanoma with BRAF^V600 mutations is not clear.

Despite these limitations, our study demonstrated one of the most promising signals of anticancer efficacy in advanced melanoma with BRAF^V600 mutations previously treated with BRAF and/or MEK inhibitors. Further investigation of the combination of vemurafenib, carboplatin, and paclitaxel in this setting is warranted.