Although only 1 percent of skin cancers are melanomas, the disease is responsible for the most skin cancer deaths. In the U.S., this year, approximately 10,000 people are expected to die from this disease.
In those patients having metastatic melanoma (stage IV), a fair percentage (~45-50%) harbor a mutation in BRAF, a serine/threonine kinase (an enzyme that catalyzes the phosphorylation of serine/threonine-based sidechain hydroxyl groups) that activates the mitogen-activated protein kinase/extracellular signal-regulating kinase (MAPK/ERK) pathway. This activation stimulates cancer cell proliferation and survival.
Of these BRAF mutations, more than 90 percent occur at codon 600. The most common of these mutations (~90%) occurs with the replacement of the wild type BRAF's valine by a glutamic acid residue (BRAF V600E). The next most common mutant is BRAF V600K, which is present in 5-6 percent of the patient population and has the wild type enzyme's valine replaced by a basic lysine residue.
Combining BRAF & MEK Inhibitors
To treat patients having a BRAF mutation, type 1 BRAF inhibitors such as dabrafenib and vemurafenib were developed. These compounds, which are ATP-competitive, bind to and block the BRAF kinase active site, thus denying access to the enzyme's phosphate source (ATP). Therapy with this type of inhibitor has been associated with enhanced tumor response and survival rates; however, treatment with these type 1 BRAF inhibitors also has given rise to the development of non-melanoma skin cancers (e.g., keratoacanthomas and well-differentiated cutaneous squamous-cell carcinomas) in about 15-30 percent of patients.
Instead of the expected reduction in phosphorylated-ERK (pERK) levels, these inhibitors increase the pERK in those wild-type (i.e., non-mutant) BRAF cells having upstream activation (e.g., up-regulated tyrosine kinases or the oncogenic RAS). This particular MAPK signaling by wild-type BRAF cells is termed “paradoxical MAPK-pathway activation.”
In a relevant study, Su, et al., (N Engl J Med 2012;366:207-15) presented data that suggested the RAF inhibitor used (vemurafenib) was not a tumorigenesis initiator, but rather accelerated the progression of those preexisting subclinical malignant lesions having a high potential for upstream MAPK signaling. They also suggested the use of a MEK (MAPK-kinase) inhibitor to circumvent this potentially fatal effect. This theory has since been validated; on Jan. 8, 2014, the FDA granted accelerated approval for the combination of dabrafenib and trametinib (a MEK inhibitor) for the treatment of unresectable or metastatic melanoma patients having the BRAF V600E or V600K mutations.
On Nov. 10, 2015, another BRAF inhibitor/MEK inhibitor combination (vemurafenib/cobimetinib) received accelerated FDA approval for the treatment of metastatic or unresectable melanoma having the BRAF V600E or V600K mutation, providing further support for this therapeutic strategy.
In 2013, a clinical study was initiated that evaluated the use of encorafenib, a BRAF inhibitor, with binimetinib, a MEK 1/2 inhibitor (NCT01909453). The lead investigator on this study is Keith Flaherty, MD, Director of the Termeer Center for Targeted Therapy at Massachusetts General Hospital and Professor of Medicine at Harvard Medical School in Boston.
The investigational compounds used in this study were encorafenib, a potent ATP-competitive BRAF inhibitor and binimetinib, an ATP uncompetitive MEK 1/2 inhibitor. Encorafenib showed remarkable selectivity, having little or no activity against a panel of more than 100 different kinases (IC50 >900 nM). Flaherty noted “the main advantage of this compound is its ability to remain bound to the target enzyme, having a dissociation half-life of greater than 24 hours, which provides prolonged inhibition of the enzyme.”
Binimetinib is a potent allosteric inhibitor of MEK (i.e., it does not compete with the binding of ATP at the active site of the enzyme). This compound has a shorter half-life than other MEK 1/2 inhibitors, and as Flaherty explained, “it is hoped that this reduced half-life would reduce toxicity in those patients having sensitivity for this therapy.”
He added that “with a compound having a short half-life, one would expect to see a more saw-toothed concentration curve over time, and currently there is some debate as to whether that profile having higher concentrations with short duration is better; additional testing clearly needs to be done.”
Trial Criteria & Endpoints
In part 1 of the Columbus trial, the inclusion criteria were as follows: patients had metastatic or locally-advanced melanoma with a BRAF V600 mutation (V600E and/or V600K); they were either untreated or had progression on/after first-line immunotherapy; and had an Eastern Cooperative Oncology Group (ECOG) performance status of 0-1. The participants were stratified according to American Joint Committee in Cancer (AJCC) stage, ECOG status, and BRAF mutation/prior first-line immunotherapy.
In the first part of the study, 577 patients were randomized to three separate study arms: COMBO450 (encorafenib 450 mg QD + binimetinib 45 mg BID, n=192); ENCO300 (encorafenib 300 mg QD, n=194); and VEM (vemurafenib 960 mg BID, n=191). The primary endpoint for the first part of the study was the comparison of the progression-free survival (PFS) (blinded radiological determination) for the COMBO450 vs. VEM arms. A secondary endpoint was the comparison of PFS for the COMBO450 and ENCO300 arms of the study. Additionally, patient-reported outcomes were obtained via the functional assessment of cancer therapy-melanoma (FACT-M) and the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ-C30).
Part 2 of the study, which is currently ongoing, has as its primary objective to assess the contribution of binimetinib in combination therapy by using lower dosages of encorafenib (300 mg QD encorafenib + binimetinib) as compared to the single agent therapy (encorafenib 300 mg QD).
For the COMBO450 group (n=192), the median age was 57 (range: 20-89), 60 percent were male, 71 percent had an ECOG status of 0, 29 percent had lactate dehydrogenase (LDH) levels greater than or equal to the upper limit of normal (ULN), 64 percent had stage IVM1c and 18 percent had stage IVM1b diseases, and 45 percent had three or more organs involved at study outset.
In the ENCO300 arm (n=194), the median age was 54 (range: 23-88), 56 percent were male, 72 percent had an ECOG status of 0, 24 percent had LDH levels ≥ ULN, 62 percent had stage IVM1c, and 20 percent had stageIVM1b diseases, while 44 percent had three or more organs involved at study outset.
For the VEM cohort (n=191), the median age was 56 (range: 21-82), 58 percent were male, 73 percent had an ECOG status of 0, 27 percent had LDH levels ≥ULN, 65 percent had stage IVM1c, and 16 percent had stage IVM1b diseases, and 46 percent had three or more organs involved at study outset. Across all arms, the BRAF mutation status ranged from 88 percent to 89 percent for the V600E mutant and from 10 percent to 12 percent for the V600K mutant (two patients in the ENCO300 arm had indeterminate BRAF mutation status).
PFS evaluations were performed both locally and centrally. For the centralized review comparing the COMBO450 and VEM regimens, the median PFS for COMBO450 was 14.9 months (95% CI range: 11.0-18.5 months), while for the VEM arm, the median PFS was 7.3 months (95% CI range: 5.6-8.2 months).
For this centralized comparison, the hazard ratio (HR) was 0.54 (95% CI range: 0.41-0.71, P < 0.001). For the centralized review comparing the COMBO450 and ENCO300 regimens, the median PFS for COMBO450 was 14.9 months (95% CI range: 11.0-18.5 months), while for the ENCO300 arm, the median PFS was 9.6 months (95% CI range: 5.6-8.2 months). For this centralized comparison, the HR was 0.75 (95% CI range: 0.56-1.00, P = 0.051). These last results did not meet preset statistical significance criteria.
When evaluating the HR values different patient subgroups, for the COMBO450 vs. VEM comparison, those having prior ipilimumab therapy had the lowest HR (0.40; 95% CI range: 0.1-1.64), followed closely by those having LDH values < ULN (0.47; 95% CI range: 0.33-0.67) and AJCC stage IVM1c disease (0.48 95% CI range: 0.34-0.69). In the comparison of the COMBO450 versus ENCO300 arms, the lowest HR was obtained for those patients having AJCC stage IVM1c disease (0.68; 95% CI range: 0.47-0.98) and those having LDH values < ULN (0.72; 95% CI range: 0.50-1.04).
For those patients in the COMBO450 group, 8 percent had complete response (CR) and 55 percent had a partial response (PR) to give an overall response rate (ORR) of 63 percent (95% CI Range: 56-70%; ORR = CR + PR). Additionally, 29 percent displayed stable disease (SD, including patients having non-target lesions and best response of non-PD/non-CR) to give a disease control rate (DCR) of 92 percent (95% CI range: 87-96%; DCR = ORR + SD), while 8 percent of these patients had progressive disease (PD, including best responding patients with no or unknown assessment).
The median duration of response (DOR) was 16.6 months (95% CI range: 12.7-20.4 months). Of the patients in the VEM group, 6 percent had CR while 35 percent had PR to give an ORR of 40 percent (95% CI range: 33-48%). In this arm of the study, 41 percent of the patients had SD to give a DCR of 82 percent (95% CI range: 75-87%). A larger portion of the patients displayed PD (18%) and the median DOR was 12.5 months (95% CI range: 6.9 months-16.9 months), which is shorter when compared to the COMBO450 group.
In the ENCO300 cohort (central review), 5 percent of the patients had CR and 45 percent had PR for an ORR of 51 percent (95% CI range: 43-58%). In this segment, 34 percent of the participants had SD to give a DCR of 84 percent (95% CI range: 78-89%), 16 percent had progressive disease, and the median DOR was 14.9 months (95% CI range: 11.0 months—not evaluable).
When asked if there were any differences observed in the results for patients having different BRAF mutations (i.e., V600E vs. V600K), Flaherty said “as had been noted in other studies, those having the V600K mutation tended to have lower response rates and shorter PFS values compared to those having the V600E mutation.” Regarding the COMBO450 regimen, he noted that “the response rates were greatest for this group whether being reviewed locally or centrally.”
To evaluate the overall safety of the respective regimens, adverse events (AE) were recorded. For grade 3/4 AEs, the following values were obtained for the different arms: COMBO450- 58; VEM- 63; ENCO300- 66.
For the COMBO450 group, the most common grade 3/4 AEs were elevated gamma-glutamyl tranferase (9) and increased blood creatine phosphokinase levels (7), while for ENCO300, the most common grade 3/4 AEs were myalgia (10), and arthralgia (9).
For the VEM group, arthralgia (6) was the most commonly encountered grade 3/4 AE. The number of deaths observed (those who died during or within 30 days after study treatment cessation) were fairly similar across all groups (COMBO450-9; VEM-10; ENCO300-7).
The AEs resulting in discontinued study participation were also fairly similar for all study cohorts (COMBO450- 13; VEM- 17; ENCO300- 14); however, the AEs leading to dose interruption did differ somewhat (COMBO450- 46; VEM- 53; ENCO300- 64). Additionally, AEs requiring dosage reductions also differed among the study arms (COMBO450- 11; VEM- 23; ENCO300- 27). These fewer AE-related dose interruptions/reductions had the overall effect of prolonging the duration of exposure and enhancing the dose intensity for those in the COMBO450 portion of the study.
For the COMBO450 cohort, the median duration of exposure (DOE) for encorafenib was 51.2 weeks (range: 0.4-116 weeks) and 50.6 weeks (range: 0.4-116 weeks) for binimetinib; while for VEM, the median DOE was 27.1 weeks (range: 0.9-121.6 weeks). In the ENCO300 cohort, the median DOE was 31.4 weeks (range: 0.1-113.3 weeks). These longer DOEs, in turn, resulted in higher dose intensities for the COMBO450 arm. For this arm, the median dose intensity for encorafenib was 100.0 percent (range: 33.3%-100.1%) and 99.6 percent for binimetinib (6.9%-100.0%).
For the VEM participants, the median dose intensity was 94.5 percent (range: 16.9%-100.0%); while for the ENCO300 cohort, the dose intensity was 86.2 percent (range: 14.8%-100.2%). As Flaherty explained, “the higher dose intensities and durations of exposure in the COMBO450 arm of the study in all likelihood relate to the improved tolerability of this combination as well as better disease control compared to BRAF inhibitor monotherapy. All of these features contributed to the higher quality of life scores, as measured by the FACT-M and the EORTC QLQ-C30 scales that were observed for patients in this portion of the study.”
When using the FACT-M scale for determining the quality of life, a HR of 0.46 was obtained for COMBO450 versus VEM (95% CI range: 0.29-0.72), while a figure of 0.48 was calculated for the COMBO450 versus ENCO300 comparison (95% CI range: 0.31-0.75). When using the EORTC QLQ-C30 scale, a HR of 0.55 was obtained for comparison of COMBO450 versus VEM (95% CI range: 0.37-0.80), while COMBO450 versus ENCO300 yielded a value of 0.45 (95% CI range: 0.31-0.65).
“The reduced toxicity observed for the COMBO450 group is most likely due to the effect of the MEK 1/2 inhibitor binimetinib blocking the paradoxical MAPK activation, which can occur with BRAF inhibitor therapy,” explained Flaherty. “A better-tolerated regimen allows patients to be on the medications for longer periods, affording the patients longer treatment periods, thus improving their quality of life.”
Richard Simoneaux is a contributing writer.