Introduction

Mutations in the v-Raf murine sarcoma viral oncogene homolog B (BRAF) gene have been found in ~4–8% of all cancers, with the greatest number found in colorectal cancer (CRC), melanoma, and non-small cell lung cancer (NSCLC)1,2,3. The most common BRAF mutation is a point mutation (T1799A) resulting in an amino acid substitution at codon 600 (V600E), which confers constitutive BRAF kinase activity2,4,5. BRAFV600E accounts for ~1–2% of NSCLCs, making it an actionable therapeutic target given the success of other therapies that target actionable mutations with similar frequencies in NSCLC (e.g., ALK, EGFR)6,7,8,9. Targeted therapeutic approaches with BRAF inhibitor monotherapy (vemurafenib and dabrafenib) demonstrated efficacy in phase 2 trials with generally acceptable toxicity in patients with BRAFV600-mutant advanced NSCLC10,11.

While BRAFV600E inhibitor monotherapy is initially effective, acquired resistance and paradoxical activation are associated with reactivation of the mitogen-activated protein kinase (MAPK) pathway and subsequent disease progression12,13,14. To delay onset of resistance, BRAF inhibitors were combined with a downstream inhibitor of the MAPK pathway, MAPK kinase (MEK)14. Trametinib, a MEK inhibitor, in combination with dabrafenib showed durable anti-tumor activity and acceptable safety in patients with BRAFV600E-mutant metastatic NSCLC in phase 2 trials15,16,17. The combination of BRAF inhibitor encorafenib plus MEK inhibitor binimetinib is being investigated in ongoing phase 2 trials in patients with BRAFV600E-mutant metastatic NSCLC5,18,19. Initial results of the PHAROS trial revealed the combination had substantial and durable anti-tumor activity and a manageable safety profile18. Based on the results from this study, in October 2023, the US Food and Drug Administration (FDA) approved encorafenib plus binimetinib for patients with BRAFV600E-mutant metastatic NSCLC20. Current guidelines recommend dabrafenib plus trametinib or encorafenib plus binimetinib as preferred first-line treatment options or as subsequent treatment for BRAFV600E-mutant metastatic NSCLC21. BRAF monotherapy could be considered in certain circumstances, such as lack of tolerability.

While there has been notable progress in effective treatments for BRAF-mutant NSCLC17,18, several uncertainties remain. Current guidelines for BRAFV600-mutant NSCLC recommend BRAF-targeted therapy in the first-line setting, but the optimal course for patients who do not tolerate or progress while on first-line BRAF and MEK inhibitor combinations remains ambiguous21,22. Second-line recommendations include immunotherapy, chemotherapy, or a combination; however, immunotherapy data are limited and conflicting for patients with BRAFV600-mutant NSCLC22,23,24. Current targeted therapeutic approaches have limited efficacy in patients with non-V600 BRAF mutations, and most clinical trials have focused primarily on the V600E mutation since its discovery11,17,18,25. There remains a need to better understand the incidence, impact, and management of brain metastases; mechanisms of resistance; optimal sequencing; and other patient- (e.g., smoking history) or disease-related factors (e.g., PD-L1 expression) that influence treatment outcomes of BRAF-mutant metastatic NSCLC. This manuscript provides a review of BRAF-mutant metastatic NSCLC and the therapeutic landscape with particular emphasis on targeted therapies for the V600E mutation.

BRAF-mutant metastatic NSCLC disease overview

Clinical characteristics

While BRAF mutations are predominantly found in adenocarcinomas (>85%), there is no clear association of BRAF mutation status with other patient characteristics, such as age, ethnicity, and sex6,8,9,26. Epidemiological patterns are difficult to identify since BRAF mutations occur in a small number of patients with advanced NSCLC5. One study reported that BRAFV600E mutations were significantly more common in females (P < 0.001), but that finding was not consistent with other studies6,10,26. While data are limited, studies have also shown that the incidence of BRAF mutations is lower in Asian patients than white patients6,26,27. Smoking history can be associated with BRAF mutations, as well as KRAS, MET, and other mutations9,28,29. However, a few studies have also suggested that BRAFV600E is less associated with smoking history than other BRAF mutations26,30,31. Therefore, all patients with advanced NSCLC regardless of smoking history should undergo broad-based mutation testing, including BRAF5,21,22. Programmed death ligand 1 (PD-L1) expression and tumor mutation burden (TMB) are potentially important indicators of immunotherapy response and are frequently elevated in BRAF-mutant NSCLC24. The prognostic implication of BRAF mutation status is inconclusive due to small patient numbers and conflicting studies8,26,28. However, a few studies reported worse outcomes with chemotherapy in patients with V600E mutations versus wild type (WT) BRAF and conflicting results for V600E versus non-V600E mutations6,31,32,33. In addition to the small patient number, discrepancies in patient characteristics and prognosis could be attributed to the heterogeneity of BRAF mutations in NSCLC33,34.

Classification of BRAF mutations

BRAF mutations are heterogenous with distinct mutation classes that each yield a functionally diverse BRAF protein, resulting in different clinical features and treatment ramifications33,35. BRAF-mutant proteins interact with and activate the MAPK pathway in various ways and to different degrees (Fig. 1)25,35,36. In the MAPK pathway, the signaling cascade begins with growth factor binding to receptor tyrosine kinase (RTK) which facilitates RAS binding GTP14,25,36. Activated RAS promotes a cascade of activation starting with RAF family kinases (ARAF, BRAF, CRAF), which form activated RAF homo- or heterodimers. RAF kinases activate MEK, which subsequently activates extracellular signal-regulated kinase (ERK). ERK has multiple downstream targets, including transcription factors that promote cell growth, proliferation, and survival. Negative feedback loops from ERK also regulate the pathway25.

Fig. 1: BRAF mutation classes and mechanism of actions for BRAF/MEK inhibitors.
figure 1

Class I and II mutations are RAS-independent, constitutively active monomers (class I) or dimers (class II). Class III mutations are RAS-dependent dimers with compromised kinase activity. Current BRAF inhibitors are effective for class I-mutant monomers. Next-generation RAF inhibitors can inhibit dimers and may inhibit class II and III mutations. P, phosphorylation.

Full size image

There are over 200 identified BRAF mutations categorized into three classes based on dimerization status, RAS-dependence, and kinase activity level5,37. Class I BRAF mutations occur on codon 600 (BRAFV600), substituting the valine to a glutamic acid, lysine, aspartic acid, arginine, or methionine (V600E, V600K, V600D, V600R, and V600M mutations, respectively) and can biochemically transform BRAF into a RAS-independent constitutively active monomer25,36. BRAFV600E is the most prevalent class I mutation and accounts for ~30–50% of all BRAF mutations in NSCLC6,7,33. Class I mutant proteins have a high level of kinase activity, and BRAFV600E is ~500-fold times more active than WT BRAF, leading to increased cellular proliferation4,36. Class I mutations all occur at codon 600 aberrantly activating monomers, but class II and III mutations occur at various other non-600 codons and form dimers5,25,36. Class II BRAF mutations span from G464 to K601, which includes the activation segment and P-loop domains of BRAF. These mutations, along with fusions and in-frame deletion mutations, are shown or predicted to be able to biochemically transform BRAF into a RAS-independent constitutively active dimer with ranges of intermediate to high kinase activity. Class III mutations, which tend to have impaired kinase activity, occur in the P-loop, catalytic loop, or DFG motif to form RAS-dependent heterodimers with CRAFWT. These mutations require upstream activation of RAS to amplify downstream signaling, similar to its normal physiological role and function.

Although constitutive activity of class I and II mutations can suppress RAS through ERK-negative feedback loop, class III mutations only mildly activate ERK, resulting in minimal negative feedback of RAS25,38. To overcome that feedback inhibition, class III BRAF mutations are often observed with concurrent RAS activating mutations25,33,38. While these coexisting mutations often occur in melanoma cancers, CRC and NSCLC cancers have fewer cases25,38: one study in NSCLC tumors reported coexisting RAS mutations in 22% of class III mutations (n = 54), including 42% of kinase-dead mutations (n = 19)33. This may be explained by basal RTK activation and subsequent RAS activity that is sufficient to support these class III mutants25.

Currently approved BRAF inhibitors effectively inhibit only class I mutant proteins and show substantially less efficacy against BRAF-mutant dimers (Fig. 1)25. BRAF monomer inhibitors binding to dimers can cause paradoxical transactivation of the unbound RAFWT protomer, enabling MEK/ERK signaling and subsequent disease progression12,13,25. Upstream inhibition may be an effective therapeutic strategy for treatment of class III mutations25,33,38. A better understanding of BRAF mutations, especially class II and III mutations, may enable the rational design of new targeted therapies and the development of next-generation drug combination strategies to treat BRAF-mutant cancers, including NSCLC25,33.

BRAF testing

Guidelines recommend that all patients with advanced non-squamous NSCLC undergo broad-based molecular testing to identify molecular drivers—including but not limited to BRAFV600 mutations21,22. Recommended and approved molecular testing assays include polymerase chain reaction (PCR) and next-generation sequencing (NGS)39,40,41. PCR offers rapid turn-around, reproducibility, specificity, sensitivity, and lower costs, but it is a single-gene assay typically limited to detection of V600E mutation39. Panel-based NGS has gained popularity for the ability to simultaneously test multiple genes, including BRAF, for both V600E and non-V600E mutations, which is more cost-effective than sequential single-gene assays and uses relatively little tumor tissue39,40. Availability of sufficient tumor tissue is a major constraint when testing for the numerous actionable mutations in NSCLC, so approaches that conserve tissue while providing a full molecular profile are being evaluated42. Liquid biopsy, a noninvasive and more rapid alternative to tissue biopsies that collects blood-based biomarkers—including circulating tumor DNA (ctDNA), can be used to detect genomic alterations5,43. Immunohistochemistry (IHC) is a highly sensitive and specific diagnostic test that uses the monoclonal antibody VE1 to detect BRAFV600E-mutant proteins39,44. While there are no current standard recommendations for IHC in BRAFV600E mutation detection, guideline recommendations for other oncogenes suggest that IHC be confirmed with a molecular test22,39. BRAF testing rates lag behind other driver oncogene testing rates, likely due to limited tissue availability and the fact that other actionable biomarkers (e.g., EGFR, ALK, PD-L1) are prioritized for testing when a sequential selective approach is used45. Given the demonstrated efficacy of BRAFV600E inhibitors in NSCLC, there is a need to improve the rate of BRAF testing so the results can be applied to clinical decision-making.

Treatment landscape

BRAF and MEK inhibitors in solid tumors

The identification of BRAF mutations, especially V600E, and their role in cancer led to the development of highly-selective BRAF inhibitors such as vemurafenib, dabrafenib, and encorafenib (Figs. 1 and 2) 46,47,48. These small-molecule inhibitors preferentially bind to the active conformation of BRAF kinase; through competitive occupation of the ATP binding pocket, the drugs stabilize the active conformation, resulting in potent inhibition of BRAFV600 47,48,49. Both vemurafenib and dabrafenib have confirmed activity against V600E, V600K, V600R, and V600D BRAF mutations 50,51,52. Encorafenib has confirmed activity against BRAFV600E and BRAFV600K mutants and BRAFWT 48,49.

Fig. 2: Timeline of key advancements in the treatment of BRAFV600E-mutant NSCLC.
figure 2

Positive data readouts for BRAF-mutant NSCLC (blue) and approved treatments for BRAFV600E-mutant NSCLC (green) are shown. aBRAF inhibitors. bMEK inhibitors. cIn 2022, dabrafenib plus trametinib was approved for patients with unresectable metastatic BRAFV600E-mutant solid tumors who progressed on previous treatments and have no acceptable alternative option. dImmunotherapy that targets PD-1. eImmunotherapy that targets PD-L1.

Full size image

BRAF inhibitor monotherapy has had varying degrees of efficacy in solid tumors53,54,55. In patients with BRAFV600E-mutant metastatic melanoma, single-agent vemurafenib was associated with a relative reduction of 63% in the risk of death and 74% in the risk of tumor progression compared with dacarbazine; median progression-free survival (PFS) was 5.3 months with vemurafenib (n = 275) and 1.6 months with dacarbazine (n = 274)53. With single-agent dabrafenib, patients with BRAFV600E-mutant metastatic melanoma (n = 76) had an objective response rate (ORR) of 59%, median PFS of 6.3 months, and median overall survival (OS) of 13.1 months56. In the dose expansion phase of a single-agent encorafenib study in patients with BRAFV600-mutant melanoma, ORR was 60.0%, median PFS was 12.4 months (95% CI, 7.4–not estimable [NE]), and median OS was NE for BRAF inhibitor-naive patients (n = 15); corresponding data for BRAF inhibitor-pretreated patients (n = 18) were 22.2%, 1.9 months (95% CI, 0.9–3.7), and 9.07 months (95% CI, 3.68–10.84)57.

A basket study evaluated vemurafenib monotherapy in patients with nonmelanoma BRAFV600-mutant cancers55. NSCLC and CRC cohorts had ORRs of 42% (95% CI, 20–67%) and 0%, respectively (Table 1). The poor clinical activity of BRAF inhibitor monotherapy in patients with CRC was consistent in additional vemurafenib (5% partial response [PR]; n = 21), dabrafenib (11% PR; n = 9), and encorafenib studies (0% PR; n = 18)54,58,59. In patients with BRAFV600-mutant CRC, addition of cetuximab, an anti-EGFR-antibody, improved ORR with vemurafenib (4%; n = 27) and encorafenib (19.5%; n = 220)55,60.

Table 1 Selected trials of BRAF and MEK inhibitors
Full size table

While BRAF inhibitor monotherapy is initially effective, acquired resistance enables reactivation of the MAPK pathway, limiting the clinical utility of monotherapy14,61. In addition, BRAF monomer inhibitors can cause paradoxical activation of the MAPK pathway in BRAFWT cells, which has been associated with hyperproliferative cutaneous events, including squamous cell carcinoma (SCC) and keratoacanthoma12,13,14,57. BRAF monomer inhibitors were combined with downstream MEK inhibitors to overcome resistance and paradoxical activation of the MAPK pathway, which increased efficacy and tolerability, resulting in several combination therapies being approved for unresectable metastatic BRAFV600-mutant melanoma (e.g., dabrafenib plus trametinib, vemurafenib plus cobimetinib, encorafenib plus binimetinib)14,62,63,64. The FDA granted accelerated approval of dabrafenib plus trametinib treatment for previously treated unresectable or metastatic solid tumors with the BRAFV600E protein in patients with no alternative treatment options65. This approval was supported by the ROAR study, which enrolled 206 patients into eight cohorts, each a different BRAFV600E-mutant rare cancer. ORR ranged from 0% for gastrointestinal stromal tumor (n = 1) to 89% for hairy cell leukemia (n = 55); ORR was ≥33% for the remaining cohorts.

The safety profile of BRAF inhibitor monotherapy was similar across solid tumors; common adverse events (AEs) included arthralgia, fatigue, rash, cutaneous events (e.g., SCC, keratoacanthoma), and gastrointestinal issues (e.g., diarrhea, nausea)10,48,53,54,57,62,64,66. Drug-specific AEs include photosensitivity with vemurafenib, pyrexia with dabrafenib, and reduced incidences of SCC and keratoacanthoma with encorafenib. Adding a MEK inhibitor resulted in a few key differences in the safety profiles62,64: pyrexia was more frequent with dabrafenib plus trametinib versus dabrafenib monotherapy15,62 and hyperproliferative cutaneous events were less common with dabrafenib plus trametinib and encorafenib plus binimetinib15,62,64.

BRAF and MEK inhibitors in metastatic NSCLC

Clinical trials with BRAF inhibitor monotherapy in patients with BRAFV600-mutant NSCLC are summarized in Table 110,11,55,66. In several trials, vemurafenib monotherapy was an effective treatment for treatment-naive (ORR: 37.5%; median PFS: 12.9 months) and previously treated patients (ORR: 37.0–44.8%; median PFS: 5.2–6.1 months) with BRAFV600-mutant NSCLC11,66. In one study, serious AEs occurred in 63% of patients, most commonly cutaneous SCC (15%) and keratoacanthoma (15%)66. Dabrafenib monotherapy was effective for previously treated patients (ORR: 33%; median PFS: 5.5 months) with BRAFV600E-mutant metastatic NSCLC; however, data were limited for treatment-naive patients because of a decision to prioritize the combination with trametinib with the expectation of improved response rates10. Pyrexia was the most common any-grade AE (36%), including grade 3 occurrences in 2% of patients. Pyrexia led to dose interruption or reduction in 11% and 4% of patients, respectively. Serious AEs occurred in 42% of patients, including pyrexia in 6% of patients.

The combination of BRAF and MEK inhibitors demonstrated synergistic anti-tumor activity with a manageable safety profile (Table 1)14,15,16,18. The multicenter, non-randomized, open-label, phase 2 trial evaluated dabrafenib plus trametinib in patients with BRAFV600E-mutant metastatic NSCLC15,16,17. At the initial data analysis with a median follow-up of 15.9 months, for treatment-naive patients (n = 36), ORR by investigator was 64% (95% CI, 46–79%), median duration of response (DOR) was 10.4 months (95% CI, 8.3–17.9), median PFS was 10.9 months (95% CI, 7.0–16.6), and median OS was 24.6 months (95% CI, 12.3–NR)16. The most common AEs included pyrexia (64%), nausea (56%), and diarrhea (36%); grade 3–4 AEs occurred in 69% of patients, including pyrexia (11%) (Fig. 3a). For previously treated patients (n = 57), with a median follow-up of 11.6 months, ORR by investigator was 63.2% (95% CI, 49.3–75.6%), median DOR was 9.0 months (95% CI, 6.9–18.3), median PFS was 9.7 months (95% CI, 6.9–19.6), and median OS was immature15. The most common AEs included pyrexia (46%), nausea (40%), and vomiting (35%); grade 3–4 AEs occurred in 49% of patients, including pyrexia (2%) (Fig. 3a). At the 5-year follow-up analysis, median PFS and OS were 10.8 months (95% CI, 7.0–14.5) and 17.3 months (95% CI, 12.3–40.2) for treatment-naive patients and 10.2 months (95% CI, 6.9–16.7) and 18.2 months (95% CI, 14.3–28.6) for previously treated patients, respectively17. The most common AE remained pyrexia (56%), and grade 3–4 AEs occurred in 66% of patients with most manageable with dose modifications. Pyrexia led to dose reduction in 11 patients (12%) and permanent treatment discontinuation in two patients (2%). The addition of the MEK inhibitor was associated with lower incidence of cutaneous SCC compared with BRAF inhibitor monotherapy (4% versus 12%)10,15. Based on these data, the FDA and European Medicines Agency approved dabrafenib plus trametinib combination for treatment of patients with BRAFV600E-mutant metastatic NSCLC67,68. A later retrospective analysis supported the use of dabrafenib plus trametinib, where the risk of death for treatment-naive patients with BRAF-mutant advanced NSCLC was significantly lower with dabrafenib plus trametinib versus platinum-based chemotherapy (HR = 0.51; 95% CI, 0.29–0.92; P = 0.03), and median OS was 17.3 months (95% CI, 14.6–NR) versus 9.7 months (95% CI, 6.4–19.6), respectively68.

Fig. 3: Adverse events experienced by ≥10% of patients with BRAF/MEK combination therapies.
figure 3

a Adverse events (AEs) shown for dabrafenib plus trametinib occurred in at least 10% of patients in combined data from interim analysis of treatment-naive and previously treated patients15,16. AEs led to dose reduction, dose interruption, and permanent treatment discontinuation in 37%, 67%, and 16% of patients. aChest pain includes musculoskeletal chest pain. b Treatment-related AEs (TRAEs) shown for encorafenib plus binimetinib occurred in at least 10% of patients18. TRAEs led to dose reduction, dose interruption, and permanent treatment discontinuation in 24%, 44%, and 15% of patients. Comparisons of safety profiles should be done cautiously since data are from different trials and reported as all-causality AEs for one treatment combination and as TRAEs for the other combination. ALP alkaline phosphatase, ALT alanine aminotransferase, AST aspartate aminotransferase, CNS central nervous system, CPK creatine phosphokinase.

Full size image

The combination of encorafenib plus binimetinib is being investigated in ongoing phase 2 trials in patients with BRAFV600E-mutant NSCLC5,18,19. PHAROS, a single-arm, open-label, multicenter trial (NCT03915951), enrolled 98 patients with BRAFV600E-mutant metastatic NSCLC (n = 59 treatment-naive, n = 39 previously treated)18. In treatment-naive patients, with a median follow-up for PFS by independent radiology review (IRR) of 18.2 months (95% CI, 16.4–22.3), ORR assessed by IRR was 75% (95% CI, 62–85%), median DOR by IRR was NE (95% CI, 23.1–NE), and median PFS by IRR was NE (95% CI, 15.7–NE). In previously treated patients, with a median follow-up for PFS by IRR of 12.8 months (95% CI, 9.0–19.8), ORR by IRR was 46% (95% CI, 30–63%), median DOR by IRR was 16.7 months (95% CI, 7.4–NE), and median PFS by IRR was 9.3 months (95% CI, 6.2–NE). OS was NE in both groups. The most frequently reported treatment-related AEs (TRAEs) were nausea (50%), diarrhea (43%), fatigue (32%), and vomiting (29%); serious TRAEs occurred in 14% of patients with the most common being colitis (3%) (Fig. 3b). All-causality pyrexia occurred in 22% of patients, and treatment-related pyrexia led to one dose interruption and no dose reductions or permanent discontinuations. The ENCO-BRAF trial (NCT04526782) includes treatment-naive and previously treated cohorts, with an estimated enrollment of 119 patients to conclude in 202619. Encorafenib plus binimetinib combination treatment was recently approved by the FDA for patients with BRAFV600E-mutant metastatic NSCLC based on the PHAROS trial20.

Primary and acquired drug resistance to BRAF-targeted therapy

Baseline concurrent mutations prior to receipt of targeted therapy have been identified in 22–30% of patients with BRAFV600E-mutant NSCLC and may be a cause of primary resistance17,61,69. Common concurrent mutations included alterations in the TP53 and SETD2 genes and the PI3K (e.g., PIK3CA E545K, PTEN R14K) and MAPK (e.g., KRAS G12C) pathways17,18,61,69. The incidence of concurrent TP53 or RAS gene family mutations was higher with BRAF class II or III mutations than class I mutations31,33,70. In several studies, the presence of a concurrent mutation in TP53, PIK3CA, KRAS, or PTEN was associated with poorer outcomes17,61,69. In a study evaluating dabrafenib plus trametinib for BRAFV600E-mutant metastatic NSCLC, patients with a concurrent mutation in the PI3K pathway (n = 4) had a median OS of 5.4 months compared with 22.7 months in patients without an identified concurrent mutation (n = 34)17. In the PHAROS trial, which evaluated encorafenib plus binimetinib for BRAFV600E-mutant metastatic NSCLC, concurrent mutations were identified, with the most common being SETD2 and TP53 (43%, each); however, there was no association between concurrent mutations and outcome18. As most of this data comes from studies with small numbers of patients, further research is necessary to understand the impact of concurrent mutations in patients with BRAFV600E-mutant NSCLC17,18,69.

The mechanisms of acquired resistance to BRAF inhibitors, alone or combined with MEK inhibitors, are poorly understood, and there is no standardized sequential treatment strategy21,61. While data are limited, acquired resistance appears to occur through bypassing or reactivating the MAPK pathway14,61. Bypassing the MAPK pathway and activating a parallel pathway (e.g., PI3K/AKT) can lead to disease progression. A PTEN frameshift mutation that could potentially activate the PI3K pathway was identified in a patient with BRAFV600E-mutant NSCLC that progressed on dabrafenib71. In preclinical BRAFV600E-mutant lung cancer models, the presence of a cooperating mutation silencing RBMS3, a regulator of the WNT pathway, promoted resistance to dabrafenib plus trametinib72.

Reactivation of the MAPK pathway can occur in BRAF-dependent or -independent manners. Resistance to targeted therapies often occurs due to secondary mutations or epigenetic changes in the target gene, and an aberrant splice variant of BRAF was identified in BRAFV600E NSCLC cells that acquired resistance73,74. However, secondary BRAF mutations may be rare, as none were discovered in several resistance studies with BRAF-mutant NSCLC and melanoma61,73,75. BRAF-independent reactivation of the MAPK pathway includes alterations either upstream or downstream of BRAF14. Mutations in RAS (NRAS, KRAS) were discovered in a few studies61,71. Strong evidence came from a study that compared ctDNA sequencing at diagnosis and disease progression for 35 patients with BRAF-mutant NSCLC who progressed on either BRAF inhibitor monotherapy or dabrafenib plus trametinib61. RAS mutations (KRAS G12V; KRAS Q61R; NRAS Q61R) were present at disease progression and not diagnosis, which suggests mutation occurred during treatment. Resistance studies of patients with BRAFV600E-mutant melanoma suggested that upregulation of RAS or overexpression of ARAF and CRAF could alleviate BRAF-dependence in tumor cells73,75. Downstream mutations in MEK1 were also identified in patients with BRAFV600E-mutant NSCLC that progressed on dabrafenib plus trametinib71. Further understanding of acquired resistance mechanisms is critical to inform optimal sequencing and providing insight for evaluation of combination approaches or next-generation target therapies.

Immunotherapy

Compared with studies with BRAF and MEK inhibitors, data concerning efficacy and safety of immunotherapy in patients with BRAFV600E-mutant NSCLC are limited; studies have not prospectively enrolled patients with BRAFV600 mutations, and immunotherapy is not specifically approved for patients with this mutation76,77,78,79. Evidence for efficacy of immunotherapy is derived from conflicting, small, retrospective studies (Table 2)23,24,77. In a multi-cohort retrospective study of immunotherapy in patients with oncogene-driven advanced lung cancer, patients (n = 10) with BRAF-mutant NSCLC in one cohort had a significantly longer median PFS (7.4 months) than patients with other common oncogene drivers (versus KRAS 2.8 months; HR 0.36, 95% CI, 0.14–0.88; P = 0.026)24. In the other cohort, PFS was longer in patients with V600E mutations (n = 30; 9.8 months) versus non-V600E mutations (n = 35; 5.4 months). However, in another retrospective study, similar PFS (2.1–3.4 months) was reported across oncogenes with immunotherapy77. In that study, two trends emerged in the BRAF cohort (n = 43). PFS was longer in patients who previously or currently smoked (4.1 months) versus had never smoked (1.9 months) and with non-V600E mutations (4.1 months) versus V600E mutations (1.8 months). According to guidelines of the European Society for Medical Oncology, patients with BRAFV600-mutant metastatic NSCLC that progresses on BRAF plus MEK inhibitor should receive immunotherapy with optional chemotherapy (in patients with smoking history) or chemotherapy with optional immunotherapy (in patients without smoking history)22.

Table 2 Efficacy data for chemotherapy and immunotherapy trials that included patients with BRAF-mutant NSCLC
Full size table

Chemotherapy

Prior to the development of targeted therapy for BRAFV600E-mutant metastatic NSCLC, platinum-based chemotherapy was the standard of care68,78. However, retrospective studies reported that patients with BRAFV600E-mutant NSCLC had poorer outcomes with platinum-based chemotherapy than those with BRAFWT NSCLC (Table 2)6,32. While those retrospective studies also reported a shorter PFS in patients with V600E mutations (4.1–5.2 months) versus non-V600E mutations (6.4–8.9 months), another retrospective study reported that carboplatin-pemetrexed in patients with treatment-naive BRAF-mutant metastatic NSCLC resulted in longer PFS in patients with class I mutations (6.2 months) versus class II or III mutations (3.3 months and 4.9 months, respectively)33. Additionally, in a multi-institutional prospective lung cancer screening project, median PFS with platinum-containing chemotherapy was longer in patients with class I mutations (11.5 months) than in those with class III mutations (5.3 months)31. Several trials demonstrated that BRAF monotherapy or BRAF plus MEK inhibitor therapy was effective in patients who had progressed on chemotherapy10,15,18,66. Chemotherapy remains a second-line recommendation for patients with a BRAFV600 mutation21,22.

Immunochemotherapy

The combination of immunotherapy and chemotherapy is approved for first-line treatment of patients with metastatic NSCLC but not specifically for patients with BRAFV600E mutations80. In a phase 3 trial of treatment-naive patients with metastatic non-squamous NSCLC, the combination of pembrolizumab, pemetrexed, and a platinum resulted in significantly longer median PFS compared with chemotherapy alone (8.8 versus 4.9 months; HR = 0.52; P < 0.001)81. In a small retrospective study in China, immunotherapy plus chemotherapy (n = 9) in treatment-naive patients with BRAFV600E-mutant advanced NSCLC resulted in a significantly longer median PFS compared with chemotherapy or targeted therapy (n = 20; 18.5 versus 4.1 months; P = 0.0098)82. This efficacy benefit with immunochemotherapy was not observed in later lines. Another retrospective study showed similar efficacy with immunotherapy-based treatments in patients with advanced NSCLC with or without BRAF mutations; median PFS was 8.4 months in both patient populations83. In the BRAF cohort, median PFS was similar for V600E and non-600E mutations (10.0 versus 8.0 months). Median PFS was longer in the first line than in subsequent treatment lines in patients with WT (12.8 versus 5.6 months) and BRAF-mutant (11.2 versus 4.0 months) NSCLC. These studies suggest that immunotherapy-based treatments are an option for patients with BRAFV600E-mutant advanced NSCLC82,83.

Emerging treatments and approaches

Immunotherapy plus targeted therapy combinations

The combination of BRAF-targeted therapy plus immunotherapy may produce a synergistic anti-tumor effect; tolerability of the combined approach will be a key consideration14. Studies have investigated various combinations of anti-PD-1/PD-L1 with BRAF and/or MEK inhibitors and reported positive outcomes in solid tumors, including BRAFV600E/K-mutant melanoma, CRC, and BRAFV600E-mutant NSCLC84,85,86,87. In a phase 1/1b, global, multicenter, open-label study, cobimetinib and atezolizumab (anti-PD-L1) were evaluated in immunotherapy-naive patients with advanced solid tumors (n = 150); patients had received a median of 5.0 prior systemic therapies, and 15% had BRAF mutations84. In patients with metastatic CRC (n = 84), ORR was 8% (95% CI, 3–16%), median PFS was 1.9 months (95% CI, 1.8–2.3), and median OS was 9.8 months (95% CI, 6.2–14.1). In patients with melanoma (n = 22), ORR was 41% (95% CI, 21–64%), median PFS was 13.3 months (95% CI, 2.8–NE), and median OS was NE (95% CI, 18.7–NE). In patients with NSCLC (n = 28), ORR was 18% (95% CI, 6–37%), median PFS was 5.8 months (95% CI, 2.7–9.2), and median OS was 13.2 months (95% CI, 9.2–NE). In the safety analysis (n = 150), the most common TRAEs were diarrhea (67%), rash (48%), and fatigue (40%); 44% of patients reported grade 3–4 TRAEs. Another trial (NCT03991819) is evaluating the combination of binimetinib and pembrolizumab in patients with EGFR WT, ALK-rearrangement–negative advanced or metastatic NSCLC with PD-L1 tumor proportion score (TPS) ≥ 50%; initial results reported that 33% of nine evaluable patients had a partial response, including one patient with BRAFV600E-mutant metastatic NSCLC85,86.

Trials have investigated the efficacy and safety of triple combination therapies for BRAF-mutant melanomas87,88,89. A phase 2, randomized trial enrolled patients with treatment-naive advanced melanoma with a V600E or V600K mutation to receive dabrafenib plus trametinib with or without pembrolizumab (triplet, n = 60; doublet, n = 60)88. When compared with the doublet therapy, triplet therapy resulted in longer median PFS (10.7 versus 16.9 months) and higher incidence of grade ≥3 TRAEs (25.0% versus 58.3%). Grade ≥3 AEs occurred in 70% of patients in the triplet arm and 45% of patients in the doublet arm. Immune-mediated AEs occurred in 15% and 52% of patients in the doublet and triplet treatment arms, respectively; pneumonitis (17%) and hypothyroidism (8%) were the most common immune-mediated AEs reported with triplet therapy. An open-label, phase 1/2 trial combined encorafenib plus binimetinib with pembrolizumab for patients with BRAFV600-mutant advanced melanoma (n = 15); ORR was 64% (95% CI, 35–87%), and 12-month PFS was 41% (95% CI, 13–68%)87. TRAEs were reported by 87% of patients; grade ≥3 TRAEs were reported in 53% of patients, with increases in aspartate aminotransferases, gamma glutamyl transferase, and creatine phosphokinase being the most common. A meta-analysis of triplet therapies compared with doublet therapy or monotherapy for melanoma revealed that triplet therapy significantly improved PFS and OS but was associated with increased frequency of immune-related AEs, including hypothyroidism, arthralgia, myalgia, alanine aminotransferase elevations, aspartate aminotransferase elevations, asthenia, and pyrexia90. Triplet therapy did not increase the overall incidence of AEs or grade ≥3 AEs. The increased incidence of AEs should be considered when determining the optimal combination of immunotherapy and targeted therapy88,90.

Next-generation BRAF inhibitors

Next-generation BRAF inhibitors target dimerization since it is an essential component of activation for WT and many mutant BRAF kinases, plays a role in resistance mechanisms to BRAF inhibitors, and is associated with AEs91. These drugs were developed following two main strategies aimed at inhibiting mutant RAF while preventing paradoxical activation and common acquired resistance mechanisms92. Type II pan-RAF inhibitors bind the active conformation of RAF monomers and dimers92,93. Despite the name of pan-RAF, at least three of these agents (naporafenib [LXH254], tovorafenib [TAK-580], belvarafenib) demonstrated poor inhibition of ARAF and potent inhibition of WT and mutant versions of CRAF and BRAF93. Paradox breakers (e.g., PLX8394, PLX7904) are more specific BRAF inhibitors that alter the dimer interface and subsequently prevent BRAF-homodimer and BRAF:CRAF heterodimer formations92,94. Data from preclinical and clinical studies suggest that BRAF non-V600 mutations could be targeted with these new inhibitors, and several are being investigated in ongoing trials to better define their efficacy and safety (Table 3)34,95,96.

Table 3 Ongoing trials and initial results of emerging BRAF inhibitors
Full size table

Unmet needs

Brain metastases

Brain metastases (BMs), a common (26% at diagnosis) challenge for patients with metastatic NSCLC, are associated with poor prognosis and quality of life97,98. One study reported that the incidence of baseline BMs in patients with BRAF-mutant metastatic NSCLC was significantly lower for class I mutations (9%) than for class II (29%; P = 0.011) or class III (31%; P = 0.007) mutations33. However, this subset of patients has been predominantly excluded from trials of targeted therapies15,16,18. In a trial evaluating vemurafenib in patients with BRAF-mutant advanced NSCLC, median PFS was 1.9 months (95% CI, 1.5–3.9) and 5.4 months (95% CI, 3.8–7.2) in patients with (n = 26) and without (n = 89) baseline BMs11. In patients enrolled in PHAROS with BRAFV600E-mutant metastatic NSCLC asymptomatic BMs, ORR was 100% (95% CI, 39.8–100.0%) for treatment-naive patients (n = 4) and 0% for previously treated patients (n = 4)18. Better understanding of the epidemiology, risk, impact, and treatment of BMs in patients with BRAF-mutant metastatic NSCLC remains an unmet need.

More robust efficacy analysis of patients with BMs has been conducted for BRAF-mutant metastatic melanoma treated with BRAF/MEK inhibitors. In a phase 3 study of patients with BRAFV600-mutant unresectable or metastatic melanoma with BMs (n = 275), systemic outcomes with dabrafenib plus trametinib were ORR of 41.5% and median PFS of 5.68 months (95% CI, 5.29–6.87)99. Treatment-naive patients had a significantly longer median PFS than previously treated patients (7.23 versus 4.96 months; P = 0.0061). An open-label, multi-cohort, phase 2 study evaluated the efficacy of dabrafenib plus trametinib in patients with BRAFV600-mutant melanoma with BMs (n = 125), including the largest cohort (n = 76) with asymptomatic BRAFV600E-mutant BMs and no prior brain-directed therapy100. For the largest cohort, intracranial response rate (IC ORR) was 58% (95% CI, 46–69%), median PFS was 5.6 months (95% CI, 5.3–7.4), and median OS was 10.8 months (95% CI, 8.7–19.6). The safety profile was consistent with those in previous dabrafenib plus trametinib studies. In a multicenter, retrospective case series investigation, encorafenib plus binimetinib combination treatment of 24 patients with BRAF-mutant metastatic melanoma BMs resulted in IC ORR of 33% and ORR of 21%101. The safety profile was consistent with what was reported in patients with melanoma without BMs.

Conclusion

BRAFV600E is an actionable mutation for metastatic NSCLC with safe and effective treatment options8,16,18. BRAF and MEK inhibitor combination therapies have demonstrated rapid and durable responses in the majority of patients15,16,18. The safety profiles of these combinations are well described, and appropriate therapy management principles should be employed for responding patients. Disease progression is inevitable, as only ~50% of patients with metastatic NSCLC receive a second-line treatment102,103. Therefore, it is prudent to use the most efficacious agents in the first-line setting. Further studies are necessary to determine optimal sequencing methods, understand resistance mechanisms, determine efficacy of treatments for BMs, and develop targeted therapies for non-V600E mutations.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.