Patterns of treatment and BRAF testing with immune checkpoint inhibitors and targeted therapy in patients with metastatic melanoma presumed to be BRAF positive.

Patients with BRAF V600 (BRAF) mutated metastatic melanoma are eligible for therapy with both immune checkpoint inhibitors and targeted therapies, making treatment choice a complex decision. The present study aimed to describe patterns of treatment with immunotherapy and targeted therapy and BRAF testing in patients with metastatic melanoma presumed to have BRAF mutations (BRAF+) in the years following the approval of the newer generation of immune checkpoint inhibitors and targeted therapies (2014-2016). Two large US commercial claims databases [Truven Health Analytics MarketScan and IQVIA Real-World Data Adjudicated Claims - USA (IQVIA RWD Adjudicated Claims - USA)] were used. Patients were presumed BRAF+ if they received at least 2 lines of therapy of which at least 1 included targeted therapy. Sequence of lines of therapy and regimens used in first (1L), second (2L), and third (3L), as well as timing of BRAF testing by sequence of therapy were described. In the Truven sample (n=162), targeted therapy was used by 66% in 1L and by 54% in 2L, and 62% had a BRAF test; in the IQVIA RWD Adjudicated Claims - USA sample (n=247), targeted therapy was used by 62% in 1L and by 50% in 2L, and 68% had a BRAF test. Among those with a claim for a BRAF test prior to 1L, over two-thirds were initiated on targeted therapy. These findings suggest that the rate of BRAF testing remained low in the years following the approval of BRAF-targeted regimens for metastatic disease. Given the recently approved adjuvant treatment options for stage III melanoma, improving the rates of BRAF testing becomes increasingly important.

Introduction

Melanoma – a type of skin cancer – constitutes a small proportion (<5%) among different skin cancers, yet accounts for the vast majority of skin-cancer-related deaths 1. An estimated 87 110 new cases of melanoma of the skin were diagnosed in the USA in 2017, with 9730 associated deaths 2. While 5-year survival rates for localized melanoma are over 90%, metastatic melanoma generally has a poor prognosis, with a 5-year survival of 15–20% 1.

For a long time, therapeutic options for patients with metastatic melanoma were limited to cytotoxic chemotherapy despite lack of evidence for overall survival benefit 3. Recent advances in molecular diagnostics have enabled identification of several oncogenes as important prognostic markers in metastatic melanoma. Mutation in the BRAF (V600) gene is the most commonly identified oncogene, present in ∼50% of cutaneous melanoma cases (BRAF+ melanoma) 4, with the remaining patients having unmutated BRAF melanoma. The BRAF mutation leads to the activation of RAF/MEK/ERK cell-signaling pathways – the protein product of this oncogene can selectively target metastasized cells and clinically undetected cells that harbor the BRAF mutation. BRAF gene mutation is an important tool for diagnosis, treatment, and predicting patient outcomes, and may have an impact on prognosis 5,6. The NCCN clinical practice guidelines recommend BRAF mutation testing in patients with unresectable or metastatic melanoma to guide treatment decisions as patients harboring a BRAF mutation may have different treatment options 7–9.

Molecularly targeted agents have been developed to specifically target BRAF-mutated melanoma. Tyrosine kinase inhibitors, vemurafenib, and subsequently, dabrafenib and trametinib monotherapy demonstrated remarkable clinical benefits for patients with advanced or metastatic BRAF-mutated melanoma 3,10,11, although response durability was variable due to drug resistance 12,13. However, using a combination of BRAF and MEK inhibitors has led to prolonged response and decreased incidence of secondary skin malignancies compared with BRAF monotherapies 11,14. In particular, combination therapy with dabrafenib and trametinib has shown durable long-term overall and progression-free survival in patients with BRAF-mutated metastatic melanoma 15.

In parallel, advances in immune-oncology, notably the discovery of immune-checkpoints- molecules responsible for dampening immune response, including programmed cell-death protein 1 (PD-1), and cytotoxic T-lymphocyte antigen-4 (CTLA-4) inhibitors, have offered more treatment options for patients with BRAF-mutated metastatic melanoma and other melanoma subtypes 16,17. CTLA-4 inhibitors, ipilimumab and tremelimumab, and PD-1 inhibitors, nivolumab and pembrolizumab, have shown survival benefit for advanced or metastatic melanoma regardless of BRAF mutation status 18,19. Despite slower initial response and higher risk for severe autoimmune toxicity compared with targeted therapies, some immunotherapies can result in durable long-term survival in patients with metastatic melanoma 20,21. Combination therapy with ipilimumab and nivolumab has shown more rapid response rates and longer overall survival compared with monotherapy 22.

With the current repertoire of mono and combination targeted therapies and immune checkpoint inhibitors options for metastatic BRAF-mutated melanoma, choosing the appropriate therapy is a complex decision. Current considerations that may influence treatment decisions include various clinical and patient considerations including disease-related symptom burden, performance status, presence of brain metastases, possible adverse events of different agents, or patient preference for oral vs. intravenous medications 23,24. However, in the absence of head-to-head comparisons between immune checkpoint inhibitors and targeted therapies many questions remain unanswered with respect to the optimal treatment sequence in metastatic melanoma 23. A better understanding of treatment patterns and BRAF testing in real-world practice is needed to identify unmet needs in this patient population. However, data regarding real-world treatment practices and patterns of BRAF testing remain scarce. Using two large commercial claims databases in the USA, this study aimed to describe patterns of BRAF testing and treatment with immune checkpoint inhibitors and targeted therapy in patients with metastatic melanoma presumed to have BRAF mutations (BRAF+) in the years following the approval of the newer generation of immune checkpoint inhibitors and targeted therapies (2014–2016).

Methods

Data sources

The study used data from two large commercial pharmacy and medical claims databases in the USA: the Truven Health Analytics MarketScan (Truven, Ann Arbor, Michigan, USA) and IQVIA (Danbury, Connecticut, USA) Real-World Data Adjudicated Claims – USA database (IQVIA RWD Adjudicated Claims – USA). As the two databases cannot be combined, analyses were performed separately on the two samples.

Truven database contains health services claims for more than 230 million individuals from ∼100 employers and several health plans. IQVIA RWD Adjudicated Claims – US contains health services claims for more than 87 million members from over 100 health plans across the US. Both databases include demographic characteristics, enrollment history, and claims for medical (provider and institutional) and pharmacy services. Truven data are representative of all census regions, predominantly the South and North Central (Midwest) regions. IQVIA RWD Adjudicated Claims – USA database is representative of the national commercially insured population. Both databases contain data from large employers that tend to have generous coverage, especially for new agents. The databases are fully compliant with the Health Insurance Portability and Accountability Act, and thus, no ethics board review was required 25.

Study design

A retrospective cohort design was employed. In both samples, patients were assigned an index date corresponding to the date of initiation of the first immune checkpoint inhibitor(s) or targeted therapy for metastatic melanoma after a diagnosis of melanoma. The baseline period was the 6 months prior to the index date and was extended until the earliest of the end of eligibility (due to disenrollment or death), or the end of data availability (30 June 2016 for Truven and 30 September 2016 for IQVIA RWD Adjudicated Claims – USA).

Study sample

Both samples (Truven and IQVIA RWD Adjudicated Claims – USA) included patients with metastatic melanoma as identified by claims for melanoma after 1 January 2014, and a systemic therapy with immune checkpoint inhibitors or targeted therapy for metastatic melanoma. Patients were included in the study if they (a) were at least 18 years old on the index date; (b) had at least one diagnosis of melanoma (International Classification of Disease, 9th ed., clinical modification code 172.xx), from 1 January 2008 to end of data availability specific to each database; (c) had initiated at least one immune checkpoint inhibitor (ipilimumab, pembrolizumab, nivolumab) or targeted therapy (vemurafenib, dabrafenib, trametinib), on or after 1 January 2014 either as monotherapy or as combination regimens; (d) had a first diagnosis of melanoma before or on the index date; (e) had continuous healthcare plan eligibility (including drug coverage) during the baseline period and at least 1 month after the index date; (f) were not enrolled in a clinical trial at any time after the melanoma diagnosis date; and (g) received two or more lines of therapy after the index date, of which at least one line included targeted therapy (i.e. patients were presumed to be BRAF+). At least two lines of treatment were required for all patients as those treated with immune checkpoint inhibitors in 1L had to have two or more lines of therapy to be identified as presumed BRAF+. A small number of patients who received combined immune checkpoint inhibitors and targeted therapy in first line (<1%) were excluded (Supplementary Fig. 1A and 1B, Supplemental digital content 1, ).

Study measures and statistical analyses

All analyses were descriptive. Frequencies and proportions were reported for categorical variables, and means (SD), medians, and ranges were reported for continuous variables. For each of the two samples, reported treatment patterns post-index date included sequence of lines of therapies as well as all treatment regimens used in first (1L), second (2L), and third (3L) lines post-index date. Identification of lines of pharmacological therapy was adapted from previously published algorithms 26–28. All agents – single or combination of multiple agents – received during the first 28 days of the line of therapy start constituted the treatment regimen. For regimens with immune check-point inhibitors (intravenous) the length of a cycle was assumed to be 21 days, while for targeted agents (oral) duration of treatment was based on the days’ supply. In both cases, gaps of more than 45 days were allowed between consecutive treatments with the same regimen. Thus, a line of therapy was considered to be discontinued if there was a gap of more than 45 consecutive days without treatment or if a new agent (i.e. new line of therapy) was initiated. In the case of immune check-point inhibitors, gaps up to 180 days were allowed if the patient received steroids. Lines of therapy were censored if the patient was still on treatment at the end of the study follow-up.

The proportion of patients presumed BRAF+ for whom BRAF testing was performed from the date of first melanoma diagnosis to 2L start, and the timing of BRAF testing by sequence of therapy were described. Identification of BRAF testing was based on procedure codes for BRAF-specific and sequencing tests [i.e. BRAF gene analysis and other multigene sequencing tests did not include immunohistochemistry (IHC)]. Procedure codes for BRAF-specific testing also captured other methods of BRAF testing including the Droplet Digital PCR test (Bio-Rad, Hercules, California, USA) and Idylla BRAF Mutation Test (Biocartis, Mechelen, Belgium). Patient characteristics during the baseline period, including demographic characteristics, underlying tumor burden, comorbidities, and healthcare utilization and costs were outlined by the type of therapy received in 1L in patients with and without a BRAF test claim before 1L initiation.

Results

A total of 2231 and 2632 patients with at least one diagnosis of melanoma and at least one immune checkpoint inhibitor administration or dispensing of targeted therapy on or after the index date were identified from the Truven database and IQVIA RWD Adjudicated Claims – USA database, respectively. From these, 162 patients from the Truven database, and 247 patients from the IQVIA RWD Adjudicated Claims – USA database with metastatic melanoma and presumed BRAF+ met the inclusion criteria and were included in the study (Supplementary Fig. 1A and 1B, Supplemental digital content 1, ). In the Truven sample, mean age was 56 years and ∼60% of the patients were men. In the IQVIA RWD Adjudicated Claims – USA sample, mean age was 53 years and ∼62% were men.

Treatment patterns

In the Truven sample, 107 (66%) of presumed BRAF+ patients were initiated in 1L on targeted therapies and 55 (34%) were initiated on immune checkpoint inhibitors (Fig. 1a). In 2L, 86 (55%) and 61 (38%) received targeted therapy and immune checkpoint inhibitors, respectively; seven (4%) were treated with combination targeted therapies and immune checkpoint inhibitors in 2L. Fifty percent of patients treated with targeted therapy in 1L switched to immune checkpoint inhibitors in 2L and 78% of patients receiving immune checkpoint inhibitors in 1L switched to targeted therapy in 2L (Fig. 1a). In the IQVIA RWD Adjudicated Claims – USA sample, 152 (62%) of presumed BRAF+ patients were initiated on targeted therapy and 95 (38%) were initiated on immune checkpoint inhibitors (Fig. 1b). In 2L, 124 (50%) and 77 (31%) received targeted therapy and immune checkpoint inhibitors, respectively; 29 (12%) were treated with combination targeted therapies and immune checkpoint inhibitors in 2L. Forty-two percent of patients treated with targeted therapy in 1L switched to immune checkpoint inhibitors in 2L and 73% of patients receiving immune checkpoint inhibitors in 1L switched to targeted therapy in 2L (Fig. 1b).

Fig. 1

Sequences of lines of pharmacological therapies post-index date: (a) Truven sample, (b) IQVIA RWD Adjudicated Claims – USA sample. I-O, immune check-point inhibitors; other, include: combinations of I-O or TT drugs with other antineoplastic agents, or other antineoplastic agents (by design not in 1L). 1L, first line of therapy post-index; 3L, third line of therapy post-index; TT, targeted therapy.

Figure 2 depicts the distribution of specific treatment regimens by line of therapy. In both Truven and IQVIA RWD Adjudicated Claims – USA samples, during the 2014–2016 study period, the most common immune checkpoint inhibitor regimen used in 1L and 2L was ipilimumab monotherapy, while pembrolizumab monotherapy was the most common immune checkpoint inhibitor regimen in 3L. Dabrafenib and trametinib combination therapy was the most commonly prescribed targeted therapy in 1L, 2L, and 3L (Fig. 2a and b).

Fig. 2

Distribution of immune checkpoint inhibitors (I-O), targeted therapy (TT), and other regimens by line of therapy: (a) Truven sample, (b) IQVIA Real-World Data Adjudicated Claims – USA sample. Other combinations include: nonstandard combinations of I-O or TT drugs, combinations of I-O or TT drugs with other antineoplastic agents, I-O in combination with TT (not in 1L by design), or other antineoplastic agents (not in 1L by design) 1. 1L, first line of therapy post-index; 2L, second line of therapy post-index; 3L, third line of therapy post-index.

BRAF testing

In the Truven sample, 100 of 162 (62%) of presumed BRAF+ patients had a BRAF test claim between melanoma diagnosis and 2L, including 57% of patients who initiated targeted therapy and 56% who initiated immune checkpoint inhibitors. The majority of those with a BRAF test claim (92%) had the test between the first melanoma diagnosis and 1L initiation (Table 1). In the IQVIA RWD Adjudicated Claims – USA sample, 168 of 247 (68%) of presumed BRAF+ patients had a BRAF test claim between melanoma diagnosis and 2L, inclusive of 65% of patients who initiated targeted therapy and 61% who initiated immune checkpoint inhibitors; 94% were tested between first melanoma diagnosis and 1L initiation (Table 1).

Table 1

Timing of BRAF testing by treatment sequence

Table 2 describes the characteristics of patients presumed to be BRAF+ with and without a BRAF test claim before 1L initiation stratified by type of therapy received in 1L, in the Truven and IQVIA RWD Adjudicated Claims – USA samples, respectively. In the Truven sample, among those with claim(s) for BRAF test, the proportion of patients with emergency department visits and renal disease during the baseline period was significantly higher in patients who initiated targeted therapy compared with those who initiated immune checkpoint inhibitors (P<0.05). Likewise, in the IQVIA RWD Adjudicated Claims – USA sample, among those with claim(s) for BRAF test, the proportion of patients with inpatient admissions was significantly higher among patients who initiated targeted therapy compared with those who initiated immune checkpoint inhibitors (P<0.05). No statistically significant differences were observed between patients treated in 1L with immune checkpoint inhibitors versus targeted therapies among those without a BRAF test claim prior to 1L.

Table 2

Patient characteristics stratified by BRAF testing and treatment used in 1L

Even though not statistically significant, a tendency towards more patients with brain metastases, anemia, immune diseases, and liver disease was observed among patients treated with targeted therapies in 1L as compared with patients treated with immune checkpoint inhibitors in 1L, especially in the Truven sample. At the same time, there were numerically more patients with cardiovascular diseases among those treated with immune checkpoint inhibitors than those treated with targeted therapies. In both samples, more than half of the patients who did not have a claim for BRAF test before 1L initiation had claim(s) for IHC test in the same period.

Overall, in the Truven sample, patients with a BRAF test claim prior to 1L initiation had more brain metastases, a higher number of metastatic sites, and higher Charlson comorbidity score than patients who did not have a claim for BRAF test prior to 1L initiation (Supplementary Table 1A, Supplemental digital content 1, ). They also had higher all-cause cost during the baseline period, and more IHC tests before 1L initiation. In the IQVIA RWD Adjudicated Claims – USA sample, patients with a BRAF test claim prior to 1L initiation had more surgeries, a higher number of metastatic sites, and more IHC tests before 1L than patients who did not have a claim for BRAF test prior to 1L initiation (Supplementary Table 1B, Supplemental digital content 1, ). They also had a tendency towards having more brain metastases, although this was not statistically significant.

Discussion

This study documents patterns of treatment and BRAF testing among patients diagnosed with melanoma and presumed BRAF+ in a real-world clinical setting. A higher proportion of patients presumed to be BRAF+ were treated with targeted therapy than immune checkpoint inhibitors in 1L and subsequent lines. Approximately two-thirds of presumed BRAF+ patients received targeted therapy in 1L and half received targeted therapy in 2L. Although recently approved targeted therapies and immune checkpoint inhibitors were commonly used; surprisingly, older agents particularly ipilimumab monotherapy, continued to be used for some patients during the study period. Only two-thirds of patients presumed to be BRAF+ had a BRAF-specific or sequencing test between melanoma diagnosis and 2L, although many of those without a BRAF-specific or sequencing test had an IHC test. Among patients who had a claim for a BRAF test, over two-thirds received targeted therapy in 1L.

Studies of treatment patterns specifically among patients with BRAF-mutant metastatic melanoma in real-world settings remain scare. To our knowledge, no study to date has examined patterns of BRAF testing among patients with metastatic melanoma in real-world settings, with the exception of one study that investigated treatment patterns and outcomes in BRAF-mutant melanoma patients with brain metastases receiving vemurafenib, a subgroup of patients that is likely not representative to all patients with metastatic melanoma 29. Furthermore, there are no head-to-head comparisons between targeted therapies and immune checkpoint inhibitors in metastatic melanoma 23, even though there is evidence showing efficacy of both targeted therapy and immune checkpoint inhibitors relative to older therapies in 1L 19,30,31. Therefore, past studies are of limited relevance to the present study that describes treatment patterns in a broader population of melanoma patients presumed BRAF+-treated with a wide range of targeted or immune checkpoint inhibitors regimens in real clinical practice.

In the absence of hard evidence of optimal treatment sequence, it remains unclear how physicians select 1L and subsequent treatments in real-world settings and which treatments are favored by the physicians at present 23. The results of the current study suggest approximately two-thirds of patients presumed to be BRAF+ receive targeted therapy in 1L and subsequent lines of therapy, while approximately one-third receive immune checkpoint inhibitors. Furthermore, some patients appear to continue to be treated with older therapies (i.e. ipilimumab, targeted agents in monotherapy) even after the approval of the new generation of immune checkpoint inhibitors and targeted therapies. A proportion of patients treated with targeted therapy in 1L continued on targeted therapy in 2L. The reasons for discontinuing or restarting a treatment regimen are not known in claims data and our study algorithm counts as a different line of therapy if the patient restarts one or more agents from the initial regimen after a gap of more than 45 days without treatment (in Truven this pattern was observed for 53.5% of patients who received targeted therapy in both 1L and 2L; in IQVIA RWD Adjudicated Claims – USA sample the pattern was observed for 30.9%). For the remaining patients in both samples, 2L included at least one new agent (i.e. either an add-on to the regimen used in 1L or a switch to a completely new regimen; for 18.6 and 30.9%, respectively, the change was from vemurafenib monotherapy in 1L to regimens with newer agents in 2L).

The choice of treatment in this patient population is likely influenced by numerous patient-specific and treatment-specific factors such as patient’s tumor and comorbidity profile, treatment adverse effects profile, as well as physician preference. The present study suggests that targeted therapy may be channeled towards more frail patients with higher comorbidity burden including those with brain metastases. Patients receiving targeted therapy had a higher comorbidity burden compared with those receiving immunotherapy in the IQVIA RWD Adjudicated Claims – USA sample; in the Truven sample, the proportion of patients with brain metastases was numerically higher among patients receiving targeted therapy compared with those receiving immune checkpoint inhibitors (all patients regardless of BRAF testing: 39 vs. 22%, P=0.02). Healthcare utilization [emergency department visits (Truven sample) and inpatient admissions (IQVIA RWD Adjudicated Claims – USA sample)] during the baseline period, which may be indicative of higher comorbidity burden, was also higher among patients who were treated with targeted therapy compared with those who received immune checkpoint inhibitors. Previous studies also suggested that the underlying comorbidity and tumor burden are important clinical factors influencing treatment choice. For example, in cases of extensive or rapidly progressing visceral metastases or high disease-related symptom burden, targeted therapy may be favored to achieve rapid response and disease stabilization 32. Similarly, in the setting of brain metastases, targeted therapy may be favored over immune checkpoint inhibitors based on the numerous studies supporting its use in this patient population 33–35. However, immune checkpoint inhibitors were also shown to have activity in this setting 36, which may partly explain why some patients with brain metastases in this study received immune checkpoint inhibitors.

In the absence of head-to-head randomized clinical trials comparing targeted therapies and immune checkpoint inhibitors, physician preference likely plays a role. It is possible that physician’s choice of immune checkpoint inhibitors instead of targeted therapy in treating BRAF+ metastatic melanoma is based on evidence of long-term disease control achieved with immune checkpoint inhibitors in some patients. Physicians may deem the benefit of achieving durable response with immune checkpoint inhibitors to outweigh risk of severe immune-related adverse events. However, recent long-term clinical trial outcome data suggest that patients treated with first-line targeted therapy may also achieve durable disease control 37. Moreover, resistance and toxicity observed with single agent targeted therapy have been overcome by combination of BRAF and MEK inhibition that has greatly improved the efficacy and safety profile of targeted therapies 11,30. A recent meta-analysis that estimated relative efficacy and safety of systemic therapies for advanced treatment-naive BRAF-mutated melanoma found no significant difference in overall survival between PD-1 and BRAF/MEK inhibitors 8. A significant advantage of BRAF/MEK inhibition compared with all other treatment strategies (i.e. targeted monotherapies, CTLA-4/GM-CSF, CTLA-4/chemo, MEK/chemo, chemo) was found for progression-free survival. Moreover, first-line use of combination BRAF/MEK targeted therapy was the most effective treatment regimen for patients with bulky or highly symptomatic disease 8. It is unclear why older therapies continue to be used after the introduction of new treatments with improved efficacy and safety profiles. One possibility is the delayed uptake of newer treatments, particularly outside academic settings. This hypothesis is supported by the trends in use of immune checkpoint inhibitor and targeted therapy regimens in the period from 2014 to 2016 in our data. For immune checkpoint inhibitors, there was a gradual shift away from ipilimumab monotherapy (100% in 2014 to <10% in 2016) to treatment with newer immune checkpoint inhibitor regimens, in particular ipilimumab+nivolumab combination. Likewise, for targeted therapy regimens, there was a shift away from monotherapy with dabrafenib (~10% in 2014 to 0% in 2016) and vemurafenib (~25% in 2014 to <15% in 2016), to combination therapy with dabrafenib+trametinib (Supplementary Fig. 2A and 2B, Supplemental digital content 1, ). Two-thirds of patients had a BRAF test claim between melanoma diagnosis and 2L, inclusive, with the majority tested prior to 1L. The remaining patients may have been tested later or may have had a BRAF test that was not reimbursed by a commercial insurer, and therefore not captured in the data. In addition, procedure codes used to identify BRAF+ status in this study were restricted to BRAF-specific or sequencing tests. However, approximately half of patients presumed to be BRAF+, who did not have claims for BRAF-specific or sequencing tests, had claims for IHC tests prior to 1L. Given the high specificity of IHC 38, it is possible that in a subset of patients the choice of treatment was guided by IHC results. Although the NCCN guidelines state that positive VE1 (an anti-BRAF V600E monoclonal antibody) IHC results are sufficient for initiating targeted therapy, it is recommend that all VE1 IHC results should be confirmed by sequencing 7.

This study was subject to common limitations of studies based on healthcare claims data, such as occasional coding errors. However, potential inaccuracies are expected to affect all groups to a similar extent. The findings are only generalizable to the population and time period studied. Although the treatment landscape in metastatic melanoma is rapidly shifting, a lag of 6–9 months exists in the database and data were available only up to 30 June 2016 (Truven) and 30 September 2016 (IQVIA RWD Adjudicated Claims – USA). In addition, identification of BRAF testing based on procedure codes has limitations and BRAF testing that was not reimbursed by the patient’s commercial insurer was not captured. Moreover, barriers to pharmacogenetic testing may exist and reimbursement practices vary across providers that require different criteria for reimbursement of pharmacogenetic tests 39–41. In some cases the insurer or the hospital may cover the cost of testing, and in some cases patients may pay out of their pocket 42. Accessibility to pharmacogenetic tests, particularly multigene tests, is made challenging by the complex and time-consuming process of ordering these tests 40,42. As it was not possible to distinguish between BRAF+ and BRAF− in the data, BRAF+ was defined based on the use of targeted therapy in at least one line of treatment. Given that only patients who received two or more lines of therapy were included, those who stopped treatment after 1L due to either death or adequate response to treatment were not observed, and therefore findings apply only to patients who receive two or more lines of therapy. As BRAF+ patients who experience progression or relapse during or after treatment with immune checkpoint inhibitors in 1L are likely to receive targeted therapy in 2L 7, this is a reasonable method of identifying BRAF+ patients.

Given the rapidly evolving therapeutic landscape of metastatic melanoma, it is important to understand treatment patterns and factors that influence treatment decisions in real-world practice, before the introduction of newer agents. This study provides real-world insight into treatment patterns in patients with metastatic melanoma that reflect the diversity of therapeutic agents used in real-world clinical practice. The multitude of effective targeted therapy and immune checkpoint inhibitors options available to patients with BRAF+ metastatic melanoma, and the complexity of the disease, may lead to variable treatment patterns and lack of consensus regarding the optimal treatment of patients with BRAF+ metastatic melanoma in real-world practice. Further studies are needed to clarify the prognostic importance of BRAF mutation status in the metastatic melanoma setting.

Supplemental Digital Content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's website, www.melanomaresearch.com.