Melanoma and Cutaneous Malignancies

A Retrospective Evaluation of Vemurafenib as Treatment for BRAF-Mutant Melanoma Brain Metastases JAMES J. HARDING,a,c FEDERICA CATALANOTTI,d RODRIGO R. MUNHOZ,a DONAVAN T. CHENG,d AMIN YAQUBIE,a NICOLE KELLY,a GREGORY C. MCDERMOTT,c ROMONA KERSELLIUS,c TAHA MERGHOUB,e MARIO E. LACOUTURE,a RICHARD D. CARVAJAL,a KATHERINE S. PANAGEAS,b MICHAEL F. BERGER,d NEAL ROSEN,a,c,f DAVID B. SOLIT,a,c,d PAUL B. CHAPMANa,c Departments of aMedicine and bEpidemiology and Biostatistics, Memorial Sloan Kettering Cancer Center, New York, New York, USA; c Department of Medicine, Weill Cornell Medical College, New York, New York, USA; dHuman Oncology and Pathogenesis Program, Memorial Hospital Research Laboratories, New York, New York, USA; eImmunology and fMolecular and Pharmacology and Chemistry Programs, Sloan-Kettering Institute, New York, New York, USA Disclosures of potential conflicts of interest may be found at the end of this article.

Key Words. Melanoma x Brain metastasis x BRAF x Vemurafenib

ABSTRACT Background. RAF inhibitors are an effective therapy for patients with BRAF-mutant melanoma and brain metastasis. Efficacy data are derived from clinical studies enriched with physiologically fit patients; therefore, it is of interest to assess the real-world experience of vemurafenib in this population. Tumor-specific genetic variants that influence sensitivity to RAF kinase inhibitors also require investigation. Methods. Records of patients with BRAF-mutant melanoma and brain metastases who were treated with vemurafenib were reviewed. Clinical data were extracted to determine extracranial and intracranial objective response rates, progression-free survival (PFS), overall survival (OS), and safety. A bait-capture, next-generation sequencing assay was used to identify mutations in pretreatment tumors that could explain primary resistance to vemurafenib. Results. Among patients with intracranial disease treated with vemurafenib, 27 were included in survival analyses and 22 patients were assessable for response. The extracranial and intracranial objective response rates were 71% and 50%, respectively. Discordant responses were observed between

extracranial and intracranial metastatic sites in 4 of 19 evaluable patients. Median PFS was 4.1 months (95% confidence interval [CI]: 2.6–7.9); median intracranial PFS was 4.6 months (95% CI: 2.7–7.9), median OS was 7.5 months (95% CI: 4.3–not reached), with a 30.4% 1-year OS rate. Outcomes were influenced by performance status. Vemurafenib was tolerable, although radiation-induced dermatitis occurred in some patients who received whole-brain radiotherapy. Adequate samples for nextgeneration sequencing analysis were available for seven patients. Melanomas categorized as “poorly sensitive” ($20% tumor growth, new lesions, or #50% shrinkage for ,4 months) harbored co-occurring mutations in genes predicted to activate the phosphatidylinositol 3-kinase-AKT (PI3K-AKT) pathway. Conclusion. Vemurafenib is highly active in BRAF-mutant melanoma brain metastases but has limited activity in patients with poor performance status.The safety and efficacy of concurrent radiotherapy and RAF inhibition requires careful clinical evaluation. Combination strategies blocking the MAPK and PI3K-AKT pathway may be warranted in a subset of patients. The Oncologist 2015;20:1–10

Implications for Practice: Vemurafenib is active for BRAF-mutant intracranial melanoma metastases in an unselected patient population typical of routine oncologic practice. Patients with poor performance status appear to have poor outcomes despite vemurafenib therapy. Preliminary data indicate that co-occurring or secondary alterations in the phosphatidylinositol 3-kinaseAKT (PI3K-AKT) pathway are involved in resistance to RAF inhibition, thus providing a rationale for dual MAPK and PI3K-AKT pathway inhibition in this patient population.

INTRODUCTION Central nervous system (CNS) involvement in patients with metastatic melanoma remains a major therapeutic challenge associated with a significant impact on quality of life and frequently results in melanoma-related death. The prognosis

for patients with melanoma that has metastasized to the brain is poor, with estimated median overall survival (OS) of 4–6 months from the time ofdiagnosis of intracranial disease [1–4]. Surgical resection and stereotactic radiosurgery (SRS) can provide

Correspondence: James J. Harding, M.D., Memorial Sloan Kettering Cancer Center, 300 East 66th Street, New York, New York 10065, USA. Telephone: 646-888-4314; E-Mail: [email protected] Received January 11, 2014; accepted for publication February 4, 2015. ©AlphaMed Press 1083-7159/2015/$20.00/0 http://dx.doi.org/10.1634/theoncologist.2014-0012

The Oncologist 2015;20:1–10 www.TheOncologist.com

©AlphaMed Press 2015

2

local control of small-volume CNS disease [5]. For patients not amenable to localized approaches, chemotherapy [6, 7], immunotherapy [8], and whole-brain radiotherapy (WBRT) [9] result in objective responses in only 7%–16% of patients. The RAF kinase inhibitors vemurafenib and dabrafenib have marked activity against BRAF-mutant melanomas, with response rates of ∼50% [10–13]. As opposed to dacarbazine, both vemurafenib and dabrafenib have been shown to extend progression-free survival (PFS) in patients with unresectable or metastatic BRAF-mutant melanoma, and vemurafenib extends their OS [12, 13]. A phase II study with dabrafenib showed an intracranial response rate of 30.8% in patients who had progressed after prior therapy for brain metastases and 39.2% in patients who had no prior therapy for brain metastases [14]. Prospective data indicate that vemurafenib is also active against melanoma brain metastasis, with a similar response rate [15, 16]. The inclusion criteria for the three prospective studies of RAF inhibitors were restrictive, limiting the size of intracranial lesions in some instances and excluding those patients with poor performance and or leptomeningeal disease. Consequently, there are few meaningful published data to guide clinicians in routine clinical practice on the use of vemurafenib in patients with advanced CNS disease. We conducted a retrospective analysis of 27 patients with BRAF-mutant melanoma and brain metastasis treated with vemurafenib at a single center. In this article, we report the clinical activity, outcomes, and toxicity of this approach in a patient population that might be more representative of the general population [1–4]. For the purposes of hypothesis generation, we also describe initial results of a next-generation sequencing assay performed on pretreatment metastatic tumors that aimed at defining somatic mutations that predict sensitivity to vemurafenib.

METHODS Patient Selection and Clinical Characteristics The database of Memorial Sloan Kettering Cancer Center (MSKCC) was queried for all patients with unresectable or metastatic melanoma with a BRAFV600 mutation who were treated with vemurafenib sometime between 2007 and 2013. Tumors were genotyped by standard amplification and sequencing of exon 15 of the BRAF gene [10] or by using a mass spectrometrybased assay (Sequenom Inc., San Diego, CA, https://www. sequenom.com) [17].We extracted demographic information, characteristics of the primary melanoma, Eastern Cooperative Oncology Group performance status (ECOG PS), lactate dehydrogenase, and extent of intracranial disease prior to initiation of vemurafenib, history of prior or concurrent intracranial therapy, and history of ipilimumab treatment. This retrospective analysis was approved by the MSKCC institutional review board.

Efficacy All patients received vemurafenib at an initial dose of 960 mg orally twice a day. Antitumor efficacy was retrospectively assessed by Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 [18]. The best overall response was determined separately for intracranial and extracranial disease

Vemurafenib for Melanoma Brain Metastases sites and calculated as the maximum change in the sum of the largest diameter of up to five target lesions. Imaging modalities used were magnetic resonance imaging (MRI) for CNS lesions and computed tomography of the chest, abdomen, and pelvis for extracranial disease. Each patient was then designated as having complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD) for both intracranial and extracranial disease at each radiographic assessment. Patients who died or were lost to follow-up prior to first assessment or who had incomplete imaging throughout the treatment period were graded as not evaluable. The extracranial and intracranial objective response rates were defined as the proportion of CR plus PR in evaluable extracranial or intracranial target lesions. Discordant responses were defined as .20% tumor growth or new tumors at intracranial sites in the setting of tumor shrinkage of .20% at extracranial sites or vice versa. Specific criteria were used for selection of intracranial target lesions. For those patients who did not receive prior intracranial therapy, target lesions were selected by the investigators on the basis of measurability per RECIST version 1.1. Lesions treated previously with resection or SRS were omitted as targets. In the case of the former, residual disease after resection was considered an acceptable target. For patients who underwent WBRT, lesions were selected based on measurability and only after progression was documented on serial imaging by the treating physician, unless otherwise noted.

Toxicity Assessment Toxicity was determined by investigator-assessed chart review and graded using the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.0.

Tumor Genotyping and Exon-Capture Sequencing In patients who had sufficient pretreatment formalin-fixed, paraffin-embedded (FFPE) tissue from either an intracranial or extracranial metastatic site and accompanying matched normal DNA obtained from peripheral blood, we profiled genomic alterations in 300 cancer-associated genes using the IMPACT assay (Integrated Mutation Profiling of Actionable Cancer Targets). This assay uses solution phase hybridizationbased exon capture and massively parallel DNA sequencing to capture all protein-coding exons and selected introns of 300 genes of interest (oncogenes,tumor suppressor genes, and members of pathways deemed actionable by targeted therapies). The analysis was performed as reported previously [17, 19]. A mean unique sequence coverage of 669 was achieved. For this report, we focused our analysis on a subset of 35 genes selected on the basis of the frequency of mutations reported in melanoma samples or on their ability to activate the MAPK and/or phosphatidylinositol 3-kinase (PI3K) signaling pathway. For the purposes of correlating somatic genetic changes, tumor sensitivity to vemurafenib was scored as highly sensitive ($50% tumor shrinkage for $7 months or any shrinkage for $12 months), poorly sensitive ($20% tumor growth, new lesions, or #50% shrinkage for ,4 months), or intermediate (not highly or poorly sensitive). This scoring system was based on the median duration of response to

OTncologist he

©AlphaMed Press 2015

®

Harding, Catalanotti, Munhoz et al.

3

vemurafenib and average tumor shrinkage according to the published data [10–13].

Table 1. Patient characteristics (n 527)

Biostatistics

Sex, n (%) Male Female Age, median (range) Primary site, n (%) Head/extremity Trunk Unknown BRAF mutational status, n (%) V600E V600K ECOG PS, n (%) 0 1 $2 LDH, U/L median (range)a Intracranial disease burden, leptomeningeal disease, n (%) Number of lesions, n (%) ,5 5–10 .10 Maximum tumor diameter, mm (range) Prior intracranial treatment, n (%) WBRTb Stereotactic radiosurgery only Craniotomy only Multimodality therapyc None Treatment with ipilimumab, n (%) None Prior Subsequent

Continuous variables were summarized with descriptive statistics, such as median and range. Categorical variables were tabulated as counts and percentages. Response rates were calculated along with a 95% confidence interval (CI). The chi-square test was used for intergroup comparisons. PFS was defined as the time from the start of vemurafenib until the date of last documented contact (censored), date of progression (overall progression or intracranial progression, whichever occurred first), or date of death. OS was defined as the time from the start of vemurafenib until the date of last documented contact (censored) or date of death. PFS and OS were estimated using Kaplan-Meier methodology with the logrank test for comparison between subgroups. A p value of ,.05 was considered significant. Genomic data are reported in a descriptive fashion. All analyses were completed using GraphPad Prism version 6 (GraphPad Software, Inc., La Jolla, CA, http://www.graphpad.com).

RESULTS Patient Characteristics Of 140 patients with unresectable or metastatic BRAF-mutant melanoma in the database, 27 (19.3%) had brain metastasis prior to the start of treatment with vemurafenib and were assessable for survival and toxicity. Clinical characteristics of the patients are reported in Table 1. The majority of the melanomas harbored a BRAFV600E mutation; 15% harbored a BRAFV600K mutation. Notably, patients had an extensive burden of intracranial disease. Eleven patients (41%) patients had $5 lesions, and 6 (22%) of these patients had .10 lesions. Three patients (11%) had leptomeningeal disease documented by cross-sectional imaging and/or cerebrospinal fluid cytology in addition to measurable targets by RECIST. The median diameter of the largest brain metastases was 19.5 mm (range: 4–56 mm). In general, patients had functional limitations due to disease, with 15 of 27 patients (56%) having an ECOG performance status of 1 and 5 patients (18.5%) having a performance status of 2 or worse. The median time from the diagnosis brain metastases to the start of vemurafenib was 6 weeks (range: 0.3–24.4 weeks) for all patients. Fifteen patients (55%) had received prior WBRT, SRS, debulking craniotomy, or a combination of these treatment modalities. No patients had been previously treated with different BRAF or MEK inhibitors. One patient received WBRT concurrently with vemurafenib therapy. Fourteen patients (52%) received ipilimumab at some time during their disease course, including five patients who received ipilimumab following progression on vemurafenib.

Antitumor Activity Of 27 patients identified, 24 patients were radiographically evaluable for extracranial response, intracranial response, or both. Supplemental online Figure 1 shows the frequency of interval imaging over the disease course for each patient until progression of disease, death, or last follow-up.Three patients

www.TheOncologist.com

Characteristic

Result 15 (56) 12 (44) 53 (25–86) 10 (37) 11 (41) 6 (22) 23 (85) 4 (15) 7 (26) 15 (55.5) 5 (18.5) 382.5 (205–742) 3 (11)

16 (59) 5 (19) 6 (22) 19.5 (4–67) 9 (33.5) 1 (3.5) 3 (11) 2 (7) 12 (44) 13 (48) 9 (33) 5 (19)

a

Only available in 10 of 27 patients. One patient received vemurafenib and concurrent WBRT. c Two patients received multimodality therapy consisting of craniotomy with stereotactic radiosurgery or WBRT. Abbreviations: ECOG PS, Eastern Cooperative Oncology Group performance status; LDH, lactate dehydrogenase; WBRT, whole-brain radiotherapy. b

who were treated at MSKCC and other hospitals had incomplete imaging throughout the course of therapy, including the one patient who received concurrent WBRT and vemurafenib therapy. Consequently, adequate assessment of best change from baseline target lesion was not possible for these patients. Of 24 patients with adequate cross-sectional imaging, 21 were evaluated for determination of the extracranial objective response rate. Three patients were excluded because measurable extracranial disease was not present. There was 1 patient with CR, 14 with PR, 5 with SD, and 1 with PD (Fig. 1A). Consistent with prior reports, the extracranial objective response rate for evaluable patients was 71% (15 of 21 ©AlphaMed Press 2015

Vemurafenib for Melanoma Brain Metastases

4

disease who received prior WBRT, progression of disease could not be confirmed prior to the start of vemurafenib; however,exclusion of these patients did not alter the intracranial response rate (50%). We did not observe a significant difference in intracranial objective response rates in patients who had received prior cranial irradiation compared with those who had not. Interestingly, 4 of 19 patients who had both extracranial and intracranial disease had discordant responses (Fig. 2). In all 4 cases, there was progression in the brain (new brain metastases or growth of target lesions of at least 20%) at the first interval assessment despite shrinkage of visceral disease.

Patient Outcomes

Figure 1. Antitumor response. (A): Extracranial objective response for 21 evaluable patients, measured as the change from baseline in the sum of the largest diameter of each extracranial target lesion. (B): Intracranial objective response for each of the 22 evaluable patients, measured as the change from baseline in the sum of the largest diameter of each intracranial target lesion. Two patients had nonmeasurable intracranial disease at baseline and are not depicted. pPatients with progression of disease as best response in extracranial (A) or intracranial (B) lesions. K denotes melanomas characterized by BRAFV600K mutation. #, Pathologic complete response. White bars represents patient who did not receive prior intracranial therapy. ‡, Patients in whom intracranial metastases were treated with whole-brain radiotherapy and did not clearly progress prior to the start of vemurafenib.

patients; 95% CI: 48%–88%); the median time to response was 7.1 weeks [12]. Twenty-two patients were radiographically evaluable for objective intracranial response. Two patients were not evaluable because follow-up brain imaging was not available. Both patients had initial evidence of extracranial response to vemurafenib. In one case, the patient died 2 months after the start of therapy, before intracranial assessment. In the second case, gastrointestinal hemorrhage occurred due to visceral progression of disease, leading to the discontinuation of vemurafenib prior to brain reimaging. The median intervals to first extracranial and intracranial response assessments were 7.5 weeks and 5.8 weeks, respectively. Twelve patients had restaging scans done simultaneously or less than 7 days apart. Assessment of intracranial disease revealed three patients with CR, eight with PR, seven with SD, and four with PD (Fig. 1B). The intracranial objective response rate (intracranial CR plus PR) for evaluable patients was 50% (11 of 22 patients; 95% CI: 29%–71%). In four patients with evaluable intracranial

The median follow-up for surviving patients was 10.8 months (range: 2.0–18.9 months). The estimated median PFS for the entire cohort was 4.1 months (Fig. 3A). The estimated median OS for all patients was 7.5 months, and the estimated 1-year OS was 30.4% (95% CI: 13.2%–50.1%) (Fig. 3B). In univariate analysis, only ECOG performance status affected OS (median OS: not reached for ECOG PS 0, 8.8 months for ECOG PS 1, 2.0 months for ECOG PS $2; p 5 .0006). The estimated median intracranial PFS was 4.6 months (95% CI: 2.7–7.9 months) and was calculated based on survival data of 26 patients; 1 patient treated with concurrent WBRT and vemurafenib was excluded from intracranial PFS analysis.

Tolerability The majority of treated patients required at least one dose modification (19 of 27 patients, 70.4%), although no patient permanently discontinued treatment because of adverse events. The spectrum and severity of adverse events to vemurafenib were similar to available prospective data in patients without brain metastases (Table 2) [12]. We noted that 3 of 10 patients who had WBRT developed either grade 2 radiation dermatitis or radiation recall (Table 2; supplemental online Fig. 2). In a patient receiving vemurafenib concurrent with WBRT, a painful cutaneous reaction began 1 week into combined therapy and reached maximum severity 3 weeks after the initiation of treatment (supplemental online Fig. 2A). The reaction, which was limited to the radiation field, was characterized by an erythematous hyperkeratotic plaque. Tortuous skin folding and hypertrophy was also recognized and consistent with acquired cutis verticis gyrata. After 3 months of topical emollients and topical high-potency corticosteroids, the dermatitis resolved completely (described in detail in [20]). A second patient started vemurafenib 4 days after completing WBRT.Within 1 week, the patient complained of pain, erythema, edema, and desquamation of the scalp and ears (supplemental online Fig. 2B). This grade 2 radiation dermatitis resolved with a vemurafenib dose modification and topical corticosteroids after 4 weeks. One additional patient developed radiation recall after 3 weeks of vemurafenib therapy, which was 6 weeks after the completion of WBRT. The reaction was characterized by erythema; hyperkeratosis, and acquired cutis verticis gyrata (supplemental online Fig. 2C) [20]. These symptoms persisted but did not worsen until she died 11 weeks after the start of vemurafenib.

OTncologist he

©AlphaMed Press 2015

®

Harding, Catalanotti, Munhoz et al.

Figure 2. Discordant intracranial and visceral response. (A): A 2by-2 table of 19 patients with extracranial and intracranial imaging demonstrating 15 concordant responses and 4 discordant responses. (B, C): Representative imaging of a discordant intracranial and extracranial response. A pretreatment magnetic resonance imaging brain scan and computed tomography (CT) scan of the abdomen demonstrate a left high frontoparietal operative cavity with enhancing tissue and mild edema, consistent with residual disease, and a large right adrenal metastasis with invasion to the inferior vena cava (B). Approximately 6 weeks later, a CT scan of the head shows disease progression with extensive edema and midline shift, whereas the retroperitoneal disease shrank significantly (C).

Genomic Profiling and Sensitivity to Vemurafenib Sufficient DNA from pretreatment metastatic melanoma sites and matched normal DNA were available for next-generation sequencing analysis in seven patients. Besides the BRAF V600E mutation, which was observed in all 7 patients, no other recurrent mutations were identified in the evaluated 300 cancer-associated genes. Mutation status of 35 genes of interest in pretreatment tumors from 7 patients is shown in Figure 4. Three patients (patients 1–3) had poorly sensitive tumors and progressed after 2–3 months of therapy. We noted that alterations in the PI3K-AKTpathway due to activating mutation in PIK3CA or PTEN inactivation were present in these tumors. The PIK3CAE545K mutation observed in patient 1 (Fig. 4B) is a hot spot mutation observed in several cancers [21]. It occurs within the helical domain (exon 9) and induces transformation of chicken embryo fibroblasts in vitro. The transforming potential of this point mutation is associated with increased

www.TheOncologist.com

5

Figure 3. Patient outcomes. Progression-free survival (A) and overall survival (B) calculated by the Kaplan-Meier method. Ticks represent censored patients (n 5 27).

lipid kinase activity and increased phosphorylated AKT [22]. In addition, a missense mutation in ERBB4 and a nonsense mutation in RB1 (R455*) that abrogates RB function [23] were observed in patient 1 (Fig. 4A).Tumors with loss of RB1 appear to be less dependent on RAF signaling for cellular proliferation, indicating a role for alterations in this pathway in RAF inhibitor resistance [24]. Highly activating ERBB4 mutations are observed in melanoma, but no evidence shows that such mutations induce resistance to RAF inhibition [25]. A PTEN mutation is due to a frameshift insertion of four base pairs (Fig. 4C) in patient 2. Using the ExPASy tool (Swiss Institute of Bioinformatics, Lausanne, Switzerland, http:// www.expasy.org), we mimicked the mutation and discovered that it leads to the generation of a new stop codon soon after insertion, generating a truncated protein. The functionality of the truncated protein requires further investigation, but the alteration is predicted to lead to an inactive protein. In patient 3, two separate alterations in PTEN were observed and have been reported previously: a nonsense mutation (PTENQ149*) and a missense mutation (PTENG165R). The former leads to a truncated protein, whereas the latter disrupts the IT loop catalytic site, resulting in a nonfunctional protein [26]. Additional alterations of unclear significance in critical genes in the PI3K-AKT pathway in this patient included PIK3CAV344G, PREX2R663S, and PREX2R1263K mutations. ©AlphaMed Press 2015

Vemurafenib for Melanoma Brain Metastases

6

Table 2. Adverse events by Common Terminology Criteria for Adverse Events version 4.0 Adverse event

Total, n (%)

Rash Fatigue Hyperproliferative skin lesions Xerosis Nausea Arthralgias Headaches Diarrhea Pyrexia Photosensitivity Seizure Radiation dermatitis Radiation recall

25 (88.8) 22 (81.5) 21 (77.8) 10 (37.0) 10 (37.0) 9 (33.0) 9 (33.0) 8 (29.6) 7 (29.2) 5 (18.5) 3 (11.1) 2 (7.4) 1 (3.7)

Grade ‡3, n (%) 6 (22.2) 0 (0) 4 (14.8) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 3 (11.1) 1 (3.7) 0 (0) 0 (0)

Three highly sensitive tumors were also identified in the cohort. Patients 5 and 6 had PR for 7 and 8 months, respectively. Patient 7 attained an intracranial complete response and an extracranial partial response to vemurafenib.The patient has been free of disease for more than 19 months following a metastasectomy at the only site of residual extracranial disease, which did not reveal viable tumor on pathologic examination. Activating mutations in the PI3K-AKT pathway were not identified in these highly sensitive tumors. Mutations of unclear significance include a splice variant of RB1, a missense mutation in NOTCH4 (G1005R), and a missense mutation in PREX2 (W375L).

DISCUSSION We have reported our single-institution experience using vemurafenib in patients with BRAF-mutant melanoma metastatic to the CNS. We observed an intracranial objective response rate of 50% among 22 evaluable patients, indicating that the activity of vemurafenib for brain metastases is meaningful and relevant in an unselected patient population typical of routine oncology practice. Our results are in line with those published by Dzienis and Atkinson in a similar series of 22 patients treated with vemurafenib [27]. Kefford and colleagues recently reported early results from a phase II trial of vemurafenib in BRAF-mutated melanoma patients with brain metastases; the intracranial response rate was 26%, as assessed by the investigators [15]. A prior phase II study of dabrafenib in patients with BRAF-mutated melanoma and brain metastasis reported an overall intracranial objective response rate of 31.4%; there was little difference between patients who progressed after prior CNS therapy and patients who were treatment na¨ıve [14]. It appears from these trials that prior treatment with radiotherapy or surgery does not significantly affect the proportion of patients who respond intracranially to RAF inhibitors. In addition, the activity of BRAF inhibitors appears to be superior to ipilimumab and temozolomide. Ipilimumab resulted in intracranial response rates of 16% and 11.7% in a phase II trial and an expanded access program, respectively [8, 28]. The results of available literature addressing

different systemic treatments for patients with metastatic melanoma and CNS involvement are summarized in Table 3. It is unclear why we observed a response rate at least twice as high as prior studies, but this could reflect differences in patient populations and bias due to small sample size. In addition, a central independent review was not performed owing to the study design, and discrepancies regarding the assessment of intracranial disease burden were reported previously [14]. In the aforementioned trial by Long et al. [14], investigator and review committee assessments were discordant in 42% of the cases, and discordances were also reported in the prospective trial by Kefford et al. [15]. These findings highlight the difficulties involved in evaluation of intracranial disease in metastatic melanoma. An appealing feature of our report is the assessment of the corresponding extracranial response rate. We observed discordant responses to vemurafenib, with intracranial growth and extracranial disease shrinkage in 4 of 19 patients. Similar findings were reported by Azer et al. in 23 patients treated with dabrafenib in phase I or II trials [29] in which intracranial and extracranial responses were concordant in 71% of the cases. It is interesting to speculate on factors that might account for this apparent observation. Pharmacokinetic variables may affect the ability of vemurafenib and dabrafenib to enter the brain in a subset of patients. In a murine model, the blood-brain barrier restricts the efflux of these compounds through the action of P-glycoprotein and breast cancer resistance protein [30, 31], and by blocking these transporters, higher intracranial concentrations of vemurafenib are obtained in vivo [32]. Moreover, it has also been suggested that size and vascular composition of brain lesions may influence the ability of drugs to penetrate the brain [33]. Intertumoral heterogeneity of mutations in BRAF and other genes also may account for varying sensitivity to RAF inhibitors among different metastatic sites. Colombino et al. showed a 20% discordance between NRAS or BRAF mutational status between the primary tumor and brain metastases [34]. However, if loss of the BRAF mutation was a common mechanism of resistance, BRAFWT tumors would be expected to be found frequently at relapse during RAF inhibitor treatment, but this does not seem to be the case. Finally, technical factors relating to the assessment of intracranial disease, such as variable image quality, cystic or necrotic lesions, or the presence of edema and hemorrhage, may also influence discordance in clinical response rates [35]. Although survival data with RAF inhibitors are favorable compared with historical controls [1–4], it is disappointing that overall survival was not longer, given the high intracranial response rate in this population. A critical observation is that patients with BRAF-mutant melanoma brain metastasis and poor ECOG performance status did poorly despite vemurafenib therapy, with median overall survival of 2 months. This suggests that patients with neurologic compromise and symptomatic CNS disease are unlikely to be rescued with RAF inhibitor therapy. Approximately 50% of patients did not achieve at least a partial intracranial response, and 4 of 22 of patients had primary progression of intracranial disease. We found that poorly sensitive melanomas had mutations that were predicted or known to activate the PI3K-AKT pathway. Such alterations were not seen in highly sensitive melanomas. Six

OTncologist he

©AlphaMed Press 2015

®

Harding, Catalanotti, Munhoz et al.

7

Figure 4. Schematic illustration of representative genes altered in the five melanoma patients analyzed by the IMPACT assay. (A): Heat map showing the mutational state of 35 selected genes in the 7 melanoma patients analyzed. All patients had a BRAFV600E mutation. (B): Representation of PIK3CAE545K mutation in DNA from FFPE tumor (left upper panel) from patient 1. Multiple reads from tumor-derived DNA show a G to A transition in 41% of reads.This indicates a glutamic acid to lysine mutation in PIK3CA. No such mutation is detected in germline DNA (left lower panel) (C). A 4-bp CGTT insertion in exon 7 of PTEN was found in 110 reads of an FFPE tumor (right upper panel) of patient 2.This frameshift insertion was not observed in the normal DNA of the same patient (right lower panel).The low frequency of the PTEN mutation observed (3.9%) is due to high stroma contamination of the FFPE tumoral sample. Abbreviations: bp, base pair; FFPE, formalin-fixed, paraffin-embedded; Ins, insertion; Pt, patient.

unique mutations in genes of the PIK3-AKT pathway were identified in the patients with poorly sensitive tumors. Two tumors contained mutations expected or known to inactivate PTEN. Mutations in the tumor suppressor PTEN are frequent in melanoma, occurring in 44% of human samples [36], and deletions in PTEN are necessary for malignant transformation in a murine model [37]. Preclinical data suggest that PTEN loss renders BRAF-mutant melanoma cell lines less sensitive to the RAF inhibitors; however, additional genetic changes (i.e., RB1 loss) are probably required for innate resistance [24]. In a small cohort of melanoma patients treated with dabrafenib, alterations in PTEN trended (p 5 .059) toward shortened PFS compared with melanomas without such mutations [38]. Two

www.TheOncologist.com

patients with poorly sensitive tumors also harbored PIK3CA mutations (the melanoma from patient 3 harbored mutations in both PIK3CA and PTEN). PIK3CA mutations co-occur with BRAF in ∼5% of melanomas [36]. PIK3CAE545K has been identified previously, is known to constitutively activate AKT signaling, and is predicted to confer resistance to RAF inhibition [22]. PIK3CAV344G has been reported in other solid tumors, although its functional relevance has not been formally explored in vitro (COSMIC database). Finally, mutations in PREX2, a gene encoding a PTEN regulator, were observed in patients with both highly and poorly sensitive tumors. This gene is mutated in 14% of melanomas, and some mutations can have oncogenic potential [39]. None of the ©AlphaMed Press 2015

Vemurafenib for Melanoma Brain Metastases

8

Table 3. Summary of available studies including patients with metastatic melanoma and central nervous system involvement treated with systemic agents Study

Year

Design

Drug

Sample

Agarwala et al. [7] Long et al. (BREAK-MB) [14]

2004 2012

Phase II Phase II

Temozolomide Dabrafenib

151 Total: 172 Cohort A, V600E: 74 Cohort A, V600K: 15 Cohort B, V600E: 65 Cohort B, V600K: 18 Total: 72 Cohort A: 51 Cohort B: 21 146 Total: 146 Cohort 1: 90 Cohort 2: 56 24 Total: 22 Cohort A: 16 Cohort B: 6

Margolin et al. [8]

Queirolo et al. [28] Kefford et al. [15]

Dummer et al. [16] Dzienis and Atkinson [27]

2012

2013 2013

2014 2014

Phase II

EAP Phase II

Pilot Retrospective

Ipilimumab

Ipilimumab Vemurafenib

Vemurafenib Vemurafenib

ORR, intracranial (%)

mOS (weeks)

6

13.9

39.2 6.7 30.8 22.2

33.1 16.3 31.4 21.9

16 5 11.7

30.0 16.0 18.0

18 20 16

28.0 27.8 23.0

50 50

NR NR

Abbreviations: EAP, Expanded Access Program; mOS, median overall survival; NR, not reported; ORR, objective response rate.

PREX2 mutations reported in this study have been described previously, and it is unclear if these mutations are implicated in RAF inhibitor resistance. Gene expression profiling indicates that melanoma brain metastases are biologically distinct from the corresponding primary tumor [40]. Others have observed, using immunohistochemistry staining, that activation of the PI3K-AKT pathway is more common in melanoma brain metastases than visceral lesions [41, 42]. Furthermore, brain metastases that progress on vemurafenib appear to exhibit AKT activation and loss of PTEN, whereas paired responding extracranial lesions show no such activation of the PI3K-AKTpathway [42]. Inhibition of AKT with or without concomitant RAF inhibition leads to growth inhibition in vitro of melanoma cell lines that are derived from RAF inhibitor-resistant intracranial disease. Taken together, these observations suggest a role for targeting both the PI3KAKT and MAPK pathways in this population. Three of 10 patients in this series who received prior or concurrent WBRT developed significant cutaneous adverse events within the irradiated field. Preclinical data indicate that RAF inhibitors may be radiosensitizers [43, 44], and several clinical anecdotes suggest that vemurafenib is a radiosensitizer capable of enhancing radiation-induced toxicities such as radionecrosis and radiation recall [20, 45, 46]. Consequently, a hurdle to combination strategies with radiotherapy may be enhanced toxicity, possibly the result of paradoxical activation of ERK signaling in normal cells in response to concurrent RAF inhibition and radiotherapy [47–49]. Although no patients were treated with concurrent stereotactic radiosurgery and vemurafenib, this may be a promising strategy to deliver combined treatment that could potentially result in fewer additive side effects. This retrospective series has several limitations. The timing of radiographic assessments was not prospectively predefined, thus inherent variability in measurement of disease among

investigators can introduce a bias with regard to outcomes, particularly for assessment of progression-free survival. Nevertheless, as observed in supplemental online Figure 1, all patients underwent regular screening, and small differences in the time to interval reimaging do not affect the estimation of response and overall survival. Another limitation is that patients were not required to progress following prior wholebrain therapy to receive vemurafenib. That being said, given the extremely low response rate to WBRT, it is highly unlikely that any of these patients would have responded; however, we excluded these few patients from the response analysis. In addition, because of the retrospective nature of this trial, symptoms were not assessed in a structured way, the doses of steroids were not uniform, and this information was missing for a number of patients. Consequently, analyses regarding some additional relevant variables involved in the clinical care of patients with CNS metastases were not possible. Finally, genomic data could not be obtained for all patients, and in this regard, results should be considered preliminary and hypothesis generating.

CONCLUSION Vemurafenib is an active and effective therapy for melanoma brain metastases characterized by a BRAFV600E or BRAFV600K mutation. Vemurafenib is associated with a higher response rate than whole-brain radiotherapy. Further investigation is required to determine whether it should be considered as a first-line therapy option for patients with BRAF-mutated brain metastases not amenable to stereotactic radiosurgical techniques and how it should be combined with radiotherapy. Clinicians should be aware that there is a signal of enhanced toxicity from radiotherapy given concomitantly with vemurafenib. We believe concomitant radiotherapy and vemurafenib should be avoided, if possible. Given the frequency of PI3K-AKT pathway

OTncologist he

©AlphaMed Press 2015

®

Harding, Catalanotti, Munhoz et al.

9

activation in melanoma brain metastases, especially in those resistant to single-agent RAF inhibitors, it seems rational to consider dual RAF and PI3K-AKT pathway inhibition in this patient cohort.

ACKNOWLEDGMENTS This work was funded by the Conquer Cancer Foundation of the American Society of Clinical Oncology (J.J.H.), the John Figgie Fund (P.B.C.), and the Melanoma Research Alliance (M.F.B., D.B.S., P.B.C.).

AUTHOR CONTRIBUTIONS Conception/Design: James J. Harding, Paul B. Chapman Provision of study material or patients: James J. Harding, Federica Catalanotti, Donavan T. Cheng, Taha Merghoub, Michael F. Berger, David B. Solit, Paul B. Chapman Collection and/or assembly of data: James J. Harding, Federica Catalanotti, Rodrigo R. Munhoz, Donavan T. Cheng, Amin Yaqubie, Nicole Kelly, Gregory C. McDermott, Romona Kersellius, Mario E. Lacouture, Richard D. Carvajal, Katherine S. Panageas, Michael F. Berger, David B. Solit, Paul B. Chapman Data analysis and interpretation: James J. Harding, Federica Catalanotti, Rodrigo R. Munhoz, Donavan T. Cheng, Amin Yaqubie, Nicole Kelly, Gregory C.

McDermott, Taha Merghoub, Mario E. Lacouture, Katherine S. Panageas, Michael F. Berger, Neal Rosen, David B. Solit, Paul B. Chapman Manuscript writing: James J. Harding, Federica Catalanotti, Rodrigo R. Munhoz, Donavan T. Cheng, Richard D. Carvajal, Katherine S. Panageas, Neal Rosen, Paul B. Chapman Final approval of manuscript: James J. Harding, Federica Catalanotti, Rodrigo R. Munhoz, Donavan T. Cheng, Amin Yaqubie, Nicole Kelly, Gregory C. McDermott, Romona Kersellius, Taha Merghoub, Mario E. Lacouture, Richard D. Carvajal, Katherine S. Panageas, Michael F. Berger, Neal Rosen, David B. Solit, Paul B. Chapman

DISCLOSURES Mario E. Lacouture: Bayer, Bristol-Myers Squibb, Genentech, EMD Serono, Merck, Novocure, Helsinn, Novartis, Boehringer Ingelheim, Reata (C/A), Berg Pharma (RF), Threshold Pharmacy, Nerre Therapeutics, Advancell (H); Richard D. Carvajal: Novartis (C/A), Novartis, Morphotek, Bristol-Myers Squibb (RF); Neal Rosen: AstraZeneca, Novartis (C/A); Paul B. Chapman: Bristol-Myers Squibb, GlaxoSmithKline, Roche/Genentech (C/A, RF). The other authors indicated no financial relationships. (C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board

REFERENCES 1. Fife KM, Colman MH, Stevens GN et al. Determinants of outcome in melanoma patients with cerebral metastases. J Clin Oncol 2004;22:1293– 1300.

12. Chapman PB, Hauschild A, Robert C et al. Improved survival with vemurafenib in melanoma with BRAF V600E mutation. N Engl J Med 2011;364: 2507–2516.

22. Kang S, Bader AG, Vogt PK. Phosphatidylinositol 3-kinase mutations identified in human cancer are oncogenic. Proc Natl Acad Sci USA 2005;102: 802–807.

2. Eigentler TK, Figl A, Krex D et al. Number of metastases, serum lactate dehydrogenase level, and type of treatment are prognostic factors in patients with brain metastases of malignant melanoma. Cancer 2011;117:1697–1703.

13. Hauschild A, Grob JJ, Demidov LV et al. Dabrafenib in BRAF-mutated metastatic melanoma: A multicentre, open-label, phase 3 randomised controlled trial. Lancet 2012;380:358–365.

23. Bartkova J, Lukas J, Guldberg P et al. The p16cyclin D/Cdk4-pRb pathway as a functional unit frequently altered in melanoma pathogenesis. Cancer Res 1996;56:5475–5483.

14. Long GV,Trefzer U, Davies MA et al. Dabrafenib in patients with Val600Glu or Val600Lys BRAFmutant melanoma metastatic to the brain (BREAKMB): A multicentre, open-label, phase 2 trial. Lancet Oncol 2012;13:1087–1095.

24. Xing F, Persaud Y, Pratilas CA et al. Concurrent loss of the PTEN and RB1 tumor suppressors attenuates RAF dependence in melanomas harboring (V600E)BRAF. Oncogene 2012;31:446–457.

3. Raizer JJ, Hwu WJ, Panageas KS et al. Brain and leptomeningeal metastases from cutaneous melanoma: Survival outcomes based on clinical features. Neuro Oncol 2008;10:199–207. 4. Davies MA, Liu P, McIntyre S et al. Prognostic factors for survival in melanoma patients with brain metastases. Cancer 2011;117:1687–1696. 5. Manon R, O’Neill A, Knisely J et al. Phase II trial of radiosurgery for one to three newly diagnosed brain metastases from renal cell carcinoma, melanoma, and sarcoma: An Eastern Cooperative Oncology Group study (E 6397). J Clin Oncol 2005;23:8870– 8876. 6. Hwu WJ, Lis E, Menell JH et al. Temozolomide plus thalidomide in patients with brain metastases from melanoma: A phase II study. Cancer 2005;103: 2590–2597. 7. Agarwala SS, Kirkwood JM, Gore M et al. Temozolomide for the treatment of brain metastases associated with metastatic melanoma: A phase II study. J Clin Oncol 2004;22:2101–2107. 8. Margolin K, Ernstoff MS, Hamid O et al. Ipilimumab in patients with melanoma and brain metastases: An open-label, phase 2 trial. Lancet Oncol 2012;13:459–465. 9. Mornex F,Thomas L, Mohr P et al. A prospective randomized multicentre phase III trial of fotemustine plus whole brain irradiation versus fotemustine alone in cerebral metastases of malignant melanoma. Melanoma Res 2003;13:97–103.

15. Kefford RF, Maio M, Arance A et al. Vemurafenib in metastatic melanoma patients with brain metastases: An open-label, single arm, phase 2, multicenter study. Presented at: Society for Melanoma Research Congress Tenth Anniversary International Congress; November 17–20, 2013; Philadelphia, PA. 16. Dummer R, Goldinger SM, Turtschi CP et al. Vemurafenib in patients with BRAF(V600) mutationpositive melanoma with symptomatic brain metastases: Final results of an open-label pilot study. Eur J Cancer 2014;50:611–621. 17. Catalanotti F, Solit DB, Pulitzer MP et al. Phase II trial of MEK inhibitor selumetinib (AZD6244, ARRY142886) in patients with BRAFV600E/K-mutated melanoma. Clin Cancer Res 2013;19:2257–2264. 18. Eisenhauer EA,Therasse P, Bogaerts J et al. New response evaluation criteria in solid tumours: Revised RECIST guideline (version 1.1). Eur J Cancer 2009;45:228–247. 19. Wagle N, Berger MF, Davis MJ et al. Highthroughput detection of actionable genomic alterations in clinical tumor samples by targeted, massively parallel sequencing. Cancer Discov 2012; 2:82–93.

10. Flaherty KT, Puzanov I, Kim KB et al. Inhibition of mutated, activated BRAF in metastatic melanoma. N Engl J Med 2010;363:809–819.

20. Harding JJ, Barker CA, Carvajal RD et al. Cutis verticis gyrata in association with vemurafenib and whole brain radiotherapy. J Clin Oncol 2014;32: e54–e56.

11. Sosman JA, Kim KB, Schuchter L et al. Survival in BRAF V600-mutant advanced melanoma treated with vemurafenib. N Engl J Med 2012;366:707–714.

21. Samuels Y, Wang Z, Bardelli A et al. High frequency of mutations of the PIK3CA gene in human cancers. Science 2004;304:554.

www.TheOncologist.com

25. Prickett TD, Agrawal NS,Wei X et al. Analysis of the tyrosine kinome in melanoma reveals recurrent mutations in ERBB4. Nat Genet 2009;41:1127– 1132. 26. Rodr´ıguez-Escudero I, Oliver MD, Andre´ s-Pons A et al. A comprehensive functional analysis of PTEN mutations: Implications in tumor- and autismrelated syndromes. Hum Mol Genet 2011;20:4132– 4142. 27. Dzienis MR, Atkinson VG. Response rate to vemurafenib in patients with B-RAF-positive melanoma brain metastases: A retrospective review. Melanoma Res 2014;24:349–353. 28. Queirolo P, Spagnolo F, Ascierto PA et al. Efficacy and safety of ipilimumab in patients with advanced melanoma and brain metastases. J Neurooncol 2014;118:109–116. 29. Azer MW, Menzies AM, Haydu LE et al. Patterns of response and progression in patients with BRAFmutant melanoma metastatic to the brain who were treated with dabrafenib. Cancer 2014;120:530–536. 30. Mittapalli RK,Vaidhyanathan S, Dudek AZ et al. Mechanisms limiting distribution of the threonineprotein kinase B-RaF(V600E) inhibitor dabrafenib to the brain: Implications for the treatment of melanoma brain metastases. J Pharmacol Exp Ther 2013;344:655–664. 31. Mittapalli RK, Vaidhyanathan S, Sane R et al. Impact of P-glycoprotein (ABCB1) and breast cancer resistance protein (ABCG2) on the brain distribution of a novel BRAF inhibitor: Vemurafenib (PLX4032). J Pharmacol Exp Ther 2012;342:33–40. 32. Durmus S, Sparidans RW, Wagenaar E et al. Oral availability and brain penetration of the B-

©AlphaMed Press 2015

Vemurafenib for Melanoma Brain Metastases

10 RAFV600E inhibitor vemurafenib can be enhanced by the P-GLYCOprotein (ABCB1) and breast cancer resistance protein (ABCG2) inhibitor elacridar. Mol Pharm 2012;9:3236–3245. 33. Eichler AF, Chung E, Kodack DP et al. The biology of brain metastases-translation to new therapies. Nat Rev Clin Oncol 2011;8:344–356. 34. Colombino M, Capone M, Lissia A et al. BRAF/ NRAS mutation frequencies among primary tumors and metastases in patients with melanoma. J Clin Oncol 2012;30:2522–2529. 35. Henson JW, Ulmer S, Harris GJ. Brain tumor imaging in clinical trials. AJNR Am J Neuroradiol 2008;29:419–424. 36. Hodis E, Watson IR, Kryukov GV et al. A landscape of driver mutations in melanoma. Cell 2012;150:251–263. 37. Dankort D, Curley DP, Cartlidge RA et al. Braf (V600E) cooperates with Pten loss to induce metastatic melanoma. Nat Genet 2009;41:544– 552. 38. Nathanson KL, Martin AM, Wubbenhorst B et al. Tumor genetic analyses of patients with

metastatic melanoma treated with the BRAF inhibitor dabrafenib (GSK2118436). Clin Cancer Res 2013;19:4868–4878. 39. Berger MF, Hodis E, Heffernan TP et al. Melanoma genome sequencing reveals frequent PREX2 mutations. Nature 2012;485:502–506. 40. Hamilton R, Krauze M, Romkes M et al. Pathologic and gene expression features of metastatic melanomas to the brain. Cancer 2013;119: 2737–2746. 41. Davies MA, Stemke-Hale K, Lin E et al. Integrated molecular and clinical analysis of AKT activation in metastatic melanoma. Clin Cancer Res 2009;15:7538–7546. 42. Niessner H, Forschner A, Klumpp B et al. Targeting hyperactivation of the AKT survival pathway to overcome therapy resistance of melanoma brain metastases. Cancer Med 2013;2:76–85. 43. Sambade MJ, Peters EC, Thomas NE et al. Melanoma cells show a heterogeneous range of sensitivity to ionizing radiation and are radiosensitized by inhibition of B-RAF with PLX-4032. Radiother Oncol 2011;98:394–399.

44. Dasgupta T, Haas-Kogan DA, Yang X et al. Genotype-dependent cooperation of ionizing radiation with BRAF inhibition in BRAF V600Emutated carcinomas. Invest New Drugs 2013;31: 1136–1141. 45. Anker CJ, Ribas A, Grossmann AH et al. Severe liver and skin toxicity after radiation and vemurafenib in metastatic melanoma. J Clin Oncol 2013; 31:e283–e287. 46. Boussemart L, Boivin C, Claveau J et al. Vemurafenib and radiosensitization. JAMA Dermatol 2013;149:855–857. 47. Su F, Viros A, Milagre C et al. RAS mutations in cutaneous squamous-cell carcinomas in patients treated with BRAF inhibitors. N Engl J Med 2012; 366:207–215. 48. Poulikakos PI, Zhang C, Bollag G et al. RAF inhibitors transactivate RAF dimers and ERK signalling in cells with wild-type BRAF. Nature 2010;464: 427–430. 49. Dent P, Yacoub A, Fisher PB et al. MAPK pathways in radiation responses. Oncogene 2003; 22:5885–5896.

See http://www.TheOncologist.com for supplemental material available online.

OTncologist he

©AlphaMed Press 2015

®