Search

9547 Background: Primary analysis of the phase 3 IMspire150 study (NCT02908672) demonstrated improved investigator-assessed progression-free survival (PFS) with first-line A + V + C vs P + V + C in patients (pts) with BRAFV600 mutation–positive advanced melanoma (hazard ratio [HR] 0.78; 95% CI, 0.63-0.97; P= 0.025). With median follow-up of 18.9 months at primary analysis, OS data were immature; interim analysis of OS at the time of the primary analysis demonstrated a trend in favor of A + V + C over P + V + C (estimated 2-year OS rate, 60.4% vs 53.1%) (Gutzmer et al. Lancet 2020;395:1835-1844). Herein, we present results from the second interim OS analysis of the IMspire150 study. Methods: IMspire150 enrolled previously untreated pts with stage IV or unresectable stage IIIc BRAFV600 mutation-positive melanoma (n = 514). Pts were randomized 1:1 to receive 28-day cycles of A + V + C (n = 256) or P + V + C (n = 258). Pts received V + C in cycle 1; A or P was added on days 1 and 15 from cycle 2 onwards. The second interim OS analysis was planned after ̃270 OS events were recorded, and was projected to have a minimally detectable difference of HR of 0.74 with a P value boundary of 0.0140. OS was estimated using the Kaplan-Meier method. Results: At data cutoff (Sept 8, 2021), 273 OS events had occurred. Median follow-up was 29.1 months (range, 1-54) for A + V + C and 22.8 months (range, 0-54) for P + V + C. A continued trend toward OS benefit in favor of A + V + C over P + V + C was observed with median OS of 39.0 vs 25.8 months, but the difference did not reach statistical significance (HR, 0.84; 95% CI, 0.66-1.06; P= 0.1432). A delayed treatment effect was observed, with landmark OS rates for A + V + C vs P + V + C of 76.1% vs 76.5% at 12 months and 61.5% vs 53.3% at 24 months. With additional follow-up, A + V + C continued to show PFS benefit over P + V + C (HR, 0.79; 95% CI, 0.64-0.97; P= 0.0224); overall response rates (66.7% vs 65.0%) and median duration of response (21.0 vs 12.6 months) remained consistent with those reported at primary analysis. No new safety signals were observed with additional follow-up. Conclusions: With further follow-up, A + V + C demonstrated a consistent, but not statistically significant, improvement in OS and continued benefit in duration of response versus P + V + C in previously untreated pts with BRAFV600 mutation–positive advanced melanoma. Clinical trial information: NCT02908672.

Purpose The treatment of advanced BRAFV600-mutated melanomas with BRAF inhibitors (BRAFi) has improved survival, but the efficacy of BRAFi varies among individuals and the development of acquired resistance to BRAFi through reactivation of mitogen-activated protein kinase (MAPK) signaling is common. We performed an exploratory, retrospective analysis to investigate the effects of BRAFV600 allelic balance, coexisting oncogene mutations, cell proliferation signaling levels, and loss of PTEN expression on progression-free survival (PFS) in patients in the phase III coBRIM study, which compared the combination of the MEK inhibitor cobimetinib with the BRAFi vemurafenib versus vemurafenib as monotherapy. Methods Baseline tumor samples from the intention-to-treat population were analyzed by targeted deep sequencing at a median coverage of 3,600× and by immunohistochemistry for cell proliferation markers, BRAFV600E, and PTEN. The association of these biomarkers with PFS was assessed by Cox proportional hazards modeling. Gene expression in relation to loss of PTEN was profiled by RNA sequencing in 205 patient samples and 42 BRAFV600-mutated melanoma cell lines. Results Neither BRAFV600 allelic balance nor coexisting mutations in the RAS/RAF/RTK pathway affected PFS in either treatment group. Increased baseline MAPK signaling and cell proliferation did not affect PFS in patients treated with cobimetinib combined with vemurafenib. PTEN loss was associated with reduced PFS in patients treated with vemurafenib alone but not in patients treated with cobimetinib combined with vemurafenib. Conclusion Deeper inhibition of the MAPK pathway through targeting of both MEK and BRAF may override the effects of tumor heterogeneity and improve PFS in all patients with advanced BRAFV600 melanoma.

Purpose We performed a retrospective exploratory analysis to evaluate the prognostic and predictive effect of two circulating biomarkers, BRAFV600 mutant circulating tumor DNA (ctDNA) and circulating hepatocyte growth factor (cHGF), in metastatic melanoma. Materials and Methods This study evaluated patients from BRIM-3, a phase III trial comparing vemurafenib and dacarbazine in 675 patients with BRAFV600 mutated advanced melanoma. ctDNA was measured using droplet digital polymerase chain reaction, and cHGF was measured by enzyme-linked immunosorbent assay. Overall survival (OS) was estimated using the Kaplan-Meier method, and hazard ratios (HRs) were estimated using Cox proportional hazards modeling. Partitioning analysis was used to group patients into risk categories. Results Patients with elevated levels of baseline BRAFV600 ctDNA had significantly shorter median OS than those with undetectable levels of ctDNA (vemurafenib arm, 9.9 v 21.4 months, respectively, and dacarbazine arm: 6.1 v 21.0 months, respectively). Median OS was also shorter in patients with high levels of cHGF compared with those with low cHGF (vemurafenib arm, 11.9 v 17.3 months, respectively, and dacarbazine arm, 6.1 v 14.4 months, respectively). In a multivariable proportional hazards model with adjustment for lactate dehydrogenase, Eastern Cooperative Oncology Group status, disease stage, and treatment, ctDNA and cHGF were both independent prognostic factors for OS, (HR, 1.75; 95% CI, 1.35 to 2.28 for high v undetectable ctDNA; HR, 1.24; 95% CI, 1.00 to 1.53 for high v low cHGF). Using partitioning analysis, we found that patients with elevated ctDNA combined with elevated cHGF constituted the highest risk group with significantly shorter OS. Conclusion Here, we report that BRIM-3 patients with high levels of ctDNA and cHGF have worse OS regardless of treatment and that these factors are independent prognostic markers for metastatic melanoma.

9508 Background: Preliminary data from the ABC (76 pts) and CheckMate 204 (94 pts) trials showed that nivo and nivo+ipi have activity in active melanoma brain metastases, with durable responses in a subset of pts. Here, we report updated 5-yr data from all pts enrolled on the ABC trial (NCT02374242). Methods: This open-label ph2 trial enrolled 3 cohorts of pts with active melanoma brain mets naïve to anti-PD1/PDL1/PDL2/CTLA4 from Nov 2014-Apr 2017. Pts with asymptomatic brain mets with no prior local brain therapy were randomised to cohort A (nivo 1mg/kg + ipi 3mg/kg, Q3Wx4, then nivo 3mg/kg Q2W) or cohort B (nivo 3mg/kg Q2W). Cohort C (nivo 3mg/kg Q2W) had brain mets i) that failed local therapy, ii) with neuro symptoms and/or iii) with leptomeningeal disease. Prior BRAF inhibitor (BRAFi) was allowed. The primary endpoint was best intracranial response (ICR) ≥wk12. Key secondary endpoints were IC PFS, overall PFS, OS, & safety. Results: A total of 76 pts (med f/u 54 mo) were enrolled; median age 59y, 78% male. For cohorts A, B and C: elevated LDH 51%, 58% and 19%; V600BRAF 54%, 56% and 81%; prior BRAFi 23%, 24%, 75%. Efficacy and toxicity are shown in the table. There were no treatment-related deaths. 1/17 deaths in cohort A & 4/16 in cohort B were due to IC progression only. Conclusions: Nivo monotherapy and ipi+nivo are active in melanoma brain mets, with durable responses in the majority of patients who received ipi+nivo upfront. A study of upfront ipi+nivo+/-SRS is underway (NCT03340129). Clinical trial information: NCT02374242. [Table: see text]

9523 Background: The phase 3 IMspire150 study showed that first-line A+V+C improved investigator-assessed PFS vs placebo (P)+V+C in BRAFV600E/K mutation–positive advanced melanoma (hazard ratio 0.78; P=.0249). Prior biomarker analyses showed that IFN-g or TMB > 10 mut/Mb were associated with greater PFS benefits with A+V+C (Lewis et al. J ImmunoTher Cancer 2020;8:A188-A189). We further evaluated the association of these biomarkers with outcomes. Methods: Exploratory recursive partitioning analysis (RPA) was used to model associations between PFS and age ( < 65 vs ≥65 y), Eastern Cooperative Oncology Group performance status (0 vs 1), liver metastases (yes vs no), metastatic sites (≤3 vs > 3), sum of longest tumor diameters ( < 44 mm vs ≥44 mm), baseline LDH (normal [n] vs elevated [e]), TMB ( < 10 vs ≥10 mut/Mb), PD-L1 (negative vs positive), and IFN-g (high [h; > Quartile 3; Q3] vs intermediate [ > Q1 and ≤Q3] vs low [≤Q1]). Time-to-event analyses were summarized using Kaplan-Meier estimates. Results: The RPA analysis included 208/256 (81.3%) patients (pts) from the A+V+C arm of IMspire150 for whom LDH, TMB, IFN-g, and PD-L1 data were available. RPA showed that LDH was associated with PFS. In pts treated with A+V+C and n-LDH, h-IFN-g signature was associated with longer PFS and higher rates of objective response (OR) and complete response (CR) vs low/intermediate (l/i) IFN-g (2-y PFS: 59% vs 38%; ORR: 77% vs 69%; CR: 38% vs 15%, respectively); TMB ≥10 mut/Mb was associated with more favorable outcomes in pts with e-LDH (Table). In contrast, neither IFN-g nor TMB discriminated PFS outcomes in n-LDH or e-LDH pt subgroups receiving P+V+C. Pts with e-LDH and TMB < 10 mut/Mb had poor PFS outcomes, with 2-y PFS rates of 9% and 3% and lower rates of OR (51% and 62%) and CR (5% and 9%) in the A+V+C and P+V+C arms, respectively. Similar trends were observed for duration of response (DOR), and for the subset of pts with BRAFV600E mutation–positive melanoma. A+V+C improved PFS vs P+V+C across all subgroups with the exception of e-LDH and TMB < 10. Conclusions: IFN-g and TMB discriminated PFS benefit in pts receiving A+V+C but not for those receiving P+V+C. Durable responses were observed for pts treated with A+V+C in the n-LDH + h-IFNg subgroups.[Table: see text]

10073 Background: In IMspire150, A+C+V significantly improved investigator-assessed progression-free survival vs placebo (Pbo)+C+V in previously untreated pts with unresectable stage IIIc/IV BRAFV600+melanoma. PRO data from this trial are now reported. Methods: 514pts were randomized 1:1 to 28-day cycles of A+C+V or Pbo+C+V. Pts received C+V in cycle 1; A or Pbo was added on days 1 and 15 from cycle 2 onward. Pts completed the EORTC QLQ-C30 questionnaire on day 1 of cycle 1 (baseline), days 1 and 15 of cycle 2 and cycle 3, day 1 from cycle 4 onward, within 28-d of treatment discontinuation, and ≤6 months post-treatment. Prespecified secondary PRO endpoints were time to confirmed deterioration (TTCD; defined as time from randomization to first ≥10-point decrease from baseline held for 2 consecutive assessments, or 1 assessment followed by death on treatment) in quality of life (QoL) and physical functioning (PF). Prespecified exploratory endpoints were TTCD in role functioning (RF); mean change from baseline and percentage of pts with a clinically meaningful change (≥10 points from baseline) in QoL, PF, and RF. Results: Questionnaire completion rates were > 80% for most of the treatment period. At baseline, mean QoL, PF, and RF scores were moderate to high (Table). At week 6 (cycle 2), following initiation of A, QoL, PF, and RF scores declined, but returned to near-baseline levels at week 10 (cycle 3) and were largely maintained until week 36 (cycle 10), when < 50% of pts contributed data. In this time, ≥56% pts in both arms did not experience a clinically meaningful deterioration in QoL, PF, and RF. TTCD on QoL (hazard ratio [HR] 1.23; 95% CI 0.90-1.67), PF (HR 1.27; 95% CI 0.93-1.74), and RF (HR 1.15; 95% CI 0.86-1.55) favored Pbo, but only one-third of pts overall experienced a confirmed deterioration. Conclusions: PRO data showed that A+C+V did not worsen QoL, PF, and RF in pts with advanced BRAFV600+melanoma, thus supporting its use as a treatment option. Clinical trial information: NCT02908672. [Table: see text]

e19289 Background: Health economic studies of anti-cancer systemic treatments typically make strong assumptions regarding the number of drug lines received after initial therapy. This may have a substantial effect on health economic outcomes and impact drug reimbursement and uptake in practice. This study aims to quantify the real-world systemic treatment patterns in 4 metastatic cancers using clinical registries to explore whether health economic assumptions are justifiable. Methods: Data from 4,431 metastatic cancer patients were extracted from Australian clinical registries: colorectal (COL, n = 3087), non-small cell lung (LUN, n = 705), pancreatic (PAN, n = 459), and melanoma (MEL, n = 180). A set of criteria was defined to consistently identify drug lines across the cancer-specific registries. These criteria were based on the type of drug (biological agent or chemotherapy), switches in chemotherapy regimen (whether a chemotherapy agent was added or removed) and the timing of such changes. Consequently, the identified drug lines provide a more detailed view on treatment patterns compared to clinical treatment lines (defined by disease progression), the latter typically including multiple drug lines. Censoring in the number of drug lines received due to ongoing treatment was accounted for in all analyses. Results: Most patients started treatment after diagnosis: 77% (COL), 89% (LUN), 56% (PAN) and 79% (MEL). For COL, LUN, PAN and MEL respectively, the proportion of patients starting a 2nd drug line was 51%, 60%, 19% and 24%, whereas 28%, 35%, 6% and 8% of patients started a 3rd drug line. In all cancers, patients were most likely to receive only a single drug line, i.e. the mode number of drug lines was 1. For all patients, the median number of drug lines were COL: 2, LUN: 2, PAN: 1, MEL: 1. For treated patients only, the median (range) number of drug lines received was COL: 2 (1-9), LUN: 2 (1-8), PAN: 1 (1-5), MEL: 1 (1-5). Conclusions: We show that patients are most likely to receive a single drug line. Considering our findings, most health economic analyses likely overestimate the intensity of drug treatment in metastatic disease, thereby underestimating the impact of initial treatment relative to downstream treatments. This is likely to bias estimates of total treatment cost, cost-effectiveness and budget impact, which will hamper the uptake of novel anti-cancer agents and may lead to suboptimal decisions regarding treatment strategies.

2033 Background: Despite increasing evidence of benefit supporting CGP in personalizing cancer therapy, its widespread uptake remains limited. Barriers include low patient understanding, unmet patient expectations related to low utility, clinician concerns over cost-effectiveness, perceived value, and discomfort in management of complex genomic results. Methods: This prospective cross-institutional demonstration study was designed to evaluate implementation of CGP in the care of adult and paediatric advanced cancer pts, incorporating pt reported outcomes (PROMs), discrete choice experiment (DCE), ongoing process optimization and clinician evaluations. DNA sequencing of FFPE tumor and matched blood was completed with CGP (PMCC Comprehensive Cancer Panel; 391 genes) via central laboratory. A tumor board reported results weekly with emphasis on therapeutic relevance. Oncologists performed consent and results delivery. Pts completed pre-and post-test surveys, including validated and study-specific questions, DCE and if eligible, semi-structured interviews. Qualitative interviews were undertaken with study clinicians and laboratory staff to evaluate processes. Results: 86% (315) of 365 enrolled pts had successful CGP; of these 63% (199) had relevant therapeutic, diagnostic or germline results. 50 (16%) had treatment change at 6m, 49 (16%) had germline mutations. 293 (88% of adult pts) completed PROMs. 17 of 19 clinicians/laboratory staff approached consented to an interview. At consent pts cited multifaceted value in testing, showed good understanding of basic concepts, but most (69%) overestimated the likelihood of result-led change. Post-test pts remained consistently satisfied with accessing CGP; valuing research contribution, taking opportunities and information for family. 21% struggled with understanding results but there were low levels of decisional regret following participation (89% had nil/mild regret). Pt-elicited preferences (via DCE) indicated priority for high rates of clinical utility and timeliness. Clinicians sited collaboration and communication as critical to delivery of CGP. Conclusions: Pts undergoing CGP are generally satisfied, and derive value on its use beyond potential therapeutic benefit. Our results suggest that to improve test utility and delivery of CGP with value to pts and investing institution, focus must be placed on addressing the additional barriers to its wider implications including efforts to improve process efficiencies, clinician genomic literacy and decision-making support.

9583 Background: The efficacy of ICIs in metastatic melanoma is well-established. However, there is limited data regarding their efficacy in in-transit melanoma metastases (ITM). This study assessed the efficacy of ICI in patients with ITM. Methods: A multisite, retrospective review of patients with ITM treated with ICI from 2004-2018. Demographic and clinicopathological factors (age, sex, primary site, AJCC version 8 stage, BRAF status, prior locoregional therapies) were collected. Objective response rate (ORR) based on a clinician-assessed best overall response, progression free survival (PFS) and overall survival (OS) were analyzed by the Kaplan-Meier method. Results: Fifty-four patients were included: 27 (50%) female; median age 69 (range 19-89); 12 (22%) stage IIIB, 40 (74%) stage IIIC and 2 (4%) stage IIID; 10(19%) BRAF mutant. Forty (74%) received single agent PD-1 inhibitor (pembrolizumab or nivolumab), 8 (15%) single agent anti-CTLA-4 (ipilimumab), 5 (9%) combination anti-PD-1/anti-CTLA-4 (ipilimumab and nivolumab or pembrolizumab) and 1 (2%) combination anti-PDL-1/MEK inhibitor (atezolizumab and cobimetinib). ORR to ICI was 54%: 14 (26%) complete responses; 15 (28%) partial responses; 9 (17%) stable disease; 16 (30%) progressive disease. Thirteen (46%) responders had only one ITM lesion. ORR was 58% for single agent anti PD-1, 38% for single agent anti-CTLA4, and 40% for anti-PD-1/anti-CTLA-4 (Table). The median follow-up was 15 months (2-46). The median PFS was 11.7 months (6.6-N/A). PFS at 1 and 2 years were 48% and 39%. Fourteen (56%) progressed locoregionally and 11 (44%) progressed distantly. OS at 1 and 2 years were 85% and 63%; the median OS was not reached. No clinicopathological features were associated with ORR. Conclusions: ICI produces objective responses in ITM and should be considered in patients with unresectable ITM or disease recurrence despite locoregional therapies. [Table: see text]

9540 Background: mMCC is a rare, highly aggressive neuroendocrine cancer with a poor prognosis. ICIs have favourable efficacy and safety in clinical trials. We outline single centre experience utilising ICIs in mMCC. Methods: Medical records of patients (pts) with mMCC treated with ICIs from Aug 2015 to Dec 2018 at Peter MacCallum Cancer Centre in Australia were retrospectively analysed. RNA sequencing and immunohistochemistry for PD-L1, CD3 and Merkel cell polyomavirus (MCPyV) on tumor samples were performed. Baseline tumor volumes and responses were assessed with FDG-PET scans using the Hicks criteria. Results: 23 pts with mMCC were treated with ICIs. Pt characteristics are summarised in Table. A median of 8 cycles (range 1 to 47) were administered, with treatment ongoing in 7 pts. Objective responses (OR) were observed in 14 pts (61%); 10 (44%) complete metabolic responses (CMR) and 4 (17%) partial metabolic responses (PMR). Median time to response was 9 weeks (range 4 to 11) and 12-month progression-free survival (PFS) rate was 32%. Increased OR were seen in pts aged less than 75 (OR 8/10, 80% vs 46%), no prior history of chemotherapy (OR 10/14, 71% vs 44%), pts with an immune-related adverse event (irAE) (OR 6/6, 100% vs 47%) and in MCPyV negative pts (OR 9/11, 82% vs 50%). Pts with a CMR had lower mean-tumor volume on baseline FDG-PET scan (CMR: 35.7mL, no CMR: 187.8mL, p value 0.05). 10 pts received radiation (RT) during ICI: 4 pts started RT concurrently (OR 75%, CMR 50%), 3 pts had isolated ICI-resistant lesions successfully treated with RT and 3 pts with multisite progression continued to progress despite RT. 6 pts (26%) had a Grade 1-2 irAE. Conclusions: ICIs showed efficacy and safety consistent with trial data. Younger age, negative MCPyV status, no prior chemotherapy, lower baseline FDG-PET tumor volume and irAEs are potentially associated with better responses. [Table: see text]

TPS9600 Background: Nivolumab combined with ipilimumab is active in melanoma brain metastases, with intracranial response rates > 55% and durable survival in treatment naïve patients (pts) (Long GV et al Lancet Onc 2018; Tawbi H et al NEJM 2018). We seek to determine if the addition of stereotactic radiotherapy (SRS) results in improved intracranial outcomes. Methods: This is a multisite, open-label, phase 2 trial in systemic treatment-naïve pts with melanoma brain metastases. Pts must have ≥1 asymptomatic brain metastases that are ≥5mm and ≤40mm as per modified RECIST 1.1, on gadolinium-enhanced MRI, and no history of previous treatment with SRS. Eligible pts are randomly assigned to either receive nivolumab plus ipilimumab with SRS or nivolumab plus ipilimumab alone. Nivolumab (1mg/kg) and ipilimumab (3mg/kg) are given every 3 weeks for 4 doses. Following induction, 480mg nivolumab is given every 4 weeks until progression, unacceptable toxicity, or a maximum of 2 years. SRS is administered as single fraction of 16-22Gy, or hypofractionated for larger lesions (24-27Gy in 3 fractions), within 7 days of immunotherapy commencement. Pts will be evaluated for intracranial and extracranial tumour response, and overall response, every 6 weeks to week 24 and 12 weekly thereafter until overall disease progression or death. The primary endpoint is neurologic specific survival (NSS) at 12 months. Secondary endpoints include intracranial response rate, intracranial PFS, overall PFS, overall progression free survival, overall survival, neurocognitive function and incidence of radiation necrosis. 109 patients in each cohort (218 total) will achieve > 80% power at the significance level (alpha) of 0.10 to detect a minimum absolute increase of 9% in the NSS rate at 12 months. Clinical trial information: NCT03340129.

Purpose To assess pharmacodynamic effects and intrinsic and acquired resistance mechanisms of the BRAF inhibitor vemurafenib in BRAFV600-mutant melanoma, leading to an understanding of the mechanism of action of vemurafenib and ultimately to optimization of metastatic melanoma therapy. Methods In the phase II clinical study NP22657 (BRIM-2), patients received oral doses of vemurafenib (960 mg twice per day). Serial biopsies were collected to study changes in mitogen-activated protein kinase (MAPK) signaling, cell-cycle progression, and factors causing intrinsic or acquired resistance by immunohistochemistry, DNA sequencing, or somatic mutation profiling. Results Vemurafenib inhibited MAPK signaling and cell-cycle progression. An association between the decrease in extracellular signal-related kinase (ERK) phosphorylation and objective response was observed in paired biopsies (n = 22; P = .013). Low expression of phosphatase and tensin homolog showed a modest association with lower response. Baseline mutations in MEK1P124 coexisting with BRAFV600 were noted in seven of 92 samples; their presence did not preclude objective tumor responses. Acquired resistance to vemurafenib associated with reactivation of MAPK signaling as observed by elevated ERK1/2 phosphorylation levels in progressive lesions and the appearance of secondary NRASQ61 mutations or MEK1Q56P or MEK1E203K mutations. These two activating MEK1 mutations had not previously been observed in vivo in biopsies of progressive melanoma tumors. Conclusion Vemurafenib inhibits tumor proliferation and oncogenic BRAF signaling through the MAPK pathway. Acquired resistance results primarily from MAPK reactivation driven by the appearance of secondary mutations in NRAS and MEK1 in subsets of patients. The data suggest that inhibition downstream of BRAF should help to overcome acquired resistance.

Purpose Imaging with [18F]fluorodeoxyglucose (FDG) –positron emission tomography (PET) allows early recognition of a response to agents that target key driver mutations in human cancer. We aimed to determine the metabolic response rate to vemurafenib in patients with advanced BRAF-mutant melanoma. Patients and Methods Baseline and day 15 FDG-PET was evaluated in 31 patients with advanced melanoma treated in a phase I study of dose escalation of vemurafenib (PLX06-02), which included four patients treated at subtherapeutic doses and 24 patients treated at 960 mg twice a day, which is the maximum-tolerated dose of vemurafenib. Results All 27 patients treated at potentially therapeutic levels had at least a partial metabolic response, and three patients achieved a complete metabolic response. In the 27 patients, there was an 80% ± 3% reduction in the maximum standardized uptake value (SUVmax) of target lesions and an 87% ± 3% decrease in the percentage of injected dose (%ID) in all identified disease sites. There was a positive correlation between %ID in all identified disease and target-lesion SUVmax (r2 = 0.66; P < .001) that indicated a significant homogeneity of the response between lesions in individual patients. Although no relationship was found between the reduction in target lesion SUVmax and best response according to RECIST (Response Evaluation Criteria in Solid Tumors), there was a trend for patients with greater reductions in uptake of FDG to have longer progression-free survival. Conclusion FDG-PET is a useful marker of an early biologic response to vemurafenib. Little variability in PET response was found between lesions in individual patients, which suggested minimal intrapatient molecular heterogeneity. FDG-PET is a useful tool for the evaluation of the biologic impact of inhibiting mutant BRAF and may allow for the more effective development of novel agents.

Background: The Master of Cancer Sciences brings together the academic strength of the University of Melbourne (UoM) and world leading content experts from the Victorian Comprehensive Cancer Centre (VCCC), to develop an innovative, interactive, evidence-based flagship educational program in cancer sciences. This program will enhance the capabilities of health professionals, in the rapidly evolving field of cancer research and clinical care. It will be wholly online and nested with qualification points at Specialist Certificate, Graduate Certificate and Masters Level to provide flexible progression and study options for practitioners in the cancer care workforce. Specific modular content from each subject will also be repurposed and repackaged as a series of derivative educational activities such as massive open online courses (MOOCs), workshops and webinars; broadening the reach of the masters programming to all. Aim: The aim of this program is to be the first cancer-specific, multidisciplinary, flexible, and wholly online master's program of its kind offered within Australia, and one of two available worldwide. Graduates will possess an unprecedented breadth of integrated cancer knowledge and skills. This initiative will unify the VCCC alliance partners as they contribute to directly supporting a world-class cancer workforce and provide best practice care to patients. Discussion: This presentation will explore the educational development process involved in developing the ten online subjects, through the engagement of content writing teams to collaborative development and repurposing modules into a series of derivative educational activities. Ultimately the graduate programs and their derivative educational activities will contribute to lifelong learning, ongoing professional development and high quality healthcare for better patient outcomes.

9522Background: The coBRIM study demonstrated that first-line C+V improved progression-free (PFS) and overall survival (OS) compared with placebo (P)+V in pts with BRAFV600-mutated advanced melanom...

Purpose To evaluate tumor response, pharmacodynamic effects, and safety of a combination of lomeguatrib (LM), an O6-methylguanine DNA-methyltransferase (MGMT) inactivator, and temozolomide (TMZ), TMZ alone, and LM/TMZ after disease progression on TMZ alone in patients with advanced melanoma. Patients and Methods Patients with unresectable stage III or IV cutaneous melanoma who had no prior systemic chemotherapy were randomly assigned to receive either 40 to 80 mg LM and 125 mg/m2 TMZ or 200 mg/m2 TMZ on days 1 through 5 of each 28-day treatment cycle. Drugs were administered orally for up to six cycles of treatment. Patients on TMZ alone were offered LM/TMZ at progression, if fit enough to receive treatment. Results One hundred four patients were enrolled, with 52 in each trial arm. Twenty-seven TMZ-treated patients received LM/TMZ after progression on TMZ. Unexpectedly, analysis of tumor biopsies showed rapid recovery of MGMT after LM/TMZ with 40 mg/d LM. Therefore, doses of LM were escalated to 60 then 80 mg/d. Tumor response rates were 13.5% with LM/TMZ and 17.3% with TMZ alone. No patient responded to LM/TMZ having progressed through TMZ. Median time to disease progression was 65.5 days for LM/TMZ and 68 days for TMZ. All treatments were well tolerated, although hematologic and gastrointestinal adverse events were common. A higher incidence of hematological adverse events was observed in the LM/TMZ combination arm. Conclusion The efficacy of LM and TMZ in the current dosing schedule is similar to that of TMZ alone. To maintain MGMT depletion in tumor dosing of LM needs to be continued beyond that of TMZ.

9512 Background: Most patients treated with BRAF inhibitors (BRAFi) +/- MEK inhibitors (MEKi) eventually progress on treatment. Along with genetic acquired resistance, epigenetic mechanisms that could be reversed after BRAFi discontinuation have been described. The purpose of this study was to analyse outcomes for patients (pts) retreated with BRAF-directed therapy. Methods: 116 pts who received BRAFi based therapy and, after a break, were re-challenged with BRAFi +/- MEKi treated at 14 centres in Europe, US, and Australia were analysed for progression free survival (PFS) and response rate (RR), as well as factors predicting overall survival (OS) (demographics, disease stage, treatment, LDH level, duration of first BRAFi treatment, reason for first BRAFi discontinuation and interval between BRAFi stop and re-challenge). Multivariate Cox regression, regression trees and Kaplan Meier method were used. Results: Median duration of 1st BRAFi +/- MEKi treatment was 9.4 months (mts) and 7.7 mts for the subsequent treatment after discontinuation (immunotherapy 72%, other 17 %, drug holiday 11%). Brain metastases were present in 51 pts (44%). RR to re-challenge with BRAFi +/- MEKi was 43%: complete response (CR) 3%, partial response (PR) 39%, stable disease 24% and progressive disease (PD) 30%, 4% missing. Of 80 pts who previously discontinued BRAFi for PD, 31 (39%) responded (30 PR and 1 CR). Median OS from retreatment was 9.8 mts. Independent prognostic factors for survival at re-challenge included number of metastatic sites (HR = 1.32 for each additional organ with metastases, p < .001), LDH (HR = 1.37 for each multiple of the upper normal limit, p < .001), while combination of BRAFi+MEKi conferred a better prognosis vs BRAFi alone (HR = 0.5, p = .006). Pts with < 3 metastatic sites treated with BRAFi and MEKi had a better survival (median OS not reached) and pts with ≥ 3 metastatic sites and raised LDH treated with BRAFi alone had the worse outcome (median OS 4 mts). Number of metastatic sites (HR = 1.44, p < .001) and combination of BRAFi and MEKi (HR = 0.45, p < .001) were independent prognostic factors for PFS (median 5.0 mts). Conclusions: Re-challenge with BRAFi +/- MEKi induces a clinically significant response and should be considered for selected cases.

9508 Background: Nivolumab (nivo) and the combination of nivo + ipilimumab (ipi) improve response rates (RR) and progression-free survival (PFS) compared with ipi alone in clinical trials of metastatic melanoma pts, but pts with untreated brain mets were excluded. Brain mets are a major cause of morbidity and mortality in melanoma and their management is critical. We sought to determine the antitumour activity and safety of nivo and nivo+ipi in pts with active melanoma brain mets (NCT02374242). Methods: This open-label, ph II trial enrolled 3 cohorts of pts naïve to anti-PD1/PDL1/PDL2/CTLA4 from Nov 2014 - Feb 2017. Pts with asymptomatic melanoma brain mets with no prior local brain therapy were randomised to cohort A (nivo 1mg/kg + ipi 3mg/kg, Q3W x4, then nivo 3mg/kg Q2W) or cohort B (nivo 3mg/kg Q2W). Cohort C (nivo 3mg/kg Q2W) had brain mets 1) that failed local therapy (new +/- progressed in previously treated met), 2) were neurologically symptomatic and/or 3) with leptomeningeal disease. Prior BRAF inhibitor (BRAFi) was allowed. The primary endpoint was best intracranial response (ICR) ≥ wk12. Secondary endpoints were best extracranial response (ECR), best overall response (OR), IC PFS, EC PFS, overall PFS, OS, and safety. Results: A total of 66 pts (med f/u 14 mo) were included in this analysis of total 76 planned; median age 60y, 77% male. For cohorts A, B and C: elevated LDH 48%, 58% and 19%; V600BRAF 44%, 56% and 81%; prior BRAFi 24%, 24%, 75%. Table shows RR, PFS and OS. ICR in cohort A treatment naïve vs prior BRAFi was 53% vs 16%. Treatment-related gd 3/4 toxicity in cohorts A, B and C were 68%, 40% and 56%, respectively. There were no treatment-related deaths. Conclusions:Nivo monotherapy and ipi+nivo and are active in melanoma brain mets. Ipi+nivo had reduced activity in pts who progressed on BRAFi. Pts with symptomatic brain mets, leptomeningeal mets or previous local therapy responded poorly to nivo alone. Clinical trial information: NCT02374242. [Table: see text]

2577 Background: CC-90003 is an irreversible inhibitor of ERK 1/2 with potent anti-proliferative activity in KRAS and BRAF mutant tumor models. We conducted a first-in-human study of CC-90003 in patients with RAS or BRAF mutant tumors. Methods: Patients received escalating doses of oral CC-90003 on a 21/28 day cycle. Standard safety (adverse events, chemistry/hematology, physical findings, ECGs and cardiac ECHO/MUGA scans) and PK parameters were assessed. Response was assessed per RECIST 1.1. A proprietary ELISA-based assay measured ERK levels unbound to CC-90003 in peripheral blood mononuclear cells. Results: Nineteen patients (median age: 60 yrs) harboring KRAS (n = 15), NRAS (n = 1), or BRAF (n = 3) mutant tumors received CC-90003 doses from 20 to 160 mg /day. The MTD was 120 mg based on the occurrence of Grade 3 transaminase elevations (n = 2) and hypertension (n = 1) observed at 160 mg (the NTD). Patients completed a median of 2 cycles (range: 1 to 5). AEs (mostly Grade 1 or 2) reported in ≥ 3 patients included constitutional (asthenia, fatigue), gastrointestinal (anorexia, nausea/vomiting, diarrhea), hepatic (transaminase elevations) and neurologic (dizziness, gait disturbance, paresthesias) toxicities. Grade 1-3 neurotoxicity was observed primarily at doses from 80 to 160 mg/day and resolved with dose reduction/interruption. PK parameters were highly variable, with AUC and Cmax increasing overall, with increasing dose. CC-90003 accumulation was observed after multiple doses. There were no objective responses. Levels of free ERK were reduced by ≥80% compared to baseline by C1D8 at doses ≥ 80 mg/day. Conclusions: ERK inhibition may be an attractive target for the management of mutant RAS or BRAF-driven tumors, however proof-of-concept demonstration for CC-90003 was limited by a lack of objective responses, an unfavorable PK profile and unanticipated neurotoxicity. Clinical trial information: NCT02313012.

9563 Background: With the evolving treatment landscape in metastatic melanoma, approaches to disease surveillance post resection in stage 3 disease requires reconsideration. We previously reported the outcomes of sub-stage-specific schedules of combined fluorodeoxyglucose-positron emission and computerized tomography (PET/CT) surveillance at high risk of relapse following surgery. The aim of this study was to provide an update on our surveillance protocol with an extended sample size and longer duration of follow-up. Methods: From 2009, patients with AJCC stage 3 melanoma underwent PET/CT scans according to pre-specified schedules based on Bayesian probabilities of sub-stage-specific relapse. Schedules were stage 3A: 6, 18 mo; stage 3B: 6, 12, 18, 24, 36, 48, 60 months; stage 3C: 6, 12, 18, 24, 36 months. Contingency tables were used to evaluate the sensitivity, specificity and predictive values of these schedules. Results: In total, 171 patients (3A: 34; 3B: 93; 3C:44) underwent 553 PET/CT scans with a median follow up of 47 months. Relapses were identified in 65 (38%) patients of which (72%) were asymptomatic at the time of radiologically documented relapse. False positive results occurred in 8%. The positive predictive value (PPV) of an individual scan for diagnosing true recurrence was 77% (64-87%). Negative scans at 6 months had negative predictive values (NPV) between 57% in Stage 3A to 69% in Stage 3B for relapse. Sensitivity and specificity of the overall approach of sub-stage-specific PET/CT surveillance for detecting disease relapse were 70% and 89%, respectively. Evaluable predictive values for detecting disease relapse were: stage 3A: PPV:56%, NPV:76%; 3B: PPV 83%, NPV 86%; stage 3C: PPV 84%, NPV 84%. 32 of 65 patients (49%; 3A: 1; 3B: 7; 3C: 2) underwent resection of relapsed disease and 10 of these patients remained free of disease with a median follow-up of 24 months. Conclusions: Sub-stage-specific PET/CT is effective in detecting asymptomatic recurrence in stage 3 melanoma, and is associated with a high rate of disease resection at relapse.

9519 Background: Simultaneous inhibition of MEK and CDK4/6 may suppress MAPK pathway activation and cell-cycle checkpoint dysregulation in NRAS-mutant melanoma, resulting in enhanced antitumor activity. Phase 1b data are reported. Methods: The phase 1b primary objective was to determine maximum tolerated dose (MTD)/RP2D. A 28-d cycle of oral ribociclib (RIBO) once daily (QD) for 21 d + oral binimetinib (BINI) twice daily (BID) for 28 d, and a 21-d cycle of RIBO QD + BINI BID, both for 14 d per cycle, were evaluated. Secondary objectives were to evaluate efficacy, safety and pharmacodynamics. Results: Based on dose escalation (van Herpen, ESMO 2015), MTD was 600mg RIBO/45mg BINI for the 21-d and 200/45 for the 28-d regimens. Due to promising activity, the 28-d cycle was selected as RP2D(unconfirmed partial response [PR] with limited follow-up occurred in 35% of pts). This finding was supported by comparable and manageable safety and the Bayesian logistic regression model.As of Jan 2017, the RP2D was received by 16 pts in phase 1b (ECOG PS 0/1/2, 63%/31%/6%; elevated lactate dehydrogenase, 44%; stage IVM1c disease, 50%; prior ipilimumab [ipi], 44%; prior anti–programmed death [PD]-1/PD-L1, 31%). Median (range) exposure was 4 (0–13) mo. Common adverse events (AEs) were increased blood creatine phosphokinase, elevated AST, peripheral edema, acneiform dermatitis, diarrhea and fatigue. Common grade 3/4 AEs were elevated AST and ALT (19%/6%), nausea (19%/0%), rash (19%/0%), vomiting (6%/6%) and neutropenia (12%/0%). Confirmed PR (cPR) occurred in 4 pts (25%; time to response, 48–168 d), stable disease in 7 pts (44%), disease progression in 3 pts (19%); 2 pts (12%) were not evaluable. Among cPR pts, 3 had prior ipi and/or anti–PD-1/PD-L1. Median progression-free survival (mPFS) was 6.7 (95% CI, 3.5–9.2) mo. Sequence analysis of synchronous non- RAS genetic alterations will be presented. Conclusions: Combined RIBO/BINI at the selected RP2D had a manageable safety profile and favorable efficacy (based on mPFS) for NRAS-mutant melanoma in phase 1b. Based on these promising data, the phase 2 expansion is underway to assess antitumor activity at the RP2D. Clinical trial information: NCT01781572.

3557 Background: Pts with BRAFm mCRC have an exceedingly poor prognosis. Unlike melanoma, BRAF inhibitor monotherapy has limited activity in BRAFm mCRC. Preclinically, BRAF inhibition results in rapid feedback activation of EGFR and ongoing tumor proliferation, which can be readily overcome by combining BRAF and EGFR inhibition. EViCT examined the safety and efficacy of combining two oral agents targeting BRAF with Vemurafenib (Vem) and EGFR with Erlotinib (Erl), in BRAFm mCRC pts. Herein, we report safety and preliminary efficacy data. Methods: EViCT had 2 parts: a Phase I dose escalation of Erl (cohort 1: 100mg qd; cohort 2: 150mg qd) together with Vem 960mg bd, to determine the maximum tolerated dose (MTD). The Phase II component involved dose expansion at MTD using a Simon 2-stage design to treat 9 pts in stage 1 and 15 pts in stage 2. Cycles were 28 days. Eligible pts had ECOG < 1, < 2 lines of systemic therapy for metastatic disease, and acceptable organ function. Staging CT scans were performed every 2 cycles, response assessed using RECIST 1.1. Primary endoint was clinical benefit rate (CR, PR and SD). A number of pharmacodynamics correlates were assessed including serial ctDNA, FDG-PET and optional tumour biopsies. Pts were treated until disease progression or toxicity requiring discontinuation. Results: Between Jul-2014 and Oct-2016, 30 BRAFm mCRC pts were enrolled. The Phase I Lead-in enrolled 4 pts in cohort 1 and 7 pts in cohort 2. There was 1 DLT (grade 3 hand-foot syndrome) in cohort 2. MTD/Recommended Phase 2 Dose was Erl 150mg qd and Vem 960mg bd, the full dose for each agent. The Phase II expansion enrolled 19 pts. Overall, 23 pts are evaluable for this interim analysis. Median age was 61 years, 11 (48%) pts were male. Most pts had 1(50%) or 2 (41%) lines of prior treatment. Overall response rate was 39% (95%CI = [20%, 61%]), including 5 (22%) confirmed PR, 4 (17%) unconfirmed PR, 3 (13%) stable disease and 11 (48%) progressive disease. Conclusions: Vem and Erl can both be given safely at their individual full doses when used in combination. In BRAFm mCRC, this combination resulted in clear clinical activity. Clinical trial information: ACTRN12614000486628.

9505Background: In CheckMate 067, NIVO (anti-PD-1) plus IPI (anti-CTLA-4) significantly improved progression-free survival (PFS) and objective response rate (ORR) vs IPI alone in pts with MEL (N En...

9536Background: V therapy has shown improved objective response rate (ORR), progression-free survival (PFS), and overall survival (OS), compared with dacarbazine (D) (BRIM-3), in pts with BRAFV600-...

9533Background: coBRIM (Larkin J et al. N Engl J Med 2014;371:1867-1876.) showed significant improvement in PFS and OS with C+V treatment for advanced BRAF-mutated melanoma patients (pts) but was a...

101Background: OX40 is a co-stimulatory receptor that is transiently expressed by T cells upon antigen recognition. In murine models, OX40 engagement by an agonist anti-OX40 antibody can promote durable tumor regression associated with co-stimulation of effector T cells and reduction of regulatory T cells. This dual mechanism of action is predicted to complement the activity of PD-L1 blockade. MOXR0916 is a humanized effector-competent agonist IgG1 monoclonal antibody (mAb) that targets OX40 and atezolizumab is an engineered humanized IgG1 mAb that targets PD-L1. Methods: A Phase I, open-label, multicenter study was conducted to evaluate the safety and pharmacokinetics (PK) of MOXR0916 and atezolizumab in patients (pts) with locally advanced or metastatic solid tumors. A 3+3 dose-escalation was conducted with a 21-day window to evaluate dose-limiting toxicity (DLT). Escalating doses of MOXR0916 in combination with a fixed 1200 mg dose of atezolizumab were administered every 3 weeks (q3w). An expansion coh...

9523Background: Immunotherapy has resulted in unprecedented improvements in survival for patients with metastatic melanoma. Despite a response rate of 10% and 40% respectively for Ipilimumab (ipi) ...

9528Background: The coBRIM study showed significant improvement in overall survival (OS), progression-free survival (PFS), and overall response rate (ORR) in patients (pts) with BRAFV600 mutation–p...

9582Background: While immunotherapy and MAPK-targeted therapies have improved patient outcome, the genomic heterogeneity of melanoma contributes to treatment resistance. Molecular approaches for mo...

9510Background: The BRIM7 study (ClinicalTrials.gov ID, NCT01271803) evaluated the safety and preliminary efficacy of C+V (Ribas A et al. Lancet Oncol. 2014;15:954-965.); a subsequent randomized ph...

9530Background: In coBRIM, C+V significantly improved PFS and OS versus placebo (P)+V in pts with BRAFV600-mutated advanced melanoma, with benefit in all prespecified subgroups evaluated (Larkin J ...

LBA1 Background: The results of a phase I study in MEL suggested complementary clinical activity with NIVO (a PD-1 checkpoint inhibitor) plus IPI (a CTLA-4 checkpoint inhibitor). Here, we report the results of a randomized, double-blind, phase III trial designed to evaluate NIVO combined with IPI or NIVO alone vs IPI alone in MEL. Methods: Treatment-naïve pts (N = 945) were randomized 1:1:1 to NIVO 1 mg/kg Q2W + IPI 3 mg/kg Q3W for 4 doses followed by NIVO 3 mg/kg Q2W, NIVO 3 mg/kg Q2W + placebo, or IPI 3 mg/kg Q3W for 4 doses + placebo, until progression or unacceptable toxicity. Pts were stratified by PD-L1 status, BRAF mutation status, and M-stage. Co-primary endpoints are progression-free survival (PFS) (reported here) and overall survival (pts continue to be followed). Secondary endpoints include objective response rate (ORR) by RECIST v1.1 and safety. Results: At a minimum follow-up of 9 months, NIVO + IPI and NIVO alone significantly improved PFS and ORR vs IPI (Table). Grade 3-4 drug-related adverse events (AEs) occurred in 55.0%, 16.3%, and 27.3% of pts in the NIVO + IPI, NIVO, and IPI arms, respectively (most commonly diarrhea [9.3%, 2.2%, 6.1%], increased lipase [8.6%, 3.5%, 3.9%], increased alanine aminotransferase [8.3%, 1.3%, 1.6%], and colitis [7.7%, 0.6%, 8.7%]). Drug-related AEs led to discontinuation in 36.4%, 7.7%, and 14.8% of pts in the NIVO + IPI, NIVO, and IPI arms, with 0, 1, and 1 drug-related deaths, respectively. Efficacy outcomes by PD-L1 status will also be presented. Conclusions: NIVO + IPI and NIVO alone had superior clinical activity vs IPI alone. The results with NIVO + IPI and NIVO alone further suggest complementary activity of the two agents. There were no new safety signals or drug-related deaths observed with the combination. Clinical trial information: NCT01844505. [Table: see text]

9020 Background: BRIM7 formed the basis for development of the VEM + COBI regimen, which, in a randomized phase III trial, confirmed the statistically significant and clinically meaningful progress...

9033 Background: Various ocular adverse events (AE) have been described with targeted agents. MEK inhibitors (MEKi) are associated with SR. We describe the clinical features of SR observed with COB...

e13557 Background: Increasing evidence suggests that epigenetic changes underlie many of the oncogenic processes in melanoma. In previous studies we have shown that epigenetic regulation of gene ex...

9006 Background: Primary analysis of coBRIM showed significant improvement in PFS and overall response rate (ORR) in BRAFV600 mutation–positive patients (pts) with advanced melanoma treated with ve...

9021 Background: The phase 3 coBRIM study showed significant improvement in progression-free survival in pts with metastatic melanoma on V + C compared with those on V and placebo (P) (HR, 0.51; 95...

e22212 Background: Ribosome biogenesis driven by RNA polymerase 1 (Pol1) is a fundamental cellular process increased in malignancy. We have shown that CX-5461, a small molecule inhibitor of Pol1, s...

TPS9081 Background: T-VEC is a herpes simplex virus-1-based oncolytic immunotherapy designed to preferentially replicate in tumors, produce GM-CSF and stimulate an anti-tumor immune response. OPTiM...

9059 Background: Immunotherapy with anti-CTLA-4 and anti-PD-1 antibodies has demonstrated overall survival benefits in patients (pts) with metastatic melanoma (MM) compared to standard therapy. An ...

9003 Background: In the modern era of melanoma treatment, approaches to imaging surveillance following surgery require reconsideration. The aim of this study was to evaluate disease sub-stage speci...

10061 Background: Deregulated acetylation of histones plays a key role in the pathogenesis of haematological and solid tumors by changing the transcription of genes involved in cell cycle control, ...

9006^ Background: BRAF mutation is a common and potent oncogenic driver of ERK signalling in melanoma. Abrogation of pERK reduces 18-F-fluorodeoxyglucose (FDG) uptake on PET with recovery in glycol...

9050 Background: The cancer testis antigen NY-ESO-1 (ESO) has been evaluated as a biomarker and a therapeutic immunologic cancer target. Preliminary data from a phase I clinical trial of ESO and IS...

9028 Background: LGX818, a potent and selective BRAF inhibitor (BRAFi) being investigated in BRAF V600 mutant melanoma, has unique biochemical properties with a dissociation half-time > 10 times longer than other BRAF inhibitors. Methods: A phase I trial of LGX818 administered orally once (qd) or twice (bid) daily in BRAF V600 tumors was initiated to define the maximum tolerated dose (MTD)/recommended phase II dose (RP2D) and to assess pharmacokinetics and clinical activity in BRAFi–naive or pretreated patients with BRAF V600 mutant advanced melanoma. Baseline assessment of biomarkers from MAPK/PI3K pathways and pharmacodynamics were also evaluated. Results: Fifty-four patients have been enrolled in the dose-escalation phase (dose levels [DLs], 50-700 mg qd [n=42] and 75-150 mg bid [n=12]). LGX818 plasma concentrations increased proportionally by dose with a mean t1/2 of 4 hours and steady state in ≈ 15 days. The MTD/RP2D (450 mg qd) was well tolerated. Seven patients had a dose limiting toxicity (DLT): 5 at qd (1 each with hand-foot skin reaction [HFSR], foot pain, fatigue, diarrhea/rash, insomnia/asthenia) and 2 at bid (1 facial paresis/confusion, 1 musculoskeletal pain/neuralgia). All DLTs were grade 3 and reversible. The most common adverse events (≥ 20%) suspected to be treatment related were cutaneous (rash, dry skin, HFSR, pruritus, keratosis pilaris, alopecia), pain in extremity, arthralgia, and fatigue. Squamous cell carcinoma was observed in 2 patients (1 naive and 1 pretreated). As of 30 Sept 2012, the preliminary efficacy (all DLs) in patients with at least 1 postbaseline tumor assessment was 16 partial responses [PRs] (67%; 12 confirmed) out of 24 BRAFi–naive patients and 2 PRs (8.3%; 1 confirmed) among 24 BRAFi–pretreated patients. Responses were seen at all DLs from 50 to 550 mg qd. Updated safety and efficacy including time to event endpoints will be reported. Conclusions: Initial results from this study identified the MTD/RP2D as 450 mg/day and provided an early sign of promising activity in advanced melanoma. Expansion cohorts are ongoing in BRAFi–naive and BRAFi–pretreated melanoma and colorectal cancer. Clinical trial information: NCT01436656.

8517 Background: Vemurafenib, a BRAF inhibitor, is associated with improved PFS and OS in patients (pts) with BRAFV600-mutant metastatic melanoma (mM). We present preliminary safety and efficacy findings from a safety study of vemurafenib in pts with unresectable stage IIIC/IV mM with BRAFV600 mutations. Methods: Pts with untreated or previously treated stage IIIC/IV BRAFV600 mutation-positive (cobas 4800 BRAF V600 Mutation Test) melanoma were enrolled. Pts received continuous oral vemurafenib 960 mg bid. Primary study endpoint was safety; efficacy (RECIST V 1.1) was a secondary endpoint. Results: Of 1,964 screened pts between Mar and Sep 2011, 914 (47%) were enrolled and 834 were evaluable for safety. Median age was 53 (21–88 years), 55% males. Median time since first mM diagnosis was 7.6 months (0–18 years). At baseline, 80% of pts had ECOG PS 0–1, 11% ECOG PS 2 (missing 9%); 27% of pts had brain metastases, and 31% had elevated LDH. Most pts had received prior systemic therapy (70%) including ipilimumab (14%), MEK and BRAF inhibitors (2%). At data cut-off (Sep 30, 2011), median treatment duration was 68 days (1–223 days) with 87% of pts still on treatment. Of 834 pts, 553 (66%) to date have reported AEs. Of 553 pts reporting AEs, 88% were related to vemurafenib, 33% Grade 3, and 1.9% Grade 4. The most common (>1%) Grade 3/4 AEs were rash (3.6%), arthralgia (3.1%), and cutaneous squamous cell carcinoma/keratoacanthoma (4.3%). Most common AEs (>10%) of any grade were arthralgia (31%), rash (29%), fatigue (22%), photosensitivity (21%), nausea (15%), and were similar irrespective of brain metastases and ECOG PS. AEs caused treatment interruption in 141 (17%) pts. Of 109 pts who discontinued treatment (13%), main reasons for withdrawal were progressive disease (60%), death (20%), AEs (6%; most commonly arthritis and abdominal pain). Tumor assessments at Week 8 of treatment were available for 302/834 (36%) pts, 61% pts achieved CR or PR, and 29% had SD. Conclusions: In a setting representative of routine clinical practice, vemurafenib is seen to be well tolerated and both safety profile and activity resemble the phase I–III data although this analysis is limited by the study duration.

8502^ Background: We previously reported results of the planned OS interim analysis for BRIM-3 (50% of the planned 196 deaths required for final analysis) at which time the independent Data Safety Monitoring Board recommended release of results due to compelling efficacy (hazard ratio [HR] for death, 0.37 [95% CI 0.26–0.55]); pV600E mutation by the cobas 4800 BRAF V600 Mutation Test were randomized (1:1) from Jan to Dec 2010 to vem (960 mg po bid) or DTIC (1000 mg/m2 IV q3w). Co-primary endpoints were OS and PFS. OS data for DTIC patients who crossed over to vem were censored at the time of crossover. Results: Median lengths of follow-up on vem and DTIC were 10.5 months (range 0.4–18.1) and 8.4 months (range

8503^ Background: Vem induces frequent clinical responses (RR >50%) and improved survival in patients (pts) with BRAF-mutated metastatic melanoma. Multiple mechanisms of escape from vem have been proposed. We performed a centralized analysis of pretreatment, cycle 1, day 15 (D15), and progression (DP) tumor samples collected during the phase II BRIM-2 trial (Sosman et al, NEJM 2012). Methods: Of 132 pts enrolled, archival tumor samples, biopsies taken at baseline (BL), D15, and DP were obtained from 84, 38, 26, and 22 pts, respectively, and analyzed by immunohistochemistry (IHC) for signaling molecules in the MAPK, PI3K/AKT, and cell cycle pathways. Genetic analyses of signaling genes were performed using Sequenom MASSarray and direct DNA sequencing. Results: High levels of ERK phosphorylation were seen at BL indicating constitutive MAPK signaling due to the BRAF mutation. In 19/22 paired samples, pERK levels were reduced following vem (BL vs D15). Mean absolute pERK reduction in pts with clinical response (RECIST criteria) to vem (n=14) was significantly greater than in non-responders (n=8; p=0.013). Baseline cytoplasmic PTEN H-score was higher in responders vs non-responders (H-score=88 vs 68; p=0.042). At progression, upregulation of pERK (H-score=181) was frequently, but not uniformly found vs D15 tumors (H-score=48.5). At progression, increases in Cyclin D1 and Ki67 were seen, without obvious changes in PTEN or pAKT. NRAS mutations occurred in 3/13 DP; 2 had paired BL samples without NRAS mutations. Only 1/82 pts had a concomitant NRAS and BRAF mutation at BL, and did not respond to vem. MAP2K1 (MEK1) codon 124 mutations occurred in 7/92 BL and 1/20 DP samples. 2 pts with BL MEK1 mutations had partial responses. Conclusions: MAPK signaling was effectively inhibited by vem early in treatment, and in the subset of patients with matched tumor samples the degree of pathway inhibition correlated with clinical response. MAPK signaling is upregulated in many lesions at progression. NRAS mutations appear to be a mechanism of acquired resistance in a few tumors (3/13), while the role of MEK1 mutations in resistance is less clear.