NSCLC Pts Research Articles

1569P - Final results of randomized phase II trial of metronomic vs weekly oral vinorelbine (OV) as first-line chemotherapy (CT) in advanced NSCLC patients unfit to platinum-based CT (P-CT): Tempo-Lung

BackgroundMetronomic chemotherapy (CT) is a therapeutic strategy based on a regular, short interval, fixed-dose schedule developed to improve disease control and reduce toxicity. Metronomic oral vinorelbine (m-OV) showed promising results. The trial randomly compared single agent m-OV vs OV standard schedule as first-line treatment in advanced NSCLC patients (pts) unfit for platinum-based CT (P-CT). MethodsCT naive pts with advanced NSCLC unfit for P-CT (Creatinine clearance <60ml/min; heart failure NYHA class II-III; hearing loss >G2; medical conditions impairing P-CT according to physician’s opinion) were 1:1 randomly allocated to arm A: m-OV 50mg x3/week (wk) or arm B: OV 60mg/m²/wk cycle 1, then 80mg/m²/wk. Primary objective: progression-free survival without grade 4 toxicity (G4PFS). Secondary objectives: disease control rate (DCR), PFS, overall survival (OS), safety, quality of life. Stratification factors: stage (IIIB vs IV), age (< 70yrs or>70 yrs) and ECOG PS (0-1 vs 2). ResultsIntention-to-treat (ITT) population included 165 pts: 83 pts (A) / 82 (B). Pt characteristics were well balanced (A vs B): mean age 76.1 vs 75.9 yrs; ECOG PS 0-1: 66% vs 62%; Stage IV: 87% vs 94% pts. Median relative dose intensity was 85% vs 69%. Primary endpoint was reached (A vs B) with median G4PFS [95%CI]: 4.0 [2.6-4.3] vs 2.2 [1.5-2.9] months (mo) (p=0.0068), HR [95%CI] = 0.63 [0.45-0.88]. Secondary efficacy end-points were as follows (A vs B): DCR 63.9% vs 63.4%; median PFS 4.3 vs 3.9mo; median OS: 7.1 vs 7.6mo. Treatment related AEs (r-AEs) A vs B were: all grade 61.4% vs 84%; total hematological AEs: 27.7% vs 55.6%. Most common (≥ 5%) grade 3-4 r-AEs (A vs B) were: febrile neutropenia 3.6% vs 6.2%; neutropenia 11% vs 52%; asthenia 4.8% vs 8.6%. One grade 5 AE occurred in each arm (febrile neutropenia, neutropenic sepsis). ConclusionsPositive results of first randomized prospective trial: m-OV significantly improved G4PFS as first-line CT in advanced unfit NSCLC pts, with reduced toxicity. Treatment efficacy was similar in both arms. m-OV is a suitable, safe option for first-line therapy of NSCLC pts unfit to P-CT. Clinical trial identificationEudraCT: 2014-003859-61. Legal entity responsible for the studyInstitut de Recherche Pierre Fabre- Pierre Fabre Médicament. FundingInstitut de Recherche Pierre Fabre- Pierre Fabre Médicament. DisclosureA. Morabito: Speaker Bureau / Expert testimony, Speaker Bureau: Pfizer; Speaker Bureau / Expert testimony, Speaker Bureau: BMS; Speaker Bureau / Expert testimony, Speaker Bureau: Boehringer Ingelheim; Speaker Bureau / Expert testimony, Speaker Bureau: MSD; Speaker Bureau / Expert testimony, Speaker Bureau: Roche; Speaker Bureau / Expert testimony, Speaker Bureau: AstraZeneca. F. Grossi: Honoraria (self): Ely Lilly; Honoraria (self): Roche; Honoraria (self): AstraZeneca; Honoraria (self): Boehringer Ingelheim; Honoraria (self): Pierre Fabre; Honoraria (self): BMS; Honoraria (self): Amgen; Honoraria (self): MSD; Research grant / Funding (institution): AstraZeneca; Research grant / Funding (institution): BMS; Research grant / Funding (institution): MSD; Advisory / Consultancy: Ely Lilly; Advisory / Consultancy: Roche; Advisory / Consultancy: Boehringer Ingelheim; Advisory / Consultancy: AstraZeneca; Advisory / Consultancy: Pierre Fabre; Advisory / Consultancy: MSD; Advisory / Consultancy: Novartis; Advisory / Consultancy: BMS; Honoraria (self): Celgene. R. Ramlau: Advisory / Consultancy: Roche; Advisory / Consultancy: Boehringer Ingelheim; Advisory / Consultancy: Pfizer; Advisory / Consultancy: Novartis. T. Ciuleanu: Advisory / Consultancy, Travel / Accommodation / Expenses: Astellas Pharma; Advisory / Consultancy, Travel / Accommodation / Expenses: Janssen; Advisory / Consultancy, Travel / Accommodation / Expenses: BMS; Advisory / Consultancy, Travel / Accommodation / Expenses: Amgen; Advisory / Consultancy, Travel / Accommodation / Expenses: Merck Serono; Advisory / Consultancy, Travel / Accommodation / Expenses: Roche; Advisory / Consultancy, Travel / Accommodation / Expenses: Pfizer; Advisory / Consultancy, Travel / Accommodation / Expenses: Lilly; Advisory / Consultancy, Travel / Accommodation / Expenses: AstraZeneca; Advisory / Consultancy, Travel / Accommodation / Expenses: Boehringer Ingelheim; Advisory / Consultancy, Travel / Accommodation / Expenses: MSD; Advisory / Consultancy, Travel / Accommodation / Expenses: Sanofi; Advisory / Consultancy, Travel / Accommodation / Expenses: Novartis; Advisory / Consultancy, Travel / Accommodation / Expenses: Servier; Advisory / Consultancy, Travel / Accommodation / Expenses: A et D Pharma; Travel / Accommodation / Expenses: Ipsen. G.L. Ceresoli: Advisory / Consultancy: Pfizer; Advisory / Consultancy: Astellas; Advisory / Consultancy: Boehringer Ingelheim. J. Bosch Barrera: Honoraria (self): Pierre Fabre; Honoraria (self): Roche; Honoraria (self): BMS; Advisory / Consultancy: Boehringer Ingelheim. P. Landreau: Full / Part-time employment: Pierre Fabre. S. Gautier: Full / Part-time employment: Pierre Fabre. C. Ta Thanh Minh: Full / Part-time employment: Pierre Fabre. A. Camerini: Speaker Bureau / Expert testimony, expert testimony: Pierre Fabre; Travel / Accommodation / Expenses: Pierre Fabre; Travel / Accommodation / Expenses: Roche. All other authors have declared no conflicts of interest.

1524P - Apatinib combined with EGFR-TKI in treating advanced non-small cell lung cancer with EGFR-TKI resistance

BackgroundEGFR-TKI have been widely used in patients (pts) with EGFR mutations in NSCLC and brought significant benefits. But resistance to EGFR-TKI is inevitable. Previous studies have shown that apatinib (a TKI against VEGFR-2) combined with EGFR-TKI might prevent progression of the disease. We conducted this trial to investigate the efficacy and safety of apatinib combined with EGFR-TKIs (including erlotinib, gefitinib, icotinib, afatinib and osimertinib) compared with traditional chemotherapy for EGFR-TKI-resistant NSCLC pts. MethodsThis study enrolled 39 advanced NSCLC pts who acquired resistance to EGFR-TKI therapy from Mar 2017 to Jan 2019. 25 pts received apatinib combined with EGFR-TKI (apatinib in start dose of 250mg+prior EGFR-TKI dose),14 pts received chemotherapy (pemetrexed or vinorelbine with platinum). ResultsIn the apatinib group, 88% (22/25) pts were available evaluated. The objective response rate was 17% (3/22) and the disease control rate was 96% (21/22). The most common adverse events in the apatinib group were diarrhea (60%,15/25), hypertension (56%,14/25), hand-foot syndrome (40%,10/25), fatigue (30%,7/23). Main grade 3 or 4 toxicities were proteinuria (12%, 3/25). Two pts with brain metastases in the apatinib group showed a decrease of metastatic lesions. The lesions of two pts who having taken the dose of 250mg apatinib and progressed decreased when they changed the dose to 500mg. In the chemotherapy group, 79% (11/14) pts were available to be evaluated. The objective response rate was 27% (3/11) and the disease control rate was 91% (10/11). No new adverse events occurred. The median length of treatment of the apatinib group was 8.7 monthsand of chemotherapy group was 4.3 months. The longest treatment period in the apatinib group was 24 months. ConclusionsApatinib combined with EGFR-TKI showed a good clinical efficacy in pts with acquired EGFR-TKI (1st 2nd or 3rd generation) resistance. Patient’s quality of life and the compliance with therapy hads been increased by use of oral drugs. We found that in the patients who were treated with erlotinib, or patients with EGFR 21 mutation, or male, their PFS tended to be prolonged compared to other patients. Clinical trial identificationChiCTR-OIN-17012051. Legal entity responsible for the studyThe authors. FundingChina Anti-Cancer Association. DisclosureAll authors have declared no conflicts of interest.

Real-World Effectiveness and Safety of Nivolumab in NSCLC pts in Japan: A Multicenter-Retrospective Observational Study

Abstract Background In December 2015, nivolumab was approved as treatment for previously treated NSCLC irrespective of PD-L1 expression or histology in Japan. There are few reports that include real-world data on large cohorts of patients from Japan. Method Data from patients who initiated nivolumab treatment from April 1, 2016 to December 31, 2016 were collected at 23 regional medical institutions across Japan. Here we report a retrospective data analysis accessing effectiveness and safety of nivolumab in 901 Japanese NSCLC pts from a multicenter, observational study; CA209-9CR (NCT03273790). Result Among 901 pts, median age of the pts treated with nivolumab was 67 ys old and proportion of 75 ys or over was 19.8%. Pts with ECOG PS 0 or 1 account for 75.8%, and pts with SQ and adeno histology were 24.5% and 67.7%, respectively. Pts with EGFR mutation account for 12.9%. Median numbers of nivolumab dose was five and nivolumab was most frequently used at 2nd line (46.8%) and followed by at 3rd line (26.5%). Rate of 1 yr overall survival (OS) was 54.3% and, median PFS was 2.1 months and overall ORR was 20.6%. There was no significant difference in effectiveness among histology. Although there was no difference in effectiveness of nivolumab by age, ECOG PS status significantly influence effectiveness of nivolumab. Analysis is ongoing to access various factors that may be associated with the effectiveness of nivolumab. Overall irAE incidence for all grade and grade 3-4 will be reported as well as categorized irAE. Further analysis is planned for safety management of irAEs and prognosis. Conclusion The effectiveness and safety of nivolumab observed in real-world data from NSCLC patients in Japan was comparable to results observed in clinical trials. ECOG PS status was identified as a factor influencing effectiveness of nivolumab.

1995P - Application of ESCAT and OncoKB scales in liquid biopsy (LB) in advanced NSCLC patients (pts): Is it feasible and reliable?

BackgroundNext generation sequencing (NGS) of cfDNA in NSCLC is entering in clinical practice. Several commercially available platforms using NGS report wide variety of somatic aberrations. Vendors also include therapeutic suggestions to guide oncologists. In addition, new levels of evidence tools are now available for tissue results, but not still used in LB. MethodsAdvanced NSCLC pts underwent commercial 73-gene cfDNA NGS analysis. ESMO Scale for Clinical Actionability (ESCAT) and OncoKB were used to grade levels of evidence for categorize aberrations and compared with variant allele frequency (VAF), treatment decisions, and vendor suggestions. Results77 samples from 73 advanced NSCLC pts (73% adenocarcinoma, 27% squamous cell carcinoma) at the time of diagnosis (49%) or during disease course (51%) were analyzed. Median turnaround time was 8 days (range 5-17), with no genotyping failures. There were 323 unique somatic alternations identified: Sequence mutations (254), amplifications (43), synonymous mutations (23), and fusions (3). Median cfDNA % was 0.7 (range 0.05-49.5); 7 samples had no detectable genetic alterations. Variants of unknown significance were 33% of point mutations. We detected 87 and 88 potentially actionable genetic alterations according to ESCAT (IA-IV) and OncoKB (1-4), respectively, and 26% received a matched targeted drug. Discrepancies between these two tools and vendor suggestions were reported in 4 cases. We performed a subset analysis of 64 samples: Median VAF was 4.37 % (range 0.16-43.05) and a VAF < 1% was reported in 16 samples; 45% of alterations (excluding amplifications) were clonal events (cfDNA% divided by VAF > 0.5). Among EGFR and ALK positive samples median VAF was 7.98% (range 0.29-41.2); 3/17 samples presented a VAF below 1%, with no detrimental effect on treatment response. ConclusionsThe application of ESCAT and OncoKB is feasible in LB. Driver mutations with low VAF are amenable to receive treatment. VAF could be included as complementary tool to grading systems to better understand the significance of aberrations.Discrepancies between vendor therapeutic suggestions and evidence-based grading systems requests caution in the use of information outside the molecular tumor board. Legal entity responsible for the studyThe authors. FundingHas not received any funding. DisclosureC.D. Rolfo: Honoraria (self), personal fees : Novartis; Honoraria (self), personal fees : MSD; Non-remunerated activity/ies, non-financial support : OncoDNA; Honoraria (self), personal fees and non-financial support : GuardantHealth. All other authors have declared no conflicts of interest.

1782P - Impact of nutritional derangement on treatment outcome in advanced non-small-cell lung cancer (A-NSCLC) patients (pts)

BackgroundWeight loss and lean body mass wasting are highly prevalent in NSCLC pts, but frequently underestimated. This study aimed to assess the prevalence of malnutrition and its correlation with outcome in A-NSCLC pts. MethodsA-NSCLC pts treated at AOUI of Verona (2016-2018) received nutritional counselling by a qualified dietitian. Nutritional Risk Screening (NRS) 2002 was used to estimate the nutritional risk. Bilateral psoas major muscles were measured at L3 vertebrae level with routine staging-computed tomography (CT). Changes in psoas muscles were evaluated using Wilcoxon signed-rank test. Data were correlated to progression-free/overall survival (PFS/OS) and response rate (ORR) using a Cox and logistic regression model. Kaplan-Meier curves were compared with Log-Rank. ResultsData from 38 pts (20 males [52.6%], 18 females [47.4%]) were gathered, with a median follow-up of 21 months (range 1-197). At baseline, 18.4% were underweight, 18.4% normal weight, 34.2% overweight and 31.6% obese. The majority (65.8%) were at risk of malnutrition (NRS≥3). At multivariate analysis, stage (HR 4.99, 95% CI 1.05-27.74, p=0.04), performance status (HR 4.99, 95% CI 1.55-16.03, p=0.007) and NRS (HR 7.61, 95% CI 1.52-38.11, p=0.01), were significant independent predictors for PFS. Pts with baseline NRS≤3 had significantly longer 1-year PFS (58.6% vs 16.7%, p=0.04) and 2-year OS (90.6% vs 68.3%, p=0.03) and a better ORR than those with NRS > 3 (66.7% vs 21.4%). A significant loss in psoas muscle mass was detected in pts treated with both immunotherapy and other therapies (p=0.01 and p=0.002, respectively). Of interest, in immunotherapy-treated pts (n=16) loss in psoas muscle mass correlated with worse ORR, PFS and OS, although differences did not reach a statistical significance. ConclusionsMalnutrition was common in A-NSCLC pts and associated with poor ORR, PFS and OS. Particularly, in pts treated with immunotherapy, muscle mass wasting seems to impact on efficacy outcome, suggesting a potential interaction between immunological and nutritional parameters. The introduction in clinical routine of a comprehensive nutritional profiling and monitoring is highly recommended in A-NSCLC. Legal entity responsible for the studyThe authors. FundingHas not received any funding. DisclosureE. Bria: Honoraria (self): MSD, Astra-Zeneca, Celgene, Pfizer, Helsinn, Eli-Lilly, BMS, Novartis, and Roche; Research grant / Funding (self): AIRC-IG 20583. M. Milella: Honoraria (self): Pfizer, EUSA Pharma, AstraZeneca. S. Pilotto: Honoraria (self): Astra-Zeneca, Eli-Lilly, BMS, Boehringer Ingelheim, Roche, MSD and Istituto Gentili; Research grant / Funding (self): AIRC-IG 20583. All other authors have declared no conflicts of interest.

1527P - Sequential treatment with afatinib followed by 3rd generation EGFR-TKI – subgroup analysis of the GIDEON trial: A prospective non-interventional study (NIS) in EGFR mutated NSCLC patients in Germany

BackgroundAfatinib and osimertinib both inhibit sensitizing EGFR mutations (deletion 19 and L858R). In addition, osimertinib is highly active against the most frequent resistance mutation (∼60%) T790M and was originally registered for these patients (pts.). Some evidence suggests that a sequential treatment approach using 1st or 2nd gen EGFR-TKIs like afatinib in first-line, followed by a 3rd gen EGFR-TKI in T790M positive pts. may lead to overall survival rates up to 80% after 4 years. Here, we report first results of prospectively collected data within the German NIS GIDEON on pts. treated with 1st line afatinib followed by osimertinib. Methods151 EGFR-mutated NSCLC pts. were treated with afatinib according to label until progression, death or treatment discontinuation. Data on effectiveness (ORR, DCR, PFS and 24 months OS rate) were prospectively assessed for pts. who have been treated with osimertinib following afatinib. ResultsA total of 30/151 (20%) pts. were treated with 1st line afatinib followed by osimertinib. 22 (73%), 7 (23%) and 1 (3%) of these pts. were initially positive for Del19, L858R and G719C, respectively. The mPFS from start of afatinib to PD under osimertinib at the time of analysis was 32.2mo, 29.8mo and NR (95% CI 16.4mo - NR) for the total population, pts. with Del19 and pts. with L858R, respectively. The 24-mo-OS rate was 89.3% for all 30 pts receiving osimertinib in a later line, and 90% and 85.7% for patients with initial Del19 and L858R, respectively. ConclusionsThese data support the evidence that pts. with acquired T790M under afatinib may benefit from sequential TKI treatment. The frequency of T790M was higher in pts. initially tested positive for Del19, which is in line with recently published data. This analysis is limited by the small patient number, explainable by the restricted availability of osimertinib during the study conduct (March 2014 – Dec 2017). Clinical trial identificationNCT02047903. Legal entity responsible for the studyBoehringer Ingelheim. FundingBoehringer Ingelheim. DisclosureW. Brückl: Advisory / Consultancy: AbbVie; Advisory / Consultancy: AstraZeneca; Advisory / Consultancy: BMS; Advisory / Consultancy: Boehringer Ingelheim; Advisory / Consultancy: Celgene; Advisory / Consultancy: Chugai; Advisory / Consultancy: Lilly; Advisory / Consultancy: MSD; Advisory / Consultancy: Pfizer; Advisory / Consultancy: Roche; Advisory / Consultancy: Stratifyer. M. Reck: Honoraria (self), Advisory / Consultancy: Abbot; Honoraria (self), Advisory / Consultancy: AstraZeneca; Honoraria (self), Advisory / Consultancy: BMS; Honoraria (self), Advisory / Consultancy: Boehringer Ingelheim; Honoraria (self), Advisory / Consultancy: Celgene; Honoraria (self), Advisory / Consultancy: MSD; Honoraria (self), Advisory / Consultancy: Lilly; Honoraria (self), Advisory / Consultancy: Novartis; Honoraria (self), Advisory / Consultancy: Pfizer; Honoraria (self), Advisory / Consultancy: Roche; Honoraria (self), Advisory / Consultancy: Merck. J. Rawluk: Honoraria (self), Advisory / Consultancy: Boehringer Ingelheim; Honoraria (self), Advisory / Consultancy: AstraZeneca; Honoraria (self), Advisory / Consultancy: Roche; Honoraria (self), Advisory / Consultancy: MSD; Honoraria (self), Advisory / Consultancy: BMS. R.M. Huber: Speaker Bureau / Expert testimony: ARIAD; Speaker Bureau / Expert testimony: AstraZeneca; Advisory / Consultancy, Speaker Bureau / Expert testimony: BMS; Advisory / Consultancy, Speaker Bureau / Expert testimony: Boehringer Ingelheim; Advisory / Consultancy: Celgene; Advisory / Consultancy: Lilly; Advisory / Consultancy: Novartis; Advisory / Consultancy, Speaker Bureau / Expert testimony: Roche; Advisory / Consultancy: Clovis Oncology; Research grant / Funding (institution): Pierre Fabre; Speaker Bureau / Expert testimony: Pfizer. C. Hoffmann: Full / Part-time employment: Boehringer Ingelheim. A. Schueler: Full / Part-time employment: Boehringer Ingelheim. All other authors have declared no conflicts of interest.

1596TiP - DUBLIN-3, a stage IIIb/IV NSCLC phase (Ph)3 trial comparing the plinabulin (P)/Docetaxel(D) combination with D Alone

BackgroundThe PD-L1 inhibitors Avelumab (AV) and Atezolizumab (AT) have proven clinical anticancer efficacy, however in NSCLC trials, Javelin Lung 200 (JAV for AV) failed, whereas OAK (for AT) met primary endpoint showing better survival vs 75mg/m2 D as comparator. The authors suggested JAV trial design factors may have played a role (Barlesi, Lancet Oncology 2018). Firstly, its open-label study design lead to more D patients (pts) dropping out prior to receiving first D dose (8% vs 1% for D vs AV). Secondly, JAV did not stratify for region, resulting in 29% vs 25% Asian pts with D vs AV; Asian pts tend to respond better to D then non-Asians. However, tumor stage also might be another factor, since advanced primary and metastatic lesions likely harbor antigens (Ags) for which immune tolerance has already been developed. Mutation burden of advanced primary and metastatic lesions show high concordance (Sherwood, J Exp & Clin Canc Res 2015). In contrast, early or novel subclonal lesions located in the lung are more likely to harbor (novel) immunogenic Ags (De Bruin, Science 2014). More pts in JAV (close to 100% (6%M0) vs OAK (70.6%) were at stage IIIb/IV (Rittmeyer, Lancet 2016; OAK Team, 2019). P is a novel Dendritic Cell (DC) modulator that is combined with D. D induces Ags that DC cells can present to CD4 and CD8 T-Cells after P stimulation (Lloyd, AACR 2016). P has favorable safety/tolerability in>500 pts and prevents D-induced-Neutropenia (N) and -Thrombocytopenia (Blayney, ASCO 2018; IASLC 2018; ESMO 2018). DUBLIN-3 may have avoided some of the JAV design limitations. Trial designDUBLIN-3 (NCT02504489), is a global, single-blinded (for pts) Ph3 study in EGFR wild-type, stage IIIb/IV NSCLC pts (target n=554) stratified for region (Asia/non-Asia), and receiving 2nd- or 3rd-line systemic therapy with D+P or D in a 1:1 ratio. Key inclusion criteria are, pts must have at least one measurable lesion located in the lung, and have failed a prior platinum-based regimen. Prior PD1/PD-L1 use is allowed. Primary endpoint is OS. Key secondary endpoints are incidence of Gr 4 N, D dose modification due to N, QoL, ORR, PFS, and DoR. The first pre-specified Interim Analysis (IA) occurred at ∼150 events and a second IA at ∼ 300 events is projected for Q4 2019. Clinical trial identificationBPI-2358-103 NCT02504489. Legal entity responsible for the studyBeyondSpring Pharmaceuticals, Inc. FundingBeyondSpring Pharmaceuticals, Inc. DisclosureR. Mohanlal: Leadership role, Travel / Accommodation / Expenses, Full / Part-time employment, Officer / Board of Directors: BeyondSpring Pharmaceuticals, Inc; Shareholder / Stockholder / Stock options: BeyondSpring, Inc. L. Huang: Leadership role, Travel / Accommodation / Expenses, Full / Part-time employment, Officer / Board of Directors: BeyondSpring Pharmaceuticals, Inc; Shareholder / Stockholder / Stock options: BeyondSpring, Inc.

1483PD - Pembrolizumab (pembro) plus platinum-based chemotherapy (chemo) in NSCLC with brain metastases: Pooled analysis of KEYNOTE-021, 189, and 407

1483PD - Pembrolizumab (pembro) plus platinum-based chemotherapy (chemo) in NSCLC with brain metastases: Pooled analysis of KEYNOTE-021, 189, and 407

Antitumor activity of TAK-788 in NSCLC With EGFR exon 20 insertions

Abstract Background We report results of a phase 1/2 open-label, multicenter study of TAK-788 (NCT02716116), an oral investigational EGFR/HER2 inhibitor. Methods Pts with advanced, previously treated NSCLC received daily TAK-788 in dose escalation and expansion cohorts based on tumor genotype. Antitumor activity was determined for pts with EGFR exon 20 insertions who received TAK-788 at the RP2D. Safety is reported for all pts across all doses and at 160 mg. Results As of 14 Sep 2018, 101 pts (median age, 61 y; female, 70%; ≥2 prior anticancer therapies, 76%; brain metastases, 53%) were treated with TAK-788 at 5-180 mg qd. RP2D was determined to be 160 mg. 28 pts with EGFR exon 20 insertions were treated with 160 mg qd during dose escalation or in expansion cohort 1 (3.6 mo on treatment; 3.8 treatment cycles [medians]); 24 pts remain on treatment. At data cutoff, best response (by RECIST v1.1) among 26 pts with ≥1 disease assessment was PR, n = 14; SD, n = 9; and PD, n = 1; 2 pts were not evaluable. 7/14 objective responses (all PR) were confirmed, with 6 awaiting confirmation and 1 unconfirmed PR at 160 mg qd; median time to response in these 14 pts was 56 days. 23/26 pts achieved disease control. 23/24 evaluable pts with EGFR exon 20 insertions treated at 160 mg qd had decreased target lesion measurements (median best percent change, -32.6% [-79.1%, 3.8%]). Most common TEAEs (≥20%) in pts treated with 160 mg qd: diarrhea (85%), rash (43%), nausea (41%), vomiting (30%), decreased appetite (28%), and stomatitis (22%); gr ≥ 3 TEAEs (≥5%): diarrhea (26%); hypokalemia, nausea, and stomatitis (7% each). Among pts treated with 160 mg qd, median dose intensity was 93%, and the rate of treatment discontinuation due to AEs was 10.7%. There is no clear trend that response to TAK-788 is enriched in any single EGFR exon 20 insertion variant. Conclusions In NSCLC pts with EGFR exon 20 insertions, TAK-788 demonstrated antitumor activity and an AE profile consistent with other EGFR TKIs.

Real-World Effectiveness and Safety of Nivolumab in Special Subgroups of NSCLC pts: A Multicenter-Retrospective Study

Abstract Background In December 2015, nivolumab was approved as treatment for previously treated NSCLC irrespective of PD-L1 expression or histology in Japan. There are few reports that include real-world data on large-scale cohorts of patients with special backgrounds from Japan. Method Data from patients who initiated nivolumab treatment from April 1, 2016 to December 31, 2016 were collected at 23 regional medical institutions across Japan. Here we report retrospective analysis of all 901 pts data and focus on effectiveness and safety in special subgroups of the Japanese NSCLC patients, from a multicenter, observational study; CA209-9CR (NCT03273790). Result Patients with EGFR mutations showed significantly shorter PFS as compared to EGFR wild-type patients. Among other special groups, 316 pts with high blood pressure, in which 272 pts was on antihypertensive, showed superior ORR as compared to 585 pts with normal pressure at baseline. Safety analysis showed higher pulmonary toxicity in patients with pulmonary infection at baseline and higher hepatotoxicity in a group of pts infected with hepatitis virus. In this presentation, further analysis on the other subgroups will be presented.

P1.16-48 Experiences of Pts on 1st Line Care (EP1C): Symptoms and Impacts of EGFR TKI Therapy on Real-World Daily Lives of NSCLC Pts

P1.16-48 Experiences of Pts on 1st Line Care (EP1C): Symptoms and Impacts of EGFR TKI Therapy on Real-World Daily Lives of NSCLC Pts

1486PD - Exposure-response analyses of ALK-inhibitors crizotinib and alectinib in NSCLC patients

BackgroundCrizotinib and alectinib are first and second generation ALK-inhibitors, respectively, indicated for the treatment of ALK+ NSCLC. At the currently used fixed doses, interpatient variability in exposure is high (40%), while the registration studies suggest that exposure might be linked to efficacy (FDA Reviews). Therefore, the aim of this study was to explore whether exposure to crizotinib and alectinib is related to efficacy in a real life patient (pt) cohort. MethodsA retrospective observational study was performed. ALK+ NSCLC pts who were treated with crizotinib or alectinib and of whom pharmacokinetic (PK) samples were drawn as part of routine care were included. Calculated minimum plasma concentration (Cmin) was taken as a measure of exposure. Efficacy endpoint was progression free survival (PFS). Univariable and multivariable exposure-response analyses were performed using previously proposed thresholds of 235 µg/L for crizotinib and 435 µg/L for alectinib (Verheijen 2017). Variables taken into account were WHO performance status (PS) and previous lines of therapy for crizotinib, and WHO PS and previous treatment with ALK-inhibitors for alectinib. ResultsIn total, 100 pts were included in this study (48 crizotinib, 52 alectinib), of whom 376PK samples were eligible for analysis (235 crizotinib, 141 alectinib). Median number of samples per pt was 3 (range 1 – 15). Median Cmin per pt was 244 µg/L (IQR 176 – 344 µg/L) for crizotinib and 517 µg/L (IQR 369 – 659 µg/L) for alectinib. For crizotinib, median PFS was 5.7 vs 17.4 months (mo) for pts with median Cmin < 235 µg/L (48%) and ≥ 235 µg/L, respectively (log rank test, p=0.08). In multivariable analysis Cmin < 235 µg/L resulted in a HR of 1.79 (95% CI 0.90 – 3.59, p=0.10). For alectinib, median PFS was 12.6mo vs not reached yet after 20mo for pts with Cmin < 435 µg/L (37%) and ≥ 435 µg/L, respectively (log rank test, p=0.04). Multivariable analysis resulted in a HR of 3.86 (95% CI 1.19 – 12.58, p=0.025). Thus far, this could not be explained by bias. ConclusionsThis study shows that exposure to alectinib is associated with efficacy in a real life patient cohort. Therefore, therapeutic drug monitoring might be appropriate to individualize treatment and improve treatment outcomes. Legal entity responsible for the studyNetherlands Cancer Institute - Antoni van Leeuwenhoek Hospital. FundingHas not received any funding. DisclosureS.A. Burgers: Advisory / Consultancy: AstraZeneca; Advisory / Consultancy: Roche Advisory / Consultancy: MSD. N. Steeghs: Research grant / Funding (institution): AstraZeneca/MedImmune; Research grant / Funding (institution): Bayer; Research grant / Funding (institution): Bristol-Myers Squibb; Research grant / Funding (institution): Novartis; Research grant / Funding (institution): GlaxoSmithKline; Research grant / Funding (institution): Pfizer; Research grant / Funding (institution): Roche; Research grant / Funding (institution): Genentech; Research grant / Funding (institution): Boehringer Ingelheim; Research grant / Funding (institution): Blueprint Medicines; Research grant / Funding (institution): AB Science; Research grant / Funding (institution): Deciphera; Research grant / Funding (institution): Merck Sharp & Dohme; Research grant / Funding (institution): Amgen; Research grant / Funding (institution): Merus; Research grant / Funding (institution): Lilly; Research grant / Funding (institution): Incyte. All other authors have declared no conflicts of interest.

P1.01-43 Programmed-Death Ligand 1 Spectrum in a Large Cohort of Genetically Characterized Non-Small Cell Lung Cancer Patients

Programmed-death ligand 1 (PD-L1) expression, assessed by immunohistochemistry (IHC), is used to select patients (pts) for checkpoint inhibitors. However, the pattern of PD-L1 expression across rare oncogenic alterations remains poorly characterized. We aimed to evaluate PD-L1 expression in a large dataset of pts with activating molecular alterations to define the potential responsiveness to immunotherapy. From 2016 to 2018, we prospectively characterized advanced NSCLC pts from a single institution. Tumor Proportion Score (TPS) for PD-L1 protein expression was evaluated by IHC (22C3 clone) using formalin-fixed paraffin-embedded tumor samples and scored into <1% (negative), 1-49% (weakly positive) and ≥50% (high). DNA and RNA were extracted and analysed by Oncomine Solid Tumour panel (22 genes) and a customized nCounter-based panel (ALK, ROS1, RET, NTRK1, METΔ14). Associations between PD-L1 expression and the most prevalent driver alterations were assessed. Smoking habit and its possible relationship with PD-L1 expression was also appraised. Statistical analyses were performed using Kruskal-Wallis and Mann-Whitney tests. TPS for PD-L1 expression was available for a total of 140 patients (pts) fully genotyped. The cohort included: 36% women; 89% adenocarcinoma; 87% smokers. TPS for PD-L1 expression was negative (<1%), weak (1-49%) and high (≥50%) in 40%, 35% and 25% of pts, respectively. Actionable drivers were found in 84 pts (60%) being KRAS (n=43.31%) the most commonly detected, followed by EGFR (n=22.16%), BRAF (n=7.5%), METΔ14 (n=8.6%), ALK (n=3.2%) and ERBB2 (n=1, 1%). Thirty (21.4%) [ME1] pts harboured TP53 co-mutations. By comparing all genotyped cohorts, METΔ14 alteration was associated with higher PD-L1 expression levels compared with other subgroups (median TPS 58.62 vs 25.26, p=0.017[ME2] ). In addition, PD-L1 expression was also higher in pts harbouring any TP53 co-mutation than those with any alteration but TP53-WT (median TPS 41.53 vs 21.45, p=0.035). When KRAS-mut, BRAF-mut and EGFR-mut were evaluated separately, PD-L1 expression was higher in TP53 mutated tumors compared to TP53 WT only in BRAF-mut (p=0.028). Finally, no significant differences were found regarding patients’ smoking status (p=0.527). Our results suggest differential expression of PD-L1 based on the presence of MET alterations and TP53 mutations and highlight the need of further characterizing PD-L1 expression across oncogenic alterations.

P1.14-47 ctDNA NGS for Guiding Crizotinib Treatment in ALK-Rearranged Advanced NSCLC Patients (Pts)

FISH and IHC are commonly used and the golden standard for testing ALK rearrangement. However, they provide no information on fusion types or mutation status of other genes, which could be important prognostic factors. In addition, a tissue biopsy is not always applicable/preferred by pts. 43 ALK-rearranged treatment naïve advanced NSCLC pts were included. ALK-rearrangement was confirmed by ctDNA NGS with a panel of 59 tumor-related genes in 25 pts (ctDNA arm) and by IHC/FISH/tissue NGS in 18 pts (control arm). All pts received crizotinib as first-line therapy. PFS was estimated using Kaplan-Meier method and compared using log-rank test. On average, Pts in ctDNA arm were significantly older at diagnosis than pts in the control arm (56 vs. 45, p = 0.02), indicating the preferability of ctDNA NGS in elderly pts. Though older, pts in ctDNA arm had a similar mPFS with pts in the control arm (8 vs. 10 mo, p = 0.9). ALK fusion types in ctDNA arm and their associated statistics were listed in Table 1. Concomitant mutations were identified in 15 genes in 9 pathways. Frequently mutated pathways include p53 (in 8 pts) and DNA repair pathways (in 3 pts). Mutations in DNA repair pathway (5 vs. 10 mo, p = 0.06), high frequency of EML4-ALK (≥1.5%) (6 mo vs. NR, p = 0.06), and bone metastasis (6 vs. NR, p = 0.05) were associated with shorter mPFS.Table 1ALK fusion types and associated statisticssALK fusion types# of pts(Median) age at diagnosis (year)(m)PFS (mo)Average # of mutations per ptV11152101.5V3106162.8V2160-1V5'165-1V714932DCTN1-ALK147-2 Open table in a new tab Information on ALK fusion types, concomitant mutations, and mutation frequencies provided by NGS could be valuable prognostic factors and deserves further investigation. ctDNA NGS could be used as an effective alternative to identify ALK+ pts, especially for elderly pts, when tissue biopsy is inapplicable or not preferred.

Exploratory Analysis of Nivolumab Effectiveness with Lab Data in NSCLC pts in Japan: A Multicenter-Retrospective Study

Abstract Background In December 2015, nivolumab was approved for previously treated advanced NSCLC irrespective of PD-L1 expression or histology in Japan. Real-world data can provide insights beyond the clinical trial, and in this analysis we study a large Japanese patient cohort. Methods Data for 901 patients who initiated nivolumab treatment for advanced NSCLC from April 1, 2016 to December 31, 2016 were collected at 23 regional medical institutions across Japan, including clinical laboratory data before and all the available data during nivolumab treatment. Here we report an exploratory analysis on the association of nivolumab effectiveness with the clinical laboratory data of the Japanese advanced NSCLC patients from a multicenter, observational study; CA209-9CR (NCT03273790). Results Analysis on association of objective response rate (ORR) with each clinical laboratory data at baseline among responders and non-responders suggested that neutrophils, lymphocytes, neutrophil/lymphocyte ratio (NLR), albumin, sodium, CYFRA, and uric acid at baseline were associated with significantly improved ORR. Further analyses on those multiple factors are planned to identify associations with improved outcome of with nivolumab treatment beyond ORR.

Tepotinib in NSCLC patients with METex14 mutations: interim results from the phase II VISION study

Abstract Background Activating MET exon 14 skipping (METex14) mutations can be detected by liquid biopsy (LBx) or tumor biopsy (TBx) and are sensitive to MET inhibition. We report interim data from the phase II VISION study of tepotinib, a highly selective MET inhibitor, in NSCLC patients with METex14 mutations (NCT02864992). Methods Patients (pts) with advanced EGFR/ALK wild-type NSCLC harboring METex14 mutation identified by RNA-based testing with LBx or TBx (both ≥60 pts; overlap of LBx/TBx anticipated), enrolled at 113 sites in 8 countries (13 sites in Japan), receive tepotinib 500mg QD until progression, intolerable toxicity or withdrawal. Primary endpoint is ORR by independent review committee (IRC); secondary endpoints include investigator assessed ORR (INV) and safety. Results To date, 90 pts are enrolled, 15 are from Japan. At data cut-off (16/10/2018), in 35 evaluable LBx pts ORR (95% CI) was 51.4% (34.0, 68.6) by IRC with a median duration of response (mDoR) (range) of 9.8 months (1.1, 18.0) and 63.9% (46.2, 79.2) by INV with a mDOR of 17.1 months (1.3, 21.5). In 41 evaluable TBx pts, ORR was 41.5% (26.3, 57.9) by IRC and 58.5% (42.1, 73.7) by INV; mDoR was 12.4 months (1.1, 18.0; IRC) and 14.3 months (1.3, 21.5; INV). Of 11 evaluable Japanese pts, 6 were METex14 positive by LBx and 9 by TBx (overlap occurred). Confirmed responses were reported in 4/6 LBx pts and 3/9 TBx pts by IRC and in 4/6 LBx pts and 4/9 TBx pts by INV. Safety was evaluable in 69 pts. Main treatment-related adverse events (TRAEs; ≥15%) of any grade were peripheral edema (47.8%), diarrhea (18.8%), nausea (15.9%). The safety profile was similar in Japanese pts. TRAEs led to permanent discontinuation in 2 (2.9%) pts (1 ILD, 1 diarrhea & nausea). No TRAEs led to death. Conclusion In NSCLC pts with METex14 mutations, tepotinib had promising clinical efficacy. Similar activity was observed in Japanese pts. The safety profile was favorable. Recruitment, including a MET amplification cohort, is ongoing.

Primary results from Japanese phase I study of pembrolizumab plus chemotherapy as front-line therapy for advanced NSCLC

Abstract Background The anti¯PD¯1 antibody pembrolizumab (pembro) exhibits robust antitumor activity in patients with advanced NSCLC. In part B¯C of KEYNOTE¯011 (NCT01905657), we evaluated the efficacy and safety of pembro + chemotherapy combinations in advanced NSCLC. Here we report primary results from both studies. This is the longest follow-up data on treatment of pembro plus standard chemotherapy regimens in Japanese NSCLC patients. Methods Chemotherapy-naive, advanced, EGFR/ALK (¯) NSCLC pts, PS 0 or 1, were treated to pembro plus either cisplatin or carboplatin + pemetrexed (PEM) (part B, non-squamous), or carboplatin and paclitaxel or nab-paclitaxel (part C, squamous) for 4 cycles followed by maintenance pembro + PEM (B), pembro (C). Results In part B, 12 pts (6 in cisplatin cohort, 6 in carboplatin cohort) had been treated; median (range) follow-up duration was 16.0 mo (3.7 ¯ 21.2 mo). 1 DLT occurred in cisplatin cohort (Grade 4 Hyponatremia) during the DLT evaluation period. At the time of data cutoff, ≥Gr 3 treatment-related AEs occurred in 75.0%, 2 treatment¯related death occurred (2 pneumonitis in carboplatin cohort). ORR was 66.7% and 80.0% in cisplatin and carboplatin cohort, respectively. ORR with cohorts combined in TPS ≥50%, TPS = 1¯49%, TPS Conclusions Pembro plus standard chemotherapy regimens is feasible and yields substantial clinical efficacy regardless of PD¯L1 status in treatment-naive advanced NSCLC. Detailed survival and safety will be presented.

479P - Safety profile of tepotinib in patients with advanced solid tumors: Pooled analysis of phase I and II data

BackgroundTepotinib is an oral, selective, potent MET tyrosine kinase inhibitor, with promising clinical activity in patients (pts) with MET+ tumors, being tested in NSCLC harboring METex14 mutations or MET amplification. We aimed to further characterize the safety profile of tepotinib. MethodsSafety data from pts receiving tepotinib 500mg once daily (QD) from five studies (NCT01014936, NCT01832506, NCT01988493, NCT02115373, NCT02864992) were pooled and summarized. Pts received tepotinib until disease progression, intolerable toxicity, or other withdrawal. Adverse events (AEs) were graded by NCI CTCAE v4.03. ResultsAs of 9/28/2018, 422 pts had received tepotinib; 228 received 500mg tepotinib monotherapy QD and were further analyzed for safety (73.2% male, 50.9%/35.5% white/Asian, median age 65.0 [range 19–89] yrs). Median treatment duration was 2.7 months (range 0.0–21.5); in NSCLC pts with METex14 mutations it was 4.1 months (NCT02864992). 219 pts (96.1%) had an AE (any cause). 122 pts (53.5%) had a Grade ≥3 any cause AE including (>3.5%): disease progression (n=18, 7.9%), hyponatremia (n=12, 5.3%), increased aspartate aminotransferase (AST) (n=12, 5.3%), increased lipase (n=11, 4.8%), peripheral edema (n=9, 3.9%) and ascites (n=9, 3.9%). The incidence rates were: 19.0, 12.9, 12.9, 12.0, 9.6 and 9.6 per 100pt-years, respectively. Treatment-related (TR)AEs are shown in the Table. 18 pts (7.9%) had a serious TRAE including (≥1%) peripheral edema (n=4, 1.8%). AEs (any cause) led to dose reduction of tepotinib in 33 pts (14.5%) and discontinuation in 49 pts (21.5%), and was due to peripheral edema in 5 (8.5%) and 7 (3.1%) pts, respectively. Of 25 AEs leading to death, two were considered treatment related by the investigator (upper GI hemorrhage and hypoglycemic coma). ConclusionsTepotinib demonstrated an acceptable safety profile across a range of tumor types and is being developed in NSCLC as monotherapy and in combination. Further characterization of peripheral edema is ongoing.Table479PTableN=228Any Grade (≥10%)Grade ≥3 (≥2%)Any TRAE, n (%)172 (75.4)52 (22.8)Peripheral edema77 (33.8)8 (3.5)Diarrhea45 (19.7)4 (1.8)Fatigue34 (14.9)3 (1.3)Nausea29 (12.7)0 (0)Decreased appetite27 (11.8)0 (0)Increased lipase13 (5.7)9 (3.9)Increased AST11 (4.8)5 (2.2) Clinical trial identificationNCT01014936, NCT01832506, NCT01988493, NCT02115373, NCT02864992. Editorial acknowledgementMedical writing assistance (funded by Merck Healthcare KGaA, Darmstadt, Germany) was provided by Matthew deSchoolmeester, PhD of Bioscript Science (Macclesfield, UK). Legal entity responsible for the studyMerck KGaA. FundingMerck KGaA. DisclosureT. Decaens: Honoraria (self), Advisory / Consultancy, Travel / Accommodation / Expenses: BMS; Honoraria (self), Travel / Accommodation / Expenses: Bayer; Honoraria (self), Advisory / Consultancy, Travel / Accommodation / Expenses: AstraZeneca; Honoraria (self), Advisory / Consultancy: IPSEN; Honoraria (self), Travel / Accommodation / Expenses: Gilead; Honoraria (self), Travel / Accommodation / Expenses: Abbvie; Honoraria (self): Sanofi; Advisory / Consultancy: SITEX; Research grant / Funding (institution): ArQule; Research grant / Funding (institution): Genoscience Pharma; Research grant / Funding (institution), Travel / Accommodation / Expenses: Roche; Travel / Accommodation / Expenses: MSD; Travel / Accommodation / Expenses: Merck Serono; Advisory / Consultancy: BTG; Advisory / Consultancy: Sirtex. R. Veillon: Advisory / Consultancy, Speaker Bureau / Expert testimony: MSD; Advisory / Consultancy: Pfizer; Advisory / Consultancy: Novartis; Speaker Bureau / Expert testimony, Research grant / Funding (institution): BMS; Speaker Bureau / Expert testimony, Research grant / Funding (institution): Roche; Research grant / Funding (institution): Takeda; Research grant / Funding (institution): Abbvie; Research grant / Funding (institution): Merck. K. Yamazaki: Honoraria (self): Chugai; Honoraria (self): Takeda; Honoraria (self): Yakult; Honoraria (self): Daiichi‐sankyo; Honoraria (self): Merck Serono; Honoraria (self): Bristol; Honoraria (self): Bayer; Honoraria (self): Eli Lilly; Honoraria (self): Taiho; Research grant / Funding (institution): BMS; Research grant / Funding (institution): Sanofi. J. Scheele: Full / Part-time employment: Merck KGaA. R. Bruns: Full / Part-time employment: Merck KGaA. K. Berghoff: Full / Part-time employment: Merck KGaA. S. Faivre: Honoraria (self), Advisory / Consultancy, Research grant / Funding (institution), Travel / Accommodation / Expenses: Blueprint; Honoraria (self), Advisory / Consultancy, Research grant / Funding (institution), Travel / Accommodation / Expenses: Bristol-Myers Squibb; Honoraria (institution), Advisory / Consultancy, Research grant / Funding (institution), Travel / Accommodation / Expenses: Bayer Pharma; Honoraria (institution), Advisory / Consultancy, Research grant / Funding (institution), Travel / Accommodation / Expenses: Eli Lilly; Honoraria (institution), Advisory / Consultancy, Research grant / Funding (institution), Travel / Accommodation / Expenses: Ipsen; Honoraria (self), Advisory / Consultancy, Research grant / Funding (institution), Travel / Accommodation / Expenses: Merck Serono; Honoraria (self), Advisory / Consultancy, Research grant / Funding (institution), Travel / Accommodation / Expenses: MSD; Honoraria (self), Advisory / Consultancy, Research grant / Funding (institution), Travel / Accommodation / Expenses: Novartis. P.P. Paik: Advisory / Consultancy: Bristol-Myers Squibb; Honoraria (self), Advisory / Consultancy: Lilly; Honoraria (self), Advisory / Consultancy: Abbvie; Honoraria (self), Advisory / Consultancy, Research grant / Funding (institution): Celgene; Research grant / Funding (institution), Travel / Accommodation / Expenses: EMD Serono; Honoraria (self): Takeda. All other authors have declared no conflicts of interest.

1576P - Efficacy results of selective AXL inhibitor bemcentinib with pembrolizumab following chemo in patients with NSCLC

1576P - Efficacy results of selective AXL inhibitor bemcentinib with pembrolizumab following chemo in patients with NSCLC

P2.01-17 CANOPY-1: Phase 3 Study of Canakinumab/Placebo+Pembrolizumab+Platinum-Chemotherapy in Untreated Stage IIIB-IV NSCLC Pts

Interleukin-1β (IL-1β) inhibition with canakinumab reduced the incidence of and mortality due to lung cancer among patients with atherosclerosis in CANTOS trial. Inhibition of IL-1β driven inflammation may lead to a tumor microenvironment more susceptible to anti-PD-(L)1 therapies. Recent studies have shown that low levels of CRP at baseline or decreased levels over time correlated with improved responses to anti-PD-(L)1 agents, providing rationale for combination of canakinumab and Pembrolizumab (PEM). CANOPY-1 (NCT03631199) is a double-blind, randomized, placebo (Pb)-controlled, phase III trial to determine efficacy and safety of PEM + platinum-based chemotherapy (Ctx) ± canakinumab in untreated stage IIIB/IIIC-IV squamous and non-squamous NSCLC pts. It is a 2 part study- In Part 1 [open-label safety run-in with 3 cohorts of ∼9 pts each to confirm recommended phase 3 canakinumab regimen], pts will receive canakinumab 200 mg s.c (Q3W) + PEM 200 mg i.v (Q3W) + platinum-based Ctx [Cohort A (non-squamous), carboplatin (CBCDA) + pemetrexed (PTX); Cohort B (non-squamous), cisplatin + PTX; Cohort C (squamous or non-squamous), CBCDA + paclitaxel]. In Part 2 [with ∼600 pts) to evaluate efficacy and safety of canakinumab combination], pts will be randomized to receive canakinumab/Pb + PEM + platinum-based Ctx (non-squamous, CBCDA or cisplatin + PTX; squamous, CBCDA + paclitaxel or nab-paclitaxel). PEM and platinum-based Ctx will be administered at their approved doses. Randomization (1:1) will be stratified by PD-L1 status, region and histology. In both parts, pts will receive 4 cycles of induction therapy (canakinumab/Pb + PEM + Ctx) followed by maintenance therapy (PEM + canakinumab/Pb +/- PTX) until progressive disease. Primary objectives: confirm recommended phase 3 regimen for canakinumab combination (Part 1), compare PFS and OS between treatment arms (Part 2). Secondary objectives (Part 1 and 2): ORR, DCR, safety, PK and DOR. Section not applicable Section not applicable