Developed in association with the European Thoracic Oncology Platform

Letter from Prof Rolf Stahel Dear Colleagues It is my pleasure to present this ETOP slide set which has been designed to highlight and summarise key findings in thoracic cancers from the major congresses in 2019. This slide set specifically focuses on the ESMO 2019 Congress and is available in 4 languages – English, French, Chinese and Japanese. The area of clinical research in oncology is a challenging and ever changing environment. Within this environment we all value access to scientific data and research which helps to educate and inspire further advancements in our roles as scientists, clinicians and educators. I hope you find this review of the latest developments in thoracic cancers of benefit to you in your practice. If you would like to share your thoughts with us we would welcome your comments. Please send any correspondence to etop@etop.eu-org. I would like to thank our ETOP members Drs Solange Peters and Martin Reck for their roles as Editors – for prioritising abstracts and reviewing slide content. The slide set you see before you would not be possible without their commitment and hard work. Finally, we are also very grateful to Lilly Oncology for their financial, administrative and logistical support in the realisation of this complex yet rewarding activity. Yours sincerely, Rolf Stahel President, ETOP Foundation Council

ETOP Medical Oncology Slide Deck Editors 2019 Focus: advanced NSCLC (not radically treatable stage III & stage IV) and related biomarker data Dr Solange Peters Multidisciplinary Oncology Center, Lausanne Cancer Center, Lausanne, Switzerland Focus: early and locally advanced NSCLC (stage I–III) and related biomarker data plus other malignancies Dr Martin Reck Department of Thoracic Oncology, Hospital Grosshansdorf, Grosshansdorf, Germany

Contents Screening and biomarkers Early stage and locally advanced NSCLC – Stages I, II and III Advanced NSCLC – Not radically treatable stage III and stage IV First line Later lines Other malignancies SCLC, mesothelioma and thymic epithelial tumors

Screening and biomarkers

LBA17: Harmonization study of tumour mutational burden determination in non-small-cell lung cancer (NSCLC) – Garrido-Martin EM, et al Study objective To evaluate the performance of two gene panel assays for the detection of TMB, and investigate the correlation of calculated TMBs with PD-L1 expression in NSCLC Methods Tumour samples (n=100) were collected from patients with early stage NSCLC from 22 hospitals in Spain TMB was assessed using two NGS panels (TSO500a and TMLb) and compared with the standard reference method (F1CDxc) TMB was calculated as the total number of mutations (synonymous and non- synonymous) per exonic Mb aSNVs and indels; bSNVs; cSNVs, indels, CNAs, select rearrangements Garrido-Martin EM, et al. Ann Oncol 2019;30(suppl):Abstr LBA17

LBA17: Harmonization study of tumour mutational burden determination in non-small-cell lung cancer (NSCLC) – Garrido-Martin EM, et al Key results TMB values for 96 early stage NSCLC tumours Total TMB (F1CDx) Total TMB (TSO500 H12O) 100 80 60 40 20 R2= 0.8775 N=96 Total TMB (F1CDx) Total TMB (TML HMS) 100 80 60 40 20 R2= 0.8119 N=96 Total TMB (TSO500 H12O) Total TMB (TML HMS) 100 80 60 40 20 R2= 0.8545 N=96 F1CDx Total TMB (muts/Mb) 100 80 60 40 20 N=96 TML TSO500 Garrido-Martin EM, et al. Ann Oncol 2019;30(suppl):Abstr LBA17

LBA17: Harmonization study of tumour mutational burden determination in non-small-cell lung cancer (NSCLC) – Garrido-Martin EM, et al Key results (cont.) In tumours with PD-L1 <1% there was a high correlation of TMB values between the reference F1CDx panel and the TSO500 and TML panels (TSO500, R2=0.9120; TML, R2=0.8768; n=55) Conclusions In NSCLC tumour samples, the detection of TMB values performed through TSO500 and TML panels was comparable with that obtained with the standard F1CDx F1DCx TSO500 TML Average of total TMB, muts/Mb 14 13 Median of total TMB, muts/Mb 10 9 Range of total TMB, muts/Mb 0–74 1–84 0–60 Total TMB ≥10 muts/Mb, % 47 42.2 46.1 Total TMB ≥13 muts/Mb, % 32.6 28.8 35.5 Total TMB ≥16 muts/Mb, % 25.9 19.2 22.1 Total TMB ≥20 muts/Mb, % 16.3 14.4 15.4 Garrido-Martin EM, et al. Ann Oncol 2019;30(suppl):Abstr LBA17

Early stage and locally advanced NSCLC – Stages I, II and III

1459PD: Efficacy of durvalumab in patients with Stage III NSCLC who experience pneumonitis (PACIFIC) – Vansteenkiste JF, et al Study objective To evaluate the impact of pneumonitis on the efficacy of durvalumab in patients with stage III NSCLC in the PACIFIC study Durvalumab 10 mg/kg q2w for up to 12 months (n=476) Key patient inclusion criteria Stage III, locally advanced, unresectable NSCLC Not progressed following platinum-based cCRT (≥2 cycles) WHO PS 0–1 (n=713) R 2:1 Stratification Age, sex, smoking history Placebo q2w for up to 12 months (n=237) Co-primary endpoints PFS (BICR by RECIST v1.1), OS Secondary endpoints ORR (BICR), DoR (BICR), TTDM (BICR), PFS2, PROs, safety Vansteenkiste JF, et al. Ann Oncol 2019;30(suppl):Abstr 1459PD

1459PD: Efficacy of durvalumab in patients with Stage III NSCLC who experience pneumonitis (PACIFIC) – Vansteenkiste JF, et al Key results In the multivariate analysis adjusted for the time-dependant occurrence of pneumonitis, the OS, PFS and TTDM were consistent with the ITT population No. of events/no. of patients (%) HR (95%CI)a for durvalumab vs. placebo Durvalumab Placebo Adjusted for the time-depended occurrence of pneumonitis ITT population OS Model 1 Model 2 183/476 (38.4) 116/237 (48.9) 0.70 (0.55, 0.88) 0.65 (0.51, 0.83) 0.68 (0.53, 0.87) - PFS 243/476 (51.1) 173/237 (73.0) 0.54 (0.45, 0.66) 0.52 (0.42, 0.64) 0.51 (0.41, 0.63) - TTDM 182/476 (38.2) 126/237 (53.2) 0.57 (0.45, 0.71) 0.53 (0.41, 0.67) 0.53 (0.41, 0.68) - aCox-proportional hazards model adjusted for time-depended occurrence of pneumonitis using two sets of covariates: model 1 (age, sex, smoking history) and model 2 (model 1 factors plus stage, histology, best response to prior therapy, WHO PS, region, race) Vansteenkiste JF, et al. Ann Oncol 2019;30(suppl):Abstr 1459PD

1459PD: Efficacy of durvalumab in patients with Stage III NSCLC who experience pneumonitis (PACIFIC) – Vansteenkiste JF, et al Key results (cont.) Conclusions In patients with unresectable, stage III NSCLC, improvements in OS, PFS and TTDM appear to be maintained with durvalumab vs. placebo, irrespective of on- study pneumonitis status There was a similar incidence of grade 3–4 AEs and deaths observed in the durvalumab compared with placebo arm; however, SAEs, discontinuations and irAEs were more common in patients with pneumonitis AEs by on-study pneumonitis status Pneumonitis present Pneumonitis absent Durvalumab (n=161) Placebo (n=58) Durvalumab (n=314) Placebo (n=176) Any grade, all causality AEs, n (%) Grade 3–4 Outcome of death Leading to discontinuation 161 (100) 49 (30.4) 9 (5.6) 41 (25.5) 58 (100) 12 (20.7) 7 (12.1) 11 (19.0) 299 (95.2) 106 (33.8) 12 (3.8) 32 (10.2) 164 (93.2) 54 (30.7) 8 (4.5) 12 (6.8) SAEs, n (%) 63 (39.1) 15 (25.9) 75 (23.9) 39 (22.2) irAEs, n (%) 68 (42.2) 17 (29.3) 48 (15.3) 2 (1.1) Vansteenkiste JF, et al. Ann Oncol 2019;30(suppl):Abstr 1459PD

Advanced NSCLC Not radically treatable stage III and stage IV First line

LBA4_PR: Nivolumab (NIVO) + low-dose ipilimumab (IPI) vs platinum-doublet chemotherapy (chemo) as first-line (1L) treatment (tx) for advanced non-small cell lung cancer (NSCLC): CheckMate 227 Part 1 final analysis – Peters S, et al Study objective To evaluate the final OS endpoint of nivolumab + low-dose ipilimumab as a 1L treatment for patients with advanced NSCLC and PDL ≥1% in CheckMate 227 Nivolumab 3 mg/kg q2w + ipilimumab 1 mg/kg q6w (n=187) PD/ toxicity/ up to 2 years PD-L1 exp. <1% (n=550) R 1:1:1 Nivolumab 360 mg q3w + histology-based chemotherapy (n=177) Key patient inclusion criteria Stage IV or recurrent NSCLC No prior systemic therapy No known sensitising EGFR/ALK alterations ECOG PS 0–1 Histology-based chemotherapy (n=186) Nivolumab 3 mg/kg q2w + ipilimumab 1 mg/kg q6w (n=396) PD-L1 exp. ≥1% (n=1189) R 1:1:1 Nivolumab 240 mg q2w (n=396) Stratification Tumour histology (NSQ vs. SQ) Histology-based chemotherapy (n=397) Primary endpoints OS in PD-L1 ≥1% population PFS in high TMB (≥10 mut/Mb) population Secondary endpoints OS and PFS in patients with PD-L1 expression <1%, OS in patients with PD-L1 ≥50% Peters S, et al. Ann Oncol 2019;30(suppl):Abstr LBA4_PR

LBA4_PR: Nivolumab (NIVO) + low-dose ipilimumab (IPI) vs platinum-doublet chemotherapy (chemo) as first-line (1L) treatment (tx) for advanced non-small cell lung cancer (NSCLC): CheckMate 227 Part 1 final analysis – Peters S, et al Key results OS: PD-L1 ≥1% OS: PD-L1 <1% Nivo + Ipi (n=396) Nivo (n=396) CT (n=397) mOS, months 17.1 15.7 14.9 HR (vs. CT)a CI 0.79 0.65, 0.96b 0.88 0.75, 1.04c Nivo + Ipi (n=187) Nivo + CT (n=177) CT (n=186) mOS, months 17.2 15.2 12.2 HR (vs. CT) CI 0.62 0.48, 0.78c 0.78 0.60, 1.02b OS, % 100 80 60 40 3 6 9 12 15 33 36 45 18 21 24 27 30 39 42 OS, % 100 80 60 40 3 6 9 12 15 33 36 45 18 21 24 27 30 39 42 40% 36% 33% 40% 35% 23% 63% 57% 56% 60% 59% 51% 20 Nivo + Ipi Nivo CT 20 Nivo + Ipi Nivo + CT CT No. at risk Time, months Time, months Nivo +Ipi 396 341 295 264 244 212 9 1 190 153 145 91 41 129 165 Nivo +Ipi 187 165 142 120 110 100 8 2 87 73 69 34 19 59 80 Nivo 396 330 299 265 220 201 10 2 176 139 129 70 36 115 153 Nivo +CT 177 159 139 119 102 88 1 78 60 48 23 9 40 67 CT 397 358 306 250 218 190 6 1 166 126 112 57 22 93 141 CT 397 358 306 250 218 190 6 1 166 126 112 57 22 93 141 aHR 0.90 (95%CI 0.76, 1.07) for Nivo + Ipi vs. Nivo; b97.72%CI; c95%CI From NEJM, Hellmann MD, et al, Nivolumab plus Ipilimumab in Advanced Non-Small-Cell Lung Cancer, doi: 10.1056/NEJMoa1910231. Copyright © (2019) Massachusetts Medical Society. Reprinted with permission from Massachusetts Medical Society. Peters S, et al. Ann Oncol 2019;30(suppl):Abstr LBA4_PR

LBA4_PR: Nivolumab (NIVO) + low-dose ipilimumab (IPI) vs platinum-doublet chemotherapy (chemo) as first-line (1L) treatment (tx) for advanced non-small cell lung cancer (NSCLC): CheckMate 227 Part 1 final analysis – Peters S, et al Key results (cont.) PD-L1 TPS ≥1% Nivolumab + ipilimumab (n=396) Chemotherapy (n=397) Nivolumab (n=396) Median PFS, months 5.1 5.6 4.2 HR (95%CI) vs. chemo 0.82 (0.69, 0.97) 0.99 (0.84, 1.17) ORR, n (%) 142 (35.9) 119 (30.0) 109 (27.5) Median DoR, months (95%CI) 23.2 (15.2, 32.2) 6.2 (5.6, 7.4) 15.5 (12.7, 23.5) PD-L1 TPS <1% Nivolumab + ipilimumab (n=187) Chemotherapy (n=177) Nivolumab + chemotherapy (n=186) 51 (27.3) 67 (37.9) 43 (23.1) 18.0 (12.4, 28.6) 4.8 (3.7, 5.8) 8.3 (5.9, 9.4) Peters S, et al. Ann Oncol 2019;30(suppl):Abstr LBA4_PR

LBA4_PR: Nivolumab (NIVO) + low-dose ipilimumab (IPI) vs platinum-doublet chemotherapy (chemo) as first-line (1L) treatment (tx) for advanced non-small cell lung cancer (NSCLC): CheckMate 227 Part 1 final analysis – Peters S, et al Key results (cont.) Median OS, months Nivo + Ipi (n=583) Chemo (n=583) HR HR (95%CI) Randomised groups Stratified PD-L1 All randomised (N=1166) PD-L1 <1% (n=373) PD-L1 ≥1% (n=793) 17.1 17.2 13.9 12.2 14.9 0.73 0.62 0.79a Additional exploratory subgroups analysesb,c Unstratified PD-L1 1–49% (n=393) PD-L1 ≥50% (n=397) 15.1 21.2 14.0 0.94 0.70 TMBd (mut/Mb) low <10 (n=380) high ≥10 (n=299) 16.2 23.0 12.6 16.4 0.75 0.68 0.25 0.5 1 2 aStratified HR 0.90 (97.72%CI 0.76, 1.07) for Nivo + Ipi vs. Nivo; bpatients not stratified by TMB or PD-L1, analyses therefore may be impacted by imbalances and should be interpreted with caution; cnot controlled by randomisation; dunstratified HR for Nivo + Ipi vs. chemo in TMB evaluable and non-evaluable patients was 0.74 (95%CI 0.61, 0.88) and 0.74 (95%CI 0.60, 0.92), respectively From NEJM, Hellmann MD, et al, Nivolumab plus Ipilimumab in Advanced Non-Small-Cell Lung Cancer, doi: 10.1056/NEJMoa1910231. Copyright © (2019) Massachusetts Medical Society. Reprinted with permission from Massachusetts Medical Society. Peters S, et al. Ann Oncol 2019;30(suppl):Abstr LBA4_PR

Improvement of OS was observed independent of PD-L1 expression LBA4_PR: Nivolumab (NIVO) + low-dose ipilimumab (IPI) vs platinum-doublet chemotherapy (chemo) as first-line (1L) treatment (tx) for advanced non-small cell lung cancer (NSCLC): CheckMate 227 Part 1 final analysis – Peters S, et al Key results (cont.) Conclusions In patients with advanced NSCLC and PD-L1 ≥1, the combination of nivolumab + ipilimumab presented with a highly manageable safety profile, and showed significant improvement in OS vs. chemotherapy Improvement of OS was observed independent of PD-L1 expression TRAEs in all patients, % Nivo + Ipi (n=576) Chemotherapy (n=570) Nivolumab (n=391) Any grade Grade 3–4 Any TRAEs 77 33 82 36 66 19 TRAEs leading to discontinuation 18 12 9 5 7 Most frequent TRAEs (≥15%) Diarrhoea Rash Fatigue Decreased appetite Nausea Anaemia Constipation Neutropenia 17 17 14 13 10 4 4 <1 2 2 2 1 <1 1 0 0 10 5 19 20 36 33 15 17 1 0 1 1 2 12 <1 10 12 11 11 7 6 3 2 <1 <1 1 <1 0 <1 <1 0 0 Treatment-related deaths 1 <1 Peters S, et al. Ann Oncol 2019;30(suppl):Abstr LBA4_PR

Comparator EGFR TKI* Gefitinib 250 mg/day or Erlotinib 150 mg/day LBA5_PR: Osimertinib vs comparator EGFR-TKI as first-line treatment for EGFRm advanced NSCLC (FLAURA): Final overall survival analysis – Ramalingam SS, et al Study objective To evaluate the final OS with osimertinib as a 1L treatment for patients with EGFRm advanced NSCLC in the FLAURA trial Key patient inclusion criteria Locally advanced or metastatic NSCLC Ex19del/L858R No prior EGFR TKI/systemic anti-cancer therapy Stable CNS metastases allowed WHO PS 0–1 (n=556) Osimertinib 80 mg/day (n=279) PD Stratification Mutation status (Ex19del/L858R) Race (Asian/non-Asian) R 1:1 Comparator EGFR TKI* Gefitinib 250 mg/day or Erlotinib 150 mg/day (n=277) PD Primary endpoint PFS (investigator-assessed RECIST v1.1) Secondary endpoints OS, ORR, DoR, DCR, depth of response, PROs, safety *Crossover to osimertinib was permitted upon central confirmation of progression and T790M positivity Ramalingam SS, et al. Ann Oncol 2019;30(suppl):Abstr LBA5_PR

LBA5_PR: Osimertinib vs comparator EGFR-TKI as first-line treatment for EGFRm advanced NSCLC (FLAURA): Final overall survival analysis – Ramalingam SS, et al Key results Osimertinib Comparator EGFR TKI PFS OS Probability of PFS 1.0 0.8 0.6 0.4 0.2 3 6 9 12 15 18 21 24 27 Time from randomisation, months Probability of OS 1.0 0.8 0.6 0.4 0.2 3 12 18 24 33 39 45 51 6 9 15 21 27 30 36 42 48 54 89% 74% 83% 59% 54% 44% Time from randomisation, months No.at risk Osimertinib 279 262 233 210 178 139 71 28 4 204 279 276 245 217 193 153 123 50 2 270 254 236 180 166 138 86 17 277 239 197 152 107 78 Comparator EGFR-TKI 37 10 2 277 263 219 182 148 121 101 40 2 252 239 205 165 138 131 110 72 17 Osimertinib Comparator EGFR TKI Median PFS, months (95%CI) 18.9 (15.2, 21.4) 10.2 (9.6, 11.1) HR (95%CI) 0.46 (0.37, 0.57) p-value 0.001 Osimertinib Comparator EGFR TKI Median OS, months (95%CI) 38.6 (34.5, 41.8) 31.8 (26.6, 36.0) HR (95%CI) 0.799 (0.641, 0.997) p-value 0.0462 Ramalingam SS, et al. Ann Oncol 2019;30(suppl):Abstr LBA5_PR

Favours comparator EGFR TKI LBA5_PR: Osimertinib vs comparator EGFR-TKI as first-line treatment for EGFRm advanced NSCLC (FLAURA): Final overall survival analysis – Ramalingam SS, et al Key results (cont.) OS across subgroups Subgroup Favours osimertinib Favours comparator EGFR TKI HR 95%CI Overall (n=556) Log-rank (primary) Unadjusted Cox PH Sex Male (n=206) Female (n=350) Age at screening <65 years (n=298) ≥65 years (n=258) Race Asian (n=357) Non-Asian (n=209) Smoking history Yes (n=199) No (n=357) CNS metastases at trial entry Yes (n=116) No (n=440) WHO PS 0 (n=228) 1 (n=327) EGFR mutation at randomisation Ex19del (n=349) L858R (n=207) EGFR mutation by ctDNA Positive (n=359) Negative (n=124) 0.799 0.641, 0.996 0.789 0.634, 0.983 0.794 0.554, 1.135 0.786 0.595, 1.037 0.723 0.539, 0.969 0.873 0.627, 1.215 0.995 0.752, 1.139 0.542 0.378, 0.772 0.699 0.485, 1.002 0.848 0.644, 1.118 0.832 0.530, 1.298 0.788 0.613, 1.014 0.927 0.629, 1.366 0.699 0.535, 0.913 0.679 0.509, 0.904 0.996 0.708, 1.404 0.773 0.601, 0.995 0.719 0.374, 1.359 HR for death (95%CI) 0.1 1.0 10.0 2.0 0.2 0.3 0.4 0.6 0.8 Ramalingam SS, et al. Ann Oncol 2019;30(suppl):Abstr LBA5_PR

LBA5_PR: Osimertinib vs comparator EGFR-TKI as first-line treatment for EGFRm advanced NSCLC (FLAURA): Final overall survival analysis – Ramalingam SS, et al Key results (cont.) In the comparator EGFR TKI arm, 85/180 (47%) patients who received subsequent treatment crossed over to osimertinib Grade ≥3 AEs that were possibly treatment-related occurred in 51/279 (18%) patients receiving osimertinib and 79/277 (29%) patients receiving a comparator EGFR TKI, with diarrhoea (3%) and anorexia (3%) being the most frequently reported grade ≥3 AEs in the osimertinib arm Conclusions In patients with EGFRm advanced NSCLC, osimertinib showed a statistically significant improvement in OS vs. comparator EGFR TKI as 1L treatment, with a favourable toxicity profile No difference in OS was seen in Asian patients and patients with Exon 21 mutation – these findings require further research Time to first subsequent therapy or death Events Median, months (95%CI) HR (95%CI) p-value Osimertinib 194 25.5 (22.0, 29.1) 0.478 (0.393, 0.581) 0.0001 Comparator EGFR TKI 242 13.7 (12.3, 15.7) Ramalingam SS, et al. Ann Oncol 2019;30(suppl):Abstr LBA5_PR

LBA78: IMpower110: Interim overall survival (OS) analysis of a Phase III study of atezolizumab (atezo) vs platinum-based chemotherapy (chemo) as first-line (1L) treatment (tx) in PD-L1–selected NSCLC – Spigel D, et al Study objective To evaluate the efficacy and safety of atezolizumab monotherapy as 1L treatment in patients with PD-L1 positive NSCLC Atezolizumab 1200 mg q3w Atezolizumab 1200 mg q3w Key patient inclusion criteria Squamous or non-squamous stage IV NSCLC Chemotherapy-naïve PD-L1 ≥1 (n=572) PD/ loss of clinical benefit Stratification Sex, ECOG PS, PD-L1 expression, histology R 1:1 NSQ: Cisplatin/carboplatin + pemetrexeda SQ: Cisplatin/carboplatin + gemcitabineb NSQ: pemetrexed SQ: BSC PD Primary endpoint OS (tested in a hierarchical manner according to PD-L1 expression status) Secondary endpoints PFS (investigator assessed), ORR, DoR aCisplatin 75 mg/m2 or carboplatin AUC6 + pemetrexed 500 mg/m2 q3w; bcisplatin 75 mg/m2 + gemcitabine 1250 mg/m2 or carboplatin AUC5 + gemcitabine 1000 mg/m2 q3w Spigel D, et al. Ann Oncol 2019;30(suppl):Abstr LBA78

LBA78: IMpower110: Interim overall survival (OS) analysis of a Phase III study of atezolizumab (atezo) vs platinum-based chemotherapy (chemo) as first-line (1L) treatment (tx) in PD-L1–selected NSCLC – Spigel D, et al Key results OS: TC3 or IC3 WT TC3 or IC3 WT* Atezolizumab (n=107) Chemotherapy (n =98) 6-mo OS, % (95%CI) 76.3 (68.2, 84.4) 70.1 (60.8, 79.4) 12-mo OS, % (95%CI) 64.9 (55.4, 74.4) 50.6 (40.0, 61.3) mOS, months (95%CI) 20.2 (16.5, NE) 13.1 (7.4, 16.5) HR (95%CI) 0.59 (0.40, 0.89) p-value 0.0106 Time, months OS, % 100 80 60 40 20 No. at risk 2 4 6 8 10 12 14 16 18 22 24 26 28 30 32 36 38 Median follow-up, 15.7 mo (range 0–35) 34 Atezolizumab 107 94 85 80 66 61 48 40 34 25 18 16 11 7 6 5 2 Chemotherapy 98 89 75 65 50 40 33 28 19 12 9 7 6 4 3 3 3 1 *PD-L1 expression evaluated with VENTANA SP142 IHC assay. TC3: TC≥ 50% PD-L1+ and IC3: IC ≥10% PD-L1+ Spigel D, et al. Ann Oncol 2019;30(suppl):Abstr LBA78

LBA78: IMpower110: Interim overall survival (OS) analysis of a Phase III study of atezolizumab (atezo) vs platinum-based chemotherapy (chemo) as first-line (1L) treatment (tx) in PD-L1–selected NSCLC – Spigel D, et al Key results (cont.) PFS: TC3 or IC3 WT TC3 or IC3 WT* Atezolizumab (n=107) Chemotherapy (n=98) 6-mo PFS, % (95%CI) 59.8 (50.4, 69.2) 38.3 (28.5, 48.1) 12-mo PFS, % (95%CI) 36.9 (27.0, 46.9) 21.6 (12.6, 30.6) mPFS, months (95%CI) 8.1 (6.8, 11.0) 5 (4.2, 5.7) HR (95%CI) 0.63 (0.45, 0.88) Time, months PFS, % 100 80 60 40 20 No. at risk 2 4 6 8 10 12 14 16 18 22 24 26 28 30 32 36 38 34 Atezolizumab 107 82 72 60 45 31 25 21 16 13 10 8 4 4 4 2 Chemotherapy 98 74 62 36 26 16 13 8 5 5 1 1 1 *PD-L1 expression evaluated with VENTANA SP142 IHC assay. TC3: TC≥ 50% PD-L1+ and IC3: IC ≥10% PD-L1+ Spigel D, et al. Ann Oncol 2019;30(suppl):Abstr LBA78

LBA78: IMpower110: Interim overall survival (OS) analysis of a Phase III study of atezolizumab (atezo) vs platinum-based chemotherapy (chemo) as first-line (1L) treatment (tx) in PD-L1–selected NSCLC – Spigel D, et al Key results (cont.) ORR and DoR: TC3 or IC3 WT Chemotherapy Atezolizumab Median DoR, months (range) NE (1.8*–29.3*) 6.7 (2.6–23.9*) Response, % 60 50 40 30 20 10 38.3% 28.6% PR CR *Censored Spigel D, et al. Ann Oncol 2019;30(suppl):Abstr LBA78

In patients with TC2/3 or IC 2/3, significant OS benefit was not met LBA78: IMpower110: Interim overall survival (OS) analysis of a Phase III study of atezolizumab (atezo) vs platinum-based chemotherapy (chemo) as first-line (1L) treatment (tx) in PD-L1–selected NSCLC – Spigel D, et al Key results (cont.) Conclusions In patients with TC3 or IC3 WT NSCLC, there were significant improvements in OS, PFS, ORR and DoR with 1L atezolizumab monotherapy compared with chemotherapy and there were no new safety signals In patients with TC2/3 or IC 2/3, significant OS benefit was not met n, % Atezolizumab (n=286) Chemotherapy (n=263) Pembrolizumab Gemcitabine Carboplatin Cisplatin Median treatment duration, months (mix-max) 5.3 (0–33) 3.5 (0–20) 2.6 (0–5) 2.3 (0–5) 2.1 (0–5) Any cause AEs 258 (90.2) 249 (94.7) Grade 3–4 AEs 91 (31.8) 141 (53.6) SAEs 81 (28.3) 75 (28.5) Grade 5 AEs 11 (3.8) 11 (4.2) AEs leading to treatment withdrawal 18 (6.3) 43 (16.3) Atezolizumab AE of special interest Grade 3–4 Atezolizumab AE of special interest 115 (40.2) 19 (6.6) 44 (16.7) 4 (1.5) AEs requiring use of corticosteroids 22 (7.7) 1 (0.4) Spigel D, et al. Ann Oncol 2019;30(suppl):Abstr LBA78

LBA81_PR: Phase II/III Blood First Assay Screening Trial (BFAST) in patients (pts) with treatment-naïve NSCLC: initial results from the ALK+ cohort – Gadgeel SM, et al Study objective To prospectively evaluate the clinical utility of predictive blood-based biomarkers in patients with treatment-naïve advanced or metastatic NSCLC treated with targeted therapy or immunotherapy ALK+ Alectinib 600 mg BID (n=87) PD/ toxicity/ withdrawal/ death Key patient inclusion criteria Unresectable, stage IIIb/IV NSCLC Treatment-naïve Measureable disease Blood sample positive for BFAST alteration ECOG PS 0–2 RET+ Alectinib 900/1200/750 mg BID PD bTMB+ Atezolizumab 1200 mg or chemotherapy PD/ loss of clinical benefit ROS1+ Entrectinib 600 mg BID PD Primary endpoint ORR (investigator assessed) Secondary endpoints ORR (BICR), DoR and PFS (investigator and BICR), safety Gadgeel SM, et al. Ann Oncol 2019;30(suppl):Abstr LBA81_PR

No baseline CNS mets (n=52) LBA71_PR: Phase II/III Blood First Assay Screening Trial (BFAST) in patients (pts) with treatment-naïve NSCLC: initial results from the ALK+ cohort – Gadgeel SM, et al Key results 6-month DoR was 90.4% (95%CI 83.7, 97.2) among confirmed responders 2-month PFS was 78.38% (95%CI 69.07, 87.69) ORR ORR: with or without CNS mets Response, % INV BICR 84.1, 96.7 78.5, 93.5 95%CI Median duration of follow-up: 12.58 months Response, % Baseline CNS No baseline (n=52) (n=35) Response, n (%) [95%CI] INV (n=87) BICR (n=87) CR 0 [0.00, 4.15] 11 (12.6) [6.48, 21.50] PR 76 (87.4) [78.50, 93.52] 69 (79.3) [69.29, 87.25] PD 1 (1.1) [0.03, 6.24] Response*, n (%) [95%CI] Baseline CNS mets (n=35) No baseline CNS mets (n=52) CR 0 [0.00, 10.00] 0 [0.00, 6.85] PR 32 (91.4) [76.94, 98.20] 44 (84.6) [71.92, 93.12] PD 1 (2.9) [0.07, 14.92] *Investigator assessed Gadgeel SM, et al. Ann Oncol 2019;30(suppl):Abstr LBA81_PR

LBA81_PR: Phase II/III Blood First Assay Screening Trial (BFAST) in patients (pts) with treatment-naïve NSCLC: initial results from the ALK+ cohort – Gadgeel SM, et al Key results (cont.) Conclusions This study assigned treatment through the results of blood-based NGS testing alone. In the cohort of patients with advanced or metastatic NSCLC identified as ALK+, treatment with alectinib resulted in an ORR of 87.4% and demonstrates the utility of blood-based NGS testing AEs, n (%) ALK+ cohort (n=87) Total number of all-grade AEs 716 Patient with at least one AE SAEs Grade 3–4 AEs Treatment-related SAEs Fatal AEs 87 (100) 21 (24) 30 (35) 5 (6) 1 (1) AEs leading to Treatment discontinuation Dose reduction Dose interruption 6 (7) 7 (8) 27 (31) Gadgeel SM, et al. Ann Oncol 2019;30(suppl):Abstr LBA81_PR

LBA82: Pembrolizumab (pembro) + chemotherapy (chemo) in metastatic squamous NSCLC: final analysis and progression after the next line of therapy (PFS2) in KEYNOTE-407 – Paz-Ares L, et al Study objective To update the final efficacy and PFS2 of pembrolizumab + chemotherapy in patients with metastatic squamous NSCLC in KEYNOTE-407 Pembrolizumab 200 mg q3w + carboplatin AUC6 q3w + paclitaxel 200 mg/m2 q3w OR nab-paclitaxel 100 mg/m2 q1w (4 cycles) (n=278) Pembrolizumab 200 mg q3w up to 31 cycles Key patient inclusion criteria Stage IV squamous NSCLC Treatment naïve No symptomatic brain metastases ECOG PS 0–1 (n=559) Stratification Taxane (paclitaxel vs. nab-paclitaxel) PD-L1 expression (TPS <1% vs. ≥1%) Geographical region (East Asia vs. rest of world) R 1:1 Placebo q3w + carboplatin AUC6 q3w + paclitaxel 200 mg/m2 q3w OR nab-paclitaxel 100 mg/m2 q1w (4 cycles) (n=281) Placebo q3w up to 31 cycles* Co-primary endpoints PFS (BIRC), OS Secondary endpoints PFS2, ORR, DoR, safety *Optional crossover to pembrolizumab (up to 35 cycles) permitted after PD Paz-Ares L, et al. Ann Oncol 2019;30(suppl):Abstr LBA82

Pembrolizumab + CT (n=278) LBA82: Pembrolizumab (pembro) + chemotherapy (chemo) in metastatic squamous NSCLC: final analysis and progression after the next line of therapy (PFS2) in KEYNOTE-407 – Paz-Ares L, et al Key results PFS2 Updated survival outcomes Pembrolizumab + CT (n=278) Placebo + CT (n=281) Median OS, months (95%CI) 17.1 (14.4, 19.9) 11.6 (10.1, 13.7) HR (95%CI) 0.71 (0.58, 0.88) Median PFS, months (95%CI) 8.0 (6.3, 8.4) 5.1 (4.3, 6.0) 0.57 (0.47, 0.69) ORR, % 62.6 38.4 Median DoR, months (range) 8.8 (1.3+ –28.4+) 4.9 (1.3+ –28.3+) Median, months (95%CI) 13.8 (12.2, 15.9) 9.1 (8.2, 10.2) HR 0.59 (95%CI 0.49, 0.72) PFS2, % No. at risk Time, months 100 80 60 40 20 Pembrolizumab + CT 278 256 227 189 157 127 98 10 4 67 39 81.7% 3 6 9 12 15 18 21 24 27 30 33 69.5% Placebo + CT 281 241 193 142 99 74 62 2 42 +No progressive disease before data cut-off date Paz-Ares L, et al. Ann Oncol 2019;30(suppl):Abstr LBA82

Pembrolizumab + CT (n=278) LBA82: Pembrolizumab (pembro) + chemotherapy (chemo) in metastatic squamous NSCLC: final analysis and progression after the next line of therapy (PFS2) in KEYNOTE-407 – Paz-Ares L, et al Key results (cont.) Conclusions In patients with metastatic squamous NSCLC, the combination of pembrolizumab with chemotherapy as 1L treatment demonstrated consistent improved outcomes in terms of OS, PFS, ORR and PFS2, with a manageable safety profile AEs, n (%) Pembrolizumab + CT (n=278) Placebo + CT (n=280) Any AE Grade 3–5 Leading to discontinuation Any treatment All treatments Leading to death Treatment-related 274 (99) 206 (74) 76 (27) 45 (16) 31 (11) 12 (4) 275 (98) 195 (70) 37 (13) 20 (7) 19 (7) 5 (2) Immune-related AEs and infusion reaction 98 (35) 37 (13) 25 (9) 9 (3) Paz-Ares L, et al. Ann Oncol 2019;30(suppl):Abstr LBA82

PD/toxicity/ loss of benefit LBA83: Final efficacy results from B-F1RST, a prospective Phase II trial evaluating blood-based tumour mutational burden (bTMB) as a predictive biomarker for atezolizumab (atezo) in 1L non-small cell lung cancer (NSCLC) – Socinski M, et al Study objective To evaluate the final efficacy of atezolizumab as a 1L treatment for patients with NSCLC Key patient inclusion criteria Stage IIIB/IVB locally advanced or metastatic NSCLC Immunotherapy naïve PD-L1 unselected Blood samples provided ECOG PS 0–1 (n=152) Atezolizumab 1200 mg q3w PD/toxicity/ loss of benefit Co-primary endpoints ORR (investigator assessed) PFS (investigator assessed) Secondary endpoints DoR, OS, safety Socinski M, et al. Ann Oncol 2019;30(suppl):Abstr LBA83

LBA83: Final efficacy results from B-F1RST, a prospective Phase II trial evaluating blood-based tumour mutational burden (bTMB) as a predictive biomarker for atezolizumab (atezo) in 1L non-small cell lung cancer (NSCLC) – Socinski M, et al Key results mOS in bTMB <16 group was 13.4 months (95%CI 11.7, 18.0) and in bTMB ≥16 group was 23.9 months (95%CI 8.8, NE) with HR 0.66 (90%CI 0.40, 1.10) ORR PFS bTMB subgroups bTMB ≥16 (n=28) bTMB <16 (n=91) mPFS, months (95%CI) 5.0 (1.6, 10.8) 3.5 (2.6, 4.3) HR (90%CI) 0.80 (0.54, 1.18) p-value 0.35 bTMB high bTMB Low ≥10 cut-off ≥16 cut-off ≥20 cut-off p<0.0001 100 p<0.0001 47.7% 80 bTMB ≥16 bTMB <16 PR 35.7% CR p=0.033 60 PFS, % ORR, % 20.4% 40 17.1% 12.6% 20 7.1% 5.5% 6.0% ITT (n=152) BEP (n=119) n=49 n=70 n=28 n=91 n=19 n=100 3 9 12 15 18 24 27 30 33 No. at risk Time, months 28 27 18 15 14 12 11 9 7 5 4 3 2 1 ≥16 <16 91 86 56 44 39 28 25 19 15 12 8 6 3 2 Socinski M, et al. Ann Oncol 2019;30(suppl):Abstr LBA83

No additional safety findings were observed LBA83: Final efficacy results from B-F1RST, a prospective Phase II trial evaluating blood-based tumour mutational burden (bTMB) as a predictive biomarker for atezolizumab (atezo) in 1L non-small cell lung cancer (NSCLC) – Socinski M, et al Key results (cont.) Conclusions In patients with NSCLC receiving 1L atezolizumab monotherapy, bTMB ≥16 confirmed favourable clinical outcomes in terms of PFS and ORR No additional safety findings were observed AEs, n (%) ITT population All grade AEs, any cause TRAEs 152 (100) 116 (76) All grade SAEs, any cause Treatment-related SAEs 81 (53) 22 (14) Grade 3–4 AEs, any cause Grade 3–4 TRAEs 86 (57) 30 (20) Grade 5 AEs, any cause Grade 5 TRAEs 10 (7) 1 (1) AEs leading to treatment discontinuation 27 (18) AEs of special interest (immune-related) 66 (43) Socinski M, et al. Ann Oncol 2019;30(suppl):Abstr LBA83

Nivolumab 480 mg q4w up to 1 year 1457PD: Efficacy evaluation of concurrent nivolumab addition to a first-line, concurrent chemo-radiotherapy regimen in unresectable locally advanced NSCLC – Results from the European Thoracic Oncology Platform (ETOP 6-14) NICOLAS phase II trial – Peters S, et al Study objective To evaluate the efficacy of nivolumab combined with 1L concurrent chemoradiotherapy in patients with unresectable locally advanced NSCLC Key patient inclusion criteria Unresectable locally advanced stage IIIA/B NSCLC Nodal status N2 or N3 ECOG PS 0–1 (n=79) Platinum-based chemotherapy* + radiotherapy 66 GY/33 fractions for 3 cycles + 4 doses nivolumab 360 mg q3w Nivolumab 480 mg q4w up to 1 year Primary endpoints Grade ≥3 pneumonitis-free rate, 1-year PFS rate Secondary endpoints Time to first grade ≥3 pneumonitis, ORR, OS, time-to-treatment failure, safety *Cisplatin + vinorelbine/etoposide/pemetrexed Peters S, et al. Ann Oncol 2019;30(suppl):Abstr 1457PD

1457PD: Efficacy evaluation of concurrent nivolumab addition to a first-line, concurrent chemo-radiotherapy regimen in unresectable locally advanced NSCLC – Results from the European Thoracic Oncology Platform (ETOP 6-14) NICOLAS phase II trial – Peters S, et al Key results PFS: Stage PFS: Histology Stage PFS events, n (%) 12-mo PFS, % (95%CI) Median, months (95%CI) Log-rank p-value IIIA 13 (46.4) 66.3 (45.2, 88.9) 27.4 (7.3, NE) 0.11 IIIB 33 (66.0) 50.0 (35.6, 62.8) 12.1 (8.9, 17.8) Histology PFS events, n (%) 12-mo PFS, % (95%CI) Median, months (95%CI) Log-rank p-value SQ 19 (67.9) 56.3 (35.9, 72.3) 13.5 (10.1, 22.0) 0.68 NSQ 26 (55.3) 54.5 (39.1, 67.5) 12.9 (7.2, NE) Time, months PFS, % 100 80 60 40 20 6 2 4 8 10 12 14 16 18 22 24 26 No. at risk 28 27 21 19 9 3 IIIA 50 46 41 38 34 30 25 17 13 11 7 1 IIIB 100 80 60 PFS, % 40 20 2 4 6 8 10 12 14 16 18 20 22 24 26 Time, months 28 27 25 23 21 19 15 11 10 9 8 1 SQ 47 43 39 35 30 27 25 19 12 10 8 6 4 3 NSQ Peters S, et al. Ann Oncol 2019;30(suppl):Abstr 1457PD

1457PD: Efficacy evaluation of concurrent nivolumab addition to a first-line, concurrent chemo-radiotherapy regimen in unresectable locally advanced NSCLC – Results from the European Thoracic Oncology Platform (ETOP 6-14) NICOLAS phase II trial – Peters S, et al Key results (cont.) Overall, pneumonitis was reported by 34 patients (7 grade 3, 1 grade 5), oesophagitis by 24 patients (5 grade 3) and dyspnoea by 27 patients (2 grade 3) In total, 240 nivolumab TRAEs were reported; 26 grade 3, 5 grade 4 and 4 grade 5 (colitis, pulmonary fibrosis, autoimmune disorder, pneumonitis). Of these, 7% (17/240 TRAEs) led to permanent discontinuation Conclusions In patients with unresectable locally advanced NSCLC combining nivolumab with concurrent chemoradiation is feasible, without any unexpected safety signal The PFS observed for combining nivolumab with concomitant definitive chemoradiation as 1L therapy compares favourably to other studies in the same patient population Peters S, et al. Ann Oncol 2019;30(suppl):Abstr 1457PD

1484PD: Final PFS, updated OS and safety data from the randomised, phase III ALEX study of alectinib (ALC) versus crizotinib (CRZ) in untreated advanced ALK+ NSCLC – Mok TSK, et al Study objective To evaluate the final efficacy and safety of alectinib as a 1L treatment for patients with ALK+ NSCLC in the ALEX study Alectinib 600 mg BID (n=152) Key patient inclusion criteria Stage IIIB/IV ALK+ NSCLC Treatment naive CNS metastases permitted if asymptomatic ECOG PS 0–2 (n=303) PD/ toxicity/ withdrawal R 1:1 Stratification ECOG PS, ethnicity, CNS metastases at baseline Crizotinib 250 mg BID (n=151) PD/ toxicity/ withdrawal Primary endpoint PFS (investigator assessed) Secondary endpoints PFS (IRC), ORR, time to CNS progression, DoR, OS, safety Mok TSK, et al. Ann Oncol 2019;30(suppl):Abstr 1484PD

PFS event-free rate: with CNS metastases 1484PD: Final PFS, updated OS and safety data from the randomised, phase III ALEX study of alectinib (ALC) versus crizotinib (CRZ) in untreated advanced ALK+ NSCLC – Mok TSK, et al 32.5 Alectinib 36 32 24 1 No. at risk Crizotinib 17 3 NE PFS event-free rate: with CNS metastases PFS, % Key results PFS Alectinib (n=152) Crizotinib (n=151) 100 80 60 40 20 HR 0.43 (95%CI 0.32, 0.58) p<0.0001 PFS, % 34.8 10.9 PFS event-free rate: without CNS metastases 6 12 18 24 30 36 42 48 57.2 Time, months No. at risk PFS, % Alectinib 152 135 113 109 98 84 81 79 61 49 39 14 3 76 69 68 NE Crizotinib 151 132 104 83 65 48 43 36 33 17 13 11 6 29 19 Alectinib 62 47 37 2 No. at risk Crizotinib 48 30 16 NE Mok TSK, et al. Ann Oncol 2019;30(suppl):Abstr 1484PD

1484PD: Final PFS, updated OS and safety data from the randomised, phase III ALEX study of alectinib (ALC) versus crizotinib (CRZ) in untreated advanced ALK+ NSCLC – Mok TSK, et al Key results (cont.) Conclusions In patients with ALK+ NSCLC, alectinib demonstrated consistent superior efficacy vs. crizotinib as 1L treatment, regardless of baseline CNS metastases and with a favourable safety profile AEs, n, % Alectinib (n=152) Crizotinib (n=151) All grade AEs 147 (97) SAEs 54 (36) 48 (32) Grade 3–5 AEs 74 (49) 83 (55) Fatal AE 6 (4) 7 (5) AEs leading to treatment discontinuation 21 (14) 22 (15) AEs leading to dose reduction 29 (19) 30 (20) AEs leading to dose interruption 38 (25) 39 (26) Mok TSK, et al. Ann Oncol 2019;30(suppl):Abstr 1484PD

Advanced NSCLC Not radically treatable stage III and stage IV Later lines

KEYNOTE-042 Pembrolizumab 200 mg q3w or platinum-based chemotherapy LBA79: Association between tissue TMB (tTMB) and clinical outcomes with pembrolizumab monotherapy (pembro) in PD-L1-positive advanced NSCLC in the KEYNOTE-010 and -042 trials – Herbst RS, et al Study objective To evaluate the association between tTMB levels and outcomes in the KEYNOTE-010 and KEYNOTE-042 trials tTMB assessment by whole exome sequencing of tumour tissue Outcomes (OS, PFS, ORR) association with tTMB assessed KEYNOTE-010 Pembrolizumab 2 or 10 mg/kg q3w or docetaxel 75 mg/m2 q3w (n=1034) Key patient inclusion criteria Advanced NSCLC PD-L1-positive PD-L1 TPS ≥1% ≥1 DNA samples KEYNOTE-042 Pembrolizumab 200 mg q3w or platinum-based chemotherapy (n=1275) aPre-specified cut-points of 175 mut/exome; assessed as continuous log10-transformed variable Herbst RS, et al. Ann Oncol 2019;30(suppl):Abstr LBA79

AUROC of ORR for tTMB (95%CI) AUROC of ORR for tTMB (95%CI) LBA79: Association between tissue TMB (tTMB) and clinical outcomes with pembrolizumab monotherapy (pembro) in PD-L1-positive advanced NSCLC in the KEYNOTE-010 and -042 trials – Herbst RS, et al Key results Association of tTMB with efficacy KEYNOTE-010 AUROC of ORR for tTMB (95%CI) Pembrolizumab 0.61 (0.50, 0.71) Chemotherapy 0.40 (0.21, 0.58) KEYNOTE-042 AUROC of ORR for tTMB (95%CI) Pembrolizumab 0.67 (0.61, 0.73) Chemotherapy 0.57 (0.50, 0.63) 1–Specificity Sensitivity 1.0 0.8 0.6 0.4 0.2 1–Specificity Sensitivity 1.0 0.8 0.6 0.4 0.2 Nominal p-valuea Pembrolizumab (n=164) Chemotherapy (n=89) OS 0.006b 0.410c PFS 0.001b 0.579c ORR 0.009b 0.330c Nominal p-valuea Pembrolizumab (n=414) Chemotherapy (n=379) OS <0.001b 0.060c PFS 0.174c ORR 0.035c aWald test; bone-sided; ctwo-sides; statistical significance determined at 0.05 level without multiplicity adjustment Herbst RS, et al. Ann Oncol 2019;30(suppl):Abstr LBA79

LBA79: Association between tissue TMB (tTMB) and clinical outcomes with pembrolizumab monotherapy (pembro) in PD-L1-positive advanced NSCLC in the KEYNOTE-010 and -042 trials – Herbst RS, et al Key results (cont.) There was no association between tTMB and PD-L1 in KEYNOTE-010 (pembrolizumab r=0.16; chemotherapy r=0.18) or KEYNOTE-042 (pembrolizumab r=0.05; chemotherapy r=0.04) Conclusions In previously treated or untreated patients with PD-L1+ NSCLC, high tTMB levels were associated with improved clinical outcomes in those receiving pembrolizumab monotherapy, suggesting a possible clinical use of tTMB as a predictive biomarker Herbst RS, et al. Ann Oncol 2019;30(suppl):Abstr LBA79

Pembrolizumab 200 mg/kg q3w + carboplatin + pemetrexed 500 mg/m2 q3w LBA80: Pembrolizumab (pembro) plus platinum-based chemotherapy (chemo) for metastatic NSCLC: tissue TMB (tTMB) and outcomes in KEYNOTE-021, 189, and 407 – Paz-Ares L, et al Study objective To evaluate the association of tTMB with efficacy of pembrolizumab + platinum- based chemotherapy vs. platinum-based chemotherapy alone tTMB assessment by whole exome sequencing of tumour tissue Outcomes (OS, PFS, ORR) association with tTMB assessed KEYNOTE-021 (Cohort C) Pembrolizumab 2 or 10 mg/kg q3w + carboplatin + pemetrexed 500 mg/m2 q3w (n=24) Key patient inclusion criteria Untreated stage IIIB/IV, squamous or non-squamous NSCLC ALK or EGFR negative ECOG PS 0–1 KEYNOTE-021 (Cohort G) Pembrolizumab 200 mg/kg q3w + carboplatin + pemetrexed 500 mg/m2 q3w (n=123) KEYNOTE-189 Pembrolizumab 200 mg q3w or placebo + carboplatin/cisplatin + pemetrexed 500 mg/m2 q3w (n=616) KEYNOTE-407 Pembrolizumab 200 mg q3w or placebo + carboplatin + paclitaxel/nab-paclitaxel q3w (n=559) Paz-Ares L, et al. Ann Oncol 2019;30(suppl):Abstr LBA80

There was no association between tTMB and efficacy LBA80: Pembrolizumab (pembro) plus platinum-based chemotherapy (chemo) for metastatic NSCLC: tissue TMB (tTMB) and outcomes in KEYNOTE-021, 189, and 407 – Paz-Ares L, et al Key results There was no association between tTMB and efficacy The ROC curve analysis for the association of tTMB with ORR suggested limited ability of tTMB to predict the treatment response, either with pembrolizumab + chemotherapy or chemotherapy alone, in all studies Pembrolizumab + chemotherapy had OS, PFS and ORR benefit vs. chemotherapy in both tTMB-high (≥175 mt/exome) and tTMB-low (<175 mut/exome) subgroups Nominal p-valuea KEYNOTE-021 C and G KEYNOTE-189 KEYNOTE-407 Pembrolizumab + CT (n=44) CT alone (n=26) Pembrolizumab + CT (n=207) Placebo + CT (n=86) Pembrolizumab + CT (n=143) Placebo + CT (n=169) ORR 0.180 0.279 0.072 0.434 0.393 0.086 PFS 0.187 0.409 0.075 0.055 0.052 0.560 OS 0.081 0.475 0.174 0.856 0.160 0.818 aP-values were values calculated using the Wald test and are one-sided for pembrolizumab + chemotherapy and two-sided for chemotherapy alone and placebo + chemotherapy Paz-Ares L, et al. Ann Oncol 2019;30(suppl):Abstr LBA80

LBA80: Pembrolizumab (pembro) plus platinum-based chemotherapy (chemo) for metastatic NSCLC: tissue TMB (tTMB) and outcomes in KEYNOTE-021, 189, and 407 – Paz-Ares L, et al Conclusions In patients with squamous and non-squamous NSCLC, no significant association between tTMB and efficacy was demonstrated for any treatment group Paz-Ares L, et al. Ann Oncol 2019;30(suppl):Abstr LBA80

Without brain metastases 1482O: Outcomes with pembrolizumab (pembro) monotherapy in patients (pts) with PD-L1–positive NSCLC with brain metastases: Pooled analysis of KEYNOTE-001, -010, -024, and -042 – Mansfield AS, et al Study objective To evaluate the efficacy and safety of pembrolizumab in patients with PD-L1- positive NSCLC and stable pretreated brain metastases at baseline in a post- hoc analysis of data pooled from KEYNOTE-001, -010, -024 and -042 Pembrolizumab* (n=199) Key patient inclusion criteria Advanced NSCLC Previously treated or treatment naive PD-L1 TPS ≥1% Data pooled from KEYNOTE-001, -010, -024 and -042 (n=3170) With brain metastases (n=293) Chemotherapy* (n=94) Pembrolizumab* (n=1754) Without brain metastases (n=2877) Chemotherapy* (n=1123) *KEYNOTE-001: pembrolizumab 2 or 10 mg/kg q3w or 10mg/kg q2w; KEYNOTE-010: pembrolizumab 2 or 10 mg/kg q3w vs. docetaxel 75 mg/m2 q3w; KEYNOTE-024: pembrolizumab 200 mg q3w vs. platinum chemotherapy; KEYNOTE-042: pembrolizumab 200 mg q3w vs. platinum chemotherapy Mansfield AS, et al. Ann Oncol 2019;30(suppl):Abstr 1482O

Without brain metastases 1482O: Outcomes with pembrolizumab (pembro) monotherapy in patients (pts) with PD-L1–positive NSCLC with brain metastases: Pooled analysis of KEYNOTE-001, -010, -024, and -042 – Mansfield AS, et al Key results With brain metastases Without brain metastases Pembrolizumab Chemotherapy PD-L1 TPS ≥50%, n 112 48 842 598 mOS, months (95%CI) HR (95%CI) 19.7 (12.1, 31.4) 0.78 (0.71, 0.85) 9.7 (7.2, 19.4) 19.4 (17.0, 22.4) 0.66 (0.58, 0.76) 11.7 (10.1, 13.1) PFS, months (95%CI) 4.1 (2.3, 10.6) 0.70 (0.47, 1.03) 4.6 (3.5, 8.4) 6.5 (6.1, 8.1) 0.69 (0.62, 0.78) 6.1 (5.8, 6.2) ORR, n (%) 38 (33.9) 7 (14.6) 321 (38.1) 156 (26.1) Median DoR, months (range) NR (4.0+–41.7+) 7.6 (2.9+–28.6+) 33.9 (1.4+–49.3+) 8.2 (1.6+–30.4+) DoR ≥12 months, n (%) 28 (91.7) 1 (25.0) 173 (74.0) 27 (38.1) PD-L1 TPS ≥1%, n 199 94 1754 1123 13.4 (10.4, 18.0) 0.83 (0.62, 1.10) 10.3 (8.1, 13.3) 14.8 (13.4, 16.1) 11.3 (10.2, 12.0) 2.3 (2.1, 3.9) 0.96 (0.73, 1.25) 5.2 (4.2, 8.3) 4.3 (4.2, 5.1) 0.91 (0.84, 0.99) 6.1 (6.0, 6.3) 52 (26.1) 17 (18.1) 452 (25.8) 249 (22.2) NR (3.3–46.2+) 8.3 (2.0+–28.6+) 30.4 (1.4+–49.3+) 8.1 (1.1+–30.4+) 36 (83.9) 3 (39.3) 245 (71.2) 39 (33.9) Mansfield AS, et al. Ann Oncol 2019;30(suppl):Abstr 1482O

1482O: Outcomes with pembrolizumab (pembro) monotherapy in patients (pts) with PD-L1–positive NSCLC with brain metastases: Pooled analysis of KEYNOTE-001, -010, -024, and -042 – Mansfield AS, et al Key results (cont.) Conclusions In patients with PD-L1+ NSCLC, the OS and PFS benefit of pembrolizumab was similar regardless of pretreated brain metastasis status at baseline and pembrolizumab improved clinical outcomes compared with chemotherapy alone in those with and without brain metastases Pembrolizumab monotherapy showed a manageable safety profile both in patients with and without brain metastases and no new safety signals were identified Pembrolizumab Chemotherapy With brain metastases (n=196) Without brain metastases (n=1743) With brain metastases (n=90) Without brain metastases (n=1066) Median treatment duration, months (range) 2.8 (0.03–39.6) 4.7 (0.03–75.9) 2.9 (0.03–29.5) 3.5 (0.03–34.8) TRAEs, % Grade 3–5 Led to death Led to discontinuation Affecting CNS 66 15 2 6 10 67 18 1 8 7 84 46 3 10 27 88 43 2 11 27 Immune-mediated AEs and infusion reactions, % 21 25 9 8 Mansfield AS, et al. Ann Oncol 2019;30(suppl):Abstr 1482O

1483PD: Pembrolizumab (pembro) plus platinum-based chemotherapy (chemo) in NSCLC with brain metastases: Pooled analysis of KEYNOTE-021, 189, and 407 – Powell SF, et al Study objective To evaluate the efficacy and safety of pembrolizumab + chemotherapy in patients with advanced/metastatic NSCLC and stable pretreated or asymptomatic brain metastases at baseline in a post-hoc analysis of data pooled from KEYNOTE- 021, -189 and -407 Pembrolizumab + Chemotherapy* (n=105) With brain metastases (n=171) Key patient inclusion criteria Advanced NSCLC Treatment naive Data pooled from KEYNOTE-021, -189 and -407 (n=1298) Chemotherapy* (n=66) Pembrolizumab + Chemotherapy* (n=643) Without brain metastases (n=1127) Chemotherapy* (n=484) *KEYNOTE-021: pembrolizumab 200 mg q3w + pemetrexed-carboplatin vs. pemetrexed-carboplatin; KEYNOTE-189: pembrolizumab 200 mg q3w + pemetrexed-platinum vs. placebo + pemetrexed-platinum; KEYNOTE-407: pembrolizumab 200 mg q3w + carboplatin-paclitaxel/nab-paclitaxel vs. placebo + carboplatin-paclitaxel/nab-paclitaxel Powell SF, et al. Ann Oncol 2019;30(suppl):Abstr 1483PD

1483PD: Pembrolizumab (pembro) plus platinum-based chemotherapy (chemo) in NSCLC with brain metastases: Pooled analysis of KEYNOTE-021, 189, and 407 – Powell SF, et al Key results Patients with brain metastases OS PFS OS, % No. at risk Time, months Pembro + CT 105 92 82 71 58 53 44 9 2 1 31 20 100 80 60 40 34.9% 62.9% 3 6 12 15 18 21 24 27 30 33 CT 66 52 37 28 19 14 8 Median, month (95%CI) 18.8 (13.8, 25.9) 7.6 (5.4, 10.9) 643 598 507 420 333 264 234 10 154 90 53.6% 70.2% 484 435 344 231 168 123 98 23 68 Median, month (95%CI) 22.5 (19.8, 25.2) 13.5 (11.3, 15.8) Events, % HR (95%CI) Pembro + CT 53.3 0.48 (0.32, 0.70) CT 74.2 PFS, % No. at risk Time, months Pembro + CT 105 77 57 37 26 21 16 2 9 3 100 80 60 40 20 6.6% 29.9% 6 12 15 18 24 27 30 CT 66 35 7 4 1 Median, month (95%CI) 6.9 (5.7, 8.9) 4.1 (2.3, 4.6) Events, % HR (95%CI) Pembro + CT 78.1 0.44 (0.31, 0.62) CT 93.9 Patients without brain metastases PFS, % No. at risk Time, months Pembro + CT 643 525 394 249 171 126 97 16 3 63 39 100 80 60 40 20 20.1% 39.2% 6 9 12 15 18 21 24 27 30 33 CT 484 340 193 96 59 37 5 13 8 Median, month (95%CI) 8.8 (8.1, 9.5) 5.3 (4.8, 6.1) OS PFS Events, % HR (95%CI) Pembro + CT 41.1 0.63 (0.53, 0.75) CT 51.7 Events, % HR (95%CI) Pembro + CT 62.5 0.55 (0.48, 0.63) CT 76.0 Powell SF, et al. Ann Oncol 2019;30(suppl):Abstr 1483PD

Safety profile was tolerable, regardless of brain metastases status 1483PD: Pembrolizumab (pembro) plus platinum-based chemotherapy (chemo) in NSCLC with brain metastases: Pooled analysis of KEYNOTE-021, 189, and 407 – Powell SF, et al Key results (cont.) Conclusions In patients with advanced squamous and non-squamous NSCLC, with or without pretreated or asymptomatic brain metastases, pembrolizumab combined with platinum-based chemotherapy improved outcomes vs. chemotherapy alone Safety profile was tolerable, regardless of brain metastases status Pembro + CT CT With brain metastases (n=102) Without brain metastases (n=640) With brain metastases (n=64) Without brain metastases (n=480) Median treatment duration, months (range) 7.0 (0.03–29.1) 6.9 (0.03–30.4) 4.2 (0.03–18.5) 4.3 (0.03–30.7) TRAEs, % Grade 3–5 Led to death Led to discontinuation Affecting CNS 88.2 59.8 5.9 25.5 32.4 94.5 50.5 2.0 21.4 36.4 82.8 45.3 1.6 10.9 17.2 90.6 46.9 1.9 8.1 33.5 Immune-mediated AEs and infusion reactions, % Grade 3–5, % 25.5 13.7 27.8 9.7 9.4 1.6 10.6 3.8 Powell SF, et al. Ann Oncol 2019;30(suppl):Abstr 1483PD

1478O & 390O: Results of the ASCEND-7 phase II study – Chow LQ, et al & Barlesi F, et al Study objective To investigate the efficacy and safety of ceritinib in patients with ALK+ NSCLC and brain or leptomeningeal metastases Arm 1: Prior brain radiotherapy and prior crizotinib (n=42) Key patient inclusion criteria NSCLC ALK+ (FISH) Active brain metastases either newly diagnosed or progressive + ≥1 extra-cranial measurable lesion WHO PS 0–2 Arm 2: No prior brain radiotherapy and prior crizotinib (n=40) Arm 3: Prior brain radiotherapy and no prior ALKi (n=12) Ceritinib 750 mg/day Arm 4: No prior brain radiotherapy and no prior ALKi (n=44) Arm 5: Leptomeningeal carcinomatosis with or without active metastases (n=18) Primary endpoint ORR (whole body; RECIST v1.1) Secondary endpoints DCR, intra- and extra-cranial response, safety Chow LQ, et al. Ann Oncol 2019;30(suppl):Abstr 1478O Barlesi F, et al. Ann Oncol 2019;30(suppl):Abstr 390O

1478O & 390O: Results of the ASCEND-7 phase II study – Chow LQ, et al & Barlesi F, et al Key results Whole body response, % (95%CI) Arm 1 (n=42) Arm 2 (n=40) Arm 3 (n=12) Arm 4 (n=44) Arm 5 (n=18) ORR 35.7 (21.6, 52.0) 30.0 (16.6, 46.5) 50.0 (21.1, 78.9) 59.1 (43.2, 73.7) 16.7 (3.6, 41.4) DCR 66.7 (50.5, 80.4) 82.5 (67.2, 92.7) 66.7 (34.9, 90.1) 70.5 (54.8, 83.2) 66.7 (41.0, 86.7) Intracranial response, % (95%CI) Arm 1 (n=28) Arm 2 (n=29) Arm 3 (n=7) Arm 4 (n=33) 39.3 (21.5, 59.4) 27.6 (12.7, 47.2) 28.6 (3.7, 71.0) 51.5 (33.5, 69.2) 5.6 (0.1, 27.3) 75.0 (55.1, 89.3) 82.8 (64.2, 94.2) 85.7 (42.1, 99.6) 75.8 (57.7, 88.9) Extracranial response, % (95%CI) 31.0 (17.6, 47.1) 42.5 (27.0, 59.1) 41.7 (15.2, 72.3) 61.4 (45.5, 75.6). 22.2 (6.4, 47.6) 69.0 (52.9, 82.4) 92.5 (79.6, 98.4) 72.7 (57.2, 85.0) 72.2 (46.5, 90.3) Chow LQ, et al. Ann Oncol 2019;30(suppl):Abstr 1478O Barlesi F, et al. Ann Oncol 2019;30(suppl):Abstr 390O ORR (CR + PR); DCR (CR + PR + SD)

No new safety signals were observed 1478O & 390O: Results of the ASCEND-7 phase II study – Chow LQ, et al & Barlesi F, et al Key results (cont.) Conclusions In patients with ALK+ NSCLC and active brain or leptomeningeal metastases, ceritinib demonstrated whole body, intra- and extracranial activity, regardless of prior treatment No new safety signals were observed Arm 1 (n=42) Arm 2 (n=40) Arm 3 (n=12) Arm 4 (n=44) Arm 5 (n=18) AEs, n (%) All AEs Grade 3–4 AEs suspected to be drug-related Grade 3–4 AEs suspected to be drug-related 42 (100) 34 (81.0) 41 (97.6) 20 (47.6) 40 (100) 32 (80.0) 40 (100) 25 (62.5) 11 (91.7) 7 (58.3) 10 (83.3) 4 (33.3) 43 (97.7) 34 (77.3) 39 (88.6) 25 (56.8) 18 (100) 16 (88.9) 9 (50.0) SAEs, n (%) All SAEs Grade 3–4 SAEs SAEs suspected to be drug-related Grade 3–4 SAEs suspected to be drug-related 28 (66.7) 21 (50.0) 4 (9.5) 3 (7.1) 17 (42.5) 10 (25.0) 3 (7.5) 2 (5.0) 4 (33.3) 4 (33.3) 2 (16.7) 1 (8.3) 15 (34.1) 12 (27.3) 3 (6.8) 3 (6.8) 11 (61.1) 1 (5.6) AEs leading to discontinuation, n (%) 2 (4.8) 2 (5.0) 2 (16.7) 7 (15.9) 6 (33.3) AEs requiring dose adjustments or interruption, n (%) 33 (78.6) 35 (87.5) 11 (91.7) 33 (75.0) 12 (66.7) Chow LQ, et al. Ann Oncol 2019;30(suppl):Abstr 1478O Barlesi F, et al. Ann Oncol 2019;30(suppl):Abstr 390O

SCLC, mesothelioma and thymic epithelial tumors Other malignancies SCLC, mesothelioma and thymic epithelial tumors

LBA89: PD-L1 expression, patterns of progression and patient-reported outcomes (PROs) with durvalumab plus platinum-etoposide in ES-SCLC: Results from CASPIAN – Paz-Ares L, et al Study objective To evaluate the patterns of progression and PROs of durvalumab + platinum- etoposide in patients with extensive-stage SCLC Durvalumab 1500 mg + etoposide-platinum* q3w up to 4 cycles (n=268) Durvalumab 1500 mg q4w PD Key patient inclusion criteria Extensive-stage SCLC Treatment naïve Asymptomatic or treated and stable brain metastases permitted WHO PS 0–1 (n=805) Etoposide-platinum* q3w up to 6 cycles (n=269) Optional PCI R 1:1:1 Durvalumab 1500 mg + tremelimumab 75 mg + etoposide-platinum* q3w up to 4 cycles Durvalumab 1500 mg q4w PD Stratification Platinum agent (carboplatin vs. cisplatin) Primary endpoint OS Secondary endpoints PFS, ORR, PROs, safety *Etoposide 80–100 mg/m2 + carboplatin AUC5–6 or cisplatin 75–80 mg/m2 Paz-Ares L, et al. Ann Oncol 2019;30(suppl):Abstr LBA89

PD-L1 was evaluable* in 277 (51.6%) patients across both arms LBA89: PD-L1 expression, patterns of progression and patient-reported outcomes (PROs) with durvalumab plus platinum-etoposide in ES-SCLC: Results from CASPIAN – Paz-Ares L, et al Key results Median OS was 13.0 months (95%CI 11.5, 14.8) for durvalumab + EP and 10.3 months (95%CI 9.3, 11.2) for EP alone (HR 0.73 [95%CI 0.591, 0.909]) PD-L1 was evaluable* in 277 (51.6%) patients across both arms 94.9% had TC PD-L1 <1% and 77.6% had IC PD-L1 <1% No significant impact of PD-L1 expression was observed on OS, PFS or ORR ITT (n=537) PD-L1 evaluable (n=277) IC <1 (n=215) IC ≥1 (n=62) TC <1 (n=263) TC ≥1 (n=14) HR (95%CI) 0.73 (0.591, 0.910) 0.65 (0.482, 0.864) 0.64 (0.462, 0.897) 0.69 (0.370, 1.283) 0.66 (0.491, 0.896) 0.46 (0.119, 1.793) 0.10 0.25 0.5 1.0 2.0 Favours durvalumab + EP Favours EP OS based on PD-L1 expression *PD-L1 assessed using the VENTANA PD-L1 (SP263) assay Paz-Ares L, et al. Ann Oncol 2019;30(suppl):Abstr LBA89

LBA89: PD-L1 expression, patterns of progression and patient-reported outcomes (PROs) with durvalumab plus platinum-etoposide in ES-SCLC: Results from CASPIAN – Paz-Ares L, et al Global health status/QoL(QLQ-C30)* Key results (cont.) 0.2 1.0 0.8 0.6 0.4 3 6 9 12 15 18 21 24 Durvalumab + EP EP HR 0.81 (95%CI 0.626, 1.054) Time to deterioration Probability of not worsening No. at risk Time from randomisation, months D + EP 238 183 125 74 41 20 12 1 EP 232 147 77 40 16 5 Physical functioning (QLQ-C30)* Cognitive functioning (QLQ-C30)* 0.2 1.0 0.8 0.6 0.4 3 6 9 12 15 18 21 24 Durvalumab + EP EP HR 0.61 (95%CI 0.472, 0.776) Probability of not worsening 0.2 Probability of not worsening 1.0 0.8 0.6 0.4 3 6 9 12 15 18 21 24 Durvalumab + EP EP HR 0.75 (95%CI 0.581, 0.970) No. at risk Time from randomisation, months Time from randomisation, months D + EP 243 173 123 73 37 19 13 1 D + EP 244 181 126 72 38 17 10 EP 240 145 75 40 16 6 EP 242 142 69 32 12 2 *Patients with baseline scores ≥10 Paz-Ares L, et al. Ann Oncol 2019;30(suppl):Abstr LBA89

Durvalumab + EP showed improved disease-progression outcomes LBA89: PD-L1 expression, patterns of progression and patient-reported outcomes (PROs) with durvalumab plus platinum-etoposide in ES-SCLC: Results from CASPIAN – Paz-Ares L, et al Key results (cont.) Conclusions In patients with extensive-stage NSCLC, a statistically significant improvement OS was observed with 1L durvalumab +EP vs. EP alone (p=0.0047), and there was no significant interaction with PD-L1 expression Durvalumab + EP showed improved disease-progression outcomes Types of progression Durvalumab + EP (n=268) EP (n=269) Total progression events, n (%) 226 (84.3) 233 (86.6) RECIST-defined progression, n (%) Target lesions Non-target lesions New lesions 192 (71.6) 115 (42.9) 66 (24.6) 111 (41.4) 194 (72.1) 106 (39.4) 61 (22.7) 127 (47.2) Death in absence of progression, n (%) 34 (12.7) 39 (14.5) Sites of new lesions (>5% patients) Durvalumab + EP (n=268) EP (n=269) New lesions, n (%) Lung Brain/CNS Liver Bone Regional Lymph nodes 111 (41.4) 23 (8.6) 31 (11.6) 15 (5.6) 12 (4.5) 15 (5.6) 127 (47.2) 41 (15.2) 31 (11.5) 24 (8.9) 19 (7.1) 12 (4.5) Paz-Ares L, et al. Ann Oncol 2019;30(suppl):Abstr LBA89

1736O: IMpower133: updated overall survival (OS) analysis of first-line (1L) atezolizumab (atezo) + carboplatin + etoposide in extensive-stage SCLC (ES-SCLC) – Reck M, et al Study objective To update the OS findings with atezolizumab + carboplatin + etoposide as a 1L treatment for patients with extensive-stage SCLC in IMpower133 Key patient inclusion criteria Measureable extensive-stage SCLC No prior systemic therapy Treated asymptomatic brain metastases were eligible ECOG PS 0–1 (n=403) Atezolizumab 1200 mg IV q3w + carboplatin + etoposide (n=201) Atezolizumab maintenance PD/ loss of benefit Stratification Sex (male vs. female) ECOG PS (0 vs. 1) Brain metastases (yes vs. no) R 1:1 Placebo + carboplatin + etoposide (n=202) Placebo PD/ loss of benefit Endpoints Updated OS in ITT, and by PD-L1 subgroups Updated DoR/ORR in ITT Updated safety Updated biomarker analysis Carboplatin AUC 5 mg/mL/min IV q3w; etoposide 100 mg/m2 IV D1–3 q3w Reck M, et al. Ann Oncol 2019;30(suppl):Abstr 1736O

1736O: IMpower133: updated overall survival (OS) analysis of first-line (1L) atezolizumab (atezo) + carboplatin + etoposide in extensive-stage SCLC (ES-SCLC) – Reck M, et al Key results OS Response Atezo + CP/ET 64.4 Time, months OS, % 100 80 60 40 20 2 4 6 8 10 12 14 16 No.at risk 18 22 26 28 30 32 24 12-month OS 18-month OS 202 183 160 131 97 74 58 49 39 Placebo + CP/ET 33 3 189 21.0% 39.0% 60.2 Placebo + CP/ET Atezo + CP/ET 201 180 159 130 109 93 86 75 61 51 28 21 8 1 187 51.9% 34.0% Response, % 20.9 21.3 6.9 3.5 1.0 Atezo + CP/ET (n=201) Placebo + CP/ET (n=202) Median OS, months (95%CI) 12.3 (10.8, 15.8) 10.3 (9.3, 11.3) HR (95%CI) 0.76 (0.60, 0.95) p-value 0.0154 Atezo + CP/ET (n=121) Placebo + CP/ET (n=130) DoR, months (range) 4.2 (1.4+–24.3+) 3.9 (2.0–24.2+) Patients with ongoing response, n (%) 11 (9.1) 3 (2.3) Reck M, et al. Ann Oncol 2019;30(suppl):Abstr 1736O

PD-L1* was evaluable in 137/403 (34%) of the ITT population 1736O: IMpower133: updated overall survival (OS) analysis of first-line (1L) atezolizumab (atezo) + carboplatin + etoposide in extensive-stage SCLC (ES-SCLC) – Reck M, et al Key results (cont.) PD-L1* was evaluable in 137/403 (34%) of the ITT population Conclusions In patients with extensive-stage NSCLC, updated survival outcomes continue to favour use of 1L atezolizumab + carboplatin + etoposide over chemotherapy alone Efficacy was not related to PD-L1 expression (1% or 5% threshold) or bTMB (10 or 16 thresholds) Median OS (months) OS HRa (95%CI) Subgroup Atezo + CP/ET Placebo + CP/ET 0.25 0.5 1.0 1.5 Hazard ratioa Favours Atezo + CP/ET Favours Placebo + CP/ET ITT (N=403) 12.3 10.3 0.76 (0.61, 0.96) ITT-BEP (n=137) Non-BEP (n=266) PDL-1 expression 1% TC or IC <1% PD-L1 (n=65) ≥1% PD-L1 (n=72) PDL-1 expression 5% TC or IC <5% PD-L1 (n=108) ≥5% PD-L1 (n=29) 9.9 14.6 10.2 9.7 9.2 21.6 8.9 11.2 8.3 10.6 9.2 0.70 (0.48, 1.02) 0.81 (0.61, 1.08) 0.51 (0.30, 0.89) 0.87 (0.51, 1.49) 0.77 (0.51, 1.17) 0.60 (0.25, 1.46) *PD-L1 assessed by VENTANA SP 263 assay of slide sections ≤1 year old; aHR stratified for ITT, unstratified for subgroups Reck M, et al. Ann Oncol 2019;30(suppl):Abstr 1736O

Pembrolizumab 200 mg q3w (n=73) LBA91_PR: A multicentre randomized phase III trial comparing pembrolizumab (P) vs single agent chemotherapy (CT) for advanced pre-treated malignant pleural mesothelioma (MPM) – results from the European Thoracic Oncology Platform (ETOP 9-15) PROMISE-meso trial – Popat S, et al Study objective To evaluate the efficacy of pembrolizumab vs. platinum-based chemotherapy in patients with advanced MPM Pembrolizumab 200 mg q3w (n=73) Key patient inclusion criteria MPM Progression on platinum-based chemotherapy ECOG PS 0–1 (n=144) Stratification Histology (epithelioid vs. non-epithelioid) R 1:1 PD (max 2 years) Chemotherapya Gemcitabine 1000 mg/m2 q3w or vinorelbine 30 mg/m2 q3w or vinorelbine 60/80 mg/m2 q3w (n=71) Primary endpoint PFS (BICR) Secondary endpoints ORR, TTF, OS, PFS (investigator-assessed), safety aCrossover to pembrolizumab was allowed at progression Popat S, et al. Ann Oncol 2019;30(suppl):Abstr LBA91_PR

LBA91_PR: A multicentre randomized phase III trial comparing pembrolizumab (P) vs single agent chemotherapy (CT) for advanced pre-treated malignant pleural mesothelioma (MPM) – results from the European Thoracic Oncology Platform (ETOP 9-15) PROMISE-meso trial – Popat S, et al Key results PFS 100 Events/N mPFS, months (95%CI) 6-month PFS, % (95%CI) CT 56/71 3.4 (2.2, 4.3) 27.4 (17.1, 38.7) Pembro 62/73 2.5 (2.1, 4.2) 25.0 (15.5, 35.6) 80 60 PFS, % HR 1.06 (95%CI 0.73, 1.53); p=0.76a 40 20 2 4 6 8 10 12 14 No. at risk Time, months CT 9 16 29 55 71 4 Pembro 11 17 33 50 73 3 6 aStratified by histological subtype Popat S, et al. Ann Oncol 2019;30(suppl):Abstr LBA91_PR

LBA91_PR: A multicentre randomized phase III trial comparing pembrolizumab (P) vs single agent chemotherapy (CT) for advanced pre-treated malignant pleural mesothelioma (MPM) – results from the European Thoracic Oncology Platform (ETOP 9-15) PROMISE-meso trial – Popat S, et al Key results (cont.) OS Deaths/N mOS, months (95%CI) 6-month OS, % (95%CI) CT 34/71 11.7 (7.4, NE) 72.9 (60.8, 81.7) Pembro 37/73 10.7 (7.6, NE) 68.5 (56.5, 77.8) 100 80 60 HR 1.04 (95%CI 0.66, 1.67); p=0.85a OS, % 40 20 2 4 6 8 10 12 14 16 No. at risk Time, months CT 39 51 58 64 71 8 15 26 1 Pembro 39 50 55 66 73 17 27 9 3 aStratified by histological subtype Popat S, et al. Ann Oncol 2019;30(suppl):Abstr LBA91_PR

LBA91_PR: A multicentre randomized phase III trial comparing pembrolizumab (P) vs single agent chemotherapy (CT) for advanced pre-treated malignant pleural mesothelioma (MPM) – results from the European Thoracic Oncology Platform (ETOP 9-15) PROMISE-meso trial – Popat S, et al Key results (cont.) ORR was significantly improved with pembrolizumab vs. chemotherapy (22% [95%CI 13, 33] vs. 6% [95%CI 2, 14], respectively; p=0.004) Median DoR was 4.6 months (95%CI 2.2, 10.3) for pembrolizumab and 11.2 months (95%CI 6.2, 15.3) for chemotherapy Median TTF was similar in the two treatment arms; 2.8 months (95%CI 2.1, 4.2) vs. 2.3 months (95%CI 2.1, 3.9) for pembrolizumab vs. chemotherapy, respectively Popat S, et al. Ann Oncol 2019;30(suppl):Abstr LBA91_PR

No new safety signals were observed LBA91_PR: A multicentre randomized phase III trial comparing pembrolizumab (P) vs single agent chemotherapy (CT) for advanced pre-treated malignant pleural mesothelioma (MPM) – results from the European Thoracic Oncology Platform (ETOP 9-15) PROMISE-meso trial – Popat S, et al Key results (cont.) The most frequent pembrolizumab-related AEs were dry skin (13.9%), pruritus (12.5%) and maculopapular rash (12.5%) Conclusions In pre-treated patients with advanced MPM, pembrolizumab failed to demonstrate PFS and OS superiority vs. platinum-based chemotherapy, despite a significantly improved ORR No new safety signals were observed AEs, n (%) Pembrolizumab (n=72) Placebo + CP/ET (n=70) Any AEs 70 (97.2) 65 (92.9) Any TRAEs 50 (69.4) 51 (72.9) Grade 3–5 TRAEs 14 (19.4) 17 (24.3) TRAEs leading to discontinuation 6 (8.3) 5 (7.1) TRAEs leading to death 1 (1.4) Popat S, et al. Ann Oncol 2019;30(suppl):Abstr LBA91_PR

Nivolumab 3 mg/kg IV q2w + ipilimumab 1 mg/kg IV q6w 1841O: 2nd/3rd-line nivolumab vs nivo plus ipilimumab in malignant pleural mesothelioma: Long-term results of IFCT-1501 MAPS2 phase 2R trial with a focus on hyperprogression (HPD) – Zalcmann G, et al Study objective To evaluate the long-term efficacy and safety of nivolumab ± ipilimumab as a 2L or 3L treatment for patients with MPM Key patient inclusion criteria Histological diagnosis of MPM Unresectable cancer with documented progression after 1 or 2 lines of chemotherapy including a pemetrexed/ platinum doublet ECOG PS 0−1 (n=125) Nivolumab 3 mg/kg IV q2w (n=63) PD/ toxicity R 1:1 Nivolumab 3 mg/kg IV q2w + ipilimumab 1 mg/kg IV q6w (n=62) PD/ toxicity Primary endpoint DCR at 12 weeks in first 54 accrued patients in each group Secondary endpoints PFS, OS, QoL, safety Zalcmann G, et al. Ann Oncol 2019;30(suppl):Abstr 1841O

1841O: 2nd/3rd-line nivolumab vs nivo plus ipilimumab in malignant pleural mesothelioma: Long-term results of IFCT-1501 MAPS2 phase 2R trial with a focus on hyperprogression (HPD) – Zalcmann G, et al Key results OS: Nivolumab OS: Nivolumab + Ipi OS, % 100 80 60 40 20 15 10 5 25 Time, months 30 35 Events=52, censored=11 OS, % 100 80 60 40 20 15 10 5 25 Time, months 30 35 Events=49, censored=13 Nivolumab (n=63) Nivolumab + Ipi (n=62) 2-years OS, % (95%CI) 25.4 (15.5, 36.6) 31.7 (20.5, 43.4) mOS, months (95%CI) 11.9 (6.7, 17.5) 15.9 (10.7, 22.2) Zalcmann G, et al. Ann Oncol 2019;30(suppl):Abstr 1841O

1841O: 2nd/3rd-line nivolumab vs nivo plus ipilimumab in malignant pleural mesothelioma: Long-term results of IFCT-1501 MAPS2 phase 2R trial with a focus on hyperprogression (HPD) – Zalcmann G, et al Key results (cont.) The correlation of HPD status with OS was evaluated according to tumour growth rate (TGR) or tumour growth kinetics (TGK) methods; despite a negative result with TGR, TGK revealed an impact of HPD on the OS in both study arms (p<0.0001) Conclusions In patients with MPM, both nivolumab or nivolumab + ipilimumab demonstrated a favourable survival outcome compared with chemotherapy In fast-progressing MPM patients who are receiving immunotherapy, it is recommended to be cautious and requires early evaluation Zalcmann G, et al. Ann Oncol 2019;30(suppl):Abstr 1841O

PD/ toxicity/ withdrawal 1843O: SAKK 17/16: Lurbinectedin as second or third line palliative chemotherapy in malignant pleural mesothelioma (MPM): A multi-center, single-arm Phase II trial – Metaxas Y, et al Study objective To evaluate the efficacy and safety of lurbinectedin in patients with progressive MPM Key patient inclusion criteria Confirmed MPM Progression after 1L platinum-pemetrexed chemotherapy ± immunotherapy ECOG PS 0–1 (n=42) Lurbinectedin 3.2 mg/m2 q3w (n=42) PD/ toxicity/ withdrawal Primary endpoint PFS at 12 weeks Secondary endpoints PFS, OS, safety Metaxas Y, et al. Ann Oncol 2019;30(suppl):Abstr 1843O

1843O: SAKK 17/16: Lurbinectedin as second or third line palliative chemotherapy in malignant pleural mesothelioma (MPM): A multi-center, single-arm Phase II trial – Metaxas Y, et al Key results PFS OS 9 6 3 12 15 18 Survival probability 1.00 0.75 0.50 0.25 Time, months No. at risk 58.4% (95%CI 41.8, 71.7) 30.54% (95%CI 17.1, 44.9) 12.7% (95%CI 4.7, 24.9) 42 4 23 1 9 6 3 12 15 18 Survival probability 1.00 0.75 0.50 0.25 Time, months No. at risk 73.8% (95%CI 57.7, 84.67) 44.9% (95%CI 28.1, 60.3) 42 25 31 40 13 5 Lurbinectedin (n=42) Simon’s first-stage 11/21 12-week PFS, % (90%CI) 52.4 (38.7, 63.5) mPFS, months (95%CI) 4.1 (2.6, 5.5) Lurbinectedin (n=42) No. of events (death) 26 mOS, months (95%CI) 11.1 (8.8, 14.7) Metaxas Y, et al. Ann Oncol 2019;30(suppl):Abstr 1843O

Grade 3–4 AEs occurred in 32/42 (76.2%) patients 1843O: SAKK 17/16: Lurbinectedin as second or third line palliative chemotherapy in malignant pleural mesothelioma (MPM): A multi-center, single-arm Phase II trial – Metaxas Y, et al Key results (cont.) Neither histological types (epithelioid and non-epithelioid) nor prior immunotherapy affected the PFS and OS outcomes Grade 3–4 AEs occurred in 32/42 (76.2%) patients Grade 3–4 TRAEs occurred in 20/42 (47.6%) patients The most frequent grade 3–4 TRAEs were neutropenia, fatigue and febrile neutropenia (23.8%, 16.7% and 9.5%, respectively) Conclusions In patients with progressive MPM, lurbinectedin showed clinical activity and had a manageable safety profile Metaxas Y, et al. Ann Oncol 2019;30(suppl):Abstr 1843O

1844O: Durable anti-tumor activity of the multi-targeted inhibitor lenvatinib in patients with advanced or metastatic thymic carcinoma; Preliminary results from a multicenter phase II (REMORA) trial – Itoh S, et al Study objective To evaluate the efficacy and safety of lenvatinib, a multi-targeted inhibitor of VEGFR kinases, in patients with advanced or metastatic thymic carcinoma Key patient inclusion criteria Advanced or metastatic thymic carcinoma Progression after ≥1 prior platinum-based chemotherapy ECOG PS 0–1 (n=42) Lenvatinib 24 mg/day PD/ toxicity Primary endpoint ORR Secondary endpoints PFS, OS, DCR, safety Itoh S, et al. Ann Oncol 2019;30(suppl):Abstr 1844O

Maximum % change from baseline 1844O: Durable anti-tumor activity of the multi-targeted inhibitor lenvatinib in patients with advanced or metastatic thymic carcinoma; Preliminary results from a multicenter phase II (REMORA) trial – Itoh S, et al Key results The median PFS was 9.3 months (95%CI 7.7, 13.9), while median OS was not reached (95%CI 16.1, NR) ORR (BICR) Maximum % change from baseline 20 10 –10 –20 –30 –40 –50 –60 –80 –70 Lenvatinib (n=42) Best overall response, n (%) PR SD PD 16 (38.1) 24 (57.1) 2 (4.8) ORR, % (90%CI) 38.1 (25.6, 52.0) DCR, % (95%CI) 95.2 (83.8, 99.4) Median DoR, n/N months (95%CI) 16/42 11.6 (5.8, 18.0) Best overall response PR (n=16) SD (n=24) PD (n=2) Itoh S, et al. Ann Oncol 2019;30(suppl):Abstr 1844O

1844O: Durable anti-tumor activity of the multi-targeted inhibitor lenvatinib in patients with advanced or metastatic thymic carcinoma; Preliminary results from a multicenter phase II (REMORA) trial – Itoh S, et al Key results (cont.) Of the patients 20/42 (47.6%) discontinued the treatment due to disease progression and 7/42 (16.7%) due to TRAEs AEs that led to discontinuation were all grade 3 and included arthralgia, intestine perforation, left ventricular dysfunction, pneumonitis and pneumothorax Conclusions In patients with advanced or metastatic thymic carcinoma, lenvatinib showed clinical efficacy with a tolerable safety profile Itoh S, et al. Ann Oncol 2019;30(suppl):Abstr 1844O