1. LUX-Lung 8 A randomised, open-label phase III trial of afatinib versus erlotinib in patients with advanced squamous cell carcinoma of the lung as second-line therapy following first-line platinum-based chemotherapy Procedure ID: XXXX – February 2016 NSCLC=non-small cell lung cancer; OS=overall survival. Soria JC et al. Lancet Oncol. 2015;16(8):897-907. European indication and usage: Afatinib is approved in a number of markets, including the EU, Japan, Taiwan and Canada under the brand name GIOTRIF ® and in the U.S. under the brand name GILOTRIF ® for use in patients with distinct types of EGFR mutation-positive NSCLC. Registration conditions differ internationally, please refer to locally approved prescribing information. Afatinib is under regulatory review by health authorities in other countries worldwide. Afatinib is not approved in other indications.

2. Contents LUX-Lung clinical trials The ErbB Family in squamous NSCLC LUX-Lung 8: study design LUX-Lung 8: efficacy LUX-Lung 8: tumour genetic analysis LUX-Lung 8: adverse events LUX-Lung 8: patient-reported outcomes Conclusions 2

3. Return to contents LUX-LUNG Clinical Trials Overview 3

4. Afatinib has been studied in a comprehensive NSCLC clinical trial programme LUX-Lung 2 Phase II Published LUX-Lung 3 Phase III – Pivotal trial Published LUX-Lung 6 Phase III – Pivotal trial Published LUX-Lung 7 Phase IIb Reported LUX-Lung 1 Phase IIB/III – Pivotal trial Published LUX-Lung 4 Phase I/II Published LUX-Lung 5 Phase III Published LUX-Lung 8 Phase III Published EGFR TKI-naive EGFR M+ EGFR TKI pre-treated Treatment beyond progression Squamous cell carcinoma chemotherapy pre-treated LL1: Miller VA et al. Lancet Oncol. 2012;13(5):528-538. LL2: Yang JC et al. Lancet Oncol. 2012;13(5):539-548. LL3: Sequist LV et al. J Clin Oncol. 2013;31(27):3327-3334. LL4: Katakami N et al. J Clin Oncol. 2013;31(27):3335-3341. LL5: Schuler M et al. J Clin Oncol 2014;32:(suppl; abstr 8019). LL6: Wu YL et al. Lancet Oncol. 2014;15(2):213-22. LL7: Park K et al. Abstract LBA2 and oral presentation. European Society for Medical Oncology (ESMO) ASIA, Singapore, 20th December 2015. LL8: Soria JC et al. Lancet Oncol. 2015;16(8):897-907. 4

5. Return to contents The ErbB Family in squamous NSCLC 5

6. ErbB Family dysregulation plays an important role in the pathogenesis of squamous NSCLC The ErbB Family of receptors is composed of 4 members: EGFR (ErbB1), HER2 (ErbB2), ErbB3, and ErbB4 While ErbB Family dysregulation in adenocarcinoma is predominantly driven by mutations in EGFR, multiple oncogenic alterations across ErbB Family receptors, including overexpression, amplification/polysomy and mutation, have been demonstrated in squamous NSCLC Hall PE et al. Future Oncol. 2015;11(15):2175-91. Frequency of ErbB aberrations in squamous NSCLC ErbB aberrationsFrequency (%) EGFR overexpression57-82 EGFR amplification/polysomy7-26 EGFRvlll mutation3-5 EGFR kinase domain mutation1-3 ErbB2 mutation/amplification4 ErbB3 overexpression28 ErbB3 mutation1-2 ErbB4 mutation1-2 6

7. Afatinib is the an irreversible ErbB Family Blocker Afatinib PROLIFERATION SURVIVAL RAS ERK AKT RAF MEK P13K mTOR EGFR/ HER2 HER2/ ErbB3 ErbB3/ ErbB4 AfatinibGefitinib Erlotinib 1.Solca F et al. J Pharmacol Exp Ther. 2012;343(2):342-350. 2.TARCEVA ® (erlotinib) Summary of Product Characteristics, 2014. Whereas erlotinib reversibly inhibits EGFR, afatinib irreversibly blocks signalling from all ErbB Family receptors 1,2 7

8. Return to contents LUX-LUNG 8 Study design 8

9. Afatinib was studied vs erlotinib in a head-to-head trial in squamous NSCLC * Including mixed histology. ECOG=Eastern Cooperative Oncology Group; EORTC=European Organisation for Research and Treatment of Cancer. Soria JC et al. Lancet Oncol. 2015;16(8):897-907. Afatinib 40 mg oral once daily n=398 Median follow-up for: Primary PFS analysis: 6.7 months Primary OS and follow-up PFS analysis: 18.4 months LUX-Lung 8: a large, multi-national, prospective, randomised phase III trial Erlotinib 150 mg oral once daily n=397 Randomization 1:1 N=795 Advanced NSCLC (Stage IIIB/IV) Squamous histology* ≥4 cycles of a first-line platinum doublet ECOG PS 0–1 Adequate organ function Progression-free survival (PFS), measured by an independent central radiology review Objective response rate Disease control rate Tumour shrinkage Patient-reported outcomes, including symptom control and health-related quality of life as measured through EORTC questionnaires Primary endpoints Secondary endpoints included Key secondary endpoint Overall survival (OS) 9

10. LUX-Lung 8 – 183 cancer centres in 23 countries 10 Soria JC et al. Lancet Oncol. 2015;16(8):897-907.

11. Patient demographics and characteristics were well balanced between the 2 treatment arms Data are n (%), unless stated otherwise. ECOG PS=Eastern Cooperative Oncology Group performance status. CR=complete response. PR=partial response. SD=stable disease. *Protocol violations. † 1 year before diagnosis. ‡ 71 (18%) versus 85 (21%) were current smokers. § Three patients in the erlotinib group had undifferentiated tumour histology but were considered to be squamous by the treating investigator. ¶ Seven patients (four in the afatinib group and three in the erlotinib group) had a best response of progressive disease on chemotherapy. Soria JC et al. Lancet Oncol. 2015;16(8):897-907. Afatinib (n=398) Erlotinib (n=397) Sex Male Female 335 (84%) 63 (16%) 331 (83%) 66 (17%) Median age (years, range)65.0 (36–84)64.0 (35–88) Age group <65 years ≥65 years 189 (47%) 209 (53%) 210 (53%) 187 (47%) Baseline ECOG PS 0 1 2* 126 (32%) 269 (68%) 3 (<1%) 134 (34%) 262 (66%) 1 (<1%) Ethnic origin Non-eastern Asian Eastern Asian 312 (78%) 86 (22%) 311 (78%) 86 (22%) Smoking status Never smoker Light ex-smoker † Current and other ex-smoker ‡ 26 (7%) 11 (3%) 361 (91%) 18 (5%) 12 (3%) 367 (92%) Median time since diagnosis (years, range)0.8 (0.2–9.3)0.7 (0.2–13.5) Tumour histology § Squamous Mixed 381 (96%) 17 (4%) 382 (96%) 15 (4%) Previous platinum doublet Carboplatin-based Cisplatin-based Other 249 (63%) 163 (41%) 5 (1%) 229 (58%) 198 (50%) 8 (2%) Clinical stage at screening IIIA IIIB IV 1 (<1%) 48 (12%) 349 (88%) 4 (1%) 48 (12%) 345 (87%) Best response to chemotherapy ¶ CR or PR SD Unknown 186 (47%) 161 (40%) 47 (12%) 185 (47%) 167 (42%) 42 (11%) 11

12. Return to contents LUX-LUNG 8 Efficacy 12

13. Afatinib significantly increased PFS vs erlotinib OS=overall survival; PFS=progression-free survival. Soria JC et al. Lancet Oncol. 2015;16(8):897-907. At the time of primary PFS analysis (2014), median PFS was 2.4 months for afatinib vs 1.9 months for erlotinib (HR 0.82; P=0.0427) Hazard ratio 0.81 (95% CI, 0.69-0.96) P=0.0103 Afatinib (n=398) Erlotinib (n=397) 1.0 0.8 0.6 0.4 0 Progression-free survival (%) Months 0369121518212427 2.6 months 0.2 1.9 months Progression-free survival by independent review (primary endpoint) At time of OS analysis 13

14. PFS benefit was consistent across clinically relevant subgroups PFS=progression-free survival. Soria JC et al. Lancet Oncol. 2015;16(8):897-907. Total605/7952·6 (2·0–2·9)1·9 (1·9–2·1)0·81 (0·69–0·96) Ethnic origin Non-eastern Asian487/623 2·5 (2·0–2·9)2·0 (1·9–2·2)0·90 (0·75–1·07) 0·0122 Eastern Asian118/1722·8 (1·9–7·4)1·9 (1·9–2·8)0·54 (0·37–0·79) Sex Male512/6662·6 (2·0–2·9)2·0 (1·9–2·3)0·83 (0·70–0·99) 0·3863 Female93/129 2·7 (1·9–3·8)1·9 (1·8–2·1)0·74 (0·49–1·11) Best response to first-line chemotherapy CR or PR287/3712·7 (2·0–3·1)1·9 (1·9–2·3)0·87 (0·69–1·10) 0·2815 SD246/3282·6 (1·9–3·3)2·0 (1·9–2·7)0·83 (0·64–1·06) Unknown65/89 2·8 (1·9–5·6)1·9 (1·8–2·8)0·54 (0·33–0·90) Interval between end of first-line and beginning of second-line <16 weeks340/4361·9 (1·9–2·4)1·9 (1·9–2·0)0·87 (0·71–1·08) 0·5556 ≥16 weeks265/3593·3 (2·7–3·7)2·2 (1·9–2·8)0·77 (0·60–0·98) Histology Squamous histology579/7632·6 (2·0–2·9)1·9 (1·9–2·1)0·82 (0·70–0·97) 0·1955 Mixed squamous histology26/32 1·9 (0·9–5·3)1·6 (0·4–2·8)0·61 (0·28–1·33) Smoking history Never smoker32/442·8 (1·9–7·4)1·8 (1·5–5·5)0·55 (0·27–1·11) 0·1506 Light ex-smoker*14/235·6 (1·0–NR)1·9 (1·6–6·3)0·44 (0·14–1·37) Current and other ex-smoker559/728 2·6 (2·0–2·8)1·9 (1·9–2·2)0·85 (0·72–1·01) ECOG PS at baseline 0207/2603·0 (2·0–4·3)1·9 (1·9–2·4)0·68 (0·51–0·89) 0·2488 1395/5312·5 (1·9–2·8)1·9 (1·9–2·3)0·87 (0·72–1·06) Age <65 years301/399 2·0 (1·9–2·7)1·9 (1·9–2·1)0·81 (0·65–1·02) 0·7657 ≥65 years304/3962·8 (2·4–3·7)2·0 (1·9–2·4)0·83 (0·66–1·04) Maintenance therapy received No566/7502·6 (2·0–2·9)1·9 (1·9–2·2)0·81 (0·68–0·95) 0·7159 Yes39/452·0 (1·7–4·3)1·9 (1·3–3·0)0·88 (0·47–1·68) P interaction 0.2514160.0625 HR (95% Cl)Events/patients Median progression-free survival, months (95% CI) AfatinibErlotinib HR=hazard ratio. CR=complete response. PR=partial response. SD=stable disease. ECOG PS=Eastern Cooperative Oncology Group performance status. NR=not reached. * 1 year before diagnosis. Favours erlotinibFavours afatinib Progression-free survival by subgroups 14

15. Afatinib significantly increased OS vs erlotinib OS=overall survival. Soria JC et al. Lancet Oncol. 2015;16(8):897-907. 19% relative reduction in the risk of death vs erlotinib (HR 0.81, 95% CI 0.691-0.946; P=0.0077) The survival benefit was evident from the first pre-specified time point measured at 6 months and maintained at 18 months OS unlikely to have been affected by subsequent therapies as a similar proportion of patients in both arms received additional treatment(s) Hazard ratio 0.81 (95% CI, 0.69-0.95) P=0.0077 Afatinib (n=398) Erlotinib (n=397) Overall survival (%) 0.2 0.4 0.6 0.8 1.0 0.0 Time (months) 7.9 months 6.8 369121530182124270 Overall survival (key secondary endpoint) 15

16. OS benefit was consistent across clinically relevant subgroups OS=overall survival. Soria JC et al. Lancet Oncol. 2015;16(8):897-907. Total 632/795 7·9 (7·2–8·7)6·8 (5·9–7·8)0·81 (0·69–0·95) Ethnic origin Non-eastern Asian 500/623 7·5 (6·7–8·4)6·7 (5·8–7·8)0·87 (0·73–1·03) 0·1101 Eastern Asian 132/172 9·6 (7·3–14·1)7·2 (4·9–8·9)0·62 (0·44–0·88) Sex Male 533/666 8·1 (7·2–9·0)6·9 (5·9–8·0)0·82 (0·69–0·97) 0·7032 Female 99/129 7·3 (5·4–10·9)5·8 (4·6–8·2)0·77 (0·51–1·14) Best response to first-line chemotherapy CR or PR 285/371 8·5 (7·3–10·7)8·4 (6·9–10·2)0·91 (0·72–1·15) 0·2584 SD 271/328 7·4 (6·5–8·7)5·8 (4·7–6·9)0·71 (0·56–0·90) Unknown 69/89 8·3 (6·3–10·9)5·8 (3·7–9·4)0·72 (0·44–1·17) Interval between end of first-line and beginning of second-line <16 weeks 366/436 6·6 (5·6–7·7)5·6 (4·9–6·4)0·78 (0·63–0·96) 0·4493 ≥16 weeks 266/359 9·2 (7·9–11·6)8·9 (7·8–10·1)0·88 (0·69–1·12) Histology Squamous histology 602/763 8·0 (7·2–8·8)6·9 (6·2–8·0)0·82 (0·70–0·96) 0·1997 Mixed squamous histology 30/32 7·4 (0·9–16·5)3·6 (1·4–5·3)0·55 (0·26–1·17) Smoking history Never smoker 34/44 7·6 (3·2–15·4)6·8 (1·8–13·2)0·77 (0·37–1·57) 0·8177 Light ex-smoker* 19/23 12·4 (4·9–17·0)8·2 (3·5–10·8)0·43 (0·16–1·12) Current and other ex-smoker 579/728 7·8 (7·1–8·7)6·8 (5·9–7·8)0·81 (0·69–0·96) ECOG PS at baseline 0 197/260 12·0 (8·0–14·1)9·6 (7·4–11·6)0·76 (0·51–1·01) 0·9088 1 431/531 6·7 (6·1–8·1)5·7 (5·1–6·7)0·80 (0·66–0·97) Age <65 years 327/399 8·4 (7·2–11·2)6·1 (5·3–7·3)0·68 (0·55–0·85) 0·0498 ≥65 years 305/396 7·4 (6·4–8·6)7·7 (6·4–8·8)0·95 (0·76–1·19) Maintenance therapy received No 599/750 7·8 (7·2–8·7)6·8 (5·8–7·8)0·79 (0·68–0·93) 0·4135 Yes 33/45 8·4 (6·5–13·8)7·0 (2·1–24·7)1·07 (0·52–2·17) Favours erlotinib 0.2514160.0625 HR (95% Cl)Events/patientsP interaction Median overall survival, months (95% CI) AfatinibErlotinib Favours afatinib HR=hazard ratio. CR=complete response. PR=partial response. SD=stable disease. ECOG PS=Eastern Cooperative Oncology Group performance status. * 1 year before diagnosis. Overall survival by subgroups 16

17. Similar proportion of patients in both arms received additional systemic treatment(s) Across all subsequent lines of therapy, n (%) Afatinib (n=392) Erlotinib (n=395) Subsequent systemic treatment 182 (46.4)192 (48.6) Chemotherapy 176 (44.9)185 (46.8) Docetaxel93 (23.7)103 (26.1) Platinum-based doublet44 (11.2)43 (10.9) Gemcitabine41 (10.5)43 (10.9) Vinorelbine37 (9.4)34 (8.6) EGFR-targeted 12 (3.1)8 (2.0) Erlotinib9 (2.3)8 (2.0) Afatinib2 (0.5)0 (0.0) Immune checkpoint inhibitor 1 (0.3)0 (0.0) Other 5 (1.3)11 (2.8) Soria JC et al. Lancet Oncol. 2015;16(8):897-907. 17

18. Afatinib provided greater and more durable tumour control than erlotinib Objective response rate (ORR) by independent review was 6% for afatinib vs 3% for erlotinib (P=0.0551) ORR by investigator review was 11% for afatinib vs 4% for erlotinib (P=0.0005) Afatinib (n=398) Erlotinib (n=397) Disease control rateDuration of response Soria JC et al. Lancet Oncol. 2015;16(8):897-907. P=0.002 Disease control rate and duration of response (secondary endpoints) 18

19. Return to contents LUX-LUNG 8 Tumour genetic analysis 19

20. Tumour samples were analysed as part of an on-going biomarker analysis Performed on 238 (~30%) patients retrospectively selected based on efficacy endpoints with an enrichment of patients with long PFS (>2 months) in both treatment arms Archival tumour tissue was assessed using the Foundation Medicine FoundationOne™ platform Analysed ~300 cancer-related genes for copy number variations, gene rearrangements, and point mutations Aimed to identify predictive marker(s) for patients benefitting on afatinib/erlotinib Estimated PFS probability 0 Time (months) 123456789101112131415 0.4 0.8 1.0 0.6 0.2 0 Group 1: Patients benefiting from afatinib/erlotinib Group 2 (control): Patients with little or no benefit from afatinib/erlotinib Soria JC et al. Oral presentation ORAL32.01 at 16th World Conference on Lung Cancer, Denver, 2015. Soria JC et al. Lancet Oncol. 2015;16(8):897-907. Primary PFS analysis 20

21. No significant differences in frequency of genetic anomalies observed between groups SV=single nucleotide variants; CNA=copy number alterations. * Tumour genetic analysis subset; † Any differences between Groups 1 and 2 should be interpreted with caution due to low numbers. A Fisher’s exact test found no significant differences between the groups Soria JC et al. Oral presentation ORAL32.01 at 16th World Conference on Lung Cancer, Denver, 2015. Soria JC et al. Lancet Oncol. 2015;16(8):897-907. Gene, % LL8 subset* (n=238) Group 1 † PFS >2 months (n=144) Group 2 † PFS ≤2 months (n=94) SVs TP5387.889.685.1 LRP1B39.538.241.5 MLL232.831.235.1 CDKN2A28.627.829.8 FAT327.726.429.8 ErbB family21.024.316.0 EGFR 5.9 6.9 4.3 HER 2/3/4 5.0/6.3/5.5 4.9/8.3/6.9 5.3/3.2/3.2 CNAs SOX2 43.749.335.1 KLHL6 40.346.530.9 PIK3CA 37.042.428.7 MAP3K13 32.836.127.7 BCL6 31.134.026.6 FGF12 28.630.625.5 ErbB family 10.1 9.710.6 EGFR 6.3 7.6 4.3 HER 2 3.8 2.1 6.4 Predefined families: any aberration ErbB29.031.924.5 FGF58.856.362.8 21

22. Specified genetic alterations not predictive of PFS benefit with afatinib vs erlotinib SV=single nucleotide variants; CNA=copy number alterations. * Tumour genetic analysis subset; † Interaction p values between yes/no on all markers >0.05 Soria JC et al. Oral presentation ORAL32.01 at 16th World Conference on Lung Cancer, Denver, 2015. Soria JC et al. Lancet Oncol. 2015;16(8):897-907. LL8 LL8 subset* (n=238) 0.81 (0.69–0.96) 0.65 (0.48–0.87) SVs EGFR No (94.1) Yes (5.9) 0.65 (0.48–0.88) 0.52 (0.13–2.01) TP53 No (12.2) Yes (87.8) 0.57 (0.24–1.37) 0.64 (0.47–0.89) LRP1B No (60.5) Yes (39.5) 0.56 (0.38–0.84) 0.77 (0.49–1.23) MLL2 No (67.2) Yes (32.8) 0.63 (0.44–0.90) 0.64 (0.38–1.08) CDKN2A No (71.4) Yes (28.6) 0.62 (0.44–0.89) 0.65 (0.37–1.13) FAT3 No (72.3) Yes (27.7) 0.58 (0.41–0.83) 0.76 (0.44–1.33) CNAs EGFR No (93.7) Yes (6.3) 0.66 (0.48–0.89) 0.67 (0.17–2.68) SOX2 No (56.3) Yes (43.7) 0.74 (0.50–1.09) 0.54 (0.34–0.86) KLHL6 No (59.7) Yes (40.3) 0.73 (0.50–1.07) 0.54 (0.33–0.88) PIK3CA No (63.0) Yes (37.0) 0.66 (0.45–0.95) 0.64 (0.39–1.07) MAP3K13 No (67.2) Yes (32.8) 0.74 (0.52–1.06) 0.49 (0.28–0.86) BCL6 No (68.9) Yes (31.1) 0.71 (0.50–1.01) 0.53 (0.30–0.93) FGF12 No (71.4) Yes (28.6) 0.69 (0.49–0.97) 0.57 (0.31–1.05) Predefined families: any aberration ERBB No (71.0) Yes (29.0) 0.68 (0.48–0.97) 0.51 (0.28–0.92) FGF No (41.2) Yes (58.8) 0.81 (0.51–1.28) 0.52 (0.35–0.76) Favours afatinibFavours erlotinib 0.51.02.01.52.50 SubgroupAberration present (%)HR (95% CI) † 22

23. Specified genetic alterations not predictive of OS benefit with afatinib vs erlotinib SV=single nucleotide variants; CNA=copy number alterations. * Tumour genetic analysis subset; † Interaction p values between yes/no on all markers >0.05. ‡ For this analysis, all available OS information was used, and data after 632 deaths were not censored. Soria JC et al. Oral presentation ORAL32.01 at 16th World Conference on Lung Cancer, Denver, 2015. Soria JC et al. Lancet Oncol. 2015;16(8):897-907. LL8 LL8 subset* (n=238) 0.82 (0.71–0.96) ‡ 0.73 (0.56–0.96) SVs EGFR No (94.1) Yes (5.9) 0.72 (0.54–0.95) 1.04 (0.33–3.25) TP53 No (12.2) Yes (87.8) 0.54 (0.24–1.24) 0.74 (0.55–0.99) LRP1B No (60.5) Yes (39.5) 0.74 (0.52–1.05) 0.73 (0.47–1.13) MLL2 No (67.2) Yes (32.8) 0.68 (0.49–0.96) 0.83 (0.52–1.32) CDKN2A No (71.4) Yes (28.6) 0.75 (0.54–1.03) 0.68 (0.40–1.14) FAT3 No (72.3) Yes (27.7) 0.69 (0.50–0.96) 0.75 (0.44–1.27) CNAs EGFR No (93.7) Yes (6.3) 0.76 (0.57–1.01) 0.42 (0.12–1.43) SOX2 No (56.3) Yes (43.7) 0.78 (0.54–1.11) 0.70 (0.46–1.07) KLHL6 No (59.7) Yes (40.3) 0.76 (0.54–1.08) 0.72 (0.46–1.12) PIK3CA No (63.0) Yes (37.0) 0.72 (0.51–1.02) 0.78 (0.50–1.23) MAP3K13 No (67.2) Yes (32.8) 0.80 (0.57–1.11) 0.66 (0.40–1.08) BCL6 No (68.9) Yes (31.1) 0.74 (0.54–1.03) 0.79 (0.47–1.32) FGF12 No (71.4) Yes (28.6) 0.76 (0.55–1.04) 0.77 (0.45–1.32) Predefined families: any aberration ERBB No (71.0) Yes (29.0) 0.74 (0.54–1.03) 0.70 (0.42–1.16) FGF No (41.2) Yes (58.8) 0.69 (0.45–1.06) 0.76 (0.53–1.08) SubgroupHR (95% CI) † Favours afatinibFavours erlotinib 0.51.02.01.52.50 Aberration present (%) 23

24. Tumour genetic analysis conclusions Squamous NSCLC has a high somatic mutation burden and complex tumour genetic alterations The frequency and pattern of these alterations in the LUX-Lung 8 subset are consistent with prior reports in patients with squamous NSCLC Specified tumour genetic alterations: ‒ are not predictive of benefit with afatinib treatment compared with erlotinib ‒ do not appear to be associated with longer PFS/OS, regardless of treatment It is unlikely that the improved survival outcomes with afatinib were driven by molecular aberrations of EGFR Survival improvements might be a result of the higher potency and broader irreversible ErbB blockade of afatinib in this setting compared with EGFR inhibition alone Soria JC et al. Oral presentation ORAL32.01 at 16th World Conference on Lung Cancer, Denver, 2015. Soria JC et al. Lancet Oncol. 2015;16(8):897-907. 24

25. Return to contents LUX-LUNG 8 Adverse events 25

26. Pattern of AEs consistent with the known profiles of both agents The overall rate of adverse events and CTCAE ≥ grade 3 events was similar for both treatments (afatinib vs erlotinib: 99% vs 97%; 57% vs 57%, respectively) The most common AEs differed between treatment arms: diarrhoea, rash/acne, stomatitis and fatigue with afatinib; rash/acne, diarrhoea, fatigue and pruritus with erlotinib Treatment-related discontinuation due to any AE was similar in both arms (20% vs 17% for afatinib vs erlotinib) The most common AEs with afatinib were predictable and generally manageable through supportive care and dose reduction Includes events that occurred in >10% of patients with grade 1-2 adverse events, or >1% patients with grade 3-5 adverse events in any treatment group. † Group term. AE=adverse event. Soria JC et al. Lancet Oncol. 2015;16(8):897-907. Afatinib (n=392)Erlotinib (n=395) Grade 1Grade 2Grade 3Grade 4Grade 1Grade 2Grade 3Grade 4 Diarrhoea 165 (42%)68 (17%)39 (10%)2 (<1%)94 (24%)28 (7%)9 (2%)1 (<1%) Rash/acne † 157 (40%)83 (21%)23 (6%)0 (0%)142 (36%)83 (21%)41 (10%)0 (0%) Stomatitis † 65 (17%)32 (8%)16 (4%)0 (0%)21 (5%)13 (3%)0 (0%) Fatigue † 33 (8%)20 (5%)6 (2%)0 (0%)24 (6%)17 (4%)7 (2%)0 (0%) Nausea 35 (9%)13 (3%)4 (1%)0 (0%)20 (5%)5 (1%)3 (<1%)0 (0%) Decreased appetite 31 (8%)16 (4%)3 (<1%)0 (0%)24 (6%)15 (4%)2 (<1%)0 (0%) Paronychia † 28 (7%)11 (3%)2 (<1%)0 (0%)9 (2%)7 (2%)1 (<1%)0 (0%) Dry skin 28 (7%)4 (1%)2 (<1%)0 (0%)34 (9%)7 (2%)0 (0%) Pruritus 22 (6%)9 (2%)1 (<1%)0 (0%)37 (9%)10 (3%)0 (0%) Vomiting 20 (5%)8 (2%)3 (<1%)0 (0%)7 (2%)4 (1%)2 (<1%)0 (0%) Dehydration 2 (<1%)5 (1%)3 (<1%)4 (1%)0 (0%) 3 (<1%)0 (0%) 26

27. Return to contents LUX-LUNG 8 Patient-reported outcomes 27

28. Assessment of Patient-Reported Outcomes Assessed using the EORTC core quality of life questionnaire, QLQ-C30, and its lung cancer-specific module, QLQ-LC13 ‒ At the first visit of each treatment course ‒ And at the end of treatment Scores were converted to a 0-100 scale and analysed in line with EORTC scoring algorithms Prespecified symptoms relevant to lung cancer (cough, dyspnoea, and pain) were analysed alongside GHS/QoL for status change, TTD, and change in scores over time ‒ Cough: Question (Q)1 from QLQ-LC13 ‒ Dyspnoea: Q3-Q5 from QLQ-LC13 ‒ Pain: Q9, Q19 from QLQ-C30 ‒ GHS/QoL: Q29, Q30 from QLQ-C30 Questionnaire completion rates were high throughout treatment ‒ Afatinib: range 77.3%-99.0%; erlotinib: range 68.7%-99.0% 28 EORTC=European Organisation for Research and Treatment of Cancer; TTD = time to deterioration. 1. Aaronson et al. J Natl Cancer Inst. 1993;85(5):365-76. 2. Bergman et al. Eur J Cancer. 1994;30A(5):635-42. 3. Gadgeel et al. ASCO 2015. Abstract 8100, Poster 425.

29. 29 Afatinib significantly improved cough and quality of life vs erlotinib Improvement=a linear transformation was applied to standardise the raw score to a range from 0 to 100. A 10-point change is accepted as the threshold for being clinically meaningful (Osoba D et al. J Clin Oncol. 1998;16(1):139-44). EORTC=European Organisation for Research and Treatment of Cancer. Soria JC et al. Lancet Oncol. 2015;16(8):897-907. Gadgeel et al. Abstract 8100 and poster 425 presented at the American Society of Clinical Oncology (ASCO) congress 2015, Chicago, US. P=0.029 P=0.061 P=0.775 P=0.041 The number of patients reporting improved GHS/QoL and cough was significantly higher with afatinib than erlotinib While more patients reported improved dyspnoea with afatinib, the different was not statistically significant as the effect was primarily concentrated in those reporting improvement in “dyspnoea walked” (34.6% vs 26.5%, P=0.022) Improvement in pain was similar across both treatment arms Afatinib (n=398) Erlotinib (n=397) Patients with improvement in symptoms (EORTC scores improved by ≥10 points)

30. 30 Afatinib significantly delayed deterioration of dyspnoea vs erlotinib TTD=time to deterioration, the time from randomisation to first appearance of a score ≥10 points lower than the baseline score. Soria JC et al. Lancet Oncol. 2015;16(8):897-907. Gadgeel et al. Abstract 8100 and poster 425 presented at the American Society of Clinical Oncology (ASCO) congress 2015, Chicago, US. Hazard ratio 0.79 (95% CI, 0.66-0.94) P=0.0078 Afatinib Erlotinib Estimated Probability 0.2 0.4 0.6 0.8 1.0 0.0 Time (months) 3691215271821240 2.6 months 1.9 Trend toward delayed TTD of cough with afatinib vs erlotinib (median 4.5 vs 3.7 months; HR 0.89, 95% CI 0.72–1.09, P=0.2562) TTD of pain was similar across both treatment arms (median 2.5 vs 2.4 months; HR 0.99, 95% CI 0.82–1.18, P=0.8690) Time to deterioration

31. 31 Mean scores for cough, dyspnoea and pain over time favoured afatinib vs erlotinib GHS=global health status. Gadgeel et al. Abstract 8100 and poster 425 presented at the American Society of Clinical Oncology (ASCO) congress 2015, Chicago, US. -20-1001020 Coughing (Q1 from QLQ-C13)604-3.50.009 Dyspnoea (Q3-Q5 from QLQ-LC13)603-3.50.002 Pain (Q9, Q19 from QLQ-C30)609-2.70.038 GHS/QoL (Q29-Q30 from QLQ-C30)602-1.6 Favours afatinibFavours Erlotinib No. of PatientsAdjusted Mean DifferenceP Value There were no significant differences between afatinib and erlotinib for changes in GHS/QoL over time but, with the exception of social functioning, changes in functional scales over time significantly favoured afatinib

32. Return to contents CONCLUSIONS 32

33. 33 Afatinib significantly increased OS vs erlotinib in squamous NSCLC Afatinib represents a new treatment option for 2nd-line squamous NSCLC patients Soria JC et al. Lancet Oncol. 2015;16(8):897-907. LUX-Lung 8 is the largest prospective trial comparing two EGFR targeting agents for second-line treatment of patients with squamous NSCLC Afatinib demonstrated superior efficacy vs erlotinib: Significantly increased PFS (median 2.6 vs 1.9 months; HR 0.81; P=0.0103) Significantly extended OS (median 7.9 vs 6.8 months; HR=0.81; P=0.0077) Significantly improved disease control rate (51% vs 40%; P=0.002) PFS and OS consistent across clinically relevant subgroups Afatinib significantly improved cough and global health status/health-related quality of life vs erlotinib Afatinib and erlotinib demonstrated a pattern of AEs consistent with the known profiles of both agents