IRESSA: A journey of experience from broad to biomarker populations Claire Watkins Global Product Statistician, AstraZeneca EFSPI meeting on Oncology Basel, 24th June 2010

A brief history of IRESSA (gefitinib) Lessons learned Outline A brief history of IRESSA (gefitinib) Lessons learned Looking to the future of biomarker targeted drug development

What is IRESSA and how does it work? http://www.egfr-info.com/EGFR-lung-cancer/

European Indication – approved June 2009 IRESSA is indicated for the treatment of adult patients with locally advanced or metastatic non‑small cell lung cancer (NSCLC) with activating mutations of EGFR‑TK.

Biomarker targeted drug The ideal Biomarker targeted drug Studies Indicated for Biomarker+

Clinical characteristics? Which biomarker? What cut-off? The reality Broad population? Biomarker targeted drug Studies Clinical characteristics? Biomarker(s)? Which biomarker? What cut-off? Indicated for Biomarker+

IRESSA - May 2001 “Dramatic” Tumour shrinkage in patient with metastatic NSCLC 7

IDEAL 1&2 – NSCLC Phase II non-comparative - 2002 250 mg 500 mg 250 mg 500 mg References Kris MG et al. JAMA 2003; 290: 2149-2158. Fukuoka M et al. J Clin Oncol 2003; 21: 2237-2246. IDEAL 2 – USA IDEAL 1 – Japan and Europe Vertical bars represent 95% CI. Kris 2003, Fukuoka 2003

Japan and US approvals Japan – full approval granted July 2002 Indication: Inoperable or recurrent non small cell lung cancer. Precautions related to Indication 1. Efficacy and safety of IRESSA in patients without previous chemotherapy regimens have not been established. 2. Efficacy and safety of IRESSA in post-operative adjuvant therapy have not been established. US – accelerated approval granted May 2003: IRESSA is indicated as monotherapy for the treatment of patients with locally advanced or metastatic non-small cell lung cancer after failure of both platinum- based and docetaxel chemotherapies. The effectiveness of IRESSA is based on objective response rates. There are no controlled trials demonstrating a clinical benefit, such as improvement in disease- related symptoms or increased survival. US – Phase III post approval pre-treated commitment studies including: ISEL – OS superiority vs placebo INTEREST – OS non-inferiority vs docetaxel IBREESE – Symptom improvement superiority vs placebo Question – what is needed from these studies to lift the conditional approval?

ISEL – reports December 2004 OS HR (95% CI) =0.89 (0.77, 1.02) p= 0.0871 by primary stratified log rank test n=1692, deaths=976 [Adjusted Cox analysis HR 0.86 (0.76-0.99) p=0.0299] TTF HR (95% CI) =0.82 (0.73, 0.93) p=0.0006 n=1316, progressions=1137 1.0 1.0 0.8 0.8 0.6 0.6 Proportion without treatment failure Proportion surviving 0.4 0.4 Reference Thatcher N et al. Lancet 2005; 366: 1527-1537. 0.2 IRESSA 0.2 Placebo 0.0 0.0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Months Months Objective Response Rate 8.0% vs 1.3%, p<0.0001 Thatcher 2005

ISEL OS subgroups by smoking status and histology Treatment by smoking interaction test p=0.047 Never smoked (n=375) Ever smoked (n=1317) 1.0 HR 0.67; 95% CI 0.49, 0.92; p=0.012 HR 0.92; 95% CI 0.79, 1.06; p=0.242 IRESSA 0.8 Placebo 0.6 Proportion surviving 0.4 0.2 0.0 2 4 6 8 10 12 14 16 2 4 6 8 10 12 14 16 Treatment by race interaction test p=0.043 Asian origin (n=342) Non-Asian origin (n=1350) 1.0 HR 0.66; 95% CI 0.48, 0.91; p=0.010 HR 0.92; 95% CI 0.80, 1.07; p=0.294 0.8 0.6 References Thatcher N et al. Lancet 2005; 366: 1527-1537. Chang A et al. J Thorac Oncol 2006; 1: 847-855. Proportion surviving 0.4 0.2 0.0 2 4 6 8 10 12 14 16 2 4 6 8 10 12 14 16 Time (months) Cox regression analysis Thatcher 2005, Chang 2006

Regulatory reactions MHLW open public mtg 17th Jan 05 FDA Advisory committee 4th March 05 MHLW open public mtg (2) 10th March 05 MHLW open public mtg (3) 17th March 05 MHLW open public mtg (4) 24th March 05 FDA restricts labelling IRESSA is indicated as monotherapy for the continued treatment of patients with locally advanced or metastatic non-small cell lung cancer after failure of both platinum-based and docetaxel chemotherapies who are benefiting or have benefited from IRESSA Japan – no change to labelling

IRESSA registration Japan EGFR biomarkers ISEL, INTEREST: Unselected trials in pre-treated setting ISEL IRESSA registration Japan INTEREST IPASS 2002 2005 2007 2009 EGFR protein expression EGFR gene copy number IPASS: Clinically selected trial in first line setting EGFR mutations

ISEL: OS by EGFR gene copy number Treatment by gene copy number interaction test p=0.047 High (+) Low (-) N=114, E=68 Cox HR=0.61 (0.36, 1.04) p=0.07 N=256, E=157 Cox HR=1.16 (0.81, 1.64) p=0.42 Percent surviving 1.0 1.0 IRESSA Placebo IRESSA Placebo 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 Reference Hirsch F et al. J Clin Oncol 2006; 24: 5034-5042. 0.0 0.0 2 4 6 8 10 12 14 16 2 4 6 8 10 12 14 16 Time (months) Time (months) OS could not be analysed by EGFR mutation status as there were only 5 mutation positive patients on placebo. The ORR was 38% in the 21 mutation positive patients treated with IRESSA Hirsch 2006

INTEREST: Phase III study of IRESSA vs docetaxel in pre-treated NSCLC Endpoints Patients Progressive or recurrent disease following CT Considered candidates for further CT with docetaxel 1 or 2 CT regimens (≥1 platinum) PS 0-2 Primary Overall survival (co-primary analysesa of non-inferiority in all patients and superiority in patients with high EGFR gene copy number) Secondary Progression-free survival Objective response rate Quality of life Disease related symptoms Safety and tolerability Exploratory Biomarkers EGFR mutation EGFR protein expression EGFR gene copy number K-Ras mutation IRESSA 250 mg/day Docetaxel 75 mg/m2 every 3 weeks 1:1 randomization Reference Kim ES et al. Lancet 2008; 372: 1809-1818. 1466 patients amodified Hochberg procedure applied to control for multiple testing CT, chemotherapy; PS, performance status; EGFR, epidermal growth factor receptor Kim 2008 15

INTEREST: OS and PFS and ORR OS: NI margin 1.154, PP population HR (96% CI) =1.020 (0.905, 1.150) n=1433, deaths=1169 Median survival: IRESSA 7.6m, Docetaxel 8.0m PFS: EFR population HR (95% CI) =1.04 (0.93, 1.18), p=0.466 n=1316, progressions=1137 Median PFS: IRESSA 2.2m, Docetaxel 2.7m 1.0 1.0 IRESSA Docetaxel IRESSA Docetaxel 0.8 0.8 0.6 0.6 Probability of progression-free survival Probability of survival 0.4 0.4 Reference Kim ES et al. Lancet 2008; 372: 1809-1818. 0.2 0.2 0.0 0.0 4 8 12 16 20 24 28 32 36 40 4 8 12 16 20 24 28 32 36 40 Months Months ORR [EFR population]: 9.1% IRESSA, 7.6% Docetaxel; p=0.3257 Kim 2008 16

INTEREST: Summary of key subgroup analyses Overall Survival ORR (%) IRESSA v. Docetaxel Progression-free Survival Overall Ever smoker Never smoker Asian Non-Asian EGFR FISH+ EGFR FISH- EGFR Mutation+ EGFR Mutation- 9.1 v. 7.6 Overall Ever smoker Never smoker 19.7 v. 8.7 Asian 6.2 v. 7.3 Non-Asian 13.0 v. 7.4 EGFR FISH+ 7.5 v. 10.1 EGFR FISH- 42.1 v. 21.1 EGFR Mutation+ 6.6 v. 9.8 EGFR Mutation- Overall Ever smoker Never smoker Asian Non-Asian EGFR FISH+ EGFR FISH- EGFR Mutation+ EGFR Mutation- References Kim ES et al. Lancet 2008; 372: 1809-1818. Douillard et al. J Clin Oncol; 26 (May 20 Suppl.): Abstract 8001. 0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0 2.5 HR (IRESSA vs docetaxel) and 95% CI EFR population HR (IRESSA vs docetaxel) and 95% CI Unadjusted analysis PP population for clinical factors ITT population for biomarker factors Adjusted analysis EFR population Kim 2008; Douillard 2010 17

Carboplatin AUC 5 or 6 and Paclitaxel 200mg/m2 3 wkly IPASS: Phase III study of IRESSA versus doublet chemotherapy in first line NSCLC Endpoints Patients Adenocarcinoma histology Never smokers or light ex-smokers* PS 0-2 Provision of tumour sample for biomarker analysis strongly encouraged Primary Progression free survival (non-inferiority) Secondary Objective response rate Quality of life Disease related symptoms Overall survival Safety and tolerability Exploratory Biomarkers EGFR mutation EGFR gene copy number EGFR protein expression IRESSA 250 mg/day Carboplatin AUC 5 or 6 and Paclitaxel 200mg/m2 3 wkly 1:1 randomization Reference Mok TS et al. N Engl J Med 2009; 361: 947-957. 1217 patients from East Asian countries *Never smokers:<100 cigarettes in lifetime; light ex-smokers: stopped 15 years ago and smoked  10 pack yrs Carboplatin/paclitaxel was offered to IRESSA patients upon progression PS, performance status; EGFR, epidermal growth factor receptor Mok 2009 18

IPASS reports September 2008, partway through the European MAA review of INTEREST

IPASS: Exceeded primary objective and demonstrated superior PFS for IRESSA versus doublet chemotherapy IRESSA Carboplatin / paclitaxel N Events 609 453 (74.4%) 608 497 (81.7%) HR (95% CI) = 0.741 (0.651, 0.845) p<0.0001 IRESSA demonstrated superiority relative to carboplatin / paclitaxel in terms of PFS Reference Mok TS et al. N Engl J Med 2009; 361: 947-957. Mok 2009 Primary Cox analysis with covariates; ITT population HR <1 implies a lower risk of progression on IRESSA HR, hazard ratio; CI, confidence interval; PFS, progression-free survival 20 20 20

Objective response rate 43% vs 32% p=0.0001 IPASS: Superior PFS and ORR with IRESSA vs doublet chemotherapy; PFS effect not constant over time Probability of PFS 1.0 Carboplatin / paclitaxel IRESSA N Events 609 453 (74.4%) 608 497 (81.7%) 0.8 HR (95% CI) = 0.741 (0.651, 0.845) p<0.0001 0.6 Median PFS (months) 4 months progression-free 6 months progression-free 12 months progression-free 5.7 61% 48% 25% 5.8 74% 48% 7% 0.4 Primary objective exceeded: IRESSA demonstrated superiority relative to carboplatin / paclitaxel in terms of PFS 0.2 Reference Mok TS et al. N Engl J Med 2009; 361: 947-957. 0.0 4 8 12 16 20 24 Months At risk : IRESSA 609 363 212 76 24 5 Carboplatin / paclitaxel 608 412 118 22 3 1 Objective response rate 43% vs 32% p=0.0001 Mok 2009 Primary Cox analysis and logistic regression with covariates; ITT population HR <1 implies a lower risk of progression on IRESSA 21 21 21

IPASS: Superior progression-free survival and response rate for IRESSA in EGFR mutation positive patients IRESSA EGFR M+ (n=132) Carboplatin / paclitaxel EGFR M+ (n=129) Probability of PFS 1.0 0.8 EGFR M+ HR=0.48, 95% CI 0.36, 0.64 p<0.0001 0.6 0.4 Objective response rate 71.2% vs 47.3% p=0.0001 0.2 Reference Mok TS et al. N Engl J Med 2009; 361: 947-957. 0.0 4 8 12 16 20 24 Time from randomisation (months) Mok 2009 M+, mutation positive 22 22 22

Objective response rate 1.1% vs 23.5% Time from randomisation (months) IPASS: Superior progression-free survival and response rate for doublet chemotherapy in EGFR mutation negative patients IRESSA EGFR M- (n=91) Carboplatin / paclitaxel EGFR M- (n=85) Probability of PFS 1.0 0.8 EGFR M- HR=2.85, 95% CI 2.05, 3.98 p<0.0001 0.6 0.4 Objective response rate 1.1% vs 23.5% p=0.0013 0.2 Reference Mok TS et al. N Engl J Med 2009; 361: 947-957. 0.0 4 8 12 16 20 24 Time from randomisation (months) Mok 2009 M-, mutation negative 23 23 23

IPASS: EGFR mutation is a strong predictor for differential PFS benefit between IRESSA and doublet chemotherapy IRESSA EGFR M+ (n=132) IRESSA EGFR M- (n=91) Carboplatin / paclitaxel EGFR M+ (n=129) Carboplatin / paclitaxel EGFR M- (n=85) Probability of PFS 1.0 0.8 EGFR M+ HR=0.48, 95% CI 0.36, 0.64 p<0.0001 EGFR M- HR=2.85, 95% CI 2.05, 3.98 Treatment by subgroup interaction test, p<0.0001 0.6 0.4 0.2 Reference Mok TS et al. N Engl J Med 2009; 361: 947-957. 0.0 4 8 12 16 20 24 Time from randomisation (months) Mok 2009 M+, mutation positive; M-, mutation negative 24 24 24

European Indication – approved June 2009 IRESSA is indicated for the treatment of adult patients with locally advanced or metastatic non‑small cell lung cancer (NSCLC) with activating mutations of EGFR‑TK.

Lessons learned Understand the biology Make friends with your translational scientists Determine whether to go down the targeted biomarker route as early as possible “The tissue is the issue” – collect as many samples as you can No sample = no biomarker Pathologists are key Conflict between push for faster studies and push for targeted healthcare Fast recruiters are not often the most experienced at sample collection A targeted drug is useless without a diagnostic Co-development has its own unique challenges Ensure an understanding of prognostic vs predictive A predictive factor cannot be identified from a single arm study A poor prognostic factor can be a good predictive factor for a new agent

Prognostic vs Predictive Not prognostic Prognostic Not predictive + - + - Predictive + - + - Blue=Experimental, Purple=comparator

Lessons learned It matters What you measure How you measure it How you define positive (cut-off) Biomarker status Positive or negative MAGIC ALGORITHM! Tissue sample Diagnostic test

It matters what you measure Gene copy number Protein expression EGFR Gene mutation

It matters how you measure it FISH CISH Fluorescence Gene copy number IHC Protein expression EGFR ARMs Gene mutation PNA-LNA PCR clamp Sequencing

It matters how you define positive (cut-off) Staining percentage # of copies Staining intensity FISH CISH Fluorescence Gene copy number IHC Pattern of copies Protein expression EGFR ARMs Gene mutation New diagnostics may use more than one biomarker to define positivity PNA-LNA PCR clamp Sequencing Type of mutation

INTEREST: Overlap of biomarkers (EGFR gene copy number by FISH, EGFR expression by IHC, EGFR mutation) EGFR expression + n=189 EGFR FISH + n=117 n=73 n=16 n=84 +++ n=24 249 patients evaluable for EGFR expression, FISH and mutations Reference Douillard et al. J Clin Oncol 2009, in press. 4 EGFR mutation + n=39 3 n=8 --- n=37 Douillard 2010 32

Lessons learned It matters What you measure How you measure it How you define positive (cut-off) Consider if there is a surrogate for the biomarker e.g. clinical characteristics, another marker Biomarker status Positive or negative MAGIC ALGORITHM! Tissue sample Diagnostic test

Overall EGFR mutation positive rate 14.8% (44/297) INTEREST: EGFR mutation appeared to be associated with some clinical characteristics % of samples EGFR mutation positive 60 50 40 30 20 10 Reference Douillard et al. J Clin Oncol 2009, in press. Male PS 2 Female Asian PS 0-1 Second-line Third-line Non-Asian Never-smoked Ever-smoked Adenocarcinoma Non- adenocarcinoma Overall EGFR mutation positive rate 14.8% (44/297) Douillard 2010

K-Ras and EGFR mutations rarely co-exist in the same tumour 5 incidences across 19 studies totalling around 3300 patients Study AstraZeneca studies INTEREST ISEL INVITE Literature Wu et al 2008 Yamamoto et al 2008 Zhu et al 2008 Do et al 2008 Sasaki et al 2008 Na et al 2007 Massarelli et al 2007 Bae et al 2007 Hirsch et al 2006 van Zandwijk et al 2007 Yokoyama et al 2006 Suzuki et al 2006 Tam et al 2006 Tomizawa et al 2005 Shigematsu et al 2005 N evaluable 275 152 90 237 86 206 200 190 133 70 115 349 150 215 120 617 K-Ras mutations N (%) K-Ras+ 49 (17.8) 12 (7.9) 24 (26.7) 9 (3.8) 26 (30.2) 30 (14.6) 25 (12.5) 21 (11.1) 17 (12.8) 16 (22.9) 6 (5.2) 21 (6.0) 6 (4.0) 21 (9.8) 4 (3.3) 50 (8.1) N evaluable 297 215 65 235 86 204 200 195 133 71 115 41 349 150 241 120 519 EGFR mutations N (%) M+ 44 (14.8) 26 (12.1) 6 (9.2) 96 (40.9) 10 (11.6) 34 (16.7) 73 (36.5) 82 (42.1) 32 (24.1) 7 (9.9) 20 (17.4) 13 (31.7) 102 (29.2) 38 (25.3) 116 (48.1) 29 (24.2) 120 (23.1) Number K-Ras+/M+ 1 3 35

Lessons learned Engage with regulators early Everyone is learning as they go along FDA in particular has stated positions that may not be practical in all cases >90% evaluable samples Prove don’t work in –ve

IPASS: Attrition factors in biomarker analysis 1217 randomised patients (100%) Reasons for samples not evaluable: Sample not available, insufficient quantity to send, cytology only, sample at another site 1038 biomarker consent (85%) 683 provided samples (56%) 565 histology 118 cytology Evaluable for: EGFR mutation: 437 (36%) EGFR gene copy number: 406 (33%) EGFR expression: 365 (30%) Reference Mok TS et al. N Engl J Med 2009; 361: 947-957. Mok 2009, Fukuoka 2009 37 37 37

Lessons learned Engage with regulators early Everyone is learning as they go along FDA in particular has stated positions that may not be practical in all cases >90% evaluable samples Prove don’t work in –ve Don’t want to do a repeat of Phase IIIs Issues of generating a strong signal in a small early study Payers are key stakeholders Randomised Phase IIs Keep an eye to the future New or revised tests, markers, tissue types Flexible consent Be aware that science will move on as your study is ongoing

Personalised Healthcare development today and in the future Predictive biomarker for IRESSA discovered by external collaborator ~7 years after start of clinical trials Took ~4.5 further years retrospective research to show significant increase in clinical benefit for those patients identified by diagnostic test Ultimately identified patients most likely to benefit offers an alternative treatment option to doublet chemotherapy in newly diagnosed advanced/metastatic NSCLC 2013 Personalised Healthcare research discovers predictive biomarker in preclinical models before start of clinical development Early engagment with payers and health authorities ensures that drug is targeted to patients likely to respond Clinical programme prospectively tailored for responders, used for co-development of drug and diagnostic Drug launched globally, linked to diagnostic

Summary IRESSA is approved in Europe for a biomarker targeted population But it took a long time to get there In future, pharmaceutical companies are unlikely to be able or willing to follow a similar development path for new agents There are several useful learnings for future biomarker targeted products Understand the science Maximise tissue samples Diagnostic is as important as the drug Pharmaceutical companies and regulators are learning about this together Engage early Considerable challenges on both sides Opportunity for collaboration

References Kris MG, Natale RB, Herbst RS, et al: Efficacy of gefitinib, an inhibitor of the epidermal growth factor receptor tyrosine kinase, in symptomatic patients with non-small cell lung cancer: A randomized trial. JAMA 290:2149-2158, 2003 Fukuoka M, Yano S, Giaccone G, et al: Multi-institutional randomized phase II trial of gefitinib for previously treated patients with advanced non-small-cell lung cancer. J Clin Oncol 21:2237-2246, 2003 Thatcher N, Chang A, Parikh P, et al. Gefitinib plus best supportive care in previously treated patients with refractory advanced non-small-cell lung cancer: results from a randomised, placebo-controlled, multicentre study (Iressa Survival Evaluation in Lung Cancer). Lancet 366: 1527–37, 2005 Chang A, Parikh P, Thongprasert S, et al: Gefitinib (IRESSA) in Patients of Asian Origin with Refractory Advanced Non-small Cell Lung Cancer: Subset Analysis from the ISEL Study. J Thoracic Oncol 1: 8: 847-855, 2006 Hirsch FR, Varella-Garcia M, Bunn PA, et al. Molecular predictors of outcome with gefitinib in a Phase III placebo-controlled study in advanced non-small-cell lung cancer. J Clin Oncol 24: 5034-5042, 2006 Kim ES, Hirsch V, Mok T, et al: Gefitinib versus docetaxel in previously treated non-small-cell lung cancer (INTEREST): a randomised phase III trial. Lancet 372:1809-1818, 2008 Douillard JY, Hirsch V, Mok T, et al: Molecular analyses from a phase III trial comparing gefitinib with docetaxel in previously treated non-small-cell lung cancer (INTEREST). J Clin Oncol 26 (May 20 Suppl): Abstract 8001, 2008 Mok T, Wu Y, Thongprasert S, et al: Gefitinib or carboplatin-paclitaxel in pulmonary adenocarcinoma NEJM 361: 947-957, 2009 Douillard J, Hirsh V, Mok T, et al: Molecular predictors of outcome with gefitinib and docetaxel in previously treated non-small-cell lung cancer: data from the randomised phase III INTEREST trial J Clin Oncol 5:744-752, 2010 Fukuoka M, Wu Y, Thongprasert S, et al. Biomarker analyses from a phase III, randomized, open-label, first-line study of gefitinib (G) versus carboplatin/paclitaxel (C/P) in clinically selected patients (pts) with advanced non-small cell lung cancer (NSCLC) in Asia (IPASS). J Clin Oncol 27 (15s suppl): Abstract 8006, 2009