2. 2 Lucio Crinò Medical Oncology Department University Hospital Perugia, Italy The optimal therapeutic algorithm for EML4-ALK + ve pts

3. “Anaplastic Lymphoma kinase”(ALK)- rearrangment 3-5% of lung adenocarcinomas ALK signal transduction¹ALK fusions² 1. Roskoski jr. Pharma research 2013 2. Peters et al. Lung Cancer 2013

4. ‘’c-ros oncogene 1” (ROS1) rearrangment 1% of lung adenocarcinomas (70% of homology with ALK) Vie di trasduzione di ROS1¹ Partner di fusione di ROS1² 1. Ou et al. Expert Anticancer Ther 2012 2. Gainor et al. Oncologist 2013

6. Timeline Mano H Cancer Discovery 2012;2:495-502

7. Clinical development of Crizotinib for ALK+ NSCLC *Cisplatin or carboplatin according to investigator’s choice ≠ Cross-over to crizotinib allowed at PD in the standard arm **Pemetrexed or docetaxel; prior chemo must have been platinum-based chemotherapy ∞ May have received Pemetrexed or Docetaxel from previous phase III PROFILE 1007 trial and discontinued treatment due to RECIST-defined progression Study Phase (planned accrual) Histology Line of therapy Study design Primary endpoint PROFILE 1014 III (334 pts) Non-squamous1st Platinum*-Pemetrexed vs Crizotinib PFS ≠ PROFILE 1007 III (318 pts) NSCLC2nd 2 nd line chemo** vs Crizotinib PFS PROFILE 1005 II (400 pts) NSCLC 3 rd or more ∞ Crizotinib monotherapy ORR ORR = overall response rate; PFS = progression-free survival; pts = patients

8. StudyNo. of patientsRR (%)PFS (mos.) A808100114360.89.7 A808100526159.8 ∞ 8.1 A808100717365 * 7.7 French Temporary Authorization for use of Crizotinib 23056.5Not reported Crizotinib for ALK+ NSCLC ∞ 259 pts evaluable for response *Independent radiologic review Camidge, et al. Lancet Oncol 2012 Kim, et al. ASCO 2012 Shaw, et al. NEJM 2013 Perol, et al. ECCO 2013

9. Tumor responses to crizotinib by patient Median time to response: 8 wk 1. Camidge et al., ASCO 2011; Abs #2501 2. Riely et al., IASLC 2011; Abs #O31.05 PROFILE 1005 2 PROFILE 1001 1

10. Study Design Key entry criteria ●ALK+ by central FISH testing ●Stage IIIB/IV NSCLC ●1 prior chemotherapy (platinum-based) ●ECOG PS 0−2 ●Measurable disease ●Treated brain metastases allowed N=318 Crizotinib 250 mg BID PO, 21-day cycle (n=159) Pemetrexed 500 mg/m 2 or Docetaxel 75 mg/m 2 IV, day 1, 21-day cycle (n=159) PROFILE 1007: NCT00932893 Endpoints ●Primary –PFS (RECIST 1.1, independent radiology review) ●Secondary –ORR, DCR, DR –OS –Safety –Patient reported outcomes (EORTC QLQ-C30, LC13) RANDOMIZERANDOMIZE CROSSOVER TO CRIZOTINIB ON PROFILE 1005 a Stratification factors: ECOG PS (0/1 vs 2), brain metastases (present/absent), and prior EGFR TKI (yes/no) a

11. a RECIST v1.1 65.3 19.5 ORR (%) ORR ratio: 3.4 (95% CI: 2.5 to 4.7); P<0.001 Crizotinib (n=173) PEM/DOC (n=174) 80 60 40 20 0 Treatment 65.7 29.3 6.9 Crizotinib (n=172) PEM (n=99) DOC (n=72) Treatment 80 60 40 20 0 ORR in PROFILE 1007

12. Crizotinib (n=172 a ) PEM (n=99 b ) DOC (n=72 b ) Events, n (%)100 (58)72 (73)54 (75) Median, mo7.74.22.6 HR c (95% CI)0.59 (0.43 to 0.80)0.30 (0.21 to 0.43) P<0.001 Probability of survival without progression (%) 100 80 60 40 20 0 0510152025 Time (months) 17293381120 993612310 7213310 No. at risk Crizotinib PEM DOC a Excludes 1 patient who did not receive study treatment; b excludes 3 patients in chemotherapy arm who did not receive study treatment; c vs crizotinib PROFILE 1007: PFS of Crizotinib vs Pemetrexed or Docetaxel

13. Shaw, et al. NEJM 2013 Survival curves from PROFILE 1007 PFSOS

14. Survival in ALK-positive NSCLC with crizotinib versus crizotinib-naive controls 0 0% 20% 40% 60% 80% 100% Overall survival (years) 1234 ALK Crizotinib (n=30) ALK Control (n=23) Median Survival, mo NR 6 1-yr Survival, % 70 44 WT/WT Control (n=125) 11 47 From 2 nd /3rd line crizotinib 2-yr Survival, % 55 12 32 HR = 0.49, p=0.02 Camidge D R, Lancet Oncol 2012; 13: 1011–19

15. PROFILE 1014 Study Design

16. Primary Endpoint Met: Crizotinib Superior to Pemetrexed-based Chemotherapy in Prolonging PFSa T. Mok – ASCO 2014

17. Secondary Endpoints: ORRa and OS

18. Crizotinib Therapy for Patients with Advanced ROS1-rearranged Non-Small Cell Lung Cancer S-HI Ou, 1 D-W Kim, 2 DR Camidge, 3 GJ Riely, 4 R Salgia, 5 GI Shapiro, 6 BJ Solomon, 7 JA Engelman, 8 JW Clark, 8 L Tye, 9 KD Wilner, 9 T Usari, 10 M Varella-Garcia, 3 K Bergethon, 11 AJ Iafrate, 8 AT Shaw 8 1 University of California at Irvine, Irvine, CA, USA; 2 Seoul National University Hospital, Seoul, South Korea; 3 University of Colorado Cancer Center, Aurora, CO, USA; 4 Memorial Sloan-Kettering Cancer Center, New York, NY, USA; 5 University of Chicago, Chicago, IL, USA; 6 Dana-Farber Cancer Institute, Boston, MA, USA; 7 Peter MacCallum Cancer Centre, Melbourne, Australia ; 8 Massachusetts General Hospital Cancer Center, Boston, MA, USA; 9 Pfizer Oncology, La Jolla, CA, USA; 10 Pfizer Oncology, Milan, Italy; 11 Duke University Medical School, Durham, NC, USA; Presented at the 15th World Conference on Lung Cancer (WCLC 2013), Sydney, Australia, October 27 -30, 2013 Abstract 2777

19. Clinical Characteristics of Patients with ROS1-positive NSCLC CharacteristicCrizotinib (N=42) Age, years Median (range)52 (31–77) Sex, n Male/female19/23 Smoking history, n (%) Never33 (79) Former 9 (21) Race, n (%) Caucasian21 (50) Asian18 (43) Other3 (7) Histology, n(%) Adenocarcinoma 41 (98) a ECOG PS, n (%) 021 (50) 120 (48) 2 1 (2) b Prior regimens for advanced/metastatic disease, n (%) None 6 (14) 1 regimen18 (43) >1 regimen18 (43) a Includes one patient with adenocarcinoma-favored, poorly differentiated NSCLC b Patient had an ECOG PS of 1 at screening and 2 on cycle 1 day 1

20. Best Tumor Responses in Evaluable Patients with Advanced ROS1-positive NSCLC a +Treatment ongoing; duration of response/SD is from first documentation of tumor response/first dose to the time of PD or death. For ongoing patients, duration of response/SD is from first documentation of tumor response/first dose to last available on-treatment scan. Duration is in weeks. a Excludes patients with early death (n=2) *This patient ALK+ Data as of April 24, 2013. Best change from baseline (%) PD Best overall response SD PR CR 36 evaluable patients; 2 CRs and 20 PRs Overall response rate: 61% (95% CI: 44–77) Disease control rate: 81% (8 weeks), 67% (16 weeks) 100 80 60 40 20 0 –20 –40 –60 –80 –100 68+ 33+84+ 47+ 44 18 45+ 41+ 29+ 28+ 16+ 12 15+ 23+ 17+ 23+ 24 81+ 7+ 15+ 65+ 7+ 60+ 21+ 5+ 68+ 12+ 19 20+ 113+ 8 *

21. Crizotinib at Perugia’s hospital: patients’ characteristics CARATTERISTICHE PAZIENTI n=46 Median age 48 (24-71) Hystological type Adenocarcinoma 46 (100.0%) Moloecular profile ALK+ ROS1+ 39 (84.7%) 7 (15.3%) Sex Male Female 18 (39.1%) 7 (60,9%) Performance status 0-1 2 ≥3 36 (78.3%) 8 (17.4%) 2 (4.3%) Smoking history Never smokers smokers 38 (82.6%) 8 (17.4%) Brain metastases 10 (21.7%) Previous lines of treatment 0 1 2 ≥3 1 (2.1%) 23 (50.0%) 12 (26.1%) 11 (23.8%) Adenocarcinoma histology Young Mainly never smokers

22. Crizotinib at Perugia’s hospital: ACTIVITY RESPONSE n=46 Best response CR PR SD PD 1 (2.2%) 32 (69.6%) 7 (15.2%) 6 (13%) ORR 71.8% Diesease control 86.9%

23. Crizotinib at Perugia’s hospital: EFFICACY Median Follow-up = 10 months (range 1-42) Survival at 1 year 75% Survival at 2 years 71% 16.9 Median PFS in PROFILE 1007 : 7.0 months

24. Emerging issues in management of crizotinib-treated, ALK-positive patients

26. OS from Start of Initial Crizotinib Treatment Ou et al, Ann Oncol 2014 Median OS (95% CI) CBPD: 29.6 months (23.1−NR) No CBPD: 10.8 months (8.9−14.7) HR=0.30 (0.19−0.46) p<0.0001 Number at risk Continued1201043061 Did not continue744080 0510152025303540 Time (months) 100 80 60 40 20 0 Shaded areas are 95% Hall-Wellner confidence bands Continued crizotinib Did not continue crizotinib Probability of survival (%)

27. Otterson, et al. ASCO 2012 a Excluding patients with target lesions only: patients could be counted more than once across organ sites Most common sites of PD in patients continuing crizotinib beyond PD

28. Study Population: ALK+ NSCLC With or Without Brain Metastases at Baseline 1. Kim D-W, et al. J Clin Oncol 2012;30(Suppl.) (abstr. 7533); 2. Shaw AT, et al. N Engl J Med 2013;368:2385–2394 Crinò et al. Poster presented at European Cancer Congress 2013 (abstract 3413) Previously untreated for BM (n=109) No BM detected (n=613) RETROSPECTIVE ANALYSIS of crizotinib-treated patients with or without asymptomatic BM at baseline (n=888) PROFILE 1005 1 (open-label, single-arm phase II) Crizotinib 250 mg BID PROFILE 1007 2 (randomized phase III vs. standard chemo) Crizotinib 250 mg BID Previously treated for BM (n=166) ●Among evaluable patients: – 20% of patients (22/109) with previously untreated asymptomatic BM had BM classified as target lesions – 11% of patients (18/166) with previously treated asymptomatic BM had BM classified as target lesions – 9% of patients (55/613) with no detectable BM at baseline developed symptomatic BM after the start of crizotinib treatment ●Median duration of crizotinib treatment similar in the three groups (22.0–29.3 weeks) 12% 19% 69%

29. ●IC and systemic DCR at 12 weeks: 56–65% in patients with BM at baseline ●IC ORR: 25% in 40 patients with BM classified as target lesions ●Systemic ORR: 46–55% across the three groups Crizotinib Antitumor Activity in Patients With or Without Brain Metastases at Baseline Previously untreated for BM (n=109) Previously treated for BM (n=166) No BM detected (n=613) nOutcome n n DCR at 12 weeks, % (95% exact CI) IC10956 (46−66) 16662 (54−70) NA Systemic10963 (54−72) 16665 (57−72) 61371 (68−75) ORR, % (95% exact CI) IC1097 (3−14) 1667 (4−12) NA Patients with target-lesion BM2218 (5−40) 1833 (13−59) NA Systemic10953 (43−63) 16646 (39−54) 61355 (51−59) Median time to tumor response (range), a weeks IC86.0 (4.9−12.4) 126.4 (5.9−17.7) NA Systemic586.1 (2.0−31.4) 776.1 (3.1−35.3) 3366.1 (3.0−49.1) Median duration of response b (range), a weeks IC826.4 (6.1−59.3) 12 NR (6.0−59.9) NA Systemic5847.9 (5.3−55.0) 7755.6 (4.4−95.3) 33649.0 (4.1−96.1) Median systemic PFS, b (95% CI), c mo1098.3 (6.7−14.0) 16613.5 (6.2−16.5) 6139.9 (8.8−12.2) NA, not applicable; NR, not reached a In patients with confirmed objective response; b Kaplan−Meier method; c Brookmeyer−Crowley method Crinò et al. Poster presented at European Cancer Congress 2013 (abstract 3413)

30. Emerging issues: brain metastases a Patients with one intracranial target lesion at baseline (n=33, 7 patients with early death/indeterminate response excluded). Costa DB, et al. Oral presentation at World Congress on Lung Cancer, October 27–30, 2013, Sydney, Australia: Abstract 2932. INTRACRANIAL DISEASE CONTROL RATE AT 12 WEEKS IN PATIENTS WITH BASELINE ASYMPTOMATIC BRAIN METASTASES 4 BEST PERCENTAGE CHANGE IN INTRACRANIAL TARGET LESIONS FROM BASELINE (%) *Patients previously treated for brain metastases. Best objective response according to RECIST. * * * * * * * * * * * * Progressive disease Stable disease Partial responseComplete response * * 40 20 0 −20 −40 −60 −80 −100

31. Mechanisms of crizotinib resistance

32. Frequencies of crizotinib resistance mechanisms in ALK+ NSCLC Unknown 25% Alk mutation 22-33%  L1196M  G1202R  S1206Y  G1269A  1151Tins  Others KIT amplification 10 % Change in driver mutations 5% Alk amplification 6-16% Increased EGFR signaling 30-35% ALK non-dominant (Bypass tracks) ALK dominant Camidge D. R. Nat. Rev. Clin. Oncol. 11, 473–481 (2014) ;

33. 2 nd generation ALK-TKIs Acquired resistance ALK translocated Crizotinib All of them seem to be very good 1) Better affinity for ALK 2) Better affinity for crizotinib resistant second-site mutated ALK 3) Improvement in pharmacokinetics to brain tissue and CSF Options at acquired resistance to Crizotinib Switch to chemotherapy Crizotinib beyond progression + CHT or Hsp90I Crizotinib beyond progression

34. Ceritinib: Highly Active Treatment Option for ALK-Positive NSCLC Presented By Howard West at 2014 ASCO Annual Meeting

35. Toxicity Challenges with Ceritinib Presented By Howard West at 2014 ASCO Annual Meeting

36. AP26113 IN CRIZOTINIB-RESISTANT ALK-REARRANGED NSCLC

37. Ceritinib, Alectinib and AP26113, show antitumor activity in ALK+ NSCLC with brain metastasis Mehra et al., ASCO (2012), abstr 3007 Gettinger et al., ESMO (2012), abstr 4390 Nishio et al., ESMO (2012), abstr 4410 6 wks BaselineBaselineLDK378LDK378 AP26113AP26113 CH5424802CH5424802BaselineBaseline 8 wks BaselineBaseline 33 wks

38. Tumor Responses to Crizotinib in ROS1- rearranged NSCLC Shaw AT et al. N Engl J Med 2014. ORR = 72% (95% CI, 58 to 84) DOR = 17.6 mos (95% CI, 14.5-not reached) PFS = 19.2 mos (95% CI, 14.4-not reached)

39. Efficacy of crizotinib in MET-amplified NSCLC a Confirmed objective responses. b Based on investigator assessment. c Two patients in the intermediate MET group had an unconfirmed PR that was not confirmed in a second assessment. Best percent change from baseline in target tumor lesions a by patient Low MET n=2 Intermediate MET n=6 High MET n=6 100 80 60 40 20 0 –20 –40 –60 –80 –100 100 80 60 40 20 0 –20 –40 –60 –80 –100 Disease progression Stable disease Partial response b Complete response b % Change From Baseline 100 80 60 40 20 0 –20 –40 –60 –80 –100 Threshold for partial response c c Presented By D. Camidge at 2014 ASCO Annual Meeting

40. Squamous cell cancer Non-squamous cancer Platinum + 3 rd generation agent EGFR WT, ALK neg EGFR mut + ALK/ROS1 rearranged EGFR-TKI ALK-TKI EGFR WT/ALK neg, clinically selected EGFR WT/ALK neg, clinically unselected Platinum + Pemetrexed Platinum-based doublet + Bevacizumab Oncogene addicted Biologically-unselected non-squamous NSCLC

42. Istituto Toscano Tumori – Livorno, Italy First-line therapy for metastatic NSCLC in 2014 Stratification for EGFR, ALK and histology EGFR Mut+ ALK-/EGFR wt non-squamous ALK-/EGFR wt squamous EGFR TKI Platinum doublet + bevacizumab OR platinum + pemetrexed +/- bevacizumab Platinum-based doublet ALK rearranged Crizotinib

43. SUMMARY Crizotinib is the first-in-class ALK-TKI inhibitor fully approved worldwide Consistent response rate, PFS >60% over 8 months, very favourable toxicity profile Evidence of clinical benefit in continuous treatment beyond smouldering progression and in brain metastases in selected patients with or without radiotherapy Significant activity in ROS1 rearranged patients