Immunoterapia Silvia Novello silvia.novello@unito.it

Lung Vaccination in Lung Cancer: ph III trials

Giaccone G et al, ESMO 2013 MAGRIT ha screenato più di 13000 pazienti

Rationale for immunotherapy in nsclc “…we show that in NSCLCs treated with pembrolizumab, elevated nonsynonymous mutation burden strongly associates with clinical efficacy. …clinical efficacy correlates with a molecular signature characteristic of tobacco carcinogen–related mutagenesis, certain DNA repair mutations, and the burden of neoantigens.” In the current study, we show that in NSCLCs treated with pembrolizumab, elevated nonsynonymous mutation burden strongly associates with clinical efficacy. Rizvi NA et al, Science April 2015

Clinical development of pd-1 and pd-l1 inhibitors Nivolumab Fully human IgG4 mAb BMS ph III Pidilizumab Humanized IgG1 mAb Cure Tech ph II Pembrolizumab Humanized IgG4 mAb Merck AMP-224 Recombinant PD-L2 Fc fusion protein GSK ph I PD-L1 BMS-936559 MEDI4736 Engineered human IgG1 mAb MedImmune MPDL3280A (Atezolizumab) Genentech MSB0010718C EMD Serono

The efficacy Finally something in Squamous and in Smokers Time to response & the duration of response Which criteria to assess the response Which Biomarkers Think about “combo” or forget “combo” (?) The toxicity profile How to inform the patient Approval issues

The efficacy Finally something in Squamous and in Smokers Time to response & the duration of response Which criteria to assess the response Which Biomarkers Think about “combo” The toxicity profile How to inform the patient Approval issues

CheckMate 057 (NCT01673867) Study Design Randomize 1:1 Stage IIIB/IV non-SQ NSCLC 1 prior PT-DC Prior maintenance therapy with pemetrexed, bevacizumab, or erlotinib alloweda Prior TKI therapy allowed for known ALK translocation or EGFR TKI for known EGFR mutation ECOG PS 0–1 N = 582 NIVO 3 mg/kg IV Q2W until PD or unacceptable toxicity n = 292 DOC 75 mg/m2 IV Q3W until PD or unacceptable toxicity n = 290 Endpoints Primary OS Secondary Investigator-assessed ORR and PFS Efficacy by tumor PD-L1 expression level Quality of life (LCSS) Pts stratified by prior maintenance therapy and line of therapy (second- vs third-line) a Not considered a separate line of therapy. ALK = anaplastic lymphoma kinase; DBL = database lock; ECOG PS = Eastern Cooperative Oncology Group performance status; EGFR = epidermal growth factor receptor; IV = intravenous; LCSS = Lung Cancer Symptom Scale; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; PD = progressive disease; TKI = tyrosine kinase inhibitor

Minimum follow-up for OS was approximately 13.2 mo Overall Survival NIVO DOC Number of pts at risk 292 232 194 169 146 123 62 32 9 290 244 150 111 88 34 10 5 50 80 Time (months) OS (%) 100 90 70 60 40 30 20 27 21 18 15 12 6 3 24 1-yr OS rate = 51% 1-yr OS rate = 39% NIVO N = 292 DOC N = 290 mOS, mo 12.2 9.4 HR = 0.73 (96% CI: 0.59, 0.89); P = 0.0015 Minimum follow-up for OS was approximately 13.2 mo Symbols represent censored observations. CI = confidence interval; HR = hazard ratio. Paz Ares L et al, ASCO 2015

Treatment Effect on OS in Predefined Subgroups Unstratified HR (95% CI) Overall 582 0.75 (0.62, 0.91) Age Categorization (years) <65 339 0.81 (0.62, 1.04) ≥65 and <75 200 0.63 (0.45, 0.89) ≥75 43 0.90 (0.43, 1.87) Gender Male 319 0.73 (0.56, 0.96) Female 263 0.78 (0.58, 1.04) Baseline ECOG PS 179 0.64 (0.44, 0.93) ≥1 402 0.80 (0.63, 1.00) Smoking Status Current/Former Smoker 458 0.70 (0.56, 0.86) Never Smoked 118 1.02 (0.64, 1.61) EGFR Mutation Status Positive 82 1.18 (0.69, 2.00) Not Detected 340 0.66 (0.51, 0.86) Not Reported 160 0.74 (0.51, 1.06) NIVO DOC 1 2 All randomized pts (NIVO, n = 292; DOC, n = 290). HR was not computed for other subsets with less than 10 pts per treatment group.

Baseline Tumor PD-L1 Expression   PD-L1 expression level 1%a 5%a 10%a Not quantifiableb Positive (≥1%) Negative (<1%) Positive (≥5%) Negativ e (<5%) Positive (≥10%) Negative (<10%) NIVO n (%) 123 (53) 108 (47) 95 (41) 136 (59) 86 (37) 145 (63) 61 (21) DOC 123 (55) 101 (45) 86 (38) 138 (62) 79 (35) 145 (65) 66 (23) 78% (455/582) of randomized pts had quantifiable PD-L1 expression a Number and percent of evaluable pts with membranous staining at the respective expression level in ≥100 tumor cells; b Number and percent of randomized pts with PD-L1 expression not quantifiable. Paz Ares L et al, ASCO 2015

Garon EB, ESMO 2014 Garon EB et al , NEJM Apr 19 2015

Keynote001: outcomes on the basis of PD-L1 staining PFS OS proportion score of at least 50% was associated with a higher response rate and longer progression- free and overall survival than was a proportion score of less than 50% in both previously untreated patients and previously treated patients, which indicates that this is a subgroup of patients in whom the PD-L1 pathway can be successfully targeted. Garon EB et al , NEJM Apr 19 2015

Phase II Trial of Pembrolizumab for Untreated Brain Metastases Consider radiation or surgery to progressing lesions Key Eligibility: Advanced NSCLC or melanoma At least 1 untreated or progressive brain metastasis 5-20mm No neurologic symptoms or steroid requirement PS 0-1 PD-L1 expression after the most recent systemic therapy Brain metastasis PD Pembrolizumab 10mg/kg q2 weeks Continue pembrolizumab if systemic control achieved Brain metastasis CR, PR, or SD Safety evaluation at 4 weeks: Brain MRI Response evaluation every 8 weeks: CT chest/abdomen/pelvis Primary Endpoint: Brain Metastasis Response Rate Secondary Endpoints: Overall response rate, safety, PFS, OS Brain metastasis response by modified RECIST allows up to 5 target lesions with diameter >5mm Data presented from an interim analysis with cut-off date June 30, 2015 Sarah B. Goldberg, et al; WCLC 2015, ORAL 31

Phase II Trial of Pembrolizumab for Untreated Brain Metastases Total evaluable patients Brain Metastasis Responses (CR + PR) Duration of BrM Response Systemic Responses N No. of patients % 95% CI Individual duration (months) 18 6* 33 0.14-0.59 3.2+, 6.0+, 6.1+, 6.6, 7.0+ 6 Median Survival 7.7 months (95% CI 3.5-NR) Sicuramente piccoli N ma c’è una certa attività e in 4 casi su 18 una RCompleta Sarah B. Goldberg, et al; WCLC 2015, ORAL 31

The efficacy Finally something in Squamous and in Smokers Time to response & the duration of response Which criteria to assess the response Which Biomarkers Think about “combo” or forget “combo” (?) The toxicity profile How to inform the patient Approval issues

CheckMate 017 (NCT01642004) - Study Design Endpoints Primary: OS Secondary: Investigator-assessed ORR Investigator-assessed PFS Correlation between PD-L1 expression and efficacy (ORR, OS, PFS), Quality of life (LCSS) NIVO 3 mg/kg IV Q2W, until PD or unacceptable toxicity n = 135 Stage IIIb/IV SQ NSCLC 1 prior platinum doublet-based chemotherapy ECOG PS 0–1 N = 272 Randomize 1:1 DOC 75 mg/m2 IV Q3W, until PD or unacceptable toxicity n = 137 DBL: December 15, 2014 At time of DBL, 199 deaths (of 272 randomized pts) were reported (86% of 231 deaths required for final analysis) The boundary for declaring superiority for OS at the pre-planned interim analysis was P <0.03 DBL, data base lock; ECOG PS = Eastern Cooperative Oncology Group performance status; IV = intravenous; LCSS = lung cancer symptom scale; PD = progressive disease; PD-L1 = programmed cell death ligand 1; SQ = squamous; Q2W = every 2 weeks; Q3W = every 3 weeks Brahamer J et al, NEJM 2015

Checkmate 017: Updated follow-up Reckamp K et al. , WCLC 2015 ORAL 02.01

Checkmate 017: Updated follow-up Reckamp K et al. , WCLC 2015 ORAL 02.01

Smoking status and response to immunotherapy in NSCLC In studies of nivolumab, a history of smoking in patients with NSCLC was associated with improved clinical response and PFS Variable ORR, % (n/N) [95% CI] P-value Smoking exposure ≤5 pack-yrs 0 (0/14) [0, 23] 0.018 >5 pack-yrs 30 (20/66) [20, 43] 14 3 1 75 28 16 12 7 ≤5 pack-yrs smokers >5 pack-yrs 20 40 60 80 100 PFS (%) 6 18 24 30 Months Since Treatment Initiation ≤5 pack-yrs smokers (mPFS 1.7 months) >5 pack-yrs smokers (mPFS 2.2 months) HR (95% CI) = 0.41 (0.22, 0.74), P = 0.003 PFS by smoking exposure HR = hazard ratio; mPFS = median progression-free survival; ORR = objective response rate; PFS = progression-free survival. HR = hazard ratio; mPFS = median progression-free survival; ORR = objective response rate; PFS = progression-free survival. Hellmann MD, et al. Poster presented at ESMO 2014 (asbtr. 1229PD).

MPDL3280A Phase Ia: Response by Smoking and Mutational Status Former / Current Smokers Never Smokers Response by Smoking Status (ORRa) Smoking Status (NSCLC; n = 53) Pts With PR, % EGFR Mutant EGFR Status (NSCLC; n = 53) Unknown Response by EGFR Status (ORRa) KRAS Status (NSCLC; n = 53) Response by KRAS Status (ORRa) KRAS Mutant 11/43 1/10 9/40 1/6 8/27 a ORR includes investigator-assessed u/c PR by RECIST 1.1. Patients first dosed at 1-20 mg/kg by Oct 1, 2012. Data cutoff: Apr 30, 2013.

Topalian S, ASCO 2015

PDL-1 expression, T cell infiltrate, smoking status and Pembrolizumab Studi di associazione (non correlazione) fra infiltrato linfocitario T e l’effetto e anche in correlaz con status tabagico Anche correlato con infiltrato B (anche su malattia resecata) Siwen Hu-Lieskovan, et al; WCLC 2015, ORAL 31.05

The efficacy Finally something in Squamous and in Smokers Time to response & the duration of response Which criteria to assess the response Which Biomarkers Think about “combo” or forget “combo” (?) The toxicity profile How to inform the patient Approval issues

Time and Durability of response with pembrolizumab in NSCLC (keynote 001) RECIST v1.1 Central Review 10 20 30 40 50 Time, weeks Individual Patients Treated With Pembro irRC Investigator Review 10 20 30 40 50 60 Time, weeks Individual Patients Treated With Pembro Partial Response Progression On Treatment Pembro 2 mg/kg Q3W Pembro 10 mg/kg Q3W Pembro 10 mg/kg Q2W 11 of 11 (100%) responses are ongoing Median duration of response not reached (median follow-up, 36 weeks) 7 of 11 (64%) responders remain on treatment Median duration of treatment: 27.1 weeks (range, 15.0+ – 48.3+) 19 of 21 (90%) responses are ongoing Median duration of response not reached (median follow-up, 36 weeks) 18 of 21 (86%) responders remain on treatmentb Median duration of treatment: 27.1 weeks (range, 6.1 – 57.1+) Garon EB et al, ESMO 2014

Time and Durability of response with NIVOLumab in NSCLC (checkmate 063) Rizvi NA et al, Lancet Oncol 2015

The efficacy Finally something in Squamous and in Smokers Time to response & the duration of response Which criteria to assess the response Which Biomarkers Think about “combo” or forget “combo” (?) The toxicity profile How to inform the patient Approval issues

Comparison: RECIST-irRC criteria

Unidimensional evaluation (RECIST) Immuno-response (irRC)

Results: Landmark Analysis Sensitivity Analysis Response at Month-3.5 # of pts # of deaths Median months from month 3.5 (95% CI) Responder 10 NR Non-responder 73 42 7.4 (4.5, 10.1) 83 of the 98 patients were alive at 3.5 months since treatment start This analysis suggested that responders had longer survival than non-responders Retrospective analysis Checkmate 063: Patients were grouped in RECIST responder and non-responder categories

FDG-PET: EORTC criteria FDG-PET as a predictive biomarker in Ph2 study (FIR) of atezolizumab. Week 6 FDG-PET: relation to response and OS (n=88) FDG-PET: EORTC criteria Whole body %ID CT target lesions 1.0 1.0 1.0 0.8 0.8 0.8 0.6 0.6 0.6 OS probability OS probability 0.4 0.4 OS probability 0.4 CMR (n=2) 0.2 PMR (n=19) 0.2 0.2 PR (n=12) SMD (n=27) Δ %ID < 0 (n=36) SD (n=46) PMD (n=40) 0.0 Δ %ID ≥ 0 (n=52) 0.0 PD (n=28) 0.0 20 40 60 80 100 300 500 20 40 60 60 Study week Study week Study week FDG-PET CT 1.0 Base SLD < median, Chg < median Base SLD < median, Chg ≥median Base SLD ≥ median, Chg < median Base SLD ≥ median, Chg ≥ median 1.0 Dati con atezolizumab in cui la risposta correla con l’attività metabolica e l’attività metabolica (anche basale) correla con la prognosi The main point to communicate for top 3 is that changes in FDG-PET at week 6, either by EORTC criteria or whole body tumor burden (%ID), are significant predictors of OS.  However, results are pretty similar when compared to CT at week 6 (right top). From bottom plots: while both baseline and week 6 changes are significant, baseline effect is stronger, both for CT and PET. 0.8 0.8 Both baseline and week 6 changes in whole body metabolic tumor burden are significant predictors of OS based on multivariate analysis 0.6 0.6 OS probability OS probability 0.4 Base %ID < median, Δ %ID < 0 Base %ID < median, Δ %ID ≥ 0 0.4 Base %ID ≥ median, Δ %ID < 0 0.2 Base %ID ≥ median, Δ %ID ≥ 0 0.2 0.0 0.0 20 40 60 80 20 40 60 80 Study week Study week Δ = change from baseline; base = baseline; CMR = complete metabolic response; PMR = partial metabolic response SMD = stable metabolic disease; PMD = progressive metabolic disease; PR = partial response; SD = stable disease Jamie Chaft et al, WCLC 2015

The efficacy Finally something in Squamous and in Smokers Time to response & the duration of response Which criteria to assess the response Which Biomarkers Think about “combo” or forget “combo” (?) The toxicity profile How to inform the patient Approval issues

……do we really have THE biomarker??

Change in baseline target lesions (%) Nivolumab activity was observed in patients with PD-L1+ tumors and also in some patients with PD-L1 tumors PD-L1 cut-off PD-L1 status Evaluable biopsies, % (n/N) ORR,a % 1% Positive 56 (38/68) 13 (5/38) Negative 44 (30/68) 17 (5/30) 5% 49 (33/68) 15 (5/33) 51 (35/68) 14 (5/35) More patients with PD-L1+ than PD-L1 tumors had a decrease in tumor burden 150 100 50 -50 -100 Positive Negative PD-L1 status (5% cut-off) Change in baseline target lesions (%) aORR of patients evaluable for PD-L1; ORR includes complete or partial responders determined by RECIST 1.0. Six patients with unconventional ‘immune-related’ responses were not included as responders. ORR = objective response rate; PD-L1 = programmed cell death-ligand 1; RECIST = Response Evaluation Criteria for Solid Tumors. Patients Brahmer JR, et al. ASCO 2014 (abstr. 8112).

Pembrolizumab activity was observed in patients with PD-L1+ tumors and also in some patients with PD-L1 tumors 9.1% pem 8.8% txt [Hanna N, JCO 2004] 7.1% txt [Shepherd FA, JCO 2000] 8.9% erlo[Shepherd FA, JCO 2005] Garon EB, ESMO 2014

The Biomarker Issue: PD-L1 & Histology CheckMate 017 [Squamous] CheckMate 057 [non-Squamous] Mettendo insieme i due studi ho un effetto differente NO effetto nello squamoso a prescindere dal cutoff utilizzato SI effetto nel non squamoso a prescindere dal cutoff utilizzato Modified – Spigel ASCO 2015; Paz-Ares ASCO 2015

Sample Source and Collection Definition of Positivity† PD-L1 as a predictive immune biomarker: assays, sample collection and analysis in NSCLC studies Pembrolizumab Merck Prototype or clinical trial IHC assay (22C3 Ab)1 Surface expression of PD-L1 on tumor specimen* Ph I: Fresh tissue Ph II/III: Archival or fresh tissue2 IHC Staining: Strong vs weak expression2 PD-L1 expression required for NSCLC for enrollment2 Note that one arm of KEYNOTE 001 trial requires PD-L1- tumors1 Tumor PD-L1 expression:1 ≥50% PD-L1+ cut-off: 32% (41/129) 1-49% PD-L1+ cut-off: 36% (46/129) Nivolumab Bristol-Myers Squibb Dako automated IHC assay (28-8 Ab)3,4 Surface expression of PD-L1 on tumor cells* Archival4 or fresh tissue IHC Staining: Strong vs weak expression3,4 Patients not restricted in PD-L1 status in 2nd- & 3rd-line4 Ph III 1st-line trial in PD-L1+3 Tumor PD-L1 expression:4 5% PD-L1+ cut-off: 49% (33/68)4 MPDL3280A Roche/Genentech Ventana automated IHC assay Surface expression of PD-L1 on TILs5 Archival or fresh tissue IHC Staining intensity (0, 1, 2, 3): IHC 3 (≥10% PD-L1+): Ph III trial5 IHC 2,3 (≥5% PD-L1+)5 IHC 1,2,3 (≥1% PD-L1+)5 IHC 1, 0, or unknown PD-L1 expression required for NSCLC for enrollment x TIL PD-L1 expression:5,6 IHC 3 (≥10% PD-L1+): 11% (6/53) PD-L1 low (IHC 1, 0): 75% (40/53) MEDI4736 AstraZeneca 1st generation or Ventana automated IHC (BenchMark ULTRA) assay (Ventana PD-L1 (SP263) clone)7,8 Surface expression of PD-L1 on TILs PhI: Fresh tissue IHC Staining intensity: Not presented to date7,8,9 TIL PD-L1 expression: PD-L1 Assay Sample Source and Collection Definition of Positivity† *Tumor cells are stained, but not TILs and other tissue areas. †Definition of PD-L1 positivity differs between assay methodologies. IHC = immunohistochemistry; NSCLC = non-small cell lung cancer; PD-L1 = programmed cell death-ligand 1; TILs = tumor-infiltrating lymphocytes. References 1. Garon EB, et al. Presented at ESMO 2014 (abstr. LBA43); 2. Rizvi NA, et al. Presented at ASCO 2014 (abstr. 8007); 3. Gettinger S et al. Poster p38 presented at ASCO 2014 (abstr. 8024); 4. Brahmer JR et al. Poster 293 presented at ASCO 2014 (abstr. 8112); 5. Rizvi NA et al. Poster presented at ASCO 2014 (abstr. TPS 8123); 6. Soria J-C, et al. ESMO 2014 (abstr. 1322P); 7. Brahmer JR, et al. Poster presented at ASCO 2014 (abstr. 8021); 8. Segal NH, et al. Presented at ASCO 2014 (abstr. 3002); 9. Segal NH, et al. ESMO 2014 (abstr. 1058PD).

IASLC:PDL-1 PROJECT Comparison of different antibodies Application to different staining platforms Intra- and inter-observer reproducibility(inter-laboratory reproducibility) Large specimens vs small specimens vs cytology from same tumors Prognostic association based on a well defined cohort Predictive association : PD1/PDL1 treated pts ATLAS

NM Hahn et al, ASCO 2015

The efficacy Finally something in Squamous and in Smokers Time to response & the duration of response Which criteria to assess the response Which Biomarkers Think about “combo” or forget “combo” (?) The toxicity profile How to inform the patient Approval issues

CheckMate 012: Nivo + Ipi [1st line] Percentage Change from Baseline Best Percentage Change in Target Lesion Tumor Burden by PD-L1 220 200 180 160 140 120 100 80 60 40 20 –20 –40 –60 –80 –100 * PD-L1 expression n (%) ≥1% PD-L1a 77 (68) <1% PD-L1a 36 (32) Unknownb 35 (24) Percentage Change from Baseline * Confirmed ongoing responses Effetto del marcatore (a differenza di quanto descritto nel melanoma) Low frequency of treatment-related grade 3–4 AEs leading to discontinuation: 3–10% (No treatment-related deaths) Rizvi N ORAL02.05

Phase Ib GP28328: ATEZO + Chemo [1st line] ATEZO no unexpected toxicities No pneumonitis or other respiratory AEs of concern ORR 50% ORR 76.5% ORR 56.3% Rizvi N ORAL02.05

E se in I linea ci dimenticassimo della combinazione? Besse B et al, ECCO 2015

The efficacy Finally something in Squamous and in Smokers Time to response & the duration of response Which criteria to assess the response Which Biomarkers Think about “combo” or forget “combo” (?) The toxicity profile How to inform the patient Approval issues

Immune Related Adverse Events (IRAEs) System Adverse Events Gastrointestinal Colitis (Diarrhea, perforation) Renal Acute Interstitial Nephritis (Increased serum Creatinine) Pulmonary Pneumonitis (dyspnea, cough) Dermatologic Dermatitis (Lichenoid/ spongiotic dermatitis, rash), Vitaligo Hepatic Hepatitis (elevated LFTs) Neurologic Central and Peripheral (Aseptic Meningitis, Guillan-Barre Syndrome, Myasthenia Gravis Endocrine Hypophysitis, thyroiditis, adrenal insufficiency Ocular Uveitis, Iritis IrAEs can manifest in nearly any organ or organ system in the body, some of which are shown here.

Immune Related Adverse Events (IRAEs) Average time to onset of irAEs is 6-12 weeks after initiation of therapy Within days of the first dose After several months of treatment After discontinuation of therapy Severity: Can be mild and asymptomatic to severe and life threatening The average time to onset of these events is 6-12 weeks, but can occur at any time, from within days of the patient receiving their first dose, until after treatment is discontinued. The majority of irAEs are quite mild, and often are asymptomatic, but they have the potential to be severe and life threatening.

Checkmate 017: safety Brahmer J et al, NEJM 2015

CheckMate 017: Updated Safety Time to Onset of First Treatment-related Select AE With Nivolumab by Category (Any Grade) Eventi maggiori concentrati nei primi 3 mesi di trattamento Pooled analyses of safety of nivolumab in metastatic SQ NSCLC is being presented at ECC (ESMO) as a poster on Sept 27, 2015 - Gettinger et al., P346 Most common treatment-related select AEs were skin (9%) and GI (8%) Median time to onset of these AEs were: hypersensitivity - 0.3 wks (range 0.3 – 0.3) GI – 3.0 wks (range 0.1 – 91.0) Renal 10.5 wks (range 3.9 – 16.6) Skin 7.3 wks (range 0.3 – 51.9) endocrine - 7.4 wks (range 6.3 – 18.1) Pulmonary 15.1 wks (range 2.6 – 85.1) Hepatic 17.6 wks (range 4.1 – 31.1) Median time to resolution of treatment-related select AEs were GI 1.7 wks (range 0.1 – 31.0); 9/11 patients resolved Endocrine NA (range 6.3 – 18.1) 2/5 patients resolved Hepatic NA (range 2.9 – 22.3+) 1/2 patients resolved Pulmonary 5.0 wks (range 0.6 – 12.1) 6/6 patients resolved Skin 2.5 wks (range 0.1 – 46.9+) 11/12 patients resolved Hypersensitivity/infusion reaction 0.3 wks (range 0.3 – 0.3) 1/1 patient resolved Renal NA (range 0.7 – 37.6) 2/4 patients resolved Pts still on study, n 131 112 85 52 Pts still on treatment, n 131 73 51 25 Total pts with first event,a n 24 6 2 1 The majority of patients who experienced treatment-related select AEs with nivolumab experienced their first event within the first 3 months of treatment Select AEs: AEs with potential immunologic etiology that require frequent monitoring/intervention. Based on December 2014 DBL. Includes events reported between first dose and 30 days after last dose of study therapy. Within each time interval, patients with ≥1 event were counted only once in each category but could be classified into more than one category Reckamp K ORAL02.01

CheckMate 017: Patient Reported Outcomes (PROs) LCSS Average Symptom Burden Index (on-treatment) LCSS-3 Item Index (on treatment) QoL misurata con LCSS a 3 o 6 items a vantaggio di nivo, anche mantenuto nel tempo Gralla R et al, WCLC 2015 ORAL31.03

CheckMate 017: Patient Reported Outcomes (PROs) Tempo alla comparsa di sintomi specifici dinuovo a vantaggio di nivo Gralla R et al, WCLC 2015 ORAL31.03

SCREENING STUDY POPULATION Treatment until progression or for 1 yr CA209-153 Study Design (N=824) SCREENING STUDY POPULATION Advanced / metastatic (stage IIIb / IV) NSCLC (NSQ and SQ) at least 1 prior systemic therapya COHORT B Discontinue treatment, and if progression occurs re-treatment allowed Enrollment period aConventional systemic therapies, excluding IO therapies Treatment until progression or for 1 yr Follow up for 5 yrs to collect survival information bPatients with CR, PR, or SD COHORT A Continue to treat to progression, unacceptable toxicity, or withdrawal of informed consent INTERVENTION Nivolumab 3 mg/kg IV q 2 weeks PATIENT SUBGROUPS SQ, PS 0-1, ≥2 prior therapies, +/- treated CNS mets SQ, PS 0-1, 1 prior therapy, +/- treated CNS mets NSQ, PS 0-1, ≥1 prior therapy, +/- treated CNS mets SQ or NSQ, PS 2, ≥1 prior therapy ENDPOINTS Safety ORR PFS DOR OS PRO PK Biomarkers RANDOMIZATION at 1 yrb Patients receive nivolumab 3 mg/kg IV (60 min) every 2 weeks until either progressive disease or unacceptable toxicity (Cohort A) or for 1 yr with the possibility of re-treatment upon disease progression (Cohort B). The primary study objective is assessment of the incidence for high grade (grade 3-4 and grade 5) treatment-related select adverse events, defined as pneumonitis, interstitial nephritis, diarrhea/colitis, hepatitis, rash, endocrinopathies, and hypersensitivity/ infusion reaction events. Study endpoints include safety and efficacy endpoints and PROs. Maen Hussein et al, WCLC 2015; ORAL 02

Summary of Adverse Events Nivolumab 3 mg/kg N = 824 Nivolumab 3 mg/kg ECOG PS 0–1 (n = 742) Nivolumab 3 mg/kg ECOG PS 2 (n = 65) Any Grade n (%) Grade 3–4 n (%) Grade 5 n (%) Grade 3–4 n (%) All adverse events 762 (93) 311 (38) 158 (19) 683 (92) 268 (36) 131 (17) 62 (95) 33 (51) 24 (37) All serious adverse events (SAEs) 309 (38) 223 (27) 257 (35) 185 (25) 42 (65) 29 (45) All select adverse events 282 (34) 37 (5) 5 (1) 253 (34) 32 (4) 3 (<1) 22 (34) 3 (5) 2 (3) All treatment-related adverse events 439 (53) 59 (7) 1 (<1) 403 (54) 52 (7) 27 (42) 4 (6) All treatment-related SAEs 23 (3) 19 (2) 1 (<1)* 18 (2) 14 (2) All treatment-related select AEs 199 (24) 20 (2) 181 (24) 16 (2) 14 (22) All AEs leading to discontinuation 87 (11) 53 (6) 34 (4) 69 (9) 42 (6) 27 (4) 16 (25) 9 (14) 7 (11) All treatment-related SAEs leading to discontinuation 12 (2) 11 (2) 9 (1) All treatment-related select AEs leading to discontinuation 11 (1) 8 (1) Su circa 800 paz “real life” PS 0-2 non sembra che la tossicità sia troppo condizionata da PS, anche se in realtà I PS 2 qui erano solo l’8% Across all patients, 93% experienced an adverse event; 38% had grade 3 or 4 events. Maen Hussein et al, WCLC 2015; ORAL 02 *One fatal event was reported as drug-related respiratory failure, with known comorbidities of lymphangitic spread of tumor, recurrent pulmonary embolism, G-bacteremia, pleural effusion, pneumothorax, or tumor progression. This patient’s death was classified as ‘Other-Multifactorial’ by the investigator.

Pembro: Immune-Related Events & Steroids KEYNOTE-001, Data from 505 pts Problematica dello steroide a prescindere NON sembra impattare su efficacia del farmaco (è RETROSPETTIVO su circa 500 casi del Keynote 001) In questo caso (per un farmaco che noi accomuniamo a nivo) la tox è sopratt tiroidea (ipo) e compare entro i primi due mesi PFS and OS and Steroids Use to Manage Immune- Mediated AEs Leighl N ORAL31.02

The efficacy Finally something in Squamous and in Smokers Time to response & the duration of response Which criteria to assess the response Which Biomarkers Think about “combo” or forget “combo” (?) The toxicity profile How to inform the patient Approval issues

The efficacy Finally something in Squamous and in Smokers Time to response & the duration of response Which criteria to assess the response Which Biomarkers Think about “combo” or forget “combo” (?) The toxicity profile How to inform the patient Approval issues

Impact of time to drug approval on potential years of life lost: the compelling need for improved trial and regulatory efficiency - Used phase III trials in incurable cancers published 2000-2015: With statistically significant increase in survival Drug approved by the FDA in that malignancy 21 drugs in 11 malignancies as illustrative examples Calculated: Life-years lost per year of approval delay: Multiplied no. deaths per year (for specific subgroups included in phase III trial) x median survival gain (yrs): - Life-years lost between drug discovery and approval: Used US dates as surrogate for worldwide - Life-years that would have been saved if time from discovery to approval reduced to < 5 years - Life-years saved by improved trial safety DJ Stewart et al, WCLC 2015

Life-years lost worldwide per year delay in drug approval Total combined life-years lost / year: Worldwide: 2,541,274 1 for every 12 seconds delay North America: 240,085 Life-years lost worldwide per year

Life-years lost worldwide from patent application to approval Total combined life-years lost: Worldwide: 31,537,958 North America: 2,956,667 Life-years lost worldwide

CheckMate 012: Nivo + Ipi [1st line] Percentage Change from Baseline Best Percentage Change in Target Lesion Tumor Burden by PD-L1 220 200 180 160 140 120 100 80 60 40 20 –20 –40 –60 –80 –100 * PD-L1 expression n (%) ≥1% PD-L1a 77 (68) <1% PD-L1a 36 (32) Unknownb 35 (24) Percentage Change from Baseline * Confirmed ongoing responses Low frequency of treatment-related grade 3–4 AEs leading to discontinuation: 3–10% (No treatment-related deaths) Rizvi N ORAL02.05

Phase Ib GP28328: ATEZO + Chemo [1st line] ATEZO no unexpected toxicities No pneumonitis or other respiratory AEs of concern ORR 50% ORR 76.5% ORR 56.3% Rizvi N ORAL02.05