Debaters For The Evening:

Similar documents
Debaters For The Evening:

Metastatic NSCLC: Expanding Role of Immunotherapy. Evan W. Alley, MD, PhD Abramson Cancer Center at Penn Presbyterian

Out of 129 patients with NSCLC treated with Nivolumab in a phase I trial, the OS rate at 5-y was about 16 %, clearly higher than historical rates.

Chemotherapy and Immunotherapy in Combination Non-Small Cell Lung Cancer (NSCLC)

Immune Checkpoint Inhibitors for Lung Cancer William N. William Jr.

Immunotherapy for NSCLC: Current State of the Art and Future Directions. H. Jack West, MD Swedish Cancer Institute Seattle, Washington, United States

Updates in Immunotherapy for Urothelial Carcinoma

Immunotherapies for Advanced NSCLC: Current State of the Field. H. Jack West Swedish Cancer Institute Seattle, Washington

Conversations in Oncology. November Kerry Hotel Pudong, Shanghai China

NSCLC: immunotherapy as a first-line treatment. Paolo Bironzo Oncologia Polmonare AOU S. Luigi Gonzaga Orbassano (To)

Immunotherapy in the clinic. Lung Cancer. Marga Majem 20 octubre 2017

Immunotherapy for the Treatment of Head and Neck Cancers. Robert F. Taylor, MD Aurora Health Care

CheckMate 012: Safety and Efficacy of First Line Nivolumab and Ipilimumab in Advanced Non-Small Cell Lung Cancer

Largos Supervivientes, Tenemos datos?

Immunotherapy for the Treatment of Head and Neck Cancers. Barbara Burtness, MD Yale University

Immune checkpoint blockade in lung cancer

Update on the development of immune checkpoint inhibitors

II sessione. Immunoterapia oltre la prima linea. Alessandro Tuzi ASST Sette Laghi, Varese

La revolución de la inmunoterapia: dónde la posicionamos? Javier Puente, MD, PhD

A Giant Leap in the Treatment Options for Advanced Bladder Cancer

Indication for- and timing of cytoreductive nephrectomy Kidney- and bladder cancer: Immunotherapy

Recent Therapeutic Advances for Thoracic Malignancies

Immunotherapy for the Treatment of Melanoma. Marlana Orloff, MD Thomas Jefferson University Hospital

Reflex Testing Guidelines for Immunotherapy in Non-Small Cell Lung Cancer

Practice changing studies in lung cancer 2017

PATIENT SELECTION CORRELATION OF PD-L1 EXPRESSION AND OUTCOME? THE ONCOLOGIST VIEW ON LUNG CANCER

Immunotherapy, an exciting era!!

Renal Cell Carcinoma: Systemic Therapy Progress and Promise

Immunotherapy in Unresectable or Metastatic Melanoma: Where Do We Stand? Sanjiv S. Agarwala, MD St. Luke s Cancer Center Bethlehem, Pennsylvania

The Immunotherapy of Oncology

Incorporating Immunotherapy into the treatment of NSCLC

Non-Small Cell Lung Cancer Webinar. Thursday, September 13, p.m. EDT

Il ruolo di PD-L1 (42%) tra la prima e la seconda linea di trattamento

Immunotherapy in Patients with Non-Small Cell Lung Cancer

Immune checkpoint inhibitors in NSCLC

Immunotherapy of Melanoma Sanjiv S. Agarwala, MD

Options for first-line cisplatin-eligible patients

Lung Cancer Immunotherapy

Immunoterapia di 1 linea Evidenze e Prospettive Future

Alessandro Inno. IRCCS Ospedale Sacro Cuore Don Calabria Negrar, Verona

Immunotherapy for the Treatment of Kidney and Bladder Cancer

Squamous Cell Carcinoma Standard and Novel Targets.

Fifteenth International Kidney Cancer Symposium November 4-5, 2016 Marriott Miami Biscayne Bay, Miami, Florida, USA

Post-ASCO Immunotherapy Highlights (Part 2): Biomarkers for Immunotherapy

Biomarkers for Cancer Immunotherapy Debate

Current Issues in Checkpoint Immunotherapy for NSCLC: A Perspective from January 2018

Development of PD-1 and PD-L1 inhibitors as a form of cancer immunotherapy: a comprehensive review of registration trials and future considerations

Prostate cancer Management of metastatic castration sensitive cancer

III Sessione I risultati clinici

Immunotherapy for Renal Cell Carcinoma. James Larkin

Checkpoint Inibitors for Bladder Cancer

ASCO 2014 Highlights*

Patient Selection: The Search for Immunotherapy Biomarkers

Melanoma: From Chemotherapy to Targeted Therapy and Immunotherapy. What every patient needs to know. James Larkin

Principles and Application of Immunotherapy for Cancer: Advanced NSCLC

Weitere Kombinationspartner der Immunotherapie

Biomarkers in Imunotherapy: RNA Signatures as predictive biomarker

Melanoma: Immune checkpoints

Update on Immunotherapy in Advanced Melanoma. Ragini Kudchadkar, MD Assistant Professor Winship Cancer Institute Emory University Sea Island 2017

Immunotherapeutic Advances in the Treatment of Metastatic Non-Small Cell Lung Cancer

PTAC meeting held on 5 & 6 May (minutes for web publishing)

Checkpoint regulators a new class of cancer immunotherapeutics. Dr Oliver Klein Medical Oncologist ONJCC Austin Health

Media Release. Basel, 21 July 2017

Updates From the European Lung Cancer Conference: Immunotherapy and Non-Small Cell Lung Cancer

Nivolumab: esperienze italiane nel carcinoma polmonare avanzato

Recent Advances in Lung Cancer: Updates from ASCO Updates from ESMO, AACR and ASCO

What we learned from immunotherapy in the past years

Genomics and Genetics in BC: Precise selection for chemotherapy and Immunotherapy. Raanan Berger MD PhD Sheba Medical Center, Israel

Immunotherapy for Genitourinary Cancers

Urothelial Cancers- New Strategies. Sandy Srinivas.MD Stanford University

Medical Treatment of Advanced Lung Cancer

Checkpoint Regulators Cancer Immunotherapy takes centre stage. Dr Oliver Klein Department of Medical Oncology 02 May 2015

New paradigms for treating metastatic melanoma

The Really Important Questions Current Immunotherapy Trials are Not Answering

Cancer Immunotherapy Patient Forum. for the Treatment of Melanoma, Leukemia, Lymphoma, Lung and Genitourinary Cancers - November 7, 2015

Medical Treatment for Melanoma Sanjiv S. Agarwala, MD

O DESAFIO DA INOVAÇÃO EM ONCOLOGIA EM PORTUGAL The Challenges of innovative oncology care in Portugal. Gabriela Sousa Oncologia Médica IPO Coimbra

Immunotherapy for Melanoma. Caroline Robert, MD, PhD Gustave Roussy and Université Paris Sud Villejuif, France

The Rationale for Immunotherapy as an Adjuvant Treatment for Locally Advanced BC

Advances in Cancer Immunotherapy for Solid Tumors Expert Perspectives on The New Data Sunday, June 5, 2016

News from ASCO. Niven Mehra, Medical Oncologist. Radboud UMC Institute of Cancer Research and The Royal Marsden Hospital

Índice. Melanoma Cáncer de Pulmón Otros tumores

Role of the Pathologist in Guiding Immuno-oncological Therapies. Scott Rodig MD, PhD

Immunotherapy in non-small cell lung cancer

Immunotherapy for Metastatic Malignant Melanoma. Dr Daniel A Vorobiof Sandton Oncology Centre Johannesburg

Plotting the course: optimizing treatment strategies in patients with advanced adenocarcinoma

Innovaciones en el tratamiento del ca ncer renal. Enrique Grande

INMUNOTERAPIA I. Dra. Virginia Calvo

Maintenance therapy in advanced non-small cell lung cancer. Egbert F. Smit MD PhD Dept Thoracic Oncology Netherlands Cancer Institute

Integrating Immunotherapy into Urologic Oncology: The New Urothelial Cancer Paradigm

IMMUNOTHERAPY FOR GASTROINTESTINAL CANCERS

Management Guidelines and Targeted Therapies in Metastatic Non-Small Cell Lung Cancer: An Oncologist s Perspective

IMMUNE CHECKPOINT BLOCKADE FOR NSCLC. Marina Chiara Garassino

Immunotherapy in the Adjuvant Setting for Melanoma: What You Need to Know

6/7/16. Melanoma. Updates on immune checkpoint therapies. Molecularly targeted therapies. FDA approval for talimogene laherparepvec (T- VEC)

Newest Oncology Agents: PD 1 Inhibitors Clinical Information and Patient Management

Fifteenth International Kidney Cancer Symposium

The role of immune checkpoint inhibitors in non-small cell lung cancer

Histology independent indications in Oncology

Pembrolizumab for Patients With PD-L1 Positive Advanced Carcinoid or Pancreatic Neuroendocrine Tumors: Results From the KEYNOTE-028 Study

Transcription:

Your Debaters For The Evening:

Current State of Immune Checkpoint Blockade in Selected Types of Solid Tumors Jeffery Weber, MD, PhD Naiyer A. Rizvi, MD

Current State of ICI in Melanoma and Bladder Cancers Jeffrey S. Weber, MD, PhD Laura and Isaac Perlmutter Cancer Center NYU Langone Medical Center

Current US/EU Approval Status of ICIs US EU PD-1 PD-L1 CTLA-4 Combo Pembrolizumab Nivolumab Atezolizumab Durvalumab Avelumab Ipilimumab Nivo + ipi Line 1L+ 2L+ 1L+ 2L+ 1L+ 2L+ 2L+ 1L+ 2L+ 1L+ 1L+ mmelanoma NSCLC chl arcc mscchn muc Merkel cell MSI-H chl=classical Hodgkin lymphoma; arcc=advanced renal cell carcinoma; mscchn=metastatic head and neck squamous cell carcinoma; muc=metastatic urothelial cancer; MSI-H=microsatellite instability-high solid tumors Product Prescribing Information Current as of June 2017

Should Immunotherapy be the First-line Therapy in all Stage IV Patients?

Overall Survival (%) Immune Checkpoint Inhibitors Provide Durable Long-term Survival for Pts with Advanced Melanoma Front-line OS = >50% for Nivolumab 100 90 80 70 60 50 N=210 IPI (Pooled Analysis) NIVO Monotherapy (Phase 1 CA209-003) NIVO Monotherapy (Phase 3 Checkmate-066) 40 30 20 10 N=107 N=1861 0 0 1 2 3 4 5 6 7 8 9 10 Years Schadendorf D, et al. J Clin Oncol. 2015; Hodi S, et al. ASCO. 2016; Atkinson V, et al. Presented at: SMR 2015; November 6-9, 2015; Boston, MA.

Long-term Outcomes with Single Agent Pembrolizumab in KEYNOTE-006 33-mo OS rate was 50% in the pooled pembrolizumab arms (n=556) 33-mo PFS rate was 31% and ORR was 42% Median duration of response was not reached for pembro (range 1.0+ to 33.8+ mo) Responses were durable in pts who completed pembro; 9.7 mo after completion of pembro with estimated PFS (95% CI) 91% (80-96) in 104 pts Pembro front-line responses were durable! Robert C, et al. ASCO. 2017.

First-line Single Agent vs. Concurrent vs. Sequential Immunotherapy?

Updated Survival Data in BMS-067 Trial of IPI vs. NIVO vs. IPI/NIVO Median OS, months (95% CI) HR (99.5% CI) vs. IPI HR (99.5% CI) vs. NIVO NIVO + IPI (N=314) NIVO (N=316) IPI (N=315) NR 0.55 (0.42 0.72)* 0.88 (0.69-1.12) NR (29.1-NR) 0.63 (0.48 0.81)* 20 (17.1-24.6) - - - *P<.0001 Larkin J, et al. AACR. 2017. Database lock: September 13, 2016. Minimum follow-up of 28 months

How Toxic is Combination Checkpoint Blockade? With an additional 19 months of follow-up, safety was consistent with the initial report 1 Patients Reporting Event, % Treatment-related adverse event (AE) Treatment-related AE leading to discontinuation NIVO + IPI (N=313) NIVO (N=313) IPI (N=311) Any Grade Grade 3-4 Any Grade Grade 3-4 Any Grade Grade 3-4 95.8 58.5 86.3 20.8 86.2 27.7 39.6 31 11.5 7.7 16.1 14.1 Treatment-related death 2 (0.6) a 1 (0.3) b 1 (0.3) b a Cardiomyopathy (NIVO+IPI, N=1); Liver necrosis (NIVO+IPI, N=1). Both deaths occurred >100 days after the last treatment. b Neutropenia (NIVO, N=1); Colon perforation (IPI, N=1) 1 Most select AEs were managed and resolved within 3-4 weeks (85-100% across organ categories) ORR was 70.7% for patients who discontinued NIVO+IPI due to AEs, with median OS not reached 1 Larkin J, et al. N Engl J Med. 2015; Larkin J, et al. AACR. 2017.

Toxicity of Immune Checkpoint Blockade Major side effects of immunotherapy are related to development of autoimmunity Immune-related adverse events, also known as immune-mediated adverse events (iraes, imaes), may occur in any organ system Certain events may be life threatening or fatal Prompt recognition of potentially severe iraes improves outcomes Early diagnosis and aggressive systemic corticosteroids are key to prevent lifethreatening consequences Combining checkpoint inhibitors with agents such as CTLA-4 inhibitors, significantly increases iraes

First-line Therapy for Melanoma is Immunotherapy for Most Patients Hard to beat the front-line median OS of 30+ months for nivolumab or pembrolizumab alone and 40+ estimated OS for IPI/NIVO 2 year OS for IPI/NIVO concurrent or sequential = 62% is unmatched by any targeted phase III trial Robert C, et al. ASCO. 2017; Larkin J, et al. AACR. 2017; Schadendorf D, et al. J Clin Oncol. 2015; Hodi S, et al. ASCO. 2016; Atkinson V, et al. Presented at: SMR 2015; November 6-9, 2015; Boston, MA.

What Should be the First Choice for Adjuvant Therapy in High-risk Melanoma?

Overall Survival: Ipilimumab Adjuvant Therapy Eggermont AMM, et al. N Engl J Med. 2016.

Dosing of Ipilimumab as Adjuvant Therapy: 3 vs. 10 mg/kg in E1609 Accrual was 1,670 including 511 IPI10, 636 HDI and 523 IPI3 pts Grade 3+ AEs were seen in 57% of IPI10 and 36% IPI3 pts; discontinuation in 53.8% at IPI10 and 35.2% at IPI3 At a median follow-up of 3.1 years, unplanned RFS analysis showed 3-year RFS of 54% (95% CI: 49, 60) at IPI10 and 56% (50, 61) at IPI3 Tarhini A, et al. ASCO. 2017.

Checkpoint Blockade for Bladder Cancer

Efficacy of PD-1 or PD-L1 Immune Checkpoint Inhibitors in Advanced Bladder Cancer Study Previous Platinum? Comparator Drug PD-L1 Assay PD-L1 (+) Expression Cutoff ORR Time to Response (mo) OS (mo) Atezolizumab (IMvigor 210) N=310 Yes No SP142 IC0: <1% (n=103) IC1: 1-5% (n=108) IC2/3: >5% (n=100) All: 15%, IC0: 8% IC1: 10%, IC2/3: 26% 2.1 (95% CI: 2-2.2) All: 7.9 Atezolizumab (IMvigor 210) N=119 No, firstline** No SP142 IC0: <1% (n=39) IC1: 1-5% (n=48) IC2/3: >5% (n=32) All: 19% IC1/2/3: 19% IC2/3: 22% Not Published All: 10.6 Avelumab (JAVELIN) N=241 Yes No Clone 73-10 >5% (n=81) PD-L1 (-): 14.7% PD-L1 (+): 25% ~2.6 PD-L1 (-) 6 month OS: 52.7% PD-L1 (+) 6 month OS: 61.2% Durvalumab (Study 1108) N=103 Yes No SP263 >25% (n=61) PD-L1 (-): 5.1% PD-L1 (+): 31.1% All: 1.41 PD-L1 (-): 2.05 PD-L1 (+): 1.41 Not Published Nivolumab (CheckMate-275) N=265 Yes No 28-8 pharmdx >1% (n=122) >5% (n=81) All: 19.6%, PD-L1 (-): 16.1% PD-L1 >1%: 23.8% PD-L1 >5%: 28.4% 1.87 (95% CI: 1.81-1.97) All: 8.74 PD-L1 (-): 59.95 PD-L1 >1%: 11.3 Nivolumab (CheckMate-032) N=78 Yes No 28-8 pharmdx <1% (n=42) >1% (n=25) All: 24.4% PD-L1 (-): 26.2% PD-L1 (+): 24% 1.5 (95% CI: 1.2-4.1) All: 9.7 PD-L1 (-): 9.9 PD-L1 (+): 16.2 Pembrolizumab (KEYNOTE-045) N=542 Yes Paclitaxel, docetaxel or vinflunine 22C3 pharmdx >10% (n=164)* All: 21.1 vs. 11.4% PD-L1 (+): 21.6 vs. 6.7% 2.1 All: 10.3 vs. 7.4 PD-L1 (+): 8 vs. 5.2 Pembrolizumab (KEYNOTE-052) N=100/374 No, firstline** Gill DM, et al. Immunotherapy. 2017. No 22C3 pharmdx >10% (n=30) All: 24% PD-L1 (+): 37% 2 (95% CI: 0.2-5) 6 month OS: 67% *Cisplatin ineligible patients; **PD-L1 expression included percentage of tumor and infiltrating immune cells with PDL1 expression

% Efficacy of PD-1 or PD-L1 Inhibitors in 2 nd Line Metastatic Bladder Cancer 25 20 Objective Response Rate 19.6% 20.4% 17.6% 21.1% 15 10 15% Historic control with chemotherapy 12% 5 0 Atezolizumab N=310 Loriot ESMO 16 Nivolumab N=265 Sharma Lancet Oncol 17 Durvalumab N=103 Powles ASCO GU 17 Avelumab N=153 Patel ASCO GU 17 Pembrolizumab N=270 Bellmunt New Engl J Med 17 Black P. Presented at AUA 2017 (Adapted from Plimack ASCO 2016).

% ICI Response According to PD-L1 Expression in Advanced Bladder Cancer 30 25 PD-L1 Positive PD-L1 Negative 20 15 10 5 0 Atezolizumab N=310 Loriot ESMO 16 Nivolumab N=265 Sharma Lancet Oncol 17 Durvalumab N=103 Powles ASCO GU 17 Avelumab N=153 Patel ASCO GU 17 Pembrolizumab N=270 Bellmunt New Engl J Med 17 Adapted from Black P. Presented at AUA 2017.

% Grade >3 Adverse Events from ICIs in Advanced Bladder Cancer 20 (2 nd line Metastatic Trials) 16 12 16% 18% 15% 8 7.5% 4 5.2% 0 Atezolizumab N=310 Loriot ESMO 16 Nivolumab N=265 Sharma Lancet Oncol 17 Durvalumab N=103 Powles ASCO GU 17 Avelumab N=153 Patel ASCO GU 17 Pembrolizumab N=270 Bellmunt New Engl J Med 17 Adapted from Black P. Presented at AUA 2017.

KEYNOTE-045 Study (NCT02256436) Overall Survival: Total 43.9% 30.7% Events, n HR (95% CI) P Pembro 155 0.73 (0.59-0.91) 0.0022 Chemo 179 Pembrolizumab Chemotherapy Median (95% CI) 10.3 mo (8.0-11.8) 7.4 mo (6.1-8.3) Bellmunt J, et al. N Engl J Med. 2017. No. at risk 270 226 194 169 147 131 87 54 27 13 4 0 0 272 232 171 138 109 89 55 27 14 3 0 0 0

Long-term Survival Follow-up of the Phase I Atezolizumab Trial Median and Landmark OS by PD-L1 Status IC0/1 (N=44) IC2/3 (N=51) All Pts* (N=95) Median OS (95% CI) 7.6 mo (4.7, 13.9) 11.3 mo (7.8, NE) 10.1 mo (7.3, 17) 1-year OS rate (95% CI) 40% (25, 56) 50% (36, 64) 46% (35, 56) 2-year OS rate (95% CI) 14% (2, 26) 43% (29, 57) 30% (20, 40) NE=Not estimable; *Efficacy-evaluable population with >12 wk f/u - A trend toward longer survival in pts with higher PD-L1 status was observed. Petrylak D, et al. Presented at ASCO GU. 2017. With extended follow-up, single-agent atezolizumab continues to be well-tolerated in this phase Ia study muc cohort Most AEs were grade 1-2, and no treatment-related deaths occurred. Incidence declined substantially (including grade 3-4) after the first year. Clinical benefit was observed in a heavily pre-treated muc population Confirmed, ORs were durable (40% ongoing), with a median DOR of 22 mo and DOR up to 33+ mo observed Median OS was ~10 mo, with 46% and 30% of patients alive after 1 and 2 years, respectively (median survival follow-up duration, ~29 mo) Supports a role for atezolizumab as new standard of care in previously treated muc Phase III study IMvigor211 (NCT02302807) is ongoing, with enrollment completed and data expected later this year.

Atezolizumab was not Superior to Chemotherapy in Cisplatin-resistant Bladder Cancer?? IMvigor211 Study Design (NCT02302807) Urothelial Cancer Progression or recurrence of urothelial cancer following a first-line platinumcontaining regimen Randomization N=932 Estimated Completion: Nov 2017 Atezolizumab SOC: Docetaxel, Paclitaxel or Vinflunine Primary Endpoints OS Secondary Endpoints ORR PFS DOR Safety Primary efficacy endpoint, OS, was to be tested in a successive fashion in study populations defined by PD-L1 expression. The first population tested was people with the highest levels of PD-L1 expression (IC2/3), followed by those with any level of PD-L1 expression (IC1/2/3), and followed by the overall study population (ITT). Statistical significance needed to be achieved in the IC2/3 population in order to evaluate the IC1/2/3 population for statistical significance, and similarly achieved in the IC1/2/3 population in order to evaluate the overall study population for statistical significance. www.clinicaltrials.com;nct02302807.

Is PD-L1 Staining Associated with Clinical Outcome in Bladder Cancer? Study Agent Companion IHC Antibody Threshold for Positivity Target Cells Assay Associated w/ Response? Powles T, et al. Nature. 2014. Atezolizumab Proprietary 5% TILs Yes Rosenberg JE, et al. Lancet. 2016. Atezolizumab SP142 5% TILs Yes Balar AV, et al. Lancet. 2017. (platinum ineligible) Atezolizumab SP142 5% TILs No Massard C, et al. J Clin Oncol. 2017. Durvalumab SP263 25% TILs & TCs Yes Sharma P, et al. Lancet Oncol. 2016. Nivolumab Dako 28-8 1% TCs No Sharma P, et al. Lancet Oncol. 2017. Nivolumab Dako 28-8 1% TCs Yes Plimack ER, et al. Lancet Oncol. 2017. Pembrolizumab 22C3 1% TILs & TCs TILs only Bellmunt J, et al. N Engl J Med. 2017. (chemo vs immuno 2 nd line) Pembrolizumab 22C3 10% TILs & TCs No

Checkpoint Inhibitor Adverse Events (occurring in 10% of those with urothelial cancer that has progressed on platinum-containing regimen) Avelumab Atezolizumab Durvalumab Nivolumab Pembrolizumab Adverse Event, (%) All Grades Grades 3-4 All Grades Grades 3-4 All Grades Grades 3-4 All Grades Grades 3-4 All Grades Grades 3-4 Gen - Fatigue 41 7 52 6 39 6 46 7 38 4.5 Gen - Periph edema 17 0.4 18 1 15 2 13 0.4 - - Gen - Pyrexia 16 1 21 1 14 1 17 0.4 14 0.8 GI - Constipation 18 1 21 0.3 21 1 16 0.4 19 1.1 GI - Nausea 24 1 25 2 16 2 22 0.7 21 1.1 GI - Vomiting 14 1 17 1 - - 12 1.9 15 0.4 GI Abdominal pain 19 2 17 4 14 3 13 1.5 13 1.1 GI Diarrhea/colitis 18 2 18 1 13 1 17 2.6 18 2.3 Resp - Cough 14 0 14 0.3 10 0 18 0 15 0.4 Resp - Dyspnea 17 2 16 4 13 2 14 3.3 14 1.9 Skin - Rash 15 0.4 15 0.3 11 1 16 1.5 20 0.4 GU - UTI 21 5 22 9 15 4 17 7 15 4.9 GU - Hematuria - - 14 3 - - - - 12 2.3 Musculoskeletal pain 25 3 15 2 24 4 30 2.6 32 3 Reduced appetite 21 2 26 1 19 1 22 2.2 21 3.8 U.S. FDA Prescribing Information

Conclusions: Checkpoint Blockade for Bladder Cancer Checkpoint inhibitors blocking the PD-1/PD-L1 axis including avelumab, nivolumab, pembrolizumab, durvalumab, and atezolizumab are well tolerated and active in metastatic bladder cancer Five drugs are approved in second-line bladder cancer; two of those are also approved first-line if cisplatin-ineligible PD-L1 staining is a poor biomarker for outcome, but may be associated with survival in some studies New immunotherapy combinations merit testing

Immunotherapy Advances in NSCLC Naiyer A. Rizvi, MD Price Family Chair, Clinical Translational Medicine Professor of Medicine Director, Thoracic Oncology Co-Director, Cancer Immunotherapy Columbia University Medical Center

OS (%) CA209-003 Five-Year Update: Phase I Nivolumab in Advanced NSCLC 100 80 Median OS (95% CI), mo Overall (N=129) 9.9 (7.8, 12.4) 60 40 20 1 y OS, 42% 2 y OS, 24% 3 y OS, 18% 5 y OS, 16% 0 0 1 2 3 4 5 6 7 8 No. at Risk Years 129 49 27 20 17 16 3 1 0 Brahmer J, et al. AACR. 2017.

KEYNOTE-024 Study Design Key Eligibility Criteria Untreated stage IV NSCLC PD-L1 TPS 50% Pembrolizumab 200 mg IV Q3W (2 years) ECOG PS 0-1 No activating EGFR mutation or ALK translocation No untreated brain metastases No active autoimmune disease requiring systemic therapy R (1:1) N=305 Platinum-Doublet Chemotherapy (4-6 cycles) PD Pembrolizumab 200 mg Q3W for 2 years Key Endpoints Primary: PFS (RECIST v1.1 per blinded, independent central review) Secondary: OS, ORR, safety Exploratory: DOR Reck a To be eligible M, et for al. crossover, N Engl PD J Med. had to 2016; be confirmed www.clinicaltrials.gov; by blinded, independent central NCT02142738. radiology review and all safety criteria had to be met.

PFS, % KEYNOTE-024: Progression Free Survival 100 90 80 70 60 50 40 30 20 10 0 No. at risk Reck M, et al. N Engl J Med. 2016. 62% 50% Events, n Median, mo HR (95% CI) P Pembro 73 10.3 0.50 Chemo 116 6.0 (0.37-0.68) 0 3 6 9 12 15 18 Time, months 154 104 89 44 22 3 1 151 99 70 18 9 1 0 48% 15% Pembrolizumab Chemotherapy <.001 Assessed per RECIST v1.1 by blinded, independent central review. Data cut-off: May 9, 2016.

O S, % 100 90 80 70 60 50 40 30 20 10 0 No. at risk KEYNOTE-024: Overall Survival 80% 72% 0 3 6 9 12 15 18 21 Time, months 154 136 121 82 39 11 2 0 151 123 106 64 34 7 1 0 Reck M, et al. N Engl J Med. 2016. Data cut-off: May 9, 2016. 70% 54% Events, n Median, mo HR (95% CI) P Pembro 44 NR 0.60 Chemo 64 NR (0.41-0.89) Pembrolizumab Chemotherapy.005

Key Eligibility Criteria: Stage IV or recurrent NSCLC No prior systemic therapy for advanced disease No EGFR/ALK mutations sensitive to available targeted inhibitor therapy 1% PD-L1 expression CheckMate-026 Study Design Nivolumab vs. Chemotherapy in First-line NSCLC (Phase III) R 1:1 Stratification Factors at Randomization: PD-L1 Expression (<5% vs. 5%) Histology (squamous vs. non-squamous) Socinski M, et al. ESMO. 2016; www.clinicaltrials.gov. Nivolumab 3 mg/kg Q2W N=271 Chemotherapy (histology dependent) Maximum of six cycles N=270 Tumor scans Q6W until week 48 then Q12W Disease Progression or Unacceptable Toxicity Disease Progression Crossover Nivolumab (optional) Primary Endpoint: PFS per BIRC ( 5% PD-L1 +) Secondary Endpoints: PFS per BIRC ( 1% PD-L1 +) OS ORR Exploratory Objective: Predictive biomarkers for outcomes with nivolumab

PFS (%) CheckMate-026: Primary Endpoint 100 80 60 40 PFS per IRRC in (>5% PD-L1+) Median PFS, months (95% CI) Nivolumab n=211 4.2 (3.0, 5.6) Chemotherapy n=212 5.9 (5.4, 6.9) 1-year PFS rate, % 23.6 23.2 HR=1.15 (95% CI: 0.91, 1.45), P=.2511 20 Nivolumab No. of patients at risk: Nivolumab 211 104 71 49 35 24 6 3 1 0 Chemotherapy 212 144 74 47 28 21 8 1 0 0 Socinski M, et al. ESMO. 2016. 0 0 Months Chemotherapy 3 6 9 12 15 18 21 24 27 All randomized patients ( 1% PD-L1+): HR = 1.17 (95% CI: 0.95, 1.43)

PFS (%) CheckMate-026: TMB Analysis High TMB Low/Medium TMB 100 90 Median PFS, months (95% CI) Nivolumab Chemotherapy n=47 n=60 9.7 (5.1, NR) 5.8 (4.2, 8.5) 100 90 Median PFS, months (95% CI) Nivolumab Chemotherapy n=111 n=94 4.1 (2.8, 5.4) 6.9 (5.5, 8.6) 80 70 HR = 0.62 (95% CI: 0.38, 1.00) 80 70 HR = 1.82 (95% CI: 1.30, 2.55) 60 60 50 Nivolumab 50 40 40 30 20 Chemotherapy 30 20 Chemotherapy 10 10 Nivolumab No. at Risk Nivolumab Chemotherapy 0 0 47 30 26 21 16 12 4 1 60 42 22 15 9 7 4 1 Adapted from Peters, et al. AACR. 2017. 3 6 9 12 15 18 21 Months 0 0 3 6 9 12 15 18 21 24 Months 111 54 30 15 9 7 2 1 1 94 65 37 23 15 12 5 0 0

PACIFIC Study: Durvalumab Monotherapy in NSCLC Phase 3, randomized, double-blind, placebo-controlled, multicenter, global study (26 countries) Patients with locally advanced unresectable NSCLC (Stage III) in a consolidation setting (N=702) Arm 1 (n=468): Durvalumab IV 10 mg/kg Q2wks for up to 12 months Absence of progression following at least 2 cycles of platinum-based chemotherapy concomitant with radiation therapy Primary endpoints PFS, OS Secondary endpoints ORR, DoR, DSR Safety/tolerability PK, immunogenicity, QOL R 2:1 Arm 2 (n=234): Placebo IV Q2wks Est. completion: 2017 FPD: Q2 14 LPCD: Q2 16 www.clinicaltrials.gov. NCT02125461; Creelan B, et al. Ann Oncol. 2015.

Response Rate (%) Ipilimumab and Nivolumab: First-line NSCLC 100 90 80 70 Unselected >1% PD-L1 >50% PD-L1 6 CRs 92 60 50 57 50 40 43 30 20 23 28 10 0 CM-012 mono (1L) CM-012 combo (1L) Hellmann, et al. Lancet Oncol. 2016.

Antonia S, et al. Lancet Oncol. 2016. Durvalumab and Tremelimumab in NSCLC

KEYNOTE-021: Cohort G Key Eligibility Criteria Untreated stage IIIB or IV nonsquamous NSCLC No activating EGFR mutation or ALK translocation Provision of a sample for PD-L1 assessment a ECOG PS 0-1 No untreated brain metastases No ILD or pneumonitis requiring systemic steroids R (1:1) a N=123 End Points Primary: ORR (RECIST v1.1 per blinded, independent central review) Key secondary: PFS Other secondary: OS, DOR, safety Exploratory: Relationship between antitumor activity and PD-L1 TPS Pembrolizumab 200 mg Q3W for 2 years + Carboplatin AUC 5 mg/ml/min + Pemetrexed 500 mg/m 2 Q3W for 4 cycles b Carboplatin AUC 5 mg/ml/min + Pemetrexed 500 mg/m 2 Q3W for 4 cycles b PD Pembrolizumab 200 mg Q3W for 2 years Langer, et al. Lancet Oncol. 2016; www.clinicaltrials.gov; NCT02039674. a Randomization was stratified by PD-L1 TPS <1% vs 1%. b Indefinite maintenance therapy with pemetrexed 500 mg/m 2 Q3W permitted.

Randomized Phase II: Carbo/Pemetrexed +/- Pembrolizumab PF S, % 100 90 80 70 60 50 77% 63% 13.0 mo 8.9 mo 40 30 Chemo 20 10 0 0 5 6 10 15 20 Time, months No. at risk PFS Pembro + Chemo 60 43 20 1 0 63 32 13 1 0 Δ26% P=.0016 Events, n HR (95% CI) Pembro + chemo 23 0.53 (0.31-0.91) Chemo alone 33 P=.0102 Langer, et al. Lancet Oncol. 2016. Pembro + Chemo Chemo Alone

Objective Response by PD-L1 TPS Langer, et al. Lancet Oncol. 2016. <1% n=21 1% n=39 1%-49% n=19 50% n=20 Pembrolizumab + Chemotherapy <1% n=23 1% n=40 1%-49% n=23 Chemotherapy Alone 50% n=17 Horizontal dotted lines represent the ORR in the total population. Assessed per RECIST v1.1 by blinded, independent central review. Data cut-off: August 8, 2016.

Overall Survival 100 90 80 70 92% 92% 75% 72% O S, % 60 50 40 30 20 10 Papadimitrakopoulou, et al. ASCO 2017. (Abstract 9094) Updated HR: 0.69 No. at risk Langer, et al. Lancet Oncol. 2016. 0 0 5 6 10 12 15 20 Time, months 60 53 33 5 0 63 57 31 6 0

Phase III First-line Combination Trials in Advanced NSCLC Treatment N Arms CheckMate227 1980 Nivolumab, Ipilimumab Nivolumab MYSTIC 1092 KEYNOTE-189 570 Impower 150 1200 Durvalumab, Tremelimumab Pembrolizumab, Pemetrexed, Carboplatin Atezolizumab, Paclitaxel/Carboplatin, Bevacizumab Durvalumab Atezolizumab, Paclitaxel/ Carboplatin Platinum-doublet Chemotherapy SOC Platinum-based Chemotherapy Pemetrexed, Carboplatin Paclitaxel/Carboplatin www.clinicaltrials.gov.

Current Status and Future Prospects for Development of Robust Prognostic and Predictive Biomarkers of Response in Selected Types of Solid Tumors

POINT: PD-L1 Staining by IHC is Generally a Useful Marker for the Benefit of PD-1 Blockade Joaquim Bellmunt, MD, PhD Associate Professor of Medicine Director, Bladder Cancer Center Dana-Farber Cancer Institute Boston, MA

The Rationale for PD-L1 Testing in Cancer PD-L1 Expression as a Predictive Biomarker Are tumors with high PD-L1 expression more sensitive to immune-mediated approaches compared with tumors that have low or no PD-L1 expression? The Prognostic Value of PD-L1 Expression Are tumors with high PD-L1 expression associated with better or worse prognosis compared with tumors that have low or no PD-L1 expression?

High PD-L1 Expression May Be Associated with Poor Prognosis Wu P, et al. PLoS ONE. 2015.

Zhao T, et al. PLoS ONE. 2017.

Considerations for PD-L1 Expression as a Predictive Biomarker and PD-L1 Testing Tumor Type PD-L1 Expression Level Cell Types That Express PD-L1 Variability Across PD-L1 Assays

PD-L1 Expression Can Vary Between Primary And Metastasis, Different Metastases, and Within One Tissue Sample Variation of PD-L1 Expression Within One Single NSCLC Tissue Sample 1 : Variation of PD-L1 Expression Between Primary Tumor Metastases 2 : Primary Melanoma Lymph Node Metastasis Brain Metastasis In 20 NSCLC samples tested with the anti PD-L1 28-8 IHC assay, discordant PD-L1 results were observed in 30% of the matched primary versus metastatic cases 3 1 McLaughlin J, et al. JAMA Oncol. 2016; 2 Madore J, et al. Pigment Cell Melanoma Res. 2015; 3 Phillips T, et al. Appl Immunohistochem Mol Morphol. 2015.

PD-L1 Expression May Predict Sensitivity to Immune-Mediated Approaches Study Cancer Type Level of PD-L1 Expression ORR, % (vs. control) Fehrenbacher 2016 1 NSCLC TC3 or IC3* TC2 or IC2* TC1 or IC1* TC0 or IC0* 37.5 (13.0) 7.7 (15.6) 14.0 (19.1) 7.8 (9.8) Massard 2016 2 Bladder PD-L1 positive** PD-L1 negative** TC 46.7 22.2 IC 55.6 12.5 TC or IC 46.4 0 Borghaei 2015 3 Non-squamous NSCLC PD-L1 10% PD-L1 5% PD-L1 1% 37 (13) 36 (13) 31 (12) Ferris 2016 4 SCCHN PD-L1 10% PD-L1 5% PD-L1 1% 27.9 (2.9) 22.2 (2.3) 17.0 (1.6) Rosenberg 2016 5 Bladder IC2/3 IC1/2/3 All 26 18 15 IC=immune cells; IHC=immunohistochemistry; TC=tumor cells *TC3 50%, TC2 5% and <50%, TC1 1% and <5%, and TC0 <1%; IC3 10%, IC2 5% and <10%, IC1 1% and <5%, and IC0 <1% **PD-L1 positive: either TC or IC 25%, PD-L1 negative: both TC and IC <25% 1 Fehrenbacher L, et al. Lancet. 2016; 2 Massard C, et al. J Clin Oncol. 2016; 3 Borghaei H, et al. N Engl J Med. 2015; 4 Ferris RL, ASCO 2016; 5 Rosenberg JE, et al. Lancet. 2016.

Advances in Other Treatment Settings in Advanced NSCLC: Pembrolizumab Immunotherapy (KEYNOTE 010 & 024) Pembrolizumab treatment setting: Indicated for first-line treatment for patients with metastatic NSCLC whose tumors express PD-L1 in 50% of cells and who do not have EGFR- or ALK-positive tumor mutations. Also indicated for patients with locally advanced or metastatic NSCLC progressing after 1 prior chemotherapy regimen and whose tumors express PD-L1 with 1% of cells. Patients with EGFR- or ALK-positive tumor mutations should also have received targeted therapy prior to treatment with pembrolizumab 1 Results from a randomized Phase III clinical trial: 2 Median OS TPS 50% 3,a Pembrolizumab 10 mg/kg (n=346) Pembrolizumab 2 mg/kg (n=345) Docetaxel (n=343) 17.3 14.9 8.2 HR (95% CI) 0.38 (0.38, 0.77) P=.0002 0.5 (0.36, 0.7) P<.0001 Median OS a 4,a Median PFS 4.0 3.9 4.0 8.5 10.4 12.7 0.71 (0.58, 0.88) P=.0008 p=0.0008 0.88 (0.74, 1.05) P=.07 0.61 (0.49, 0.75) P<.0001 0.79 (0.66, 0.94) P=.004 a Primary endpoint 0 4 8 12 16 20 Time (months) 1 FDA Prescribing Information; 2 Herbst RS, et al. Lancet 2016; 3 Reck M, et al. N Engl J Med. 2016.

Summary of FDA-Approved and Investigational PD-L1 Assays in Urothelial Carcinoma* Ab clone/ epitope Pembrolizumab 1 22C3 Atezolizumab 2,3 SP142 Nivolumab 4 28-8 Durvalumab 5 SP263 Avelumab 6 73-10 Cell type scored TCs and ICs ICs TCs TCs or ICs TCs Scoring method FDA status for urothelial carcinoma PD-L1 thresholds under evaluation CPS: % of PD-L1 positive TCs and ICs relative to the total number of tumor cells NA 1% 10% % of PD-L1 expressing ICs Complementary IC2/3 ( 5%), IC1 ( 1% but <5%), IC0 (<1%) % of PD-L1 expressing TCs NA 1% 5% % of PD-L1 expressing TCs or ICs NA 25% % of PD-L1 expressing TCs NA 5% * No head-to-head studies have been conducted and direct comparisons cannot be made between these studies. 1 Bellmunt J, et al. N Engl J Med. 2017; 2 Loriot Y, et al. Poster presentation at ESMO 2016. Abstract 83P; 3 Ventana. Roche receives FDA Approval for novel PD-L1 biomarker assay [press release]. May 18, 2016; 4 Sharma P, et al. Lancet Oncol. 2017; 5 Powles T, et al. Poster presentation at ASCO GU 2017. Abstract 286; 6 Patel M, et al. Poster presentation at ESMO 2017. Abstract 777PD.

Rosenberg JE, et al. Lancet. 2016. PD-L1 Antibody Atezolimumab in Cisplatin-resistant Bladder Cancer

Phase II CheckMate 275 Study in Chemo-resistant Bladder Cancer: Nivolumab is Active All PD-L1 <1% PD-L1 1% PD-L1<5% PD-L1 5% Outcome, % N=265 n=143 n=122 n=184 n=81 Confirmed ORR by BIRC a 19.6 16.1 23.8 15.8 28.4 95% CI 15.0 24.9 10.5 23.1 16.5 32.3 10.8 21.8 18.9 39.5 Median PFS in months (95% CI) 2.00 (1.87 2.63) 1.87 (1.77 2.04) 3.55 (1.94 3.71) Median OS in months (95% CI) 8.74 (6.05 NR) 5.95 (4.30 8.08) 11.30 (8.74 NR) Responders (N=52) Time to response: 1.9 months (1.6 5.9) Duration of response: NR (7.4-NR) Ongoing responders at time of response: 40/52 (77%) Safety: No new safety profile compare to prior reports a BIRC, blinded independent review committee Sharma P, et al. Lancet Oncol. 2017.

Clinical Response to Durvalumab Monotherapy in UC Correlates with PD-L1 Expression Antitumour Activity of Durvalumab per BICR in the Primary Efficacy Population of the UC Cohort, including the Second-line or Greater ( 2L) Post-platinum Subgroup ORR, n (%) (95% CI) ORR, n (%) (95% CI) Total PD-L1 High PD-L1 Low/Neg ORR by BICR assessment Primary Efficacy Population (103pts, 13 weeks follow-up) N=103 N=61 N=39 21 (20.4) (13.1, 29.5) 19 (31.1) (19.9, 44.3) ORR by BICR assessment ( 2L) post-platinum subgroup 2 (5.1) (0.6, 17.3) N=94 N=58 N=33 19 (20.2) (12.6, 29.8) 18 (31.0) (19.5, 44.5) 1 (3.0) (0.1, 15.8) Clinical activity seen in all patient subgroups Greater efficacy occurred in patients with PD-L1 high expression Efficacy is still observed in the PD- L1 low/neg group consistent with the level seen with SoC PD-L1 SP263 assay is especially helpful in informing patients on the likelihood of response to durvalumab Powles, et al. Presented at 2017 ASCO GU.

KEYNOTE-052: Pembrolizumab as 1 st -Line Therapy for Cisplatin-Ineligible Bladder Cancer Objective Response (by PD-L1 Subgroups) CPS <1% N=33 CPS 1% to <10% N=33 CPS 10% N=30 N = 100 n % (95% CI) n % (95% CI) n % (95% CI) ORR (24%) 6 18% (7-36%) 5 15% (5-32%) 11 37% (20-56%) Complete response 1 3% (0.1-16%) 0 4 13% (4-31%) Partial response 5 15% (5-32%) 5 15% (5-32%) 7 23% (10-42%) Stable disease 3 9% (2-24%) 5 15% (5-32%) 7 23% (10-42%) Excluding those with CPS unknown *CPS=Combined positive score for PD-L1 positive cells (tumor, immune cells) Median time to response: 2.0 months (range, 1.9-4.8) Median duration of response : Not reached (range, 1.4+ to 9.8+ months) Duration of response rate 6 months: 83% The PD-L1 high expression cut point was determined to be CPS 10% (PD-L1 positive tumor, immune cells) and enriched for response; this cut point will be validated in the remaining patients in the cohort (n=274 additional patients) Balar A, et al. Ann Oncol. 2016. Data cutoff date: June 1, 2016

Overall Survival, (%) Adapted from 1 Bellmunt J, et al. N Engl J Med. 2017; 2 Bellmunt J, et al. Oral presentation at SITC 2016; 3 http://www.mercknewsroom.com/news-release/oncologynewsroom/mercks-keytruda-pembrolizumab-significantly-improves-overall-survival. Overall Survival, (%) KEYNOTE-045: Overall Survival* No. at Risk Pembro Chemo 100 90 80 70 60 50 40 30 20 10 0 Overall Survival: Total 1 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (months) 270 272 Median OS, months (95% CI) Pembrolizumab 10.3 (8.0 11.8) Chemotherapy 7.4 (6.1 8.3) HR: 0.73; 95% CI, 0.59 0.91; P=.002 226 194 169 147 131 87 54 27 13 4 0 0 232 171 138 109 89 55 27 14 3 0 0 0 100 90 80 70 60 50 40 30 20 10 0 No. at Risk Pembro 74 Chemo 90 Overall Survival: CPS 10% 2 Median OS, months (95% CI) Pembrolizumab 8.0 (5.0 12.3) Chemotherapy 5.2 (4.0 7.4) HR: 0.57; 95% CI, 0.37 0.88; P=.0048 3 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (months) 60 51 42 35 31 18 12 7 3 0 0 0 76 51 36 28 24 16 8 4 1 0 0 0 *Assessed per RECIST v1.1 by blinded, independent central review. Data cutoff date: Sep 7, 2016. CPS is the % of PD-L1 positive tumor cells and tumor-infiltrating immune cells relative to the total number of tumor cells.

PD-L1 Expression as a Predictor of Checkpoint Blockade Sensitivity in UC Powles, et al. Nature. 2014. Phase I Atezolizumab Rosenberg, et al. Lancet. 2016. Phase II Atezolizumab Balar, et al. Lancet. 2017. Phase II Atezolizumab Massard, et al. J Clin Oncol. 2016. Phase I Durvalumab Sharma, et al. Lancet Oncol. 2016. Phase I/II Nivolumab Sharma, et al. Lancet Oncol. 2017. Phase I/II Nivolumab Plimack, et al. Lancet Oncol. 2017. Phase I Pembrolizumab Bellmunt, et al. N Engl J Med. 2017. Phase III Pembrolizumab 5/8 studies reported positive association with PD-L1 staining

Key Diagnostic Challenges in UC In-clinic use of PD-L1 expression is likely to differ across lines of therapy Multiple diagnostic assays & algorithms used in clinical development: confusion regarding the impact of the test used

UC: SP263 Uses Tumor and Immune Cell Scores Durvalumab Tumour Cell (TC) Area TC area with PD-L1 expression IC area with PD-L1 expression Immune Cell (IC) area Definition Tumour Cell: Proportion of tumour cells with membrane staining for PD-L1 at any intensity above background staining Immune Cell: Proportion of tumour associated immune cells with staining for PD-L1 at any intensity above background staining OR Assay SP263 Cut-offs for PD-L1 High TC 25% or IC 25% https://www.accessdata.fda.gov/cdrh_docs/pdf16/p160046c.pdf

UC: SP142 Uses Tumor and Immune Cell Scores Atezolizumab Tumour Area Definition The proportion of tumour area occupied by PD- L1 expressing tumour-infiltrating immune cells of any intensity IC area with PD-L1 expression Assay Cut-offs for PD-L1 High SP142 5% https://www.accessdata.fda.gov/cdrh_docs/pdf16/p160002c.pdf

UC: 22C3 Uses Combined Proportion Score Pembrolizumab Tumour Cell (TC) Area TC area with PD-L1 expression Definition The percentage of PD-L1 expressing tumour and infiltrating immune cells relative to the total number of immune cells. IC area with PD-L1 expression + Assay 22C3 10% Cut-offs for PD-L1 High https://www.accessdata.fda.gov/cdrh_docs/pdf15/p150013s001c.pdf

Overall Survival (Probability) Overall Survival (Probability) CheckMate 025: OS by PD-L1 Expression in arcc 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 Overall Survival: PD-L1 1% Median OS, months (95% CI) Nivolumab 21.8 (16.5 28.1) Everolimus 18.8 (11.9 19.9) Everolimus 0 3 6 9 12 15 18 21 24 27 30 33 Time (months) Nivolumab No. of patients at risk Nivolumab 94 86 79 73 66 58 45 31 18 4 1 0 Everolimus 87 77 68 59 52 47 40 19 9 4 1 0 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 Overall Survival: PD-L1 <1% Median OS, months (95% CI) Nivolumab 27.4 (21.4 NE) Everolimus 21.2 (17.7 26.2) HR (95% CI): 0.79 (0.53 1.17) HR (95% CI): 0.77 (0.60 0.97) Everolimus Time (months) Nivolumab 0 3 6 9 12 15 18 21 24 27 30 33 No. of patients at risk Nivolumab 276 265 245 233 210 189 145 94 48 22 2 0 Everolimus 299 267 238 214 200 182 137 92 51 16 1 0 Adapted from Motzer, et al. N Engl J Med. 2015; Sharma P, et al. Oral presentation at ESMO 2015. Abstract 3LBA. Based on data cut-off of June 2015.

PFS PFS IMmotion 150: IRF-Assessed PFS ITT Population 1% of IC expressing PD-L1 Atezo + bev (n=101) Atezo (n=103) Sunitinib (n=101) mpfs, mos (95% CI) Stratified HR* (95% CI) 11.7 (8.4 17.3) 1.00 (0.69 1.45) P=.982 6.1 (5.4 13.6) 1.19 (0.82 1.71) P=.358 8.4 (7.0 14.0) -- Atezo + bev (n=50) Atezo (n=54) Sunitinib (n=60) mpfs, mos (95% CI) Stratified HR* (95% CI) 14.7 (8.2 25.1) 0.64 (0.38 1.08) P=.095 5.5 (3.0 13.9) 1.03 (0.63 1.67) P=.917 7.8 (3.8 10.8) -- 100 100 80 80 60 60 40 40 20 20 0 No. at Risk Atezo + bev Atezo Sunitinib 0 3 6 9 12 15 18 21 24 27 30 33 36 Time (months) 101 73 62 55 48 40 34 21 13 5 1 1 103 59 43 35 31 29 24 14 10 4 2 1 101 69 53 37 30 26 22 11 7 4 2 0 No. at Risk Atezo + bev Atezo Sunitinib 0 3 6 9 12 15 18 21 24 27 30 33 36 Time (months) 50 36 31 26 24 22 19 12 7 3 1 1 54 29 19 15 14 13 13 7 6 3 1 60 40 29 21 16 13 12 6 3 1 1 *Compared with sunitinib. P values for descriptive purposes only and not adjusted for multiple comparisons. Adapted from McDermott D, et al. Poster presentation at ASCO GU 2017. Abstract 431.

Phase II Study of Atezolizumab + Bevacizumab vs. Sunitinib Phase I study: ORR 40% Phase II: PFS did not pan out but response seen in both PD-L1 high/low Overall survival a better metric Subsets of high PD-L1 cancers? McDermott D, et al. IMmotion150 biomarkers: AACR. 2017.

PD-L1 in SCCHN Treated with Nivolumab Exploratory analysis: Patients with a tumor PD-L1 expression level of 1% or more may have a greater magnitude of effect from nivolumab therapy than those whose PD-L1 level was less than 1%. Ferris RL, et al. N Engl J Med. 2016.

Conclusions The rationale includes potential prognostic and predictive value of PD-L1 testing PD-L1 positivity enriches for clinical benefit for selected drugs, diseases and settings IHC is unreliable for measuring PD-L1 expression Need to look beyond a single static biomarker Ergo, PD-L1 staining is not a useful helpful biomarker setting-based

Current Status and Future Prospects for Development of Robust Prognostic and Predictive Biomarkers of Response in Selected Types of Solid Tumors

COUNTERPOINT: PD-L1 Staining is Not All It s Purported to Be Jeffrey S. Weber, MD, PhD Laura and Isaac Perlmutter Cancer Center NYU Langone Medical Center

PD-L1 tumor staining makes little difference in melanoma, bladder, and RCC, but has some impact in lung cancer; with combined therapy, it is not useful

Nivolumab vs. DTIC-OS by PD-L1 Status in Front-line Melanoma Atkinson, et al. SMR. 2015.

PD-L1 Staining is Associated with Superior OS and PFS with Pembrolizumab in Melanoma, but Does Not Rule Out Benefit PD-L1 positive PD-L1 positive PD-L1 negative PD-L1 negative No. at risk PD-L1 positive PD-L1 negative 344 107 201 30 154 22 132 18 118 16 77 10 58 7 43 4 22 2 20 2 9 0 0 0 0 0 No. at risk PD-L1 positive PD-L1 negative 344 107 320 83 283 67 254 60 231 51 175 35 125 23 93 18 46 11 34 8 17 1 0 0 0 0 Daud A, et al. J Clin Oncol. 2016.

Larkin J, et al. AACR. 2017. PD-L1 Staining Falls Out as a Factor with Combined Checkpoint Blockade in Melanoma

No Impact of PD-L1 Staining on Outcome in Squamous NSCLC Treated with Nivolumab Brahmer J, et al. N Engl J Med. 2015.

No Impact of PD-L1 Staining on OS for Atezolizumab in Second-line NSCLC vs. Chemotherapy Rittmeyer A, et al. Lancet. 2017.

Rizvi N, et al. J Clin Oncol. 2016. No Impact of PD-L1 Expression on Outcome with Front-line NIVO + Chemotherapy in NSCLC

No Impact of PD-L1 Staining on Efficacy of Pembrolizumab in Refractory SCCHN PD-L1-positive patients PD-L1-negative patients PD-L1-positive patients PD-L1-negative patients Bauml J, et al. J Clin Oncol. 2017.

Benefit of Nivolumab Compared to Everolimus in Renal Cell Cancer is Independent of PD-L1 Staining Kaplan-Meier Curve for Overall Survival, According to Programmed Death 1 Ligand (PD-L1) Expression Level Motzer R, et al. N Engl J Med. 2015.

Little Impact of PD-L1 Staining on Survival in Renal Cell Cancer Patients Treated with Atezolizumab McDermott D, et al. J Clin Oncol. 2016.

No Impact of PD-L1 Staining on OS in Refractory Bladder Cancer Patients Treated with Atezolimumab Balar A, et al. Lancet. 2017.

Almost All Cells were PD-L1 Positive in Relapsed/Refractory Hodgkin s Lymphoma - Not Useful Chen, et al. J Clin Oncol. 2017.

Presented By Noah Hahn at 2015 ASCO Annual Meeting PD-L1 Status Summary

What are the Facts? In non-squamous NSCLC, clear association of PD-L1 staining level and OS benefit from nivolumab versus chemotherapy 1 However, in all other histologies in phase II/III trials, there is unclear benefit of PD-L1 staining With combination immune checkpoint blockade, PD-L1 staining is not a useful biomarker The utility of a predictive biomarker is to inform who not to treat: by that criterion, PD-L1 staining falls short Choosing patients by PD-L1 staining will increase response rates PD-L1 staining can be difficult to quantitate, may vary from tumor to tumor, and may vary over time within a tumor, and is inducible Ergo, PD-L1 staining is not a useful biomarker 1 Borghaei, et al. N Engl J Med. 2015.

Emerging Concepts of Combined Immune Checkpoint Blockade

POINT: The Efficacy of CTLA-4 + PD-1 is Superior to PD-1 Blockade Alone Robert L. Ferris, MD, PhD UPMC Endowed Professor and Chief Division of Head and Neck Surgery Associate Director for Translational Research Co-Leader, Cancer Immunology Program University of Pittsburgh Cancer Institute Pittsburgh, PA

Therapies to Drive an Immune Response Vaccines Adoptive T-cell therapies CAR-T TIL therapy Cytokines TLR agonists Agonist Abs (4-1BB, OX-40) Checkpoint blockade (Abs blocking CTLA4, PD-1, PD-L1)

Multiple Immuno-Inhibitory Pathways Regulate T-cell Tolerance and T-cell Exhaustion Potent ability to limit T-cell function Numerous therapeutic opportunities Many unanswered questions Freeman GJ, Sharpe AH. Nat Immunol. 2012. Not Pictured: TIGIT, B7-H4, B7-H5/VISTA

Activating and Inhibitory Co-Receptors CD28 OX40 GITR CD137 CD27 Activating Receptors HVEM Inhibitory Receptors CTLA-4 LAG-3 PD-1 TIM-3 BTLA VISTA Ipilimumab Tremelimumab Pembrolizumab Nivolumab Durvalumab Mellman I, et al. Nature. 2011. Agonistic antibodies T-cell stimulation Blocking antibodies

Proportion Alive Pooled OS Including EAP Data: 4846 pts 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 Median OS (95% CI): 9.5 (9.0 10.0) 3-year OS Rate (95% CI): 21% (20 22%) 0.2 0.1 0.0 Ipilimumab CENSORED 0 12 24 36 48 60 72 84 96 108 120 Months Patients at Risk Ipilimumab 4846 1786 612 392 200 170 120 26 15 5 0 Schadendorf D, et al. J Clin Oncol. 2015.

CheckMate-141 Study Design: Phase III Trial of Nivolumab in Recurrent/Metastatic SCCHN Key Eligibility Criteria R/M SCCHN of the oral cavity, oropharynx, larynx, or hypopharynx ECOG PS 0 1 Not amenable to curative therapy Progression 6 months of last dose of platinum-based therapy Documentation of p16 to determine HPV status No active CNS metastases Stratification factor Prior cetuximab treatment R 2:1 Randomized 360/360 Nivolumab 3mg/kg IV Q2W Investigator s Choice Methotrexate 40mg/m² IV weekly Docetaxel 30mg/m² IV weekly Cetuximab 400mg/m² IV once, then 250mg/m² weekly Primary endpoint OS Other endpoints PFS ORR Safety DOR Biomarkers Quality of life Ferris RL, et al. N Engl J Med. 2016.

Overall Survival 90 80 70 60 50 40 30 20 10 Overall Survival (% of patients) 100 Ferris RL, et al. N Engl J Med. 2016. 0 Nivolumab Investigator s Choice Median OS, mo (95% CI) Nivolumab (n=240) 7.5 (5.5 9.1) Investigator s Choice (n=121) 5.1 (4.0 6.0) HR (97.73% CI) 0.70 (0.51 0.96) 1-y OS 1-y Rate OS Rate (95% (95% CI) CI) 36.0% 36.0% (28.5-43.4) (28.5-43.4) 16.6% (8.6-26.8) 0 3 6 9 12 15 18 Months No. at Risk Nivolumab 240 167 109 52 24 7 0 Investigator s Choice 121 87 42 17 5 1 0

Overall Survival (% of patients) Overall Survival By PD-L1 Expression PD-L1 Expression 1% PD-L1 Expression <1% Treatment Arm Median OS, mo (95% CI) HR (95% CI) Treatment Arm Median OS, mo (95% CI) HR (95% CI) Nivolumab (n=88) 8.7 (5.7 9.1) Investigator s Choice (n=61) 4.6 (3.8 5.8) 0.55 (0.36 0.83) Nivolumab (n=73) 5.7 (4.4 12.7) Investigator s Choice (n=38) 5.8 (4.0 9.8) 0.89 (0.54 1.45) 100 90 80 70 60 50 40 30 20 10 0 Nivolumab Investigator s Choice 0 3 6 9 12 15 18 Months Nivolumab Investigator s Choice No. at Risk Nivolumab 88 67 44 18 6 0 73 52 33 17 8 3 0 Investigator s Choice 61 42 20 6 2 0 38 29 14 6 2 0 0 Ferris RL, et al. N Engl J Med. 2016. Nivolumab Investigator s Choice 100 90 80 70 60 50 40 30 20 10 0 Nivolumab Investigator s Choice 0 3 6 9 12 15 18 Months

CheckMate-067: Study Design Randomized, double-blind, phase III study to compare NIVO + IPI or NIVO alone to IPI alone Previously untreated, unresectable or metastatic melanoma N=945 Randomize 1:1:1 N=314 Stratify by: PD-L1 expression* BRAF status AJCC M Stage N=315 N=316 *Verified PD-L1 assay with 5% expression level was used for the stratification of patients; validated PD-L1 assay was used for efficacy analyses. **Patients could have been treated beyond progression under protocol-defined circumstances. NIVO 1 mg/kg + IPI 3 mg/kg Q3W x 4 then NIVO 3 mg/kg Q2W NIVO 3 mg/kg Q2W + IPImatched placebo IPI 3 mg/kg Q3W x4 + NIVO-matched placebo Treat until progression** or unacceptable toxicity Wolchok JD, et al. ASCO. 2015.

Response to Treatment NIVO + IPI (N=314) NIVO (N=316) IPI (N=315) ORR, % (95% CI) 57.6 (52-63.2) 43.7 (37.1-49.3) 19 (14.9-23.8) Two-sided P value vs IPI <.001 <.001 - Best Overall Response - % Complete Response 11.5 8.9 2.2 Partial Response 46.2 34.8 16.8 Stable Disease 13.1 10.8 21.9 Progressive Disease 22.6 37.7 48.9 Unknown 6.7 7.9 10.2 Duration of Response (months) Median (95% CI) NR (13.1, NR) NR (11.7, NR) NR (6.9, NR) *By RECISTv1.1; NR, not reached Wolchok JD, et al. ASCO. 2015.

PFS by PD-L1 Expression Level (1%) *Per validated PD-L1 immunohistochemical assay based on PD-L1 staining of tumor cells in a section of at least 100 evaluable tumor cells Wolchok JD, et al. ASCO. 2015.

Safety Summary 67.5% of patients (81/120) who discontinued the NIVO + IPI due to treatmentrelated AEs developed a response Patients Reporting Event, % Treatment-related adverse event (AE) Treatment-related AE leading to discontinuation Treatment-related death* NIVO + IPI (N=313) NIVO (N=313) IPI (N=311) Any Grade Grade 3-4 Any Grade Grade 3-4 Any Grade Grade 3-4 95.5 55.0 82.1 16.3 86.2 27.3 36.4 29.4 7.7 5.1 14.8 13.2 0 0.3 0.3 *One reported in the NIVO group (neutropenia) and one in the IPI group (cardiac arrest). Wolchok JD, et al. ASCO. 2015.

CTLA4 + PD-1 Targeting in Lung Cancer: CheckMate-012 NIVO 3 mg/kg Q2W + IPI 1 mg/kg Q12W (N=38) NIVO 3 mg/kg Q2W + IPI 1 mg/kg Q6W (N=39) NIVO 3 mg/kg Q2W (N=52) Confirmed ORR, % (95% CI) 47 (31-64) 39 (23-55) 23 (13-37) Median Duration of Response, months (95% CI) Median Length of Follow-up, months (range) Best Overall Response, % NR (11.3, NR) NR (8.4, NR) NR (5.7, NR) 12.9 (0.9-18) 11.8 (1.1-18.2) 14.3(0.2-30.1) Complete Response 0 0 8 Partial Response 47 39 15 Stable Disease 32 18 27 Progressive Disease 13 28 38 Unable to Determine 8 15 12 Median PFS, months (95% CI) 8.1 (5.6-13.6) 3.9 (2.6-13.2) 3.6 (2.3, 6.6) 1-year OS rate, % (95% CI) NC 69 (52-81) 73 (59-83) Antonia SJ, et al. J Thorac Oncol. 2016; Hellmann MD, et al. Lancet Oncol. 2017. NC= Not calculated (when >25% of patients are censored); NR= Not reached Combination data based on a February 2016 database lock, monotherapy data based on a March 2015 database lock except for OS data, which are based on an August 2015 database lock

Durvalumab (Anti PD-L1) + Tremelimumab (Anti CTLA-4) Trials in SCCHN Phase II HAWK Phase II CONDOR Setting 2L SCCHN post plat in R/M setting 2L SCCHN post plat in R/M setting Regimen PD-L1 status Rationale Durvalumab Durvalumab + Tremelimumab Durvalumab Tremelimumab + N=112 N=120 N=60^ N=60^ Accelerated approval of the Monotherapy in PD-L1+ Zandberg Accelerated approval of the Combination in PD-L1 Establishes individual component contribution to combination in PD-L1 Siu Phase III EAGLE www.clinicaltrials.gov. 2L SCCHN post plat 1L pts who progressed within 6 mo of multimodal tx w/pt in the locally advanced setting Durvalumab + Tremelimumab Durvalumab SOC + N=100 + N=100 + N=100 N=140 N=Adaptive 140* N=140 Confirmatory trial Combination approval in all-comers Ferris and Licitra

Potential of IO Therapies: 1L Trials with IO/IO Combinations Primary Endpoints Phase III CheckMate 651 No prior systemic therapy, platinum sensitive Nivolumab + Ipilimumab SOC (EXTREME) PFS, OS Phase II CheckMate 714 No prior therapy, platinum sensitive/refractory Nivolumab + Ipilimumab Nivolumab ORR Phase III KESTREL EAGLE No prior CT/IO, platinumsensitive (KESTREL) or refractory (EAGLE) Durvalumab + Tremelimumab Durvalumab SOC (EXTREME) PFS, OS Phase III KEYNOTE-048 1L+ R/M No prior systemic therapy, platinum sensitive Pembrolizumab + Platinum/5-FU Pembrolizumab SOC (EXTREME) PFS, OS www.clinicaltrials.gov.

Conclusions Immune checkpoint therapy, specifically anti-pd1 or -CTLA-4 monotherapy, improves survival in patients with metastatic cancers Several questions remain: Does combining two different checkpoint inhibitors add benefit? YES Do all patients benefit from combination therapy? NO Do we need to incorporate biomarkers for patient selection? YES Do we need to consider acute and chronic AE profile, and patient PS? YES

Emerging Concepts of Combined Immune Checkpoint Blockade

COUNTERPOINT: The Efficacy of CTLA-4 + PD-1 is NOT Superior to PD-1 Blockade Alone Joaquim Bellmunt, MD, PhD Associate Professor of Medicine Director, Bladder Cancer Center Dana-Farber Cancer Institute Boston, MA

PD-1 Blockade Alone CTLA-4 + PD-1 Slide courtesy of A. Ribas

PD-1 Blockade Alone Slide courtesy of A. Ribas CTLA-4 + PD-1 Matching the images using the ipi+nivo combo approach

Efficacy in 611 Patients in KEYNOTE 001 ORR: 33% ORR in previously untreated: 45% Pembrolizumab provided an ORR of 25% to 52% in the initial melanoma expansion cohorts of KEYNOTE-001, irrespective of dosing schedule or prior ipilimumab status. Ribas A, et al. JAMA. 2016. Central radiology review by RECIST v1.1 Data as of October 18, 2014; median follow-up: 21 months

Pembrolizumab Treatment-related AEs with Incidence >5% Adverse Event, % Total N=411 Any Grade Grade 3/4 Fatigue 36 2 Pruritus 24 <1 Rash 20 <1 Diarrhea 16 <1 Arthralgia 16 0 Nausea 12 <1 Vitiligo 11 0 Asthenia 9 0 Cough 9 0 Similar safety profiles in IPI-N and IPI-T patients Analysis cut-off date: October 18, 2013. Adverse Event, n (%) Total N=411 Any Grade Grade 3/4 Myalgia 9 0 Headache 8 <1 Hypothyroidism 8 <1 Decreased appetite 7 <1 Dyspnea 7 <1 Chills 6 0 Pyrexia 6 0 ALT increased 5 <1 Total 83 12 Ribas A, et al. JAMA. 2016

CheckMate-067: Study Design Randomized, double-blind, phase III study to compare NIVO + IPI or NIVO alone to IPI alone N=314 NIVO 1 mg/kg + IPI 3 mg/kg Q3W x 4 then NIVO 3 mg/kg Q2W Previously untreated, unresectable or metastatic melanoma N=945 Randomize 1:1:1 Stratify by: BRAF status AJCC M stage Tumor PD-L1 expression <5% vs 5%* N=316 NIVO 3 mg/kg Q2W + IPImatched placebo Treat until progression** or unacceptable toxicity N=315 IPI 3 mg/kg Q3W x4 + NIVO-matched placebo *Verified PD-L1 assay with 5% expression level was used for the stratification of patients; validated PD-L1 assay was used for efficacy analyses. **Patients could have been treated beyond progression under protocol-defined circumstances. Wolchok JD, et al. ASCO. 2015. The study was not powered for a comparison between NIVO and NIVO+IPI

Overall Survival Results From a Phase III Trial of Nivolumab Combined with Ipilimumab in Treatment-naïve Patients with Advanced Melanoma CheckMate 067 NIVO+IPI and NIVO significantly improved OS and PFS vs. IPI alone in patients with untreated advanced melanoma In descriptive analyses, NIVO+IPI resulted in numerically higher OS, PFS and ORR vs. NIVO alone For NIVO+IPI, median DOR and time to subsequent therapy are still not reached Larkin J, et al. AACR 2017. Abstract CT075.

Dr. Larkin s Comments about KEYNOTE-067 The trial isn t combined vs. sequential Just combo vs. single agents with crossover outside the study Study not designed or powered for comparison between nivo monotherapy vs. combination Number of events at this first OS analysis less then anticipated 112

Combination ICI Safety Summary With an additional 19 months of follow-up, safety was consistent with the initial report NIVO+IPI (N=313) NIVO (N=313) IPI (N=311) Patients reporting event, % Any Grade Grade 3-4 Any Grade Grade 3-4 Any Grade Grade 3-4 Treatment-related adverse event (AE) Treatment-related AE leading to discontinuation 95.8 58.5 86.3 20.8 86.2 27.7 39.6 31.0 11.5 7.7 16.1 14.1 Treatment-related death, n (%) 2 (0.6) a 1 (0.3) b 1 (0.3) b Most select AEs were managed and resolved within 3-4 weeks (85 100% across organ categories) ORR was 70.7% for pts who discontinued NIVO+IPI due to AEs, with median OS not reached a Cardiomyopathy (NIVO+IPI, n=1); Liver necrosis (NIVO+IPI, n=1). Both deaths occurred >100 days after the last treatment. b Neutropenia (NIVO, n=1); colon perforation (IPI, n=1). Larkin J, et al. New Engl J Med 2015.

Progression-Free Survival (Intent-to-Treat) In the randomized phase III KEYNOTE-006 study, pembrolizumab had fewer toxicities and significantly improved overall survival compared with ipilimumab. Robert C, et al. N Engl J Med. 2015. Wolchok JD, et al. ASCO. 2015.

Checkmate 016: Phase 1b Trial of Nivolumab + Ipilimumab in Metastatic RCC N=175* Key Eligibility Criteria Advanced or metastatic clear cell RCC KPS 80% No active CNS metastases Available tumor tissue (archival or recent acquisition) For NIVO3 + IPI1 and NIVO1 + IPI3 expansion arms and NIVO3 + IPI3 addition arm: Treatment naïve Stopped due to toxicity Nivolumab 3 mg/kg (IV) + ipilimumab 1 mg kg (IV) Q3W x 4 Nivolumab 1 mg/kg (IV) + ipilimumab 3 mg/kg (IV) Q3W x 4 Nivolumab 3 mg/kg (IV) + ipilimumab 3 mg/kg (IV) Q3W x 4 Nivolumab 3 mg/kg (IV) Q2W Primary Outcome Measures: Safety and tolerability Secondary Outcome Measures: ORR, DOR *Estimated study enrollment including arm that terminated due to toxicity. There are 47 patients continuing in each remaining arm. One prior adjuvant or neoadjuvant therapy for localized or locally advanced RCC is allowed provided recurrence occurred 6 months after the last dose of the adjuvant or neoadjuvant therapy. Only prior cytokine-based treatment for metastatic RCC (eg, interferon-alpha [IFN-α] or interleukin 2 [IL-2]) as prior therapy is allowed. AEs, SAEs, and laboratory abnormalities. By RECIST v1.1. www.clinicaltrials.gov. NCT01472081; Hammers HJ, et al. Poster presentation at ESMO 2016. Abstract 1062P.

Overall Survival (Probability) Overall Survival of Nivolumab + Ipilimumab in Advanced/Metastatic RCC (Checkmate 016) 1.0 0.9 0.8 0.7 0.6 0.5 0.4 Median OS, months (95% CI) NIVO3 + IPI1 NR NIVO1 + IPI3 32.6 (25.99 NR) 0.3 0.2 0.1 0.0 0 3 6 9 12 15 18 21 24 27 30 33 36 Time (months) Number of patients at risk NIVO3 + IPI1 47 47 44 41 38 35 35 27 17 14 13 6 0 NIVO1 + IPI3 47 45 43 41 39 38 37 29 17 15 12 7 0 Adapted from Hammers HJ, et al. Poster presentation at ESMO 2016. Abstract 1062P.

In Select Nivo-treated Pts, Median Survival Not Reached; How Much Crossover After Nivo? Motzer RJ, et al. N Engl J Med. 2015. Escudier B, et al. Eur Urol. 2017.

Phase II Study of Front Line Therapy With Nivolumab and Salvage Nivolumab + Ipilimumab in Patients With Advanced Renal Cell Carcinoma. HCRN: GU16-260 Nivolumab 240 mg IV Q2W x 6 then initial disease assessment PR or CR Part A Continue nivolumab 360 mg IV Q3W until PD, toxicity or 84 weeks PD or SD @ 12 months Part B Re-induce nivolumab 3 mg/kg and ipilimumab 1 mg/kg Q3W x 4 (must complete by week 16) Toxicity that requires discontinuation of study drug or 84 weeks of treatment completed continue being followed per protocol PD Evaluate for Part B Nivolumab 360 mg Q3W starting week 13-19 PR, CR or SD Continue nivolumab until PD, toxicity or 48 weeks PD Off study www.clinicaltrials.gov; NCT03117309.

DANUBE (Durvalumab +/- Tremelimumab) in 1 st Line UBC Randomization Stratification Factors 1. Cisplatin eligibility (eligible vs. ineligible) 2. PD-L1 status (positive vs. negative) 3. Visceral metastasis (presence or absence; i.e., bone, lung or liver) TCC of the urothelium (renal pelvis, ureters, urinary bladder and urethra) Treatment-naïve patients Unresectable/stage IV www.clinicaltrials.gov; NCT02516241. Randomization N=650 Estimated Timelines Estimated Completion: September 2019 Estimated Primary Data: March 2018 Durvalumab + tremelimumab N=217 Durvalumab Monotherapy N=217 SOC N=217 Primary Endpoints PFS & OS (Combo vs. SOC) Secondary Endpoints PFS & OS (Single agents vs. SOC) PFS (PD-L1+ and PD-L1-) OS (Combo vs. SOC) ORR (Combo vs. SOC) FACT-BL Immunogenicity PK (C max and C trough )

Checkmate 032: Phase 1/2 Trial of Nivolumab vs Nivolumab + Ipilimumab in muc: Overall Survival and Progression-Free Survival Overall Survival (Probability) Progression-Free Survival (Probability) Median OS, months (95% CI) Median PFS, months (95% CI) NIVO 3 + IPI 1 7.3 (5.6 11.4) NIVO 3 + IPI 1 2.6 (1.4 3.9) NIVO 1 + IPI 3 10.2 (4.5 NR) NIVO 1 + IPI 3 4.3 (1.6 8.2) No. at risk NIVO 1 + IPI 3 NIVO 3 + IPI 1 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0 3 6 9 12 15 18 21 24 27 Time (months) 26 21 17 7 3 3 3 2 1 0 104 84 59 45 38 27 8 0 0 0 No. at risk NIVO 1 + IPI 3 NIVO 3 + IPI 1 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0 3 6 9 12 15 18 21 24 27 Time (months) 26 15 6 1 1 1 1 1 0 0 104 41 27 20 19 8 4 0 0 0 Median follow-up times for NIVO 3 + IPI 1 arm is 16.7 months, and NIVO 1 + IPI 3 arm is 7.8 months Diamonds are censored patients. Sharma P, et al. Oral presentation at SITC 2016. Abstract 449.

KEYNOTE-045 Study (NCT02256436) Overall Survival: Total 43.9% 30.7% Events, n HR (95% CI) P Pembro 155 0.73 (0.59-0.91).0022 Chemo 179 Pembrolizumab Chemotherapy Median (95% CI) 10.3 mo (8.0-11.8) 7.4 mo (6.1-8.3) No. at risk 270 226 194 169 147 131 87 54 27 13 4 0 0 272 232 171 138 109 89 55 27 14 3 0 0 0 Bellmunt J, et al. N Engl J Med. 2017.

Can We Select Patients More Likely to Respond to PD-1/L1 Blockade? 1. Pre-existing T cell infiltration and adaptive PD-L1 expression 2. TCR clonality 4. Mutational load Rizvi, et al. Science. 2015. Le, et al. NEJM. 2015. Tumeh, et al. Nature. 2014. Herbst, et al. Nature. 2014. Tumeh, et al. Nature. 2014. 3. IFN signature by expression profiling 2.4 2.2 2.0 1.8 1.6 1.4 1.2 1.0 0.8 Nonresponder Responder Best Overall Response, RECISTv1.1 Ribas, et al. ASCO. 2015. 5. Transcriptome Hugo, et al. Cell. 2016.

Conclusions PD-1/PD-L1 blockade therapy should be used as single agent in patients who have a chance of responding to this therapy Combination therapies with PD-1/PD-L1 blockade should only be used in patients with a low likelihood of a tumor response to single agent therapy

Should PD-1 blockade be used alone or in combination?