roche analyst event

51st ASCO Annual Meeting, Chicago

Roche Analyst Event Sunday, 31 May 2015

Agenda

Welcome Karl Mahler, Head of Investor Relations Oncology strategy and outlook Daniel O’Day, Chief Operating Officer, Roche Pharmaceuticals Working along the cancer immunity-cycle: Strategies and new agents Ira Mellman, gRED: Ph.D., Vice President, Cancer Immunology, Genentech William Pao, pRED: M.D., Ph.D., Global Head Oncology Discovery and Translational Area, Roche ASCO 2015 Roche highlights: Setting new standards, developing combinations Sandra Horning, M.D., Chief Medical Officer and Head Global Product Development Growing importance of molecular information in cancer immunotherapy Garret Hampton, Ph.D., Vice President, Oncology Biomarker Development and Companion Diagnostics Summary and Q&A Karl Mahler, Head of Investor Relations

2

Oncology strategy and outlook

Daniel O’Day Chief Operating Officer, Roche Pharmaceuticals

3

Roche oncology: Continued sales growth A portfolio of differentiated medicines

4

CHF m

Sales at 2014 average exchange rates

0 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998

24,000

16,000 18,000

14,000

22,000 20,000

12,000 10,000 8,000

6,000 4,000 2,000

10 medicines 16 tumor types 42 indications

Oncology continues to become more complex Science, clinical practice evolving with understanding of disease biology

5

KRAS EGFR

Unknown

Classification of lung adenocarcinomas

>1 mutated gene 3% MEK1 <1%

NRAS 1% MET 1% PIK3CA

1% BRAF 2% HER2 3%

EGFR (other) 4%

ALK 8% EGFR

17%

KRAS 25%

No oncogenic driver detected

35%

2014 2004 Today

Com

plex

ity

PD-L1 positive (illustrative)

Pao & Girard. Lancet Oncol 2011; Johnson, et al. ASCO 2013

Roche’s oncology strategy has remained constant Today’s presentations highlight the results of our strategy across key programs

Personalized healthcare

•  Differentiation via biomarkers

•  Experts integrated in R&D

Higher certainty of patient benefit: Better for patients and for reimbursement

6

Follow the science

•  Bringing our best minds together

•  Science driven programs

Strong portfolio & development

•  Comprehensive pipeline and active partnering

•  Smart trial design

7

Strategy in action - Immunotherapy

Perspectives in key franchises

Summary

Strategy into action: Learning loop Comprehensive and scientifically driven strategies – crafted by our best minds

•  Ideation

•  Decision making

•  Quick & efficient pursuit of promising ideas

...and we dedicate budget for rapid proof of concept

studies (Ph1)

8

Research (gRED, pRED, Chugai &

external)

Development (early and late)

Clinical Insights & Biomarkers (internal and external)

Partnering (Genentech and Roche)

We bring our best scientists & clinicians

together for:

Immunotherapy & Combinations

Atezolizumab (aPD-L1, MPDL3280A) strategy Biomarker selection and combinations

Our vision: Bringing the potential for transformative benefit to a broad patient population either in mono therapy or combinations

Diagnostic selection to identify patients most likely to benefit from

atezolizumab as a monotherapy

Improve survival using well-tolerated combinations with existing and

emerging therapies

PD-L1+ (tumor – infiltrating

immune cells)

PD-L1+ (tumor cells)

9

Atezolizumab in 2/3L NSCLC (POPLAR) Clinical outcomes correlate with PD-L1 expression

10

Overall Survival

TC2/3 or IC2/3 37%*

TC0 and IC0 32%*

TC1/2/3 or IC1/2/3 67%*

TC3 or IC3 16%*

0 0.5 1 1.5

0.46

0.63

1.12

0.56

1.93

HRa

In favor of docetaxel In favor of atezolizumab

0 0.5 1 1.5

0.57

0.87

1.17

0.70

1.87

HRa


Progression Free Survival

a = Unstratified HR; * = eligible patient population Spira A. et al, ASCO 2015

ITT n=287 0.98 0.77

Atezolizumab strategy: Combinations Scientifically driven strategy that combines with chemo, targeted therapy, and other immunotherapy

11

Checkpoint Inhibitors Anti-PD-L1, Anti-TIGIT, IDOi, Anti-OX40

Eliminate M2 Macrophages Anti-CSF-1R

pRED gRED

Tumor-targeted cytokines Anti-CEA-IL2v FP Anti-FAP-IL2v FP APC stimulators

Anti-CD40 (agonist)

Costimulatory antibodies Anti-OX40 (agonist) T cell bispecifics

Anti-CEA/CD3 TCB

Both

Priming & activation

Antigen presentation

Antigen release

T cell infiltration

T cell Trafficking

Cancer T cell recognition

T cell killing

Clinical phase molecules

Chen & Mellman (2013) Immunity

Atezolizumab strategy: Programs by disease

LUNG 2L+ single-arm

Ph2 (x2) 2L+ rand Ph2 and

rand Ph3 1L single-agent

Ph3 (x2) 1L w/chemo,

Avastin Ph3 (x3)

BLADDER 2L+ & 1L cis-inel

single-arm Ph2

2L+ rand Ph3

Adjuvant Ph3 BREAST

1L w/Abraxane Ph3

KIDNEY

1L w/Avastin Ph2

1L w/Avastin Ph3

15 Roche / Genentech Phase 2 or 3 Studies

12

Focused strategy: Going deep in diseases where we have strong scientific rationale

Atezolizumab strategy: Summary Comprehensive, science-driven program

Immune Doublets Examples: •  Loc adv/ metastatic solid tumors with

ipilimumab or Interferon alfa-2b (Ph Ib) •  Combination with anti-CSF-1R (Ph Ib) •  Combination with anti-CD40 (Ph Ib) •  Combination with Incyte’s IDOi (Ph Ib) •  Combination with anti-OX40 (Ph Ib) •  Combination with CEA-IL2v (Ph Ib) •  Combination with Celldex anti-CD27

Dx + Monotherapy Examples: • mRCC (Ph II) • mUBC (Ph II) •  2L mNSCLC (Ph II ‘BIRCH’, ‘POPLAR’, ‘FIR’,

Ph III ‘OAK’) • Dx+ 1L NSCLC (2x: Sq/NSq, Ph III) • Dx+ mUBC (Ph III)

Combination with Chemotherapy

Examples: •  1L NSCLC Squamous and Non-

Squamous with platinum doublets (x3, Ph III)

•  1L TNBC with Abraxane (Ph III)

Combination with Targeted Therapy

Examples: • mRCC with Avastin (Ph II) • mRCC with Avastin (Ph III) •  EGFR+ NSCLC w/Tarceva (Ph Ib) •  ALK+ NSCLC w/alectinib (Ph Ib) •  Solid tumors with Avastin (Ph Ib) •  Solid tumors w/cobimetinib (Ph Ib) •  Lymphoma with Gazyva (Ph Ib)

Strategic pillars of our PD-L1 development approach

Immune Doublets

Chemotherapy Combinations

13

Dx+ Monotherapy

Targeted Therapy

Combinations

14



Summary

Hematology strategy on track Strong pipeline and programs across key indications

15

Compound Combination Indication P 1 P 2 P 3

Gazyva +chemo GREEN CLL

Gazyva +chemo GOYA aNHL

Gazyva +chemo GADOLIN iNHL

Gazyva +chemo GALLIUM 1L FL

Gazyva +aPD-L1 R/R FL

Gazyva +aPD-L1 aNHL

venetoclax* +Rituxan MURANO R/R CLL

venetoclax +Gazyva CLL14 CLL

venetoclax R/R CLL 17p

venetoclax +Rituxan/+BR R/R FL

venetoclax +Rituxan/Gazyva+chemo 1L aNHL

venetoclax +BR R/R NHL

venetoclax +bortezomib+chemo R/R MM

venetoclax +chemo AML

polatuzumab +Rituxan/Gazyva ROMULUS NHL

polatuzumab +Gazyva/BR R/R FL

polatuzumab +Gazyva/BR aNHL

Biosimilars delayed to 2017

Data presented at ASCO

* venetoclax in collaboration with AbbVie BR = bendamustine+Rituxan venetoclax (Bcl2 inhibitor); polatuzumab vedotin (CD79b ADC)

HER2 franchise outlook Complete portfolio of innovative drugs

16

2017

Established standard of care New standard of care Trials in progress

Adjuvant BC

Herceptin + chemo

Herceptin sc + chemo (HannaH)

Herceptin & Perjeta + chemo (APHINITY)

1st line mBC

Herceptin + chemo

Herceptin & Perjeta + chemo (CLEOPATRA)

2nd line mBC Xeloda + lapatinib Kadcyla (EMILIA)

2016 2012 2013 2014 2015 2011 2019 2018 2020

Kadcyla (KATHERINE) Kadcyla & Perjeta (KAITLIN)

Biosimilar launch (EU)

Neoadjuvant BC

Herceptin + chemo (NOAH)1

Herceptin & Perjeta + chemo (Neosphere, Tryphaena)

Kadcyla & Perjeta (KRISTINE)

Kadcyla & Perjeta (MARIANNE)

HER2 franchise: Perjeta commercial uptake strong

1,600

1,800

2,000

2,200

0 Q1 Q4 Q2 Q3 Q4 Q1 Q3

Perjeta Herceptin

Kadcyla

2014 2015 2013

Perjeta US Neoadj Approval

Perjeta Cleopatra

OS Readout

Next Milestone: eBC readout (APHINITY) in 2016

NEOSPHERE OS Data at this

ASCO CHFm

17

Avastin franchise Continues to be the most studied molecule at ASCO

Continued growth fueled by recent launches in cervical, ovarian cancers

Avastin Mesothelioma data presented at ASCO 2015

18

0

1,000

2,000

3,000

4,000

5,000

6,000

7,000

2011 2012 2013 2014

CHFm

Sales at 2014 average exchange rates

19



Summary

Collaborations: Zelboraf with Plexxikon; Cobimetinib iwith Exelixis; Alectinib with Chugai, Gazyva with Biogen Idec; Kadcyla with ImmunoGen

ASCO 2015: Roche highlights

•  More than 275 abstracts, Avastin most quoted drug

•  Potential improvements on current standard of care: –  Cobimetinib / Zelboraf in 1L skin cancer (targeted therapies) –  Alectinib in 2L ALK mutated lung cancer

•  New treatment options: –  Atezolizumab in 2L lung cancer (vs chemo) –  Atezolizumab combinations with chemo in lung and TNBC –  Avastin in mesothelioma –  Gazyva in R/R indolent NHL

•  Several filings resulting from data presented at the meeting

20

Roche’s oncology strategy has remained constant Today’s presentations highlight the results of our strategy across key programs

Personalized healthcare

•  Differentiation via biomarkers

•  Experts integrated in R&D

Higher certainty of patient benefit: Better for patients and for reimbursement

21

Follow the science

•  Bringing our best minds together

•  Science driven programs

Strong portfolio & development

•  Comprehensive pipeline and active partnering

•  Smart trial design

Working along the cancer immunity-cycle: Strategies and new agents I

22

Ira Mellman Vice President, Cancer Immunology, Genentech

The renaissance of immunotherapy is a revolution for cancer patients

23

Early data suggests that anti-PD-L1/PD-1 is active across a wide range of tumor types

Broad activity but only subset of patients benefit

ORR ~10-30% Melanoma

Bladder cancer

Head & Neck cancer

Breast cancer

Hodgkin’s lymphoma

Lung cancer

Renal cancer

Most patients will need biomarker selection and combination therapy: Strategy now supported by clinical results

Using patient data to understand cancer immunity-cycle

Progressive disease (PD) Why do many patients not respond? •  No pre-existing immunity?

Cha

nge

in s

um o

f lon

gest

dia

met

ers

(SLD

) fr

om b

asel

ine

(%)

100 80 60 40 20 0

-20 -40 -60 -80

-100 0 21 42 63 84 105 126 147 168 189 210 231 252 273 294

Time on study (days)

Complete response Partial response Stable disease Progressive disease

Monotherapy durable responses (PR/CR) What are the drivers of single agent response? How can PRs be enhanced to CRs? •  Insufficient T cell immunity? •  Multiple negative regulators?

Stable disease (SD) What combinations will promote PRs & CRs? •  Insufficient T cell immunity? •  Multiple negative regulators?

Powles et al., Nature 2014

Atezolizumab Phase 1 data: Urothelial bladder cancer patients

Combinations of immunotherapeutics Increasing response rates & benefit by keeping cancer immunity-cycle turning

26 Chen & Mellman (2013) Immunity

Genentech/Roche programs *Partnered programs

1

2

3

4

5

6

7

Cancer antigen presentation

(dendritic cells/APCs)

Priming and activation (APCs & T cells)

Trafficking of T cells to tumors

(CTLs)

Infiltration of T cells into tumors

(CTLs, endothelial cells)

Recognition of cancer cells by T cells (CTLs, cancer cells)

Killing of cancer cells (Immune and cancer cells)

Release of cancer cell antigens (Immune and cancer cells)

Avastin®

vaccines Anti-CD40

Anti-OX40 Anti-CD27* Anti-CTLA4*

Anti-PD-L1/PD-1 Anti-TIGIT Anti-OX40 (Treg)

Anti-CSF-1R IDO inhibitor Anti-CEA-IL2v

T cell bispecifics ImmTacs

IDO (indoleamine di-oxygenase) A suppressor of effector T cells

IFNg-mediated up-regulation of

tumor PD-L1

Shp-2

MAPK"PI3K

pathways"

IFNg-mediated up-regulation of

tumor IDO

Kynurenine Tryptophan

Shp-2

MAPK"PI3K

pathways"

Adaptive expression of PD-L1 Adaptive expression of IDO

IDO

T cell suppression

Pre-clinical data shows promising activity for IDO inhibition (GDC-0919)

28

•  Plasma kynurenine decrease

•  T regulatory cells decrease

•  T effector cells show increased proliferation & CTL function

•  Enhanced anti-tumor efficacy with anti-PD-L1 and anti–CTLA4

Anti-PD-L1 + IDOi

IDOi Anti-PD-L1 control

Ongoing studies •  Phase 1a (GDC-0919) •  Phase 1b combination (GDC-0919 + atezolizumab) •  Phase 1b combination (INCB024360 + atezolizumab)

Tum

or v

olum

e (m

m3 )

500

1000

1500

2000

Day

0 40 30 50 20 10 0

GDC-0919 in collaboration with NewLink; INCB024360 in collaboration with Incyte

TIGIT A second potent negative regulator of T cell responses

29

Tumor cell or DC

PVR

ITIM

100nm 1nm

TIGIT CD226

ITIM ITAM

2

T cell

3

1

Competes with CD226 for PVR

Disrupts CD226 activation

Directly inhibits T cells in cis

1

2

3

•  Human and murine tumor-infiltrating CD8+ T cells express high levels of TIGIT

•  Antibody co-blockade of TIGIT and PD-L1 elicits tumor rejection in preclinical models

•  TIGIT selectively limits the effector function of chronically stimulated CD8+ T cells

•  TIGIT interacts with CD226 in cis and disrupts CD226 homodimerization

TIGIT=T cell immunoreceptor with Ig and ITIM domains

TIGIT aTIGIT and aPD-L1 combination effective in PD-L1 non-responsive model

30

Tumor cell or DC

PVR

ITIM

100nm 1nm

TIGIT CD226

ITIM ITAM

2

T cell

3

1

Competes with CD226 for PVR

Disrupts CD226 activation

Directly inhibits T cells in cis

1

2

3

TIGIT

Anti-PD-L1 + TIGIT

Control Anti-PD-L1

10000

1000

100

10 0 10 20 30 40 60

Day

Med

ian

tum

or v

olum

e (m

m3 )

Complete remission (CR)

31

Elevated expression of TIGIT by T cells in human cancer supports combination with atezolizumab

Johnson et al (2014) Cancer Cell

TIGI

T/CD

3ε ra

tio

0.05

0.5

0.2

0.1

1.00

CD3ε

50

200

1000

5000

2 10 50 500

Tumor Normal

ρ=0.82

Lung squamous cell carcinoma

TIGIT

OX40 exhibits a dual mechanism of action Promote antigen dependent effector T cell activation and T regulatory cell inhibition

32

Effector T cell

TCR

Antigen Presenting Cell

OX40

OX40L

IFN-γ

OX40 engagement on Treg cells

Treg

OX40L

OX40L

OX40

Primary Tumor Challenge (EMT6)

10 0 20 30 40 50 0

1000

2000

3000

Tum

or v

olum

e (m

m3 )

Day

Control

Anti-mouse OX40

Adapted from Nature Rev Immunol. 4:420 (2004)

Increase in T effector cells by aOX40 may create need to combine with aPD-L1

anti-OX40

anti-PD-L1 (or IDOi)

PD-L1 increase

IFN-γ"

OX40L

OX40

CD8 T cell receptor

HLA

PD-1

PD-L1

Tumor cell

T cell

IFNγ

Promising pre-clinical efficacy of an aOX40/aPD-L1 combination model

0 20 40 600

500

1000

1500

2000

2500

day

Tum

or V

olum

e (m

m3 ) anti%PD%L1*

control*

Day 0 10 20 30 40 50

0

500

1000

1500

2000

2500

day

Tum

or V

olum

e (m

m3 ) !!!!!!!!!!!!!!!!!!!!!!!anti&OX40!

!!!!!!!!!!!!!!!!!control!

Tum

or v

olum

e (m

m3 )

Tum

or v

olum

e (m

m3 )

Day

anti-OX40 control

anti-PD-L1 control

0 20 40 600

500

1000

1500

2000

2500

day

Tum

or V

olum

e (m

m3 )

anti%OX40*+*anti%PD%L1*

Tum

or v

olum

e (m

m3 )

Day

anti-OX40 + anti-PD-L1

Jeong Kim et al. AACR 2015

Ongoing studies •  Phase 1a (MOXR0916)

•  Phase 1b combination (MOXR0916 + atezolizumab)

•  Planned (MOXR0916 + GDC-0919)

MOXR0916 = anti-OX40

Combinations with chemotherapy Induced inflammation & antigen release may enhance atezolizumab efficacy

35

Chemotherapy Vaccines

1

2

3

4

5

6

7

Cancer antigen presentation

(dendritic cells/APCs)

Priming and activation (APCs & T cells)

Trafficking of T cells to tumors

(CTLs)

Infiltration of T cells into tumors

(CTLs, endothelial cells)

Recognition of cancer cells by T cells (CTLs, cancer cells)

Killing of cancer cells (Immune and cancer cells)

Release of cancer cell antigens (Immune and cancer cells)


Combinations may extend the benefit of aPD-L1 Preclinical synergy with selected chemo and targeted therapies

Dosing

0

500

1000

1500

2000

2500

3000

0 6 12 16 22 29 36

Tum

or V

olum

e, m

m3

Day

aPD-L1

combo

Control

CT26 08-1033O

0 10 20 30 40 500

500

1000

1500

2000

2500

DayTu

mor

Vol

ume

(mm

3)

controlGDC-0973 (7.5mg/kg)aPDL1GDC-0973 (3mg/kg) + aPDL1

MEKi

CT26 (Krasmut)

Dosing

Control aPD-L1

aPD-L1/ MEKi

Oxaliplatin chemotherapy MEKi targeted therapy

Tum

or v

olum

e, m

m3

Tum

or v

olum

e, m

m3

Oxaliplatin

-100

Combinations with chemotherapy appear to extend the benefit of atezolizumab in NSCLC patients

Liu et al. ASCO 2015

Chemotherapy can promote Th1-type inflammation in tumors

0 42 84 126 168 210 252 294 0 42 84 126 168 210 252 294


Chan

ge in

Sum

of L

arge

st D

iam

eter

s fro

m B

asel

ine

(%)

0 42 84 126 168 210 252 294 336

CR/PR (n=3) SD (n=2) PD (n=0)

CR/PR (n=9) SD (n=1) PD (n=1)

-90 -80 -70 -60 -50 -40 -30 -20 -10

10 20 30 40 50 60 70 80 90

100

0

Atezolizumab + carboplatin/paclitaxel

Atezolizumab + carboplatin/pemetrexed

Atezolizumab + carboplatin/nab-paclitaxel

CR/PR (n=8) SD (n=4) PD (n=1)

Strategic vision: Lead by developing best-in-class combination therapies

Combinations with immunotherapies aCD40 ✔ Vaccines, Oncolytic Viruses ✔ ✔ aCTLA4 ✔ aOX40 ✔ aCD27 ✔ aCEA/FAP-IL2v ✔ T Cell Bispecifics ✔ ImmTACs Planned IDOi ✔✔ aCSF1R ✔ aTIGIT Planned Cytokines, anti-cytokines ✔ Combinations with other agents aVEGF ✔ EGFRi ✔ ✔ ALKi Planned BRAFi ✔ MEKi ✔ BTKi ✔ aCD20 ✔ aHER2 ✔ Chemo ✔ ✔ HDAC ✔ A2V ✔

Internal combo ✔ ✔

Partnered external combo

1

2

3

4

5

6

7

Priming & activation anti-CEA-IL2v anti-FAP-IL2v anti-OX40 anti-CTLA4* anti-CD27* anti-cytokine

Antigen presentation anti-CD40 IMA942 vaccine* oncolytic viruses* neo-epitope vaccine

Antigen release Pro-inflammatory EGFRi ALKi BRAFi MEKi Chemo BTKi HDAC

T cell infiltration anti-VEGF anti-Ang2/VEGF

Cancer cell recognition anti-CEA/CD3 TCB anti-CEA-IL2v FP anti-HER2/CD3 anti-CD20/CD3 anti-FAP-IL2v FP ImmTAC*

Cancer cell killing atezolizumab anti-OX40 anti-CSF-1R IDOi* anti-TIGIT IDO/TDO TDO

T cell trafficking

Clinical development Preclinical development

* Partnered projects

Immuno- suppression

Working along the cancer immunity-cycle: Strategies and new agents II

William Pao, M.D., Ph.D. Global Head Oncology Discovery and Translational Area, Roche pRED

39

Cancer immunotherapy at pRED Distinct MoAs with potential for combinations

Designed to amplify immune response

First anti-tumor T-cell engager from Roche to enter clinical trials

Best/first in class immuno-activating Ab

First antibody shown to eliminate tumor- associated macrophages

Macrophage

aCEA-IL2v FP aCEA/CD3 TCB aCD40 aCSF-1R

40

Proof of biological activity established in rare disease (PVNS)

Novel recombinant immunocytokine with superior safety, PK, and tumor targeting

Best-in-class TCB platform

Activates antigen presenting cells to jumpstart adaptive immunity

Tumor-associated macrophages A promising new target in oncology

41

Emactuzumab (aCSF-1R) Proof of biological activity shown in PVNS*

*PVNS - pigmented villonodular synovitis, emactuzumab = anti-CSF-1R (RG7155)

CSF1R Inhibition with Emactuzumab in Locally Advanced Diffuse-type Tenosynovial Giant Cell Tumors of the soft tissue: a phase 1, dose escalation and expansion study, Cassier P. et al. (2015), Lancet Oncology, in press

Emactuzumab Phase I data: PVNS

Chan

ge in

sum

of l

onge

st d

iam

eter

s

from

bas

elin

e %

Time from baseline (days)

Patients %

Objective response 24/28 86

Complete response 2/28 7

Partial metabolic response 24/26 92

Clinically progression-free 27/28 96

Clinical benefit

Each color represents a patient

42

Emactuzumab: Elimination of tumor-associated macrophages in solid tumors – phase I data

43 atezolizumab = anti-PD-L1 (MPDL3280A) Gomez-Roca, CA: Phase I study of RG7155, a novel anti-CSF1R antibody, in patients with advanced/metastatic solid tumors, 2015 ASCO

Oral presentation on Mon 1 Jun; abstract #3005

•  Proof of biological activity in wide range of tumor types •  Potentially synergistic with atezolizumab

•  Phase I combination study: dose escalation with atezolizumab ongoing •  No severe immune-related AE reported to date •  Multiple extensions to start Q3 2015

RG7155 plus Paclitaxel RG7155 monotherapy

% C

hang

e fro

m b

asel

ine

mac

roph

age

cov

erag

e (IH

C)

Depletion of CSF-1 dependent macrophages in solid tumors

Macrophages suppress T cell proliferation in vitro

Tumor targeted IL2v cytokine fusion proteins A platform to amplify and activate immune effectors in tumors

Mechanism of Action

Features

“IL-2v”

IL2v – interleukin-2 variant

Novel tumor-targeted, engineered IL-2 variant (IL-2v)-based immunocytokines for immunotherapy of cancer: Christian Klein et al., ASH, 2013, Blood 122 (21), 2278-2278

44

•  Binds to target to deliver novel variant of IL-2 to the tumor

•  Recruits killer T and Natural Killer (NK) cells to the tumor

•  Compared to standard IL-2-based therapy, it shows:

− Superior expansion of immune effectors

− Less activation of suppressive T-cells

− Better tolerability

− Better tumor targeting

aCEA-IL2v – phase I data Tumor targeting and intra-tumoral T cell activation

CEA - carcinoembryonic antigen; *Quantitative data from 2 CEA-ve patients pending, no tumor uptake on visual assessment *Schellens, Jan H. M. et al., CEA-targeted engineered IL2: Clinical confirmation of tumor targeting and evidence of intratumoral immune activation; 2015 ASCO

Increased CD8 densities and CD8:CD4 ratio confirm intra-tumoral immune activation at doses ≥ 20 mg

CD8

dens

ity (c

ells/

mm

2 )

aCEA-IL2v is an ideal combination partner for atezolizumab: Phase I combination starting

baseline (BL) 4d p.1. CEA-IL2v dose baseline (BL) 4d p.1. CEA-IL2v dose

CD8:

CD4

Ratio

45

Oral poster presentation on Saturday 30 May; abstract # 3016*

CEA-IL2v tumor accumulation shown in 4/4 patients with CEA+ tumors

89Zr-CEA-IL2v-PET/CT

Baseline

CEA

+ve

N

SCLC

C

RC

C1D5 C1D2

FDG-PET/CT

aFAP-IL2v: A complementary approach targeting tumor stroma

FAP – fibroblast-activation protein

FAP staining

•  FAP-IL2v phase I start: planned in 2015

Parameter CEA-IL2v FAP-IL2v Target cell Cancer cells

(CEA) Tumor stromal cells (FAP)

Target expression

Heterogeneous Homogeneous

Target density Medium High Indications Mainly GI

indications: GC, PaC, CRC

Very broadly expressed: High expression in e.g. H&N, sq NSCLC, BC

46

BITE

ü High interest, outstanding clinical efficacy with CD19 CARs in hematology

û Activity associated with high toxicity

û Challenging manufacturing and regulatory processes

Potency üüü ü üüü Long half-life û üüü üüü Differentiation of high vs low antigen expressing cells

ü ü üüü

“1:1” format

“2:1” format

T cell engaging therapies in development CAR T cells versus T cell bispecific antibodies

47

Engineered T cells (CARS) T cell engaging bispecific antibodies

CEA-TCB: First “ 2:1” T cell bispecific in the clinic Targets CEA-expressing tumors to induce localized T cell-mediated killing

CD3e T cell engagement Low affinity

High avidity binding to tumor

antigen

Inert, engineered Fc region

48

Mechanism of Action

Features

•  Binds T cells and tumor cells simultaneously •  Results in T cell activation/proliferation and killing of tumor cells •  Does not require MHC:peptide complex presentation by tumor

cells •  T cell engagement independent of specificity and activation

status

•  Entry into human – study started in Dec 2014 •  Combination with IL2v & PD-L1 planned early in clinical

development •  Broad, early pipeline of several TCBs addressing solid &

hematological malignancies

CEA-TCB: Novel mode of action From poorly inflamed to highly inflamed tumors

•  Recruits new T cells and/or expands pre-existing ones •  Induces T cell re-localization from tumor periphery to tumor bed •  Activates intra-tumor T cells to become differentiated towards effector memory phenotype

49

Intravital two-photon imaging 24 h after CEA-CD3 therapy. Red: Tumor cells (LS174T RFP). Green: Human T cells

Control CEA-CD3

Bacac M. et al., submitted

Roche cancer immunotherapy pipeline

Abundant opportunities for combinations

Checkpoint Inhibitors Anti-PD-L1, Anti-TIGIT, IDOi, Anti-OX40

Eliminate M2 Macrophages Anti-CSF-1R

pRED gRED

Tumor-targeted cytokines Anti-CEA-IL2v FP Anti-FAP-IL2v FP APC stimulators

Anti-CD40 (agonist)

Costimulatory antibodies Anti-OX40 (agonist) T cell bispecifics

Anti-CEA/CD3 TCB

Both

Priming & activation

Antigen presentation

Antigen release

T cell infiltration

T cell Trafficking

Cancer T cell recognition

T cell killing

Clinical phase molecules

50 Chen & Mellman (2013) Immunity

Cancer immunotherapy program growing strongly

51

Compound Combination Indication Phase 1 Phase 2 Phase 3

aPD-L1

Mono +Tarceva +chemo +Avastin+chemo

Lung cancer

aPD-L1 Mono Bladder cancer

aPD-L1 Mono +Avastin Renal cancer

aPD-L1 +Zelboraf +Zelboraf+cobimetinib Melanoma

aPD-L1

Mono +Avastin +cobimetinib +ipilimumab +IFN alfa-2b +aCD40 +aOX40 +aCSF-1R +aCEA-IL2v FP

aPD-L1 +Avastin+FOLFOX Colorectal cancer

aPD-L1 Mono +Gazyva Hematology

aPD-L1 Mono +chemo Triple neg. breast cancer

aCSF-1R Mono +aCD40 Solid tumors

aCEA-IL2v FP Mono Solid tumors

aFAP-IL2v FP Mono Solid tumors

aOX40 Mono Solid tumors

aCEA/CD3 TCB Mono Solid tumors

IDO Mono Solid tumors

IDO (Incyte)* +aPD-L1 Solid tumors

Varlilumab (Celldex)* +aPD-L1 Solid tumors

Study ongoing Study imminent * External collaboration

Solid tumors

52

ASCO 2015 Roche highlights: Setting new standards, developing combinations

Sandra Horning, M.D. Chief Medical Officer and Head Global Product Development

53

Cancer immunotherapy Roche program update Atezolizumab in lung cancer Atezolizumab in bladder cancer

Targeted therapies and combinations

Alectinib in ALK-positive lung cancer Cobimetinib + Zelboraf in melanoma Kadcyla, Perjeta, Herceptin in HER2+ breast cancer

Hematology Gazyva in NHL

Agenda Setting new standards, developing combinations

Roche cancer immunotherapy program

54

aPD-L1 NSCLC (Dx+)

aPD-L1 2/3L NSCLC

aPD-L1+Avastin 1L Renal aPD-L1

1/2L Bladder

Phase II Phase III Phase I aPD-L1

2/3L NSCLC aPD-L1**

2/3L Bladder aPD-L1+Avastin+chemo**

1L non sq NSCLC aPD-L1+chemo** 1L non sq NSCLC

aPD-L1+chemo** 1L sq NSCLC

aPD-L1** 1L non sq NSCLC (Dx+)

aPD-L1** 1L sq NSCLC (Dx+) aPD-L1+chemo**

1L TNBC

aPD-L1** Adjuvant MIBC (Dx+)

aPD-L1** tba

aPD-L1+Avastin** 1L RCC

Status as at May 31, 2015

IDO Solid tumors

aPD-L1+ipilimumab Solid tumors

aPD-L1+aCD40 Solid tumors

aPD-L1+IFN-alfa Solid tumors

aPD-L1+aCSF-1R** Solid tumors

aPD-L1+aCEA-IL2v FP** Solid tumors

aPD-L1+aOX40** Solid tumors

aPD-L1** tba

aPD-L1** tba

aPD-L1** tba

aPD-L1** tba

aPD-L1** tba

NME** tba

aPD-L1 trials

NMEs monotherapy

2015 readout expected

Study start in 2015

Immune doublets

**

üData at ASCO 2015

üü

aOX40 Solid tumors

aCEA/CD3 TCB Solid tumors

aPD-L1+Zelboraf Melanoma

aPD-L1+Tarceva NSCLC

aPD-L1+Avastin+chemo Solid tumors

aPD-L1 + Gazyva R/R FL / aNHL

aCSF-1R Solid tumors

aCEA-IL2v FP Solid tumors

aPD-L1+Avastin Solid tumors

aPD-L1+cobimetinib Solid tumors

aPD-L1+Zelboraf+cobi** Melanoma

aPD-L1+lenalidomide** MM

aPD-L1+chemo Solid tumors

aPD-L1 Solid tumors üü

ü

üü

Biomarker development for atezolizumab PD-L1 expression on different cell types

55

IC Score PD-L1 staining TC Score PD-L1 staining

IC3 IC ≥ 10% TC3 TC ≥ 50%

IC2 IC ≥ 5% and < 10% TC2 TC ≥ 5% and < 50%

IC1 IC ≥ 1% and < 5% TC1 TC ≥ 1% and < 5%

IC0 IC < 1% TC0 TC < 1%

PD-L1 expression on immune cells (IC) and tumor cells (TC)

Gettinger S. et al, ASCO 2015

Assessed by immunohistochemistry using SP142, a sensitive and specific proprietary monoclonal antibody assay against PD-L1

Tumor cell staining Immune cell staining

Atezolizumab clinical program in 2/3L NSCLC Breakthrough designation in PD-L1+ patients

56

ORR Data at ASCO

All comers 2/3L mNSCLC n=1100

PD-L1 stratified

atezolizumab

1200 mg IV Q3 weeks

docetaxel

75 mg/m2 IV Q3 weeks

OAK Phase III

OS Data expected 2016

PD-L1-selected mNSCLC n=138

atezolizumab

1200 mg IV Q3 weeks

POPLAR Phase II OS Interim data at ASCO Final data in H2 2015

All comers 2/3L mNSCLC n=287

PD-L1 stratified

atezolizumab

1200 mg IV Q3 weeks

docetaxel

75 mg/m2 IV Q3 weeks

BIRCH Phase II PD-L1-selected

mNSCLC n=667 atezolizumab

1200 mg IV Q3 weeks ORR Data in Q3 2015

Note: Atezolizumab (Anti-PD-L1) is listed as MPDL3280A in clinicaltrials.gov mNSCLC = metastatic Non Small-Cell Lung Cancer

Primary end-point: FIR Phase II

atezolizumab

1200 mg IV Q3 weeks

Atezolizumab in 2/3L NSCLC (POPLAR) Clinical outcomes correlate with PD-L1 expression

57

Overall survival

TC2/3 or IC2/3 37%*

TC0 and IC0 32%*

TC1/2/3 or IC1/2/3 67%*

TC3 or IC3 16%*

0 0.5 1 1.5

0.46

0.63

1.12

0.56

1.93

HRa


0 0.5 1 1.5

0.57

0.87

1.17

0.70

1.87

HRa


Progression-free survival

a = unstratified HR; * = eligible patient population Spira A. et al, ASCO 2015

ITT n=287 0.98 0.77

Atezolizumab in 2/3L NSCLC (POPLAR) Overall survival correlates with PD-L1 expression

58

•  Final POPLAR and BIRCH read-outs in H2 15

•  Phase III (OAK) in all comers (stratified by PD-L1 expression): Data in 2016

a = Unstratified HR Spira A. et al, ASCO 2015

9.1 mo (7.4, 12.8)

Not Reached (11.0, NE)

9.7 mo (6.7, 11.4)

9.7 mo (8.6, 12.0)

TC1/2/3 or IC1/2/3 TC0 or IC0

Atezolizumab + chemotherapy in 1L NSCLC (Ph1) Well tolerated with no unexpected toxicities

59

atezolizumab 15mg/kg IV q3w+ carboplatin q3w + nab-paclitaxel q1w

atezolizumab 15mg/kg IV q3w+ carboplatin q3w + pemetrexed q3w

atezolizumab 15mg/kg IV q3w+ carboplatin q3w + paclitaxel q3w

1L NSCLC n=37

E

D

C

4-6 cycles

atezolizumab

atezolizumab +/- pemetrexed

maintenance

Treat to PD or

loss of clinical benefit

Treat to PD or


Treat to PD or


atezolizumab

Liu S. et al, ASCO 2015

•  The addition of atezolizumab to standard 1L chemotherapy was well tolerated throughout the course of treatment, including maintenance therapy, with no unexpected toxicities

•  No autoimmune renal toxicity or pneumonitis was observed

Atezolizumab + chemotherapy in 1L NSCLC (Ph1) Efficacy unrelated to PD-L1 expression

60 ORR across all arms = 67% (48-82) CR = complete response; PR = partial response; SD = stable disease; PD = progressive disease Liu S. et al, ASCO 2015

Arm C, n=5 Arm D, n=12 Arm E, n=13

-100

0 42 84 126 168 210 252 294 0 42 84 126 168 210 252 294


Chan

ge in

Sum

of L

arge

st D

iam

eter

s fro

m B

asel

ine

(%)

0 42 84 126 168 210 252 294 336

CR/PR (n=3) SD (n=2) PD (n=0)

CR/PR (n=9) SD (n=1) PD (n=1)

-90 -80 -70 -60 -50 -40 -30 -20 -10

10 20 30 40 50 60 70 80

0

CR/PR (n=8) SD (n=4) PD (n=1)

atezolizumab + carboplatin/paclitaxel

atezolizumab + carboplatin/pemetrexed

atezolizumab + carboplatin/nab-paclitaxel

ORR (95% CI)

60% (19-92)

75% (45-93)

62% (33-83)

90 100

Atezolizumab phase III program in 1L NSCLC Developing combinations, building on predictive PD-L1 biomarker

61

Study Indication Treatment arms Primary completion*

Combination studies – All comers (PD-L1 subgroup analysis)

GO29436 n=1200

Non-squamous Atezo + Carbo + Pac Atezo + Carbo + Pac + Avastin Carbo + Pac + Avastin

2017 (PFS)

GO29537 n=550

Non-squamous Atezo + Carbo + Nab-pac Carbo + Nab-pac 2017 (PFS)

GO29437 n=1200

Squamous Atezo + Carbo + Pac Atezo + Carbo + Nab-pac Carbo + Nab-pac

2017 (PFS)

Monotherapy studies – PD-L1 Selected

GO29431 n=400

Non-squamous Atezo Cis or Carbo + Pem 2017 (PFS)

GO29432 n=400

Squamous Atezo Cis or Carbo + Gem 2017 (PFS)

Atezo=atezolizumab; Carbo=Carboplatin; Pac=Paclitaxel; Nab-pac= Nab-paclitaxel; Cis=Cisplatin; Pem=Pemetrexed; Gem=Gemcitabine Note: Atezolizumab (Anti-PD-L1) is listed as MPDL3280A in clinicaltrials.gov `

* Outcome studies are event driven, timelines may change, OS endpoint included for all studies

Atezolizumab in 2L UBC (Ph1) ORR correlates with PD-L1 expression

62 UBC = urothelial bladder Cancer; IC = immune cells; ORR = overall response rate Petrylak D. et al, ASCO 2015

•  Response rates correlate with PD-L1 status; 20% CR in IHC2/3 •  Responses observed with visceral metastases: 38% RR in IHC2/3 •  Well tolerated with 5% Gr3/4 immune-related adverse events

IC0 IC1 IC2 IC3

Atezolizumab responses by PD-L1 biomarker status

PD-L1 IHC n=87

ORR, % (95% CI)

IC3 (n=12) 67% (35, 90) 50% (35, 65)

IC2 (n=34) 44% (27, 62)

IC1 (n=26) 19% (7, 39) 17% (7, 32)

IC0 (n=15) 13% (2, 40)

Atezolizumab in 2L UBC (Ph1) Survival correlates with PD-L1 expression

63

Overall survival

•  PD-L1 is a predictive biomarker for ORR, PFS and OS in 2L bladder cancer •  Phase II study (NCT02108652) in 1/2L bladder: Read-out H2 2015 •  Phase III studies in 2/3L bladder and high risk muscle invasive bladder after cystectomy (adjuvant)

started in 2015

Petrylak D. et al, ASCO 2015

7.6 mo (95% CI, 4.7-NE)

Not Reached (95% CI, 9.0-NE)

Survivala

N = 92 IC2/3 n = 48

IC0/1 n = 44

OS Median OS

(range) Not reached (1 to 20+ mo)

8 mo (1 to 15+ mo)

1-y survival (95% CI)

57% (41-73)

38% (19-56)

IC2/3

IC0/1






64

Results from P2 US and Global Studies

•  Alectinib, next generation, brain penetrant ALK inhibitor

•  Two P2 studies (global and US) show high systemic and CNS response rates in crizotinib-failed patients

•  Median PFS: 8.9m & 6.3m (global & US study, respectively)

•  Good tolerability with minimal dose modifications

•  Final duration of response and PFS results expected in H2 2015

Alectinib in ALK+ NSCLC after crizotinib failure Strong efficacy in two phase 2 studies

65

ORR CR

0

10

20

30

40

50

60

70

80

All (n=87)

Measurable CNS disease

(n=16)

All (n=122)

Measurable CNS disease

(n=16)

US Study Global Study

46%

69%

50% 57%

13% 20%

All (n=69)

Ou S-H. et al, ASCO 2015; Gandhi L. et al, ASCO 2015

Overall and CNS Response Rates

•  Both global and US studies show high ORR (57-69%) in measureable CNS disease; disease control rates 80-100%

•  Median duration of response in measureable CNS disease was 10.3 months in the global study

•  Filing planned 2015

•  Global phase III study in 1L ALK+ NSCLC (Alex) on-going: alectinib vs crizotinib

Alectinib in ALK+ NSCLC after crizotinib failure Excellent CNS disease control

66

Measureable CNS responses in global study

Day 6 Day 87

Ou S-H. et al, ASCO 2015; Gandhi L. et al, ASCO 2015

•  Median PFS of ~1 year in Cobimetinib + Zelboraf arm with longer follow-up (14.2 m) •  OS data not yet mature •  Filed in US and EU

Cobimetinib+Zelboraf in Melanoma (coBRIM) Update confirms benefit in BRAFV600+ patients

67

Progression free survival

Orphan disease designation

ORR = overall response rate; CR = complete response; PR = partial response; DoR = duration of response Larkin J. et al, ASCO 2015

HR = 0.58 (0.46-0.72)

12.3m 7.2m

Zelboraf + Cobimetinib

(n = 247)

Zelboraf + Placebo

(n = 248)

% ORR 69.6 50.0

% CR 15.8 10.5

Median PFS 12.3 7.2

Median DoR in months 13.0 9.2

Cobimetinib+Zelboraf in Melanoma (BRIM7) Phase I update shows strong efficacy

68

•  Two-year survival rate demonstrates benefit of cobimetinib + Zelboraf

•  Projected 28.5 m median OS compares to historical 13.6 m with Zelboraf alone

•  Results show importance of combination vs sequential use of targeted therapy in melanoma

CR = complete response; PR = partial response; SD = stable disease; PD = progressive disease Pavlick A. et al, ASCO 2015

Vem prog Naive

Time (months) 1 5 9 13 17 21 25 29 33 37

Zelboraf Progressors (N=66) BRAF inhibitor

−naive (n=63)

Zelboraf-Progressors

(n=66)

ORR% 87.3 15.2

CR% 15.9 1.5

mPFS (m) 13.8 2.8

mOS (m) 28.5 8.4

1-yr OS% 82.5 34.7

2-yr OS% 61.1 15.1

Overall survival

Kadcyla + Perjeta in 1L HER2+ mBC (MARIANNE) Non-inferior PFS for Kadcyla or Kadcyla + Perjeta

69

•  K and K + P achieved non-inferiority but not superiority to H + T

•  Addition of Perjeta to Kadcyla did not improve PFS

•  Kadcyla was better tolerated and maintained health-related quality of life for a longer duration

Herceptin + docetaxel (H+T) (8 mg/kg LD then 6 mg/kg + 100 or 75 mg/m2 q3w)

OR Herceptin + paclitaxel (H+T)

(4 mg/kg LD then 2 mg/kg + 80 mg/m2 qw)

Kadcyla + placebo (K) (3.6mg/kg + 840 mg LD then 420 mg q3w)

Kadcyla + Perjeta (K+P) (3.6mg/kg + 840 mg LD then 420 mg q3w)

Previously untreated

HER2+mBC n=1092

Ellis P. et al, ASCO 2015

Median PFS (m)

HR vs H+T

HR vs

K

13.7

14.1

15.2

0.91 p=0.31

0.87 p=0.14

0.91

1:1:1

Perjeta + Herceptin + docetaxel

Perjeta + Herceptin in HER2+ eBC (NEOSPHERE) 5yr follow-up shows lasting benefit in neoadjuvant setting

70

•  PFS and DFS benefit of neoadjuvant Perjeta + Herceptin + docetaxel - despite identical adjuvant therapy

•  No new safety concerns with longer follow-up

•  APHINITY adjuvant study results in 2016

FEC = 5-fluorouracil, Epirubicin, Cyclophosphamide; pCR = pathologic complete response; PFS = progression free survival; DFS = disease free survival Gianni L. et al, ASCO 2015 * PT arm also received docetaxel q3w X 4

Herceptin+docetaxel (H+T) n=107

Perjeta+Herceptin+docetaxel (P+H+T)

n=107

Perjeta+Herceptin n=107

Perjeta+docetaxel n=96

neoadjuvant

FEC q3 w x 3* + Herceptin

adjuvant

PFS, DFS pCR

S U R G E R Y

P + H + T (n=107)

H + T (n=107)

Events PFS 17 19

5-year PFS 86% 81%

HR 0.69 (0.34-1.40)

Events DFS 15 (14.9%) 18 (17.5%)

5-year DFS 84% 81%

HR 0.60 (0.28-1.27)

•  Achievement of tpCR (ITT analysis of all patients) reduced risk of PFS (HR = 0.54)

•  Results support pCR as an earlier indicator of benefit

pCR endpoint

HER2 franchise outlook Complete portfolio of innovative drugs

71

2017

Established standard of care New standard of care Trials in progress

Adjuvant BC

Herceptin + chemo

Herceptin sc + chemo (HannaH)

Herceptin & Perjeta + chemo (APHINITY)

1st line mBC

Herceptin + chemo

Herceptin & Perjeta + chemo (CLEOPATRA)

2nd line mBC Xeloda + lapatinib Kadcyla (EMILIA)

2016 2012 2013 2014 2015 2011 2019 2018 2020

Kadcyla (KATHERINE) Kadcyla & Perjeta (KAITLIN)

Biosimilar launch (EU)

Neoadjuvant BC

Herceptin + chemo (NOAH)1

Herceptin & Perjeta + chemo (Neosphere, Tryphaena)

Kadcyla & Perjeta (KRISTINE)

Kadcyla & Perjeta (MARIANNE)

•  APHINITY, KRISTINE results in 2016

•  NEOSPHERE update ASCO 2015

72






Hematology franchise A broad portfolio of innovative drugs

73

* venetoclax in collaboration with AbbVie BR = bendamustine+Rituxan venetoclax (Bcl2 inhibitor); polatuzumab vedotin (CD79b ADC)

Biosimilars delayed to 2017

Breakthrough therapy designation for venetoclax

in R/R CLL 17p

Compound Combination Indication P 1 P 2 P 3

Gazyva +chemo GREEN CLL

Gazyva +chemo GOYA aNHL

Gazyva +chemo GADOLIN iNHL

Gazyva +chemo GALLIUM 1L FL

Gazyva +aPD-L1 R/R FL

Gazyva +aPD-L1 aNHL

venetoclax* +Rituxan MURANO R/R CLL

venetoclax +Gazyva CLL14 CLL

venetoclax R/R CLL 17p

venetoclax +Rituxan/+BR R/R FL

venetoclax +Rituxan/Gazyva+chemo 1L aNHL

venetoclax +BR R/R NHL

venetoclax +bortezomib+chemo R/R MM

venetoclax +chemo AML

polatuzumab +Rituxan/Gazyva ROMULUS NHL

polatuzumab +Gazyva/BR R/R FL

polatuzumab +Gazyva/BR aNHL

Data presented at ASCO

Second positive readout for Gazyva GADOLIN in Rituxan-refractory iNHL

74

Rituxan + CHOP

Gazyva + CHOP Front-line DLBCL

(aggressive NHL) n=1418

PFS Trial to continue until 2016

Rituxan q2mo x 2 years

Gazyva q2mo x 2 years

Rituxan + CHOP or Rituxan or Rituxan + bendamustine

CR, PR

Gazyva + CHOP or Gazyva + CVP or

Gazyva + bendamustine 1L iNHL n=1401

Induction Maintenance

PFS Enrollment complete Q1 2014 Data expected 2017

GOYA: Ph III 1L Diffuse Large B-cell Lymphoma (DLBCL)

GALLIUM: Ph III 1L Indolent NHL (iNHL)

In collaboration with Biogen Idec CHOP=Cyclophosphamide, Doxorubicin, Vincristine and Prednisone; CVP=Cyclophosphamide, Vincristine and Prednisolone

CLL11: Ph III Chronic Lymphocytic Leukemia (CLL)

chlorambucil

Rituxan + chlorambucil

Gazyva + chlorambucil 1L CLL n=781

PFS Approved in Q4 2013

Primary end-point:

GADOLIN: Ph III Recurrent Indolent NHL (iNHL)

Gazyva q2mo x 2 years

CR, PR, SD

bendamustine Rituxan-refractory iNHL

n=411

Induction Maintenance PFS Data at ASCO 2015 Gazyva + bendamustine

Gazyva in Rituxan-refractory iNHL (GADOLIN) Significant PFS improvement achieved

75

•  PFS significantly prolonged by independent (HR=0.55) and investigator (HR 0.52) assessment •  Median PFS was ~ doubled by investigator assessment (29 versus 14 months) •  OS data immature •  Filing planned in 2015

14.9m

not reached

IFR-assessed PFS

a = Stratified HR Sehn L. et al, ASCO 2015

OS

HRa = 0.55 (0.40-0.74)

Setting new standards, developing combinations Driven by the breadth of our portfolio

76

venetoclax

alectinib cobimetinib

CSF-1R

CEA-IL2v

anti-OX40

anti-CD40

IDO/TDO

CEA-CD3

Targeted combinations approved Chemotherapy combinations approved Roche combinations in trials Chemotherapy combinations in trials

NMEs to be filed soon

Launched portfolio

Immunotherapy portfolio

atezolizumab

chemo

Growing importance of molecular information in cancer immunotherapy

77

Garret Hampton, Ph.D. Vice President, Oncology Biomarker Development and Companion Diagnostics

Personalized healthcare A cornerstone of our strategy

78

> 60% of current drugs in our development portfolio have a companion diagnostic

Diagnostics Pharma

Right patient – right drug

4 of 5 BTD were enabled by

having a Dx that identified patients most

likely to benefit

Molecule Dx

Gazyva CLL

Alectinib ALK+ NSCLC

Atezolizumab PD-L1+ NSCLC

Atezolizumab PD-L1+ UBC

Venetoclax 17p- CLL

Oncology is evolving More complex science with better patient outcomes

79

Overall survival in EGFR mut+ lung cancer Disease heterogeneity in lung cancer

Time (months) 0 3 6 9 12 15 18 21 24 27 30 33 36

0

0.2

0.4

0.6

0.8

1.0

PFS

prob

abili

ty

Erlotinib (n = 86) Chemotherapy (n = 87) HR 0.37 (95% CI 0.25–0.54);

log-rank p < 0.0001

86 63 54 32 21 17 9 7 4 2 2 0 0 87 49 20 8 5 4 3 1 0 0 0 0 0

Erlotinib Chemo

Number at risk

Unknown

MET splice site MET

Amplification

KRAS NRAS

ROS1 Fusions

KIF5B-RET

EGFR

ALK Fusions

HER2 BRAF

PIK3CA AKT1 MAP2K1

PD-L1 expression

Rosell et al. (2012) The Lancet

Biomarker discovery and development at Genentech/Roche

80

Disease Technology

Therapeutics

•  Biomarker prevalence •  Prognostic significance •  Subsets of unmet need •  Pipeline opportunities •  Dx hypotheses

Understand disease biology

Identify biomarkers

•  Pharmacodynamic •  Predictive •  Response surrogate •  Resistance (innate/acquired)

Define diagnostic endpoints

•  Companion Dx endpoints •  Response surrogates

•  Multiplex assays •  Real-time Dx •  Monitoring

Develop new technologies

Predictive markers – tissue staining for PD-L1 Cancer and/or immune cell expression of PD-L1 can correlate with clinical benefit

81

PD-L1+ (immune cells)

PD-L1+ (tumor cells)

Tumor type n % of PD-L1–positive (immune cells)

% of PD-L1–positive (tumor cells)

NSCLC 184 26 24

RCC 88 25 10

Melanoma 59 36 5

HNSCC 101 28 19

Gastric 141 18 5

CRC 77 35 1

Pancreatic 83 12 4

Bladder cancer 205 27 11

Herbst et al (2014) Nature

Patient selection enriches for benefit PD-L1-selected lung (TC & IC) and bladder cancer (IC)

82

PD-L1 IC staining Score IC ≥ 10% IC3

IC ≥ 5% and < 10% IC2 IC ≥ %1 and < 5% IC1

IC < 1% IC0

PD-L1 TC staining Score TC ≥ 50% TC3

TC ≥ 5% and < 50% TC2 TC ≥ 1% and < 5% TC1

TC < 1% TC0

Bladder cancer: Overall survival*

Time (months)

Ove

rall

surv

ival

0 3 6 8 11 12 19 0

20

40

60

80

100

1 2 4 5 7 9 10 13 14 15 16 17 18 20 21

Median OS 7.6 mo (95% CI, 4.7-NE)

Median OS Not Reached (95% CI, 9.0-NE)

IC2/3

IC0/1

+ Censored

Lung cancer: Survival hazard ratio*

0.1 1

In favor of docetaxel

0.77

1.12

0.63

0.56

0.46

Hazard Ratioa

In favor of atezolizumab

TC3 or IC3

TC2/3 or IC2/3

TC1/2/3 or IC1/2/3

TC0 and IC0

ITT (N = 287)

0.2 1 2

* Monotherapy data

Using patient data to understand cancer immunity-cycle

83

Atezolizumab Phase 1 data: Urothelial bladder cancer patients

Progressive disease (PD) Why do many patients not respond? •  No pre-existing immunity?

Cha

nge

in s

um o

f lon

gest

dia

met

ers

(SLD

) fr

om b

asel

ine

(%)

100 80 60 40 20 0

-20 -40 -60 -80

-100 0 21 42 63 84 105 126 147 168 189 210 231 252 273 294


Complete response Partial response Stable disease Progressive disease

Monotherapy durable responses (PR/CR) What are the drivers of single agent response? How can PRs be enhanced to CRs? •  Insufficient T cell immunity? •  Multiple negative regulators?

Stable disease (SD) What combinations will promote PRs & CRs? •  Insufficient T cell immunity? •  Multiple negative regulators?

Powles et al., Nature 2014

Biomarker discovery and development Integral part of the cancer immunotherapy learning organization

84

Research (gRED, pRED, Chugai & external)

Development (early and late)

Clinical Insights & Biomarkers

(internal and external)

Partnering (Genentech and Roche)

Cancer Immunotherapy Committee

Combination with Avastin Increasing response rates by keeping cancer immunity-cycle turning

85


Anti-VEGF

Patie

nt 1

On-treatment biopsies show absence of T cells in tumor bed before and after treatment with PD-L1

Lack of T cell infiltration may limit response to checkpoint inhibitors

Pre-treatment

Herbst et al. (2014) Nature

On-treatment (week 6)

CD8

Patie

nt 2

CD8

CD8 CD8

Combination with Avastin Increased T cell infiltration and clinical activity

86

CD

31

(vas

cula

ture

) C

D8

(T

cel

ls)

Pre-treatment Avastin Avastin + aPD-L1

Combination regimen benefits most patients irrespective of PD-L1 status

On-treatment biopsies show increased infiltrate and reduction in tumor vasculature

Chan

ge in

sum

of l

arge

st d

iam

eter

s fro

m b

asel

ine

(%)


PR SD PD

100 90 80 70 60 50 40 30 20 10

0 -10 -20 -30 -40 -50 -60 -70 -80 -90

-100

IHC 3

IHC 0 IHC 0

IHC 1

IHC 1 IHC 0 IHC 0 IHC 1

IHC 0

0 42 84 126 168 210 252 294 336 378

Discontin.

Sznol et al. ASCO GU 2015

Phase 2 in RCC ongoing (n= 300); Phase 3 initiated

Clinical data guides combinations Identifying potentially synergistic combinations

87

Anti-CSF-1R

Anti-PD-L1

Tumor associated macrophages

Complementary mechanisms of action to enhance benefit from immunotherapies

Myeloid signature associated with lack of response to atezolizumab in bladder cancer

Boxplot of Te↵, Th2 and Myeloid signature by response group. Red: PD; Black: SD; Blue: CR/PR.

●

●

PD SD CR/PR

−6−4

−20

Teff signature

Expression

N=23 N=20 N=15

●

●

●

●

●

●

●

●

●

●

●

●●

●●●

●

●

●

●

●

●

●

●

●

●

●●●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

PD SD CR/PR

−18

−16

−14

−12

−10

Th2 signature

Expression

N=23 N=20 N=15

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●●

●

PD SD CR/PR

−20

24

Myeloid signature

Expression

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●●

●

●

●●●

●

●

●

●

●

●

●

N=23 N=20 N=15

30

Myeloid signature: IL1B, IL8, CCL2

P=0.01

−2 −1.6 −0.67 0.22 1.1 1.6 2

2 1 0 -1 -2

cbind(1:nc)

Indication

Mel

anom

a

SCC

NSC

LC

RC

C

Aden

oN

SCLC

Blad

der

TNBC

HER

2+BC

CR

C

HR

+BC

Th17 Signature

Treg Signature

Th2 Signature

Teff Signature

B cell Signature

IB Signature

IFNg Signature

NK cell Signature

APC Signature

Myeloid Signature

Immune marker expression (illustrative)

Progressive disease

Phase I combination study of anti-CSF-1R and atezolizumab ongoing

Roche and FMI will innovate together Exclusive development of immunotherapy panel

88 Schumacher and Schreiber (2015) Science Yadav et al (2014) Nature

Tumor profiling with RNA sequencing

Predictive biomarkers & new combinations

Mutation-load and neo-antigen discovery

Boxplot of Te↵, Th2 and Myeloid signature by response group. Red: PD; Black: SD; Blue: CR/PR.

●

●

PD SD CR/PR

−6−4

−20

Teff signature

Expression

N=23 N=20 N=15

●

●

●

●

●

●

●

●

●

●

●

●●

●●●

●

●

●

●

●

●

●

●

●

●

●●●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

PD SD CR/PR

−18

−16

−14

−12

−10

Th2 signature

Expression

N=23 N=20 N=15

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●

●●

●

PD SD CR/PR

−20

24

Myeloid signatureExpression

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●

●●●

●

●

●

●

●

●

●

●

●●

●

●

●

●

●

●●

●

●

●●●

●

●

●

●

●

●

●

N=23 N=20 N=15

30

Myeloid signature: IL1B, IL8, CCL2 .

P=0.01

−2 −1.6 −0.67 0.22 1.1 1.6 2

2 1 0 -1 -2

cbind(1:nc)

Indication

Mel

anom

a

SCC

NSC

LC

RC

C

Aden

oN

SCLC

Blad

der

TNBC

HER

2+BC

CR

C

HR

+BC

Th17 Signature

Treg Signature

Th2 Signature

Teff Signature

B cell Signature

IB Signature

IFNg Signature

NK cell Signature

APC Signature

Myeloid Signature

Immune marker expression (illustrative)

Progressive disease

Identifying immunogenic mutations by structure-guided algorithms

Our vision for cancer immunotherapy A personalized approach to treatment

89 *Hypothetical algorithm

*Evaluate tumor: Is the tumor inflamed?

1

Strong PD-L1

Tx with aPD-L1/PD-1

2

Weak PD-L1

Are suppressive myeloid cells present?

3

No PD-L1

IDO/kyneurinin expressed?

4 No identifiable immune targets

Chemo+ Atezolizumab

Anti-CSF-1R+ Atezolizumab

IDOi+ Atezolizumab

Are T cells at tumor periphery? MHC loss?

Tumor antigen expression?

No T cells?

Antigen experienced?

1 2 3

Avastin+ Atezolizumab

No identifiable immune targets

Chemo+ Atezolizumab

4

T cell BiSpecific+

Atezolizumab

Anti-OX40+ Atezolizumab

Inflamed Y

Non-Inflamed N

Patient journey tomorrow Bridging comprehensive testing and drug development

90

Clinical decision

Patient care

Comprehensive testing

Molecular monitoring

R&D insights

Summary and Q&A

Karl Mahler Head of Investor Relations, Roche

91

Summary Skin cancer

cobimetinib + Zelboraf: Ph III (coBRIM) in 1L BRAF+ mM; PFS & biomarker update Filed in 2014

cobimetinib + Zelboraf: Ph Ib (BRIM7) in BRAF+ mM; OS update

Lung cancer

alectinib: two Ph II in 2L ALK+ NSCLC Filing in 2015

Atezolizumab (anti-PDL1): POPLAR, COMBO, FIR Discuss filing 2015

Avastin: mesothelioma Discuss filing 2015

Bladder cancer

Atezolizumab (anti-PDL1): Ph I update in bladder Discuss filing 2015

Breast cancer

Kadcyla: Ph II (ADAPT) neoadjuvant 12 weeks HER2+ HR+ BC

Kadcyla + Perjeta: Ph III (MARIANNE) in 1L HER2+ mBC

Herceptin + Perjeta: Ph II (NEOPSPHERE) in neoadjuvant HER2+ BC Approved in US, filed in EU

Avastin + Letrozole: Ph III (CALGB40503) in 1L HR+ mBC

Hematology

Gazyva: Ph III (GADOLIN) in R/R iNHL Filing 2015

Venetoclax: Ph I in R/R NHL and R/R MM BTD: Filing 2015 in 17p CLL depleted

92

Cancer immunotherapy: Encouraging early data in monotherapy across various cancer types

93

NSCLC 2L ORR range mPFS range mOS* range Prevalence

Un-stratified ~ 15% ~ 3 m ~ 9-11 m

Stratified 40-45% ~ 6-8 m Not reached ~20-30%

* Comparisons based on publicly available competitor data sets. Significant limitations with these cross-trial comparisons; significant variability in baseline population demographics further limit conclusions Note: CheckMate 057 2L+ non-squamous NSCLC trial was stopped early because the study met its OS endpoint

Bladder 2L ORR range mPFS range mOS range Prevalence

Stratified 25-45% ~ 2-5.5 m Not reached ~30%

Range across various studies

Range across various studies

Cancer immunotherapy: Still many open questions

94

to (step 4) and infiltrate the tumor bed (step 5), specifically recog-nize and bind to cancer cells through the interaction between itsT cell receptor (TCR) and its cognate antigen bound to MHCI(step 6), and kill their target cancer cell (step 7). Killing of the can-cer cell releases additional tumor-associated antigens (step 1again) to increase the breadth and depth of the response in sub-sequent revolutions of the cycle. In cancer patients, the Cancer-Immunity Cycle does not perform optimally. Tumor antigensmaynot be detected, DCs and T cells may treat antigens as self ratherthan foreign thereby creating T regulatory cell responses ratherthan effector responses, T cells may not properly home totumors, may be inhibited from infiltrating the tumor, or (mostimportantly) factors in the tumor microenvironment might sup-press those effector cells that are produced (reviewed by Motzand Coukos, 2013).

The goal of cancer immunotherapy is to initiate or reinitiate aself-sustaining cycle of cancer immunity, enabling it to amplifyand propagate, but not so much as to generate unrestrainedautoimmune inflammatory responses. Cancer immunotherapiesmust therefore be carefully configured to overcome the negativefeedback mechanisms. Although checkpoints and inhibitors arebuilt into each step that oppose continued amplification and can

dampen or arrest the antitumor immune response, the mosteffective approaches will involve selectively targeting the rate-limiting step in any given patient. Amplifying the entire cyclemay provide anticancer activity but at the potential cost ofunwanted damage to normal cells and tissues. Many recent clin-ical results suggest that a common rate-limiting step is the im-munostat function, immunosuppression that occurs in the tumormicroenvironment (Predina et al., 2013; Wang et al., 2013).Initiating Anticancer Immunity: Antigen Release,Presentation, and T Cell PrimingAttempts to activate or introduce cancer antigen-specific T cells,as well as stimulate the proliferation of these cells over the last 20years, have led to mostly no, minimal or modest appreciableanticancer immune responses. The majority of these effortsinvolved the use of therapeutic vaccines because vaccines canbe easy to deploy and have historically represented an approachthat has brought enormous medical benefit (reviewed by Pal-ucka and Banchereau, 2013). Yet, cancer vaccines were limitedon two accounts. First, until recently, there was a general lack ofunderstanding of how to immunize human patients to achievepotent cytotoxic T cell responses. This limitation reflectscontinued uncertainties concerning the identities of antigens to

Figure 1. The Cancer-Immunity CycleThe generation of immunity to cancer is a cyclic process that can be self propagating, leading to an accumulation of immune-stimulatory factors that in principleshould amplify and broaden T cell responses. The cycle is also characterized by inhibitory factors that lead to immune regulatory feedback mechanisms, whichcan halt the development or limit the immunity. This cycle can be divided into seven major steps, starting with the release of antigens from the cancer cell andending with the killing of cancer cells. Each step is described above, with the primary cell types involved and the anatomic location of the activity listed. Ab-breviations are as follows: APCs, antigen presenting cells; CTLs, cytotoxic T lymphocytes.

2 Immunity 39, July 25, 2013 ª2013 Elsevier Inc.

Immunity

Review

How can we convert unresponsive tumors into responsive tumors?

Can we modulate / monitor T-cell responses?

What are the factors affecting T-cell infiltration?

Can we create a personalized immunotherapy paradigm?

Can we convert an unresponsive tumor to a responsive one?

What influences durability of response?

Immunosuppression

T Cell Trafficking

Antigen Presentation

Biology

Biomarker/ Tools

Indication/ Patient pop.

Therapy Efficacy/ Safety

Doing now what patients need next

roche analyst event

Documents