Roche Pharma Day 2015

Roche Pharma Day 2015 Hematology franchise overview Cynthia Perettie Global Hematology Franchise Head David Traub Lifecycle Leader anti-cd20 Franchise Pharmaceuticals Division

Hematology indications still represent large unmet need 40% DLBCL FL CLL of DLBCL patients relapse and succumb to their disease FL Still no cure, many patients experience multiple relapses before succumbing to their disease CLL Still no cure, many patients achieve incomplete responses and require chronic treatment, many experience multiple relapses AML/MDS SOC in AML has made little progress in current SOC (PI and IMIDs) AML/MDS decades, we believe MM Hemophilia are hampered by combinations will significantly improve outcomes for patients MM Highly variable disease with driver mutations, associated with significant side effects, combinations with targeted therapies will advance SOC Hemophilia Existing therapies burden on therapy and compliance 2

Roche hematology expanding within and beyond NHL/CLL Phase I Phase II Phase III Registration Approved NHL MabThera SC Gazyva Type II anti-cd20 NHL CLL CLL NHL CLL venetoclax CLL* MM AML NHL idasanutlin AML AML/MDS polatuzumab NHL atezolizumab NHL MDS MM MM NHL combos AML doublets NHL AML HEMOPHILIA 4 early assets, incl. CD20/CD3 ACE910 Hemophilia Heme indications ADC: Antibody-drug Conjugate *Co-development molecule with Abbvie 3

MabThera SC NHL launch ongoing, strong uptake in most markets First EU launches in 2014, ongoing or imminent in further countries Encouraging initial uptake in majority of markets, comparable to Herceptin SC Slower conversion in countries with strong incentives to use IV (Germany) or limited reimbursement (UK) 4

Advancing the standard of care in NHL & CLL Gazyva: Designed to be the superior acd20 CLL 11 Ph III 1L Chronic Lymphocytic Leukemia (CLL) 1L CLL n=781 Gazyva + chlorambucil Rituxan + chlorambucil chlorambucil Launched in Q4 2013 GADOLIN Ph III R/R Indolent NHL (inhl) GOYA Ph III 1L Diffuse Large B-cell Lymphoma (DLBCL) Rituxanrefractory inhl n=411 Front-line DLBCL (aggressive NHL) n=1418 INDUCTION Gazyva + bendamustine bendamustine Gazyva + CHOP Rituxan + CHOP CR, PR,SD MAINTENANCE Gazyva q2mo x 2 years Filed Q3 2015 Data expected 2016 GALLIUM Ph III 1L Indolent NHL (inhl) 1L inhl n=1407 INDUCTION Gazyva + CHOP or Gazyva + CVP or Gazyva + bendamustine Rituxan + CHOP or Rituxan or Rituxan + bendamustine CR, PR,SD MAINTENANCE Gazyva q2mo x 2 years Rituxan q2mo x 2 years Data expected 2017 In collaboration with Biogen Idec CHOP=Cyclophosphamide, Doxorubicin, Vincristine and Prednisone; CVP=Cyclophosphamide, Vincristine and Prednisolone 5

GADOLIN Strong data for GAZYVA in refractory inhl IRF-Assessed PFS MabThera-refractory inhl (n=396) B median PFS 14.9 mos Gazyva+B median PFS not reached (Inv-PFS of 30 mos) PFS HR 0.55, p=0.00011 Gazvya provides substantial benefit in inhl patients who did not achieve adequate disease control from a previous MabTherabased regimen FDA Priority Review granted (10/24/2015) B = bendamustine 6

1L DLBCL GOYA to include data across cell of origin subtypes ~50% Elevated bcl-2 Double Positive Bcl-2 and myc* Double Hit Bcl-2 and myc Segment % of incidence Median OS [months] All Comers 100% >60 GCB 55% >60 ABC** 30% 30-40 ~50% ABC Unclassified GCB Double Positive 18-33% (46% of ABC) 24 Double-Hit 5-10% 12 ** worse outcomes in ABC may be driven by higher proportion of double positive patients ~30% ~15% ~55% Lenalidomide (RELEVANCE) Ibrutinib (PHOENIX) Gazyva/Polatuzumab (GOYA) REFERENCES Johnson N et al, JCO 2012 Hu S et al, Blood 2013 MAIN, GATHER data Iqbal J et al. JCO, 2006 Iqbal J et al. Clin Can Res 2011 Davis et al. Nature 2010 Haberman T et al, JCO 2006 Thieblemont C et al. JCO 2011 * 46% of ABC patients are double positive (Hu S et l. Blood, 2013) Source: Roche internal data presented during REFORCE F2F Feb-2015 7

Development vision: Raising the bar for curative therapies in NHL + + Anti CD20 Chemo Biologic modifier Pursue compelling combinations with internal and external molecules MabThera or GAZYVA Eliminate or replace with ADC Add a targeted agent CD20/CD3 TCB polatuzumab NME ADC venetoclax atezolizumab 8

Phase Ib/II studies in NHL 14 unique combinations Drugs Disease Status Doublet Anti-CD20 + venetoclax FL ongoing Doublet Anti-CD20 + atezolizumab FL/DLBCL ongoing Doublet Anti-CD20 + idasanutlin FL/DLBCL FPI 2015 Doublet Anti-CD20 + polatuzumab vedotin FL/DLBCL ongoing Doublet + Chemo Anti-CD20 + venetoclax + CHOP DLBCL ongoing Doublet + Chemo Anti-CD20+ venetoclax + bendamustine FL ongoing Doublet + Chemo Anti-CD20 + polatuzumab vedotin + CHP DLBCL ongoing Doublet + Chemo Anti-CD20 + polatuzumab vedotin + bendamustine FL ongoing Doublet + Chemo Anti-CD20 + atezolizumab + CHOP DLBCL FPI 2015 Doublet + Chemo Anti-CD20 + atezolizumab + bendamustine FL/DLBCL FPI 2015 Triplet Anti-CD20 + atezolizumab + lenalidomide FL/DLBCL FPI 2015 Triplet Anti-CD20 + atezolizumab + polatuzumab vedotin FL/DLBCL FPI 2016 Triplet Anti-CD20 + polatuzumab vedotin + venetoclax FL/DLBCL FPI 2015 Triplet Anti-CD20 + polatuzumab vedotin + lenalidomide FL/DLBCL FPI 2015 9

AML MM NHL CLL Development plan: Venetoclax A new cornerstone of the hematology franchise Compound Combination Indication P 1 P 2 P 3 venetoclax* +Rituxan MURANO R/R CLL venetoclax +Gazyva CLL14 1L CLL venetoclax Monotherapy R/R CLL 17pdel Filing Q4 15 venetoclax Monotherapy after ibrutinib/idelalisib tx R/R CLL venetoclax +Rituxan R/R CLL and SLL venetoclax +Rituxan +/- B CONTRALTO R/R FL (inhl) venetoclax +Rituxan/Gazyva+chemo CAVALLI 1L anhl venetoclax + Rituxan + B R/R NHL venetoclax +bortezomib+dexamethasone R/R MM venetoclax + hypomethylating agents + LDAC *venetoclax (Bcl2 inhibitor) in collaboration with AbbVie B = bendamustine LDAC = low dose Ara C AML 10

Relative frequency of leukemic cells MRD negative status as surrogate end-point for long-term remission and/or cure MRD (Minimal Residual Disease) detection (illustrative) MRD as prognostic factor: CLL8 study Early relapse Late relapse Peripheral blood MRD level <10-4 10-4 to <10-2 10-2 Cytomorphology detection limit Immunophenotypic and PCR detection limit Bone marrow Cure Induction Maintenance Follow-up time J Clin Oncol. 2012 Mar 20;30(9):980-8 11

CLL: Two different treatment paradigms Short course combinations that induce deep responses followed by long treatment-free remissions Chronic treatments with agents that are effective, safe, and convenient MRD-negative responses followed by long remissions Long remissions from safe, tolerable, chronic therapy 12

CLL: Two different treatment paradigms Venetoclax 1 R- Venetoclax 8 Gazyvabendamustine 7 Gazyvachlorambucil 2 R-chlorambucil 2 R-FC 3 Ibrutinib 4 Idelalisib 5 Line R/R R/R 1L 1L 1L 1L 1L R/R R/R R/R N 78 49 158 238 233 408 31 61 54 30 ORR 77% 86% 78.5% 75.5% 65.9% 90% 71% 67% 56% 90% CR 23% CR/CRi 41% CR/CRi 32.3% MRDnegative BM: 75% BM: 55% (15/20) (6/11)* 53% (ITT) PB: 58.9% BM: 27.8% (ITT) 22.2% CR/CRi PB: 31% (41/132) BM: 17% (15/88) 8.3% CR/CRi PB: 2% (3/150) BM: 3% (2/72) R-Benda- Ibrutinib 6 44% 10 % 3% 4% 10% PB: 63% (90/143) Not reported Not reported Not reported MRD: minimal residual disease; R/R: relapsed/refractory; 1L: first-line; BM: bone marrow; PB: peripheral blood * MRD tests performed in local unvalidated laboratories in a small number of patients; in patients with a CR who have been tested References: 1. John Seymour, EHA 2014 2. Goede et al. J Clin Oncol 2013; 31:suppl; abstr 7004 (presentation update) 3. Böttcher et al. J Clin Oncol 2012 ;30:980-988 4. Byrd et al. Blood (ASH Annual Meeting Abstracts) 2012 120: Abstract 189 5. Flinn et al. Hematol Oncol 2013; 31 (Suppl. 1): Abstract 297 6. Brown et al. Haematologica 2012; 97(s1) : Abstract 0543 7. Stilgenbauer et al., ASH 2015 (GREEN subgroup analysis) 8. Ma Shuo et al., ASH 2015 (abstract 80273) 13

Hemophilia A: ACE910 addressing major medical need in highly focused market FVIII market ($6.1bn in 2012) Bypassing agent market ($2.1bn) Others Hemofil M ReFacto AF/Xyntha $5.3 bn $5.5 bn Recombinate Helixate Humate P $6.0 bn $6.1 bn 6% $8.4 bn FEIBA VH NovoSeven $1.9 bn $1.7 bn $2.1 bn $2.1 bn 3% $2.6 bn 2009 2010 2011 2012 2018 2009 2010 2011 2012 2018 Current FVIII treatments Limited half-life of only 8-12 hrs Frequent IV injections Induce neutralizing antibodies, which inhibit their function Current bypassing treatments Much shorter half-life of ~4-6 hrs. Multiple frequent IV infusions Long infusion times (30+mins) for FEIBA Unstable efficacy compared to FVIII Company reported sales EvaluatePharma consensus analyst estimates 14

Development plan: ACE910 Changing the standard of care in hemophilia A 2015 2016 2017 2018 Japanese studies Chugai OLE Inhibitor Non-interventional Inhibitor ( 12 yrs) Weight based QW dosing Non-inhibitor ( 12 yrs) Weight based QW and Q2W dosing Pediatrics Inhibitor (Q4W dosing study planned) Phase I Phase II Phase III Patient transfer QW=weekly dosing; Q2W= dosing every 2 weeks; Q4W=monthly dosing; PK=pharmacokinetic study; OLE=open label extension 15

Roche data highlights at ASH 2015 Orlando, 5-8 Dec Gazyva combinations Rituxan refractory NHL: Phase 3 GADOLIN update 1L or R/R CLL: Phase 3 GREEN subgroup analysis with bendamustine 1L CLL: Phase 3 CLL14 safety run-in, combo with venetoclax venetoclax R/R CLLp17 del: Phase 2 data presentation venetoclax combinations 1L AML: Phase 1b with decitabine or azacitidine R/R NHL: Phase 1 + Rituxan + bendamustine 1L or R/R CLL: Phase 1b with Gazyva Roche curtain raiser highlighting key abstracts on new data at ASH to be released today Planned presentations 16

Roche Pharma Day 2015 Molecular Information Garret Hampton VP, Oncology Biomarker Development Pharmaceuticals Division

Personalized healthcare A cornerstone of the Genentech / Roche strategy Diagnostics Pharma Molecule 60% of pipeline programs are being developed with companion diagnostics Actemra (Systemic sclerosis) Venetoclax (R/R CLL 17p) Atezolizumab (NSCLC) ACE 910 (Hemophilia) Esbriet (IPF) Lucentis (DR) Atezolizumab (Bladder) Alectinib (ALK+ NSCLC) Gazyva (1L CLL) 4 out of 9 BTDs enabled by a Dx that identified patients most likely to benefit 2

Significant advances in cancer biology Multiple molecular subsets of disease HER2 BRAF PIK3CA AKT1 MAP2K1 ALK Fusions EGFR PD-L1 Expression Unknown KIF5B-RET ROS1 Fusions NRAS KRAS MET Amplification MET splice site 3

Molecular Information in oncology Combination of molecular and patient data will enable change in R&D and clinical practice Patients matched to clinical trials Smarter, more efficient R&D Database & Analytics interface Better patient care Treatment plan selected Molecular understanding of cancer Patient outcomes 4

Multiple capabilities required Comprehensive tumor analysis and longitudinal assessment Comprehensive tumor analysis Continuous monitoring over time DNA-mutation & CNVs mrnaexpression Proteinexpression Cell free tumor DNA Imaging Ex: EGFR, BRAF DNA sequencing Cell signatures, targets RNA sequencing PDL1, other CI targets Multiplex IHC Ex: EGFR, BRAF Blood DNA sequencing Ex: ImmunoPet Imaging 5

FMI R&D collaboration Our partnership with FMI is key to informing R&D by leveraging molecular information Comprehensive tumor analysis Continuous monitoring over time DNA-mutation & CNVs mrnaexpression Proteinexpression Cell free tumor DNA Imaging Ex: EGFR, BRAF DNA sequencing Cell signatures, targets RNA sequencing PDL1, other CI targets Multiplex IHC Ex: EGFR, BRAF Blood DNA sequencing Ex: ImmunoPet Imaging 1 2 2 3 Molecular Information from FMI database and GNE/Roche clinical trials (FM1 & FM1 heme) Immunotherapy panel development (DNA & RNA) Development of a blood-based molecular assays 6

1 Roche / FMI database queries Example: PIK3CA mutations in multiple cancers Clinical trial data 50,000+ patient data in FMI database Roche database FMI clinical database 7

1 Implementation of FMI panels in development Focus on high-value clinical samples Clinical trial data 50,000+ patient data in FMI database Roche database FMI clinical database Current prioritization*: 1. Trials with post-progression biopsy (with archival baseline samples) 2. Phase 3 trial** n>300 3. Phase 2 trial* n>100 4. Phase 1b combos ** Positive trial, evidence of activity or potential identification of subset / high value for informing our pipeline (i.e., importance of disease setting / indication / treatment) 8

FMI R&D collaboration Immunotherapy R&D collaboration Comprehensive tumor analysis Continuous monitoring over time DNA-mutation & CNVs mrnaexpression Proteinexpression Cell free tumor DNA Imaging Ex: EGFR, BRAF DNA sequencing Cell signatures, targets RNA sequencing PDL1, other CI targets Multiplex IHC Ex: EGFR, BRAF Blood DNA sequencing Ex: ImmunoPet Imaging 2 2 Immunotherapy panel development (DNA & RNA) 9

2 Checkpoint inhibitors Most effective in inflamed tumors Inflamed TILs PD-L1 expression CD8+ T cells Genomic instability Pre-existing immunity Non-inflamed 10

2 PD-L1 IHC: Staining for TCs and ICs Assay sensitivity critical in detecting both cell types Immune cells (ICs) Tumor cells (TCs) Tumor and immune cells (TCs and ICs) WCLC 2015 e.g. bladder 1 IMvigor 210 ECC 2015, 2 POPLAR ECC 2015 e.g. NSCLC 11

Overall Survival 2 Patient selection enriches for benefit PD-L1 selected lung (TC & IC) and bladder cancer (IC) Lung cancer: Survival hazard ratio* Bladder cancer: Overall survival* TC3 or IC3 TC2/3 or IC2/3 TC1/2/3 or IC1/2/3 TC0 and IC0 0.49 0.54 0.59 1.04 100 80 60 Median OS Not Reached (95% CI, 7.6-NE) ITT N=287 0.73 0.2 1 2 0.1 Hazard Ratio In favor of atezolizumab In favor of docetaxel 40 20 0 IC2/3 IC0/1 + Censored 0 Median OS 6.7 mo (95% CI, 5.7-8.0) 1 2 3 4 5 6 7 8 9 10 11 Time (months) 12 * Monotherapy data ECC 2015 12

Overall Survival Change in sum of longest diameters (SLD) from baseline (%) 2 Benefit is not the same for every patient Some patients with high expression of PD-L1 do not benefit why? 100 80 60 40 20 0 Bladder cancer: Overall survival IC2/3 IC0/1 + Censored 0 Median OS 6.7 mo (95% CI, 5.7-8.0) 1 2 3 4 Median OS Not Reached (95% CI, 7.6-NE) 5 6 7 8 9 10 11 Time (months) 12 100 80 60 40 20 0-20 -40-60 -80-100 Why do these patients progress? Why do these patients respond? 0 21 42 63 84 105 126 147 168 189 210 231 252 273 294 Time on study (days) Example: Atezolizumab phase 1 data in urothelial bladder cancer patients 13

Biomarkers for cancer immunotherapy Key platforms for discovery and development Many types of data will be needed to inform patient care including: Proteinexpression mrnaexpression DNA-mutation & CNVs Cell free tumor DNA Other Dx & patient data PDL1, other CI targets IHC Cell signatures, targets RNA sequencing Ex: EGFR, BRAF DNA sequencing Ex: EGFR, BRAF Blood DNA sequencing Ex: Imaging, outcomes etc EMRs PDL-1 IHC and multiplex IHC 1 2 Immune cell types and signatures Mutation burden & neo-epitope prediction 14

Melanoma SCC NSCLC RCC Adeno NSCLC Bladder TNBC HER2+BC CRC HR+BC 1 Gene expression & combination hypotheses Understanding the biology of immune cells in tumors enables combination hypotheses Indication cbind(1:nc) Myeloid Signature APC Signature NK cell Signature IFNg Signature IB Signature B cell Signature Teff Signature Th2 Signature Treg Signature Th17 Signature 2 1.6 0.67 0.22 1.1 1.6 2 2 1 0-1 -2 15

Melanoma SCC NSCLC RCC Adeno NSCLC Bladder TNBC HER2+BC CRC HR+BC 1 Gene expression & combination hypotheses Understanding the biology of immune cells in tumors enables combination hypotheses Myeloid signature (macrophages) associated with lack of response to atezolizumab in bladder cancer Hypothesis: Anti-CSF-1R removes macrophages which may enable atezolizumab activity Indication cbind(1:nc) Myeloid signature: IL1B, IL8, CCL2 Myeloid Signature APC Signature NK cell Signature IFNg Signature IB Signature B cell Signature Teff Signature Th2 Signature Treg Signature Th17 Signature P=0.01 Anti-CSF-1R 2 1.6 0.67 0.22 1.1 1.6 2 2 1 0-1 -2 Anti-PD-L1 PD = Progressive disease SD = Stable disease CR/PR = Complete/partial response Tumor associated macrophages 16

1 Unlocking full value of CI through combinations Broadest industry portfolio in oncology Priming & activation anti-cea-il2v FP anti-ox40 anti-cd27* (Celldex) entinostat* (Syndax) T cell infiltration anti-vegf (Avastin) anti-ang2/vegf (vanucizumab) Antigen presentation T-Vec oncolytic viruses* (Amgen) INFa anti-cd40 CMB305 vaccine* (Immune Design) Clinical development Preclinical development Established therapies * Partnered or external Chen and Mellman. Immunity 2013; CI=cancer immunotherapy Antigen release EGFRi (Tarceva) ALKi (Alectinib) BRAFi (Zelboraf) MEKi (Cotellic) anti-cd20 (Gazyva) anti-her2 (Herceptin; Kadcyla; Perjeta) various chemotherapies lenalidomide rociletinib* (Clovis) T cell killing Cancer T cell recognition anti-cea/cd3 TCB anti-cd20/cd3 TCB anti-her2/cd3 TCB ImmTAC* (Immunocore) anti-pdl1 (atezolizumab) anti-csf-1r (emactuzumab) anti-ox40 IDOi IDOi* (Incyte) CPI-444* (Corvus) anti-tigit IDO1/TDOi* (Curadev) 17

1 Gene expression as a predictive biomarker Presence of INFg-producing CD8 T-cells predicts benefit in NSCLC treated with atezolizumab INFLAMMED PD-L1 expression IFN-g producing CD8+ T cells Genomic instability Pre-existing immunity Patients with a high tumor IFNg-associated gene signature derive OS benefit from atezolizumab in NSCLC (POPLAR) Schmid et al ECC 2015 18

Biomarkers for cancer immunotherapy Key platforms for discovery and development Many types of data will be needed to inform patient care including: Proteinexpression mrnaexpression DNA-mutation & CNVs Cell free tumor DNA Other Dx & patient data PDL1, other CI targets IHC Cell signatures, targets RNA sequencing Ex: EGFR, BRAF DNA sequencing Ex: EGFR, BRAF Blood DNA sequencing Ex: Imaging, outcomes etc EMRs 1 2 Immune cell types and signatures Mutation burden & neo-epitope prediction 19

PFS / OS # of mutations 2 Mutation burden is clinically important Tumors with higher numbers of mutation tend to be more sensitive to anti-pdl1 / anti-pd1 2 High 1 Low e.g. Lung tumors Mutation burden High Low Tumor cell 1 Tumor cell 2 Time Hypothetical partly based on: Rivzi et a l Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer, Science, 2015; Snyder et al Genetic basis for clinical response to CTLA-4 blockade in melanoma, NEJM 2014 20

Conclusions Diagnostics Pharma R&D insights Transformational time Molecular Information Personalised healthcare 21

Roche Pharma Day 2015 Biosimilar market in context Fermin Ruiz de Erenchun Global Head Biologic Strategy Team Pharmaceuticals Division

Biosimilars: Ten years in the making Regulatory environment Summary 2

Biosimilars: Ten years in the making 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 EU pioneered the biosimilar concept Six products approved, including the first mab biosimilar (infliximab) Uptake did not achieve saving expectations WHO leading global efforts; many emerging countries implemented WHO biosimilar guidelines as a reference Differential adoption of WHO biosimilar guidelines led to registration of Non-Comparable Biotherapeutic products (NCBs)*, driven by: Capacity issues at National Regulatory Agencies Local economic development policies US published final biosimilars guideline FDA pathway operating, 3 applications pending, one approval Biosimilar entry timelines delayed (incl. Herceptin & MabThera) *For definition and industry position on NCBs please refer to IFPMA Policy Statement: http://www.ifpma.org/uploads/media/non-comparable_biotherapeutic_products English 01.pdfr 3

Sales in CHFm Current biosimilar trends So far, sales have not achieved initial expectations 400 350 300 250 200 150 100 50 0 MAT 870 CHFm (June 2015) (CAGR 25.5%) 80 219 254 0.6 6.3 95 267 315 57.0 119 317 MAT Jun 2013 MAT Jun 2014 MAT Jun 2015 377 Infliximab Somatropin Filgrastim Epo *Excludes US as no biosimilars have been approved in the US so far (Omnitrope was approved under the 505(b) pathway) IMS Health; MAT=moving annual total 4

Generics vs biosimilars Clear divide in uptake; complex market drivers Market share 100% 80% 60% Somatropin EPO 2 1 Driven by price and patient offering 9 innovators, one biosimilar Efficacy visible only longer term No switching 40% 20% Filgrastim 1 Payer driven: 7 biosimilars Efficacy visible immediately High turnover of patients Diovan (Novartis) Zyprexa (Eli Lilly) 0% Year 0 Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 3 3 Small molecule Virtually disappear Sources: IMS Biosimilar Dashboard, IMS & Roche analysis 1 Volume market share based on EU5 average; 2 Volume market share based on average of France & Germany EPO; 3 Data based on % remaining sales in EU 5

Biosimilar penetration of accessible market Despite 10 years of experience in the EU, uptake of biosimilars differs across countries Reference: Assessing biosimilar uptake and competition in European markets. Report by the IMS Institute for Healthcare Informatics. HGH=human growth hormone; EPO=Erythropoietin; G-CSF=Granulocyte-colony stimulating factor 6

Anti-TNF market is not a good analogue for oncology Infliximab biosimilar could expand beyond it s accessible market and obtain market share from Enbrel and Humira Remicade Biosimilar accessible market SC Anti-TNF Biological products used to treat RA, IDB & Psoriasis Other biologics Anti TNF IV Anti-TNF 7

Small molecules and biologics Not all the same Small molecule policies allow substituting patients only the price counts For biologics, European Medicine Agency (EMA) does not provide guidance on interchangeability and substitution Most countries in Europe have specific policies in place to distinguish between small molecule and biologic medicines Biologics must be prescribed by brand name Laws against substitution Switching remains a physician s choice 8

Payer environment is one of multiple drivers for biosimilar uptake Therapeutic area Innovation and change in SoC Payer environment Drivers? Physician s support 9

Biosimilars: Ten years in the making Regulatory environment Summary 10

Establishment of biosimilar guidelines has increased driven by WHO efforts 2010 2015 Biosimilar pathways in place Biosimilar pathways under development 11

Requirements and study designs are different for biosimilars vs innovator products Aspects of development Patient population Clinical design Biosimilar Sensitive and homogeneous (patients are models) Comparative versus innovator, normally equivalence Innovator product Any Superiority vs standard of care (SoC*) Study endpoints Safety Immunogenicity Sensitive, clinically validated PD markers Similar safety profile to innovator; no new findings Similar immunogenicity profile to innovator Clinical outcomes data or accepted/established surrogates (e.g. OS and PFS) Acceptable benefit/risk profile versus SoC* Acceptable risk/benefit profile versus SoC* Extrapolation Possible if justified Not allowed * In some cases SoC may not exist 12

How should extrapolation risk be managed? The regulator s perspective Clinical Non-clinical Analytical 13

How should extrapolation risk be managed? The physicians perspective Clinical Clinical I would like to see a phase III trial for each indication Non-clinical Analytical 14

What is the right patient population to establish clinical similarity to Herceptin? Topic PK PD Clinical efficacy/safety Immunogenicity Metastatic population (advanced BC) Affected by patient`s status & tumor burden Difficult to select homogeneous group Need to control and stratify for multiple factors (e.g. prior use of chemotherapy, performance status ) Population with heterogeneous characteristics affecting final clinical outcome Immune system affected by performance status and concomitant chemotherapies received Neoadjuvant/Adjuvant population (early BC) Homogeneous population can be selected Clinically validated PD marker not available Populations less likely to be confounded by baseline characteristics and external factors Immune system impaired during chemotherapy cycles, but likely to recover to normal status thereafter 15

The regulatory thinking is evolving The Herceptin case Initiation of clinical trial(s) mbc 2010 2011 2012 2013 2014 Celltrion Mylan Pfizer Samsung ebc Amgen Celltrion Pfizer mbc Phase III Start Date Regulatory Filing ebc Phase III Start Date Celltrion Q2 2010 Q1 2014 Mylan Q4 2012 Pfizer Q4 2013 Q2 2014 Samsung Q2 2014 Amgen Q4 2013 16

Biosimilars: Ten years in the making Regulatory environment Summary 17

Generics, biosimilars: Not all the same Small molecules: Policies allow fast penetration of generics Biosimilars: Countries in Europe have specific policies in place to distinguish between small molecule and biologic medicines Biologics must be prescribed by brand name, laws against automatic substitution, switching remains a physician s choice, EMA - no guidance on interchangeability After 10 years of experience in the EU, uptake of biosimilars differ heavily across countries Regulatory environment: Still evolving with authorities in the process of finally establishing frameworks; case by case decisions likely 18