15 December 2016 EMA/197343/2017 Committee for Medicinal Products for Human Use (CHMP)

Assessment report

Alecensa

International non-proprietary name: alectinib

Procedure No. EMEA/H/C/004164/0000

Note

Assessment report as adopted by the CHMP with all information of a commercially confidential nature deleted.

30 Churchill Place ● Canary Wharf ● London E14 5EU ● United Kingdom Telephone +44 (0)20 3660 6000 Facsimile +44 (0)20 3660 5520 Send a question via our website www.ema.europa.eu/contact An agency of the European Union

Table of contents

1. Background information on the procedure ...... 8 1.1. Submission of the dossier ...... 8 1.2. Steps taken for the assessment of the product ...... 9

2. Scientific discussion ...... 10 2.1. Problem statement ...... 10 2.1.1. Disease or condition ...... 10 2.1.2. Epidemiology ...... 10 2.1.3. Biologic features, Aetiology and pathogenesis ...... 11 2.1.4. Clinical presentation, diagnosis ...... 11 2.1.5. Management ...... 11 2.2. Quality aspects ...... 13 2.2.1. Introduction...... 13 2.2.2. Active Substance ...... 14 Manufacture, characterisation and process controls ...... 14 Specification ...... 15 Stability...... 16 2.2.3. Finished Medicinal Product ...... 17 Description of the product and Pharmaceutical development ...... 17 Manufacture of the product and process controls ...... 18 Product specification ...... 19 Stability of the product ...... 19 Adventitious agents ...... 20 2.2.4. Discussion on chemical, pharmaceutical and biological aspects...... 20 2.2.5. Conclusions on the chemical, pharmaceutical and biological aspects ...... 20 2.2.6. Recommendation(s) for future quality development ...... 20 2.3. Non-clinical aspects ...... 21 2.3.1. Introduction...... 21 2.3.2. Pharmacology ...... 21 2.3.3. Pharmacokinetics ...... 26 2.3.4. Toxicology ...... 29 2.3.5. Discussion on non-clinical aspects ...... 36 2.3.6. Conclusion on the non-clinical aspects ...... 40 2.4. Clinical aspects ...... 41 2.4.1. Introduction...... 41 2.4.2. Pharmacokinetics ...... 41 2.4.3. Pharmacodynamics ...... 52

Assessment report EMA/197343/2017 Page 2/122 2.4.4. Discussion on clinical pharmacology ...... 54 2.4.5. Conclusions on clinical pharmacology ...... 55 2.5. Clinical efficacy ...... 56 2.5.1. Dose response study...... 56 2.5.2. Main studies ...... 58 2.5.3. Discussion on clinical efficacy ...... 93 2.5.4. Conclusions on the clinical efficacy ...... 96 2.6. Clinical safety ...... 96 2.6.1. Discussion on clinical safety ...... 108 2.6.2. Conclusions on the clinical safety ...... 110 2.7. Risk Management Plan ...... 111 2.8. Pharmacovigilance ...... 114 2.9. New Active Substance ...... 114 2.10. Product information ...... 115 2.10.1. User consultation ...... 115 2.10.2. Additional monitoring ...... 115

3. Benefit-Risk Balance ...... 115 3.1. Therapeutic Context ...... 115 3.1.1. Disease or condition ...... 116 3.1.2. Available therapies and unmet medical need ...... 116 3.1.3. Main clinical studies ...... 116 3.2. Favourable effects ...... 116 3.3. Uncertainties and limitations about favourable effects ...... 117 3.4. Unfavourable effects ...... 117 3.5. Uncertainties and limitations about unfavourable effects ...... 118 3.6. Effects Table ...... 118 3.7. Benefit-risk assessment and discussion ...... 119 3.7.1. Importance of favourable and unfavourable effects ...... 119 3.7.2. Balance of benefits and risks ...... 120 3.7.3. Additional considerations on the benefit-risk balance ...... 120 3.8. Conclusions ...... 121

4. Recommendations ...... 121

Assessment report EMA/197343/2017 Page 3/122 List of abbreviations

AJCC American Joint Committee on Cancer ALK anaplastic lymphoma kinase ALT alanine aminotransferase AST aspartate aminotransferase AUC area under the plasma concentration-time curve AUC0-last area under the plasma concentration-time curve from time 0 to the last measurable concentration AUCinf, AUC0- area under the plasma concentration-time curve from time 0 to infinity AUCR area under the plasma concentration-time curve ratio AUCss,12h steady-state area under the plasma concentration-time curve from time 0 to 12 hours BA Bioavailability BCRP breast cancer resistance protein BE Bioequivalence BID twice daily BOR best overall response BSEP bile salt export pump

Caverage average concentration CDOR central nervous system duration of response CI confidence interval CMA Critical material attribute

Cmax maximum observed plasma concentration CNS central nervous system

CORR central nervous system objective response rate

CPK creatine phosphokinase CPP Critical process parameter CR complete response CrCL creatinine clearance CSF cerebrospinal fluid CSR clinical study report

Css, max steady-state Cmax

Css, trough steady-state Ctrough pre-dose plasma concentration at the end of a C trough dosing interval CV coefficient of variation DCR disease control rate DDI drug-drug interaction DHMA Danish Health And Medicines Authority

Assessment report EMA/197343/2017 Page 4/122 DLTs dose limiting toxicities DoE Design of experiments DOR duration of response DSC Differential Scanning Calorimetry DVS Dynamic vapour sorption ECG Electrocardiogram Eastern Cooperative Oncology Group ECOG PS performance status EGFR epidermal growth factor receptor EMA European Medicines Agency echinoderm microtubule associated protein-like EML4 4 EPAR European public assessment report ER exposure response EU European Union FaSSIF fasted state simulated intestinal fluid FDA Food and Drug Administration FeSSIF fed state simulated intestinal fluid FMEA Failure mode effects analysis GC-MS Gas chromatography mass spectrometry GCP Good Clinical Practice GI gastrointestinal GLP Good Laboratory Practice GMR geometric mean ratio hERG human Ether-à-go-go-Related Gene HGFR hepatocyte growth factor receptor HPLC High performance liquid chromatography HR heart rate HS Head Space IC Ion chromatography

IC50 50% of maximum inhibitory concentration ICH International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use IDR intrinsic dissolution rate ILD interstitial lung disease IR Infrared IRC independent review committee IV Intravenous KRAS Kirsten rat sarcoma LD Laser diffraction LDPE Low density polyethylene LFT liver function test M/P metabolite to parent molar exposure ratio M4 RO5468924

Assessment report EMA/197343/2017 Page 5/122 MDZ Midazolam MET mesenchymal epithelial transition factor MPA Swedish Medical Products Agency MRP2 multidrug resistance-associated protein 2 MTD maximum tolerated dose NDA New Drug Application NMR Nuclear Magnetic Resonance NMT Not more than NSCLC non-small cell lung cancer OAT organic anion transporter OATP organic anion transporting polypeptide OCT organic cation transporter ORR objective response rate OS overall survival PAR Proven Acceptable Range PBPK physiologically based pharmacokinetic PD Pharmacodynamic PFS progression free survival P-gp p-glycoprotein PK Pharmacokinetic PK/PD pharmacokinetic/pharmacodynamic PO oral administration popPK population pharmacokinetic PPI proton pump inhibitor PR partial response PSD Particle Size Distribution QbD Quality by design time between the start of the Q wave and end QT of the T wave QTcF QT interval corrected using Fridericia’s formula RE response evaluable RECIST response evaluation criteria in solid tumors RH Relative Humidity RON recepteur d’origine nantais ROS1 ros proto oncogene 1 RP2D recommended phase 2 dose RRT Relative retention time SAE serious adverse event SD stable disease SLS Sodium lauryl sulfate SmPC Summary of Product Characteristics

T1/2 half-life TGA Thermo-Gravimetric Analysis

Assessment report EMA/197343/2017 Page 6/122 TKI tyrosine kinase inhibitor time to maximum observed plasma T max concentration TQT thorough QT/QTc analysis TSE Transmissible Spongiform Encephalopathy ULN upper limit of normal US United States USP United States Pharmacopoeia UV Ultraviolet V/F apparent volume of distribution XRPD X-Ray Powder Diffraction XRSCD X-ray single crystal diffraction

Assessment report EMA/197343/2017 Page 7/122 1. Background information on the procedure

1.1. Submission of the dossier

The applicant Roche Registration Limited submitted on 8 September 2015 an application for marketing authorisation to the European Medicines Agency (EMA) for Alecensa, through the centralised procedure falling within the Article 3(1) and point 3 of Annex of Regulation (EC) No 726/2004. The eligibility to the centralised procedure was agreed upon by the EMA/CHMP on 26 February 2015.

The applicant applied for the following indication:

Alecensa is indicated for the treatment of adult patients with anaplastic lymphoma kinase (ALK)-positive, locally advanced or metastatic non-small cell lung cancer (NSCLC) who have progressed on or are intolerant to crizotinib.

The legal basis for this application refers to:

Article 8.3 of Directive 2001/83/EC - complete and independent application

The application submitted is composed of administrative information, complete quality data, non-clinical and clinical data based on applicants’ own tests and studies and/or bibliographic literature substituting/supporting certain test(s) or study(ies).

Information on Paediatric requirements

Pursuant to Article 7 of Regulation (EC) No 1901/2006, the application included an EMA Decision(s) CW/1/2011 on the granting of a class waiver.

Information relating to orphan market exclusivity

Similarity

Pursuant to Article 8 of Regulation (EC) No. 141/2000 and Article 3 of Commission Regulation (EC) No 847/2000, the applicant did not submit a critical report addressing the possible similarity with authorised orphan medicinal products because there is no authorised orphan medicinal product for a condition related to the proposed indication.

New active Substance status

The applicant requested the active substance Alectinib contained in the above medicinal product to be considered as a new active substance, as the applicant claims that it is not a constituent of a medicinal product previously authorised within the European Union.

Scientific Advice

The applicant received Scientific Advice from the CHMP on 20 March 2014, 25 September 2014 and 25 June 2015. The Scientific Advice pertained to insert quality, non-clinical and clinical aspects of the dossier.

Assessment report EMA/197343/2017 Page 8/122 1.2. Steps taken for the assessment of the product

The Rapporteur and Co-Rapporteur appointed by the CHMP were:

Rapporteur: Filip Josephson Co-Rapporteur: Sinan B. Sarac

• The application was received by the EMA on 8 September 2015.

• The procedure started on 1 October 2015.

• The Rapporteur's first Assessment Report was circulated to all CHMP members on 21 December 2015. The Co-Rapporteur's first Assessment Report was circulated to all CHMP members on 18 December 2015. The PRAC Rapporteur's first Assessment Report was circulated to all PRAC members on 05 January 2016.

• During the meeting on 28 January 2016, the CHMP agreed on the consolidated List of Questions to be sent to the applicant. The final consolidated List of Questions was sent to the applicant on 29 January 2016.

• The applicant submitted the responses to the CHMP consolidated List of Questions on 15 July 2016.

• The Rapporteurs circulated the Joint Assessment Report on the applicant’s responses to the List of Questions to all CHMP members on 23 August 2016.

• During the CHMP meeting on 15 September 2016, the CHMP agreed on a list of outstanding issues to be addressed in writing by the applicant on 19 September 2016.

• The applicant submitted the responses to the CHMP List of Outstanding Issues on 15 November 2016.

• The CHMP and PRAC Rapporteurs circulated the Joint Assessment Report on the applicant’s responses to the List of Questions to all CHMP and PRAC members on 30 November 2016.

• During the meeting on December 2016, the CHMP, in the light of the overall data submitted and the scientific discussion within the Committee, issued a positive opinion for granting a conditional marketing authorisation to Alecensa on 15 December 2016.

Assessment report EMA/197343/2017 Page 9/122 2. Scientific discussion

2.1. Problem statement

2.1.1. Disease or condition

Alecensa as monotherapy is indicated for the treatment of adult patients with anaplastic lymphoma kinase (ALK) positive advanced NSCLC previously treated with crizotinib.

Non-small cell lung cancer (NSCLC) is the leading cause of cancer-related mortality worldwide and represents a major health problem.

Anaplastic lymphoma kinase, a receptor tyrosine kinase, was first identified as a fusion protein resulting from chromosomal translocation in the majority of anaplastic large cell lymphomas (ALCL). When fused to other proteins, ALK becomes constitutively active, leading to increased catalytic kinase function, signal transduction activity, and oncogenic function. ALK has since been linked with many different fusion partners in different tumour types (Soda et al 2007).

The frequency of ALK gene rearrangements in patients with NSCLC (referred to as ALK positive NSCLC from here onwards) is relatively low; it is present in approximately 2 to 7% of tumours tested. However, considering the high incidence of lung cancer, this small percentage translates into about 60,000 patients annually worldwide (Kwak et al 2010, Shaw et al 2013, Weickhardt et al 2012).

ALK+ NSCLC exhibit an inversion in chromosome 2 that juxtaposes the 5' end of the echinoderm microtubule-associated protein-like 4 (EML4) gene with the 3' end of the ALK gene, resulting in the novel fusion oncogene EML4-ALK. EML4-ALK fusion, and ALK fusions with other genes (Kinesin Family member 5B [KIF5B]), KLC1 and others define a distinct clinicopathologic subset of NSCLC.

2.1.2. Epidemiology

Lung cancer has been among the most common cancers in the world for several decades. The 2012 worldwide estimates of cancer incidence and mortality by GLOBOCAN, indicate a total of 1.8 million new lung cancer cases and 1.6 million lung cancer related deaths, accounting for 13.0% of all cancer cases (except non-melanoma skin cancers) and 19.4% of all cancer deaths (except non-melanoma skin cancers). Furthermore, lung cancer incidence rates were two-fold higher in males compared to females (1,241,601 and 583,100, respectively). In 2013, the estimated number of lung cancer related deaths is 159,480 in the United States (Siegel et al 2013) and 269,610 in the European Union (Malvezzi et al 2013).

The two most prevalent sub-types of lung cancer are small cell lung cancer and non-small cell lung cancer (NSCLC). Approximately 85% of all lung cancers are NSCLC, which is frequently further subdivided into non- squamous carcinoma (including adenocarcinoma, large-cell carcinoma, and other cell types) and squamous cell (epidermoid) carcinoma.

Assessment report EMA/197343/2017 Page 10/122 2.1.3. Biologic features, Aetiology and pathogenesis

ALK gene rearrangements are largely mutually exclusive with EGFR or KRAS mutations (Gainor et al 2013), consistent with the notion that ALK gene rearrangements defines a unique molecular subset of NSCLC. In these patients, ALK gene rearrangements serve as a key and strong oncogenic driver for NSCLC and represent a critical therapeutic target susceptible to targeted ALK kinase inhibition.

2.1.4. Clinical presentation, diagnosis

Patients with ‘ALK-positive’ tumours (tumours harbouring a rearranged ALK gene/fusion protein) tend to have specific clinical features, including never or light smoking history, younger age, NSCLC with adenocarcinoma histology, and are highly sensitive to therapy with ALK inhibitors.

2.1.5. Management

Systemic chemotherapy is a cornerstone in the management of locally advanced or metastatic NSCLC. Standard first-line treatment typically consists of a platinum-based doublet (cisplatin or carboplatin in combination with other chemotherapy agents), unless a patient has a known mutation, candidate for targeted therapy (NCCN Guidelines v3.2012, Vansteenkiste et al 2013). The outcomes with this type of chemotherapy remain poor, with response rates of 15 to 35% and median progression-free survival (PFS) and overall survival (OS) of 4 to 7 months and 10 to 16 months, respectively (Scagliotti et al 2008, Ciuleanu et al 2009, Ettinger et al 2010, Paz-Ares et al 2012).

Pemetrexed/cisplatin (or carboplatin) combination therapy and carboplatin/paclitaxel with (or without) bevacizumab represent a therapeutic option in patients with advanced non-squamous NSCLC. These regimens are commonly used based on phase III randomised trials (NCCN version 1.2015)

Recently, pemetrexed-based regimens have been used in adenocarcinomas (if patients are not candidate for targeted therapy), because taxane-based regimens are associated with more toxicity (e.g. neurotoxicity).

The outcomes with second-line chemotherapy are dismal, with response rates of less than 10%, median PFS of 2 to 3 months, and median OS of 5 to 8 months (Shepherd et al 2000, Fossella et al 2000, Hanna et al 2004).

The ALK-fusion protein is an appropriate target for the treatment of patients with NSCLC harboring ALK gene rearrangements as demonstrated in nonclinical NSCLC models and in clinical trials with crizotinib and ceritinib, both of which have received marketing authorization in ALK-positive NSCLC in the EU in 2012 and 2015, respectively.

Crizotinib, an inhibitor of receptor tyrosine kinases including ALK, hepatocyte growth factor receptor (HGFR or hepatocyte growth factor receptor [c-Met]), recepteur d’origine nantais (RON) and ROS proto-oncogene 1 (ROS1), was the first approved TKI for the treatment of ALK-positive NSCLC. In the EU, crizotinib received conditional approval for patients with previously treated, ALK-positive, advanced NSCLC (Xalkori SmPC).

Assessment report EMA/197343/2017 Page 11/122 About the product

Alectinib is a highly selective and potent ALK and RET tyrosine kinase inhibitor. In preclinical studies, inhibition of ALK tyrosine kinase activity led to blockage of downstream signalling pathways including STAT 3 and PI3K/AKT and induction of tumour cell death (apoptosis).

Alectinib demonstrated in vitro and in vivo activity against mutant forms of the ALK enzyme, including mutations responsible for resistance to crizotinib. The major metabolite of alectinib (M4) has shown similar in vitro potency and activity.

Treatment with Alectinib should be initiated and supervised by a physician experienced in the use of anticancer medicinal products.

A validated ALK assay is necessary for the selection of ALK-positive NSCLC patients. ALK-positive NSCLC status should be established prior to initiation of Alectinib therapy.

The recommended dose of Alectinib is 600 mg (four 150 mg capsules) taken twice daily with food (total daily dose of 1200 mg).

Treatment with Alectinib should be continued until disease progression or unacceptable toxicity.

If a planned dose of Alectinib is missed, patients can make up that dose unless the next dose is due within 6 hours. Patients should not take two doses at the same time to make up for a missed dose. If vomiting occurs after taking a dose of Alectinib, patients should take the next dose at the scheduled time.

Management of adverse events may require dose reduction, temporary interruption, or discontinuation of treatment with Alectinib. The dose of Alectinib should be reduced in steps of 150 mg twice daily based on tolerability. Alectinib treatment should be permanently discontinued if patients are unable to tolerate the 300 mg twice daily dose.

Dose modification advice is provided in Tables 1 and 2 below.

Table 1 Dose reduction schedule

Dose reduction schedule Dose level

Starting dose 600 mg twice daily First dose reduction 450 mg twice daily Second dose reduction 300 mg twice daily

Table 2 Dose modification advice for specified Adverse Drug Reactions (see sections 4.4 and 4.8)

CTCAE grade Alectinib treatment ILD/pneumonitis of any severity grade Immediately interrupt and permanently discontinue Alectinib if no other potential causes of ILD/pneumonitis have been identified. ALT or AST elevation of Grade ≥ 3 (> 5 times ULN) Temporarily withhold until recovery to baseline or with total bilirubin  2 times ULN  Grade 1 (≤ 3 times ULN), then resume at reduced dose (see Table 1). ALT or AST elevation of Grade ≥ 2 (> 3 times ULN) Permanently discontinue Alectinib. with total bilirubin elevation > 2 times ULN in the absence of cholestasis or haemolysis Bradycardiaa Grade 2 or Grade 3 (symptomatic, may Temporarily withhold until recovery to  Grade 1 be severe and medically significant, medical (asymptomatic) bradycardia or to a heart rate of intervention indicated) ≥ 60 bpm. Evaluate concomitant medicinal products known to cause bradycardia, as well as anti- hypertensive medicinal products. If a contributing concomitant medicinal product is

Assessment report EMA/197343/2017 Page 12/122 identified and discontinued, or its dose is adjusted, resume at previous dose upon recovery to  Grade 1 (asymptomatic) bradycardia or to a heart rate of ≥ 60 bpm. If no contributing concomitant medicinal product is identified, or if contributing concomitant medicinal products are not discontinued or dose modified, resume at reduced dose (see Table 1) upon recovery to ≤ Grade 1 (asymptomatic) bradycardia or to a heart rate of ≥ 60 bpm. Bradycardiaa Grade 4 (life-threatening consequences, Permanently discontinue if no contributing concomitant urgent intervention indicated) medicinal product is identified. If a contributing concomitant medicinal product is identified and discontinued, or its dose is adjusted, resume at reduced dose (see Table 1) upon recovery to  Grade 1 (asymptomatic) bradycardia or to a heart rate of ≥ 60 bpm, with frequent monitoring as clinically indicated. Permanently discontinue in case of recurrence. CPK elevation > 5 times ULN Temporarily withhold until recovery to baseline or to ≤ 2.5 times ULN, then resume at the same dose.

CPK elevation > 10 times ULN or second occurrence Temporarily withhold until recovery to baseline or to ≤ of CPK elevation of > 5 times ULN 2.5 times ULN, then resume at reduced dose as per Table 1.

ALT = alanine aminotransferase; AST = aspartate aminotransferase; CPK = creatine phosphokinase; CTCAE = NCI Common Terminology

Criteria for Adverse Events; ILD = interstitial lung disease; ULN = upper limit of normal a Heart rate less than 60 beats per minute (bpm).

2.2. Quality aspects

2.2.1. Introduction

The finished product is presented as hard capsules containing 150 mg of alectinib hydrochloride as active substance.

Other ingredients are:

Capsule content: lactose monohydrate, hydroxypropylcellulose, sodium lauryl sulfate, magnesium stearate, carmellose calcium;

Capsule shell: hypromellose, carrageenan, potassium chloride, titanium dioxide (E171), maize starch, carnauba wax;

Printing ink: red iron oxide (E172), yellow iron oxide (E172), indigo carmine aluminum lake (E132), carnauba wax, white shellac, glyceryl monooleate.

The product is available in aluminium/aluminium perforated blisters as described in section 6.5 of the SmPC.

Assessment report EMA/197343/2017 Page 13/122 2.2.2. Active Substance

General information

The chemical name of alectinib hydrochloride (HCl) is 9-ethyl-6,6-dimethyl-8-[4-(morpholin-4-yl)piperidin-1- yl]-11-oxo-6,11-dihydro-5H-benzo[b]carbazole-3-carbonitrile hydrochloride corresponding to the molecular formula C30H35ClN4O2. It has a relative molecular mass of 519.08 g/mol and the following structure:

Figure 1. Chemical structure of alectinib hydrochloride.

The active substance is an achiral molecule. Its molecular structure has been confirmed by elemental analysis, IR, NMR (1H and 13C) and UV spectroscopy, and mass spectrometry.

Alectinib HCl is a white to yellow white powder or powder with lumps. The pKa value is 7.05 for the free base form. It is slightly hygroscopic and has low solubility in aqueous buffers across the pH range. Therefore particle size is a critical material attribute and it is controlled in the active substance specification. The higher solubility observed in fed state simulated intestinal fluid (FeSSIF) than in fasted state simulated intestinal fluid (FaSSIF), suggests a potential food effect (see clinical section).

Alectinib HCl exhibits polymorphism. Multiple crystalline solid forms were identified in screening studies. An amorphous form of alectinib HCl was also observed. The forms have been studied by differential scanning calorimetry (DSC), X-ray powder diffraction (XRPD), dynamic vapour sorption (DVS), X-ray single crystal diffraction (XRSCD) and thermogravimetric analysis (TGA). Alectinib HCl Form I has been confirmed to be the thermodynamically stable solid form under ambient conditions and was the form selected for development and commercial use. Form I is consistently produced by the proposed commercial manufacturing process and does not change upon storage. Although formation of an amorphous fraction can occur during milling of alectinib HCl, it is controlled to acceptable levels by the milling process (see manufacture section).

Manufacture, characterisation and process controls

Alectinib HCl active substance is manufactured through a six-step synthesis from four starting materials which are controlled by acceptable specifications. This is followed by a seventh process step in which the active substance is milled to achieve desired particle size distribution.

The process has been described in detail (e.g. details of order of charging, amounts of ingredients, agitation, temperatures and in-process controls). Adequate in-process controls are applied during the synthesis. The specifications and control methods for intermediate products, starting materials and reagents have been presented. The characterisation of the active substance and its impurities are in accordance with the EU guideline on chemistry of new active substances. Potential and observed impurities (including related substances, genotoxic compounds, reagents, solvents and elemental impurities) have been discussed for each step of the

Assessment report EMA/197343/2017 Page 14/122 synthesis in relation to their origin, occurrence, possible fate products and purging capacity of the manufacturing process. All introduced materials for the manufacture of alectinib as well as compounds formed and having the possibility to be formed during manufacture or during storage, have been screened for genotoxicity structures using in silico evaluation of structures (using two complementary (Q)SAR mutagenicity prediction methods (DEREK and MultiCase)), chemical reasoning, and analytical testing. These include the starting materials and their impurities (observed and potential), reagents, solvents, process intermediates, potential degradation products, and reasonably expected reaction by-products involved in the synthesis of alectinib HCl. Several compounds were identified as genotoxic or potentially genotoxic. The potential for carry-over of these impurities was evaluated. It was demonstrated that the highest carry-over level is well below the ICH M7 LTL limit (LT 1-10 years) for individual mutagenic impurities. Batch analyses for the key impurities corroborating the efficient purge of these impurities in the manufacturing process have been presented. In addition, during commercial manufacture, each key impurity is controlled with an acceptable specification. Overall, the carry-over and purging studies performed have led to an acceptable control strategy for these compounds.

Risk assessments were performed to determine the impact of material and process inputs. These were followed by an FMEA (Failure Mode Effect Analysis) exercise to determine critical process parameters (CPPs) and a final risk assessment to summarize quality attribute/process parameter functional relationships and finalize the specifications for starting materials and isolated intermediates. The specifications for all materials used during the manufacturing process have been provided and are considered satisfactory. Critical material attributes and CPPs have been identified and are included in the control strategy. Rationales supporting the commercial process and proven acceptable ranges (PARs) (with indication of targets) have been described and justified based in development results.

As mentioned above, Form I is consistently produced by the commercial manufacturing process. Formation of an amorphous fraction can occur during milling of alectinib HCl (Form I). The particle size distribution and the amorphous content in the milled alectinib HCl are controlled by the milling process. It has been demonstrated that the amorphous content is adequately controlled when the process is operated within the established PARs. The available development data, the proposed control strategy and batch analysis data from commercial scale batches fully support the proposed PARs. The amorphous content, at levels formed during the milling process, was demonstrated to have no impact on in vitro dissolution. The polymorphic form of alectinib HCl is qualitatively confirmed at release by the XRPD identity test included in the active substance specification. Stability studies also confirmed that Form I is chemically stable when stored under accelerated and long term conditions (see stability section).

The active substance is packaged in double LDPE-bag in a heat sealed aluminium bag which complies with the EC directive EC 10/2011 as amended and Ph. Eur. 3.1.4.Specification

The active substance specification includes tests for description (colour, appearance-visual), identification (IR, XRPD, IC), related substances (HPLC), assay (HPLC), hydrogen chloride (IC), water content (KF), residual solvents (HS GC-MS, IC), residue on ignition (Ph Eur), particle size distribution (LD), microbial quality (Ph Eur).

The specification has been established based on recommendations in ICH Q3A (R2), Q3C (R5), Q3D, Q6A, and M7; consideration of analytical data from development and commercial (for Japanese market) batches of alectinib HCl; and stability data for the primary batches. The acceptance criteria, as well as the rationale for

Assessment report EMA/197343/2017 Page 15/122 inclusion or exclusion of individual tests have been provided in the dossier and are considered acceptable. The levels of residual solvents are set in accordance with ICH Q3C. A fluorinated solvent used in the manufacturing process is not ICH classified. The specification limit established for this solvent was derived from toxicological studies based on an acceptable daily exposure (ADE) of 50 g using published data from the European Chemical Agency (ECHA) database.

The specification for the particle size distribution (PSD) of alectinib HCl was based on representative batches of milled drug substance, knowledge of the relationship between PSD of the drug substance and in vitro dissolution performance of the finished product, stability results for the PSD, and manufacturing considerations.

A justification for not including amorphous content in the active substance specification was provided and was accepted.

The omission of elemental impurities in the active substance specification was justified in line with ICH Q3D.The analytical methods used have been adequately described and (non-compendial methods) appropriately validated in accordance with the ICH guidelines. Satisfactory information regarding the reference standards used for assay testing has been presented.

Batch analysis data from 21 pilot and commercial scale batches of the active substance manufactured by the intended manufacturers according to the commercial synthetic route has been provided. These batches were used for development, commercial (for Japanese market), primary stability, clinical and toxicology studies. The results are within the specifications and consistent from batch to batch.

Stability

Stability data on three commercial scale batches of active substance from the proposed manufacturers stored in a container closure system representative of that intended for the market for 12 months under long term conditions at 30 ºC / 75% RH and for 6 months under accelerated conditions at 40 ºC / 75% RH according to the ICH guidelines were provided. Supportive stability data from one commercial scale batch stored for 24 months at 25°C / 60%RH and for 6 months at 40°C / 75%RH was also provided.

The following parameters were tested: description, identification (XRPD), related substances, assay, hydrogen chloride, water content, ethyl chloride, PSD and microbial quality. The batches have been analysed with the same methods as those proposed for release testing, except for XRPD and PSD testing for which older methods were used.

The long term and accelerated data provided confirms the stability of the drug substance at the studied conditions. Apart for a yellowing of the powder with time, no significant changes were seen for any parameter. All samples remained within the proposed acceptance criteria when stored at both long term and accelerated conditions.

Photostability testing following the ICH guideline Q1B was performed on one batch. Samples were tested for appearance, identification (XRPD), related substances, water content and assay. The photostability sample showed an increase in degradation products with a decrease in the assay. Yellowing was observed in the exposed sample. Other tests (XRPD and water content) did not show any significant change. Therefore it was concluded that the active substance should be protected from light by its container closure system.

Samples were also exposed to stress conditions (elevated temperature, elevated temperature and humidity, acidic and basic conditions, and oxidative conditions) Several degradation products were observed. The

Assessment report EMA/197343/2017 Page 16/122 results The results from these studies demonstrate the stability indicating nature of the release HPLC method.

The stability results justify the proposed retest period and storage conditions

2.2.3. Finished Medicinal Product

Description of the product and Pharmaceutical development

The proposed commercial formulation is an immediate release hard capsule for oral administration containing 150 mg of alectinib (as hydrochloride) and standard excipients. The capsules are size 1, white, with “ALE” printed in black ink on the cap and “150 mg” printed in black ink on the body. They are packaged in aluminium/aluminium perforated blisters. The composition of the capsules is presented in section 6.1 of the SmPC and in chapter 2.2.1 of this report.

Different formulations were used during clinical studies. Two dosage strengths, 20 mg and 40 mg, were initially developed for first-in-human study. Since a dose of 600 mg twice a day was clinically required, a higher strength capsule for Phase II and Phase III studies was developed to reduce the number of capsules needed to achieve higher doses and increase patient compliance and convenience. This resulted in a 150 mg capsule formulation which was designed to match the dissolution characteristics of the initial 40 mg capsule. Comparable release characteristics of the 40 mg capsule and the proposed 150 mg formulation was demonstrated by the in vitro dissolution studies. The proposed 150 mg formulation was used in the Phase I/II studies NP28673 (global) and NP28761 (US and Canada), and the ongoing global Phase III study BO28984.

The development of the product followed a classic development approach and adopted elements of Quality by Design (QbD) and risk-based methodology.

As described above, alectinib is a poorly water soluble free base with a measured pKa of 7.05. The mono HCl salt (Form I) was selected for development based on higher solubility in comparison to other salts and maximum absorbable dose. To enhance dissolution performance, the active substance is milled and the particle size distribution is controlled. Form I is the only solid form used in non-clinical (toxicological) and clinical studies conducted throughout development. Form I is consistently produced by the commercial drug substance manufacturing process. A small amount of amorphous content has been observed in the milled drug substance, which was shown to have a negligible impact on clinical bioavailability. The amount of amorphous content is controlled by the milling process parameters to acceptable levels.

The excipients examined during formulation development were selected based on prior use in approved products manufactured by wet granulation and the physicochemical properties of the active substance. These were evaluated regarding compatibility. The results indicated that alectinib HCl was compatible with all of the excipients tested.

Potential excipients were also screened for their suitability to increase the active substance solubility and possibly in vivo exposure for alectinib HCl. Sodium lauryl sulfate (SLS) showed significantly higher solubility enhancement compared to the other excipients and was selected as wetting agent. Based on early in vitro dissolution studies the level of SLS in the dosage form was selected for clinical formulations. This results in a total daily intake of 600 mg of SLS at the efficacious dose of alectinib (600 mg twice a day [BID]), which is higher than the amount included in medicinal products approved within the EU. The safety concerns

Assessment report EMA/197343/2017 Page 17/122 regarding the proposed high amount of SLS were discussed with the applicant as part of a scientific advice procedure (see non-clinical section). The applicant was advised to reduce the amount of SLS in the formulation proposed for marketing. In order to explore the possibility of alternate formulations with lower SLS content, the applicant conducted a bioequivalence study (NP29040) to compare the clinical formulation proposed for commercial use (alectinib hard capsules, 150 mg with 50% SLS) and three lower SLS formulations (see clinical section). In order to reduce the intake of SLS, the applicant has been recommended to introduce an alternative capsule formulation for oral administration containing a lower amount of SLS post-approval.

From the polymers investigated the binder hydroxypropylcellulose (HPC) was selected Lactose monohydrate is added as a water-soluble diluent. Carboxymethylcellulose calcium was selected as disintegrant. Magnesium stearate is used as a lubricant at a level typical in solid dosage formulations. The list of excipients included in the commercial formulation is included in section 6.1 of the SmPC and in paragraph 2.1.1 of this report. All excipients are well known pharmaceutical ingredients and their quality is compliant with Ph. Eur standards. There are no novel excipients used in the finished product formulation.

The discriminatory power of the dissolution method has been demonstrated with regard to different critical material attributes, CPPs and the effect of stressed storage conditions.

The manufacturing process involves conventional pharmaceutical operations: sieving and blending of drug substance and internal phase excipients, high-shear wet granulation, wet sieving, fluid-bed drying, dry sieving, blending of the dried granules and external phase excipients and encapsulation. Early process development was based on informal risk assessments to identify potential CPPs and potential CMAs, which served to guide subsequent experimental studies. As mentioned above, the active substance particle size distribution was found to have a strong effect on dissolution and therefore it is controlled by a milling process and a specification was established to ensure a satisfactory dissolution performance. With regards to the process parameters, the higher-shear wet granulation step is the most critical process steps that impacts product quality and manufacturability.

The primary packaging is aluminium/aluminium perforated blisters. The material complies with Ph.Eur. and EC requirements. The choice of the container closure system has been validated by stability data and is adequate for the intended use of the product.

Manufacture of the product and process controls

The finished product is manufactured using conventional pharmaceutical wet granulation and drying processes. Specifically, the manufacturing process consists of nine main steps: sieving, blending, wet granulation, wet sieving, drying, dry sieving, blending, encapsulation and packaging. The process is considered to be a standard manufacturing process. Optimisation of the manufacturing process was performed based on experience gained from risk assessments and DOE studies.

The in-process controls, frequency of testing, and acceptance criteria for the capsules are adequate for this type of manufacturing process.

Six process evaluation batches were manufactured to confirm the proposed control strategy and to demonstrate the reproducibility and robustness of the manufacturing process. These batches were manufactured at commercial production scale at the proposed commercial manufacturing site. Three registration batches were manufactured during process development prior to the manufacture of the process evaluation batches, also at full commercial scale, for stability evaluation and batch analyses. It has been

Assessment report EMA/197343/2017 Page 18/122 demonstrated that the manufacturing process is capable of producing the finished product of intended quality in a reproducible manner.

Product specification

The specifications for the drug product at release and shelf-life has been established based on ICH Q3B (R2), Q6A and M7 requirements, taking into consideration analytical data from clinical and primary stability batches at release and during stability studies, as well as the USP, Ph. Eur. and JP monographs for capsules.

The finished product release and shelf-life specifications include appropriate tests for this kind of dosage form. They include: description (visual), description of capsule content (visual), identification of alectinib hydrochloride (HPLC, IR), content per capsule (HPLC), degradation products (GC-MS, HPLC), uniformity of dosage units (Ph Eur), dissolution (HPLC), microbial limits (Ph Eur) and water content (KF). The finished product is released on the market based on the above release specifications, through traditional final product release testing.

Although the product is formulated as an immediate-release dosage form, the active substance has low solubility across the entire physiological pH range (pH 2.0–6.8). Therefore a two-point dissolution specification has been established. The dissolution specification time points and Q values were established based on statistical analysis of the dissolution data of representative finished product batches at release and stability.

The analytical methods used have been adequately described and appropriately validated in accordance with the ICH guidelines. Satisfactory information regarding the reference standard used for assay testing has been presented.

Batch analysis results are provided for 35 batches confirming the consistency of the manufacturing process and its ability to manufacture to the intended product specification. These correspond to 6 pilot scale batches and 3 primary stability batches manufactured at the proposed commercial manufacturing site; and 26 development and clinical batches manufactured at an alternative site.

Stability of the product

Stability data of three commercial scale batches of finished product stored under long term conditions for 24 months at 30°C / 75% RH and for up to 6 months under accelerated conditions at 40 ºC / 75% RH according to the ICH guidelines were provided. The stability batches are representative of those proposed for marketing and were packed in the primary packaging proposed for marketing.

Samples were tested for appearance (description of capsules and capsule content), identification of alectinib HCl (by IR), content per capsule of alectinib HCl (by HPLC), degradation products, dissolution, microbial limits and water content according to the specification presented above.

All parameters remained within specification during the stability testing at all storage conditions. No significant changes were observed.

In addition, one batch was exposed to light as per the ICH Guideline on Photostability Testing of New Drug Substances and Products. Results demonstrated that the finished product is not light sensitive.

Based on available stability data, the proposed shelf-life of 2 years with the storage instructions “store in the original package in order to protect from moisture” as stated in the SmPC (section 6.3) is acceptable.

Assessment report EMA/197343/2017 Page 19/122 Adventitious agents

Apart from lactose monohydrate, no other excipients derived from animal or human origin have been used.

It is confirmed that the lactose is produced from milk from healthy animals in the same condition as those used to collect milk for human consumption and that the lactose has been prepared without the use of ruminant material other than calf rennet according to the Note for Guidance on Minimising the Risk of Transmitting Animal Spongiform Encephalopathy Agents Via Human and veterinary medicinal products.

2.2.4. Discussion on chemical, pharmaceutical and biological aspects

Information on development, manufacture and control of the active substance and finished product has been presented in a satisfactory manner. The active substance has low aqueous solubility across the pH range. Therefore, suitable specifications for particle size distribution of the drug substance and in vitro dissolution have been established to ensure drug product performance. Alectinib HCl Form I is consistently produced by the proposed commercial manufacturing process. Although formation of an amorphous fraction can occur during milling of alectinib HCl (Form I), the level of amorphous content is suitably controlled by the milling process parameters to ensure dissolution performance.

In order to increase the solubility of alectinib HCl, the anionic surfactant SLS is used as a wetting agent in the formulation. The recommended daily dose of Alecensa is 1200 mg (4 capsules twice a day). This results in a daily intake of SLS of 600 mg, which is higher than approved medicines in the EU. Although at present the benefit/risk is positive, the applicant is recommended to introduce an alternative formulation containing a lower amount of SLS post-approval.

The results of tests carried out indicate consistency and uniformity of important product quality characteristics, and these in turn lead to the conclusion that the product should have a satisfactory and uniform performance in clinical use.

2.2.5. Conclusions on the chemical, pharmaceutical and biological aspects

The quality of this product is considered to be acceptable when used in accordance with the conditions defined in the SmPC. Physicochemical and biological aspects relevant to the uniform clinical performance of the product have been investigated and are controlled in a satisfactory way. Data has been presented to give reassurance on TSE safety.

2.2.6. Recommendation(s) for future quality development

In the context of the obligation of the MAHs to take due account of technical and scientific progress, the CHMP recommends the following points for investigation:

The applicant is recommended to introduce an alternative capsule formulation for oral administration containing a lower amount of SLS post-approval.

Assessment report EMA/197343/2017 Page 20/122 2.3. Non-clinical aspects

2.3.1. Introduction

2.3.2. Pharmacology

Primary pharmacodynamic studies

In an enzyme assay, alectinib inhibited ALK with an IC50 of 1.9 nM. In addition RET is inhibited with an IC50 of 4.8 nM. No kinase inhibitory effect was seen with a large panel of other tyrosine kinases. This is in contrast to other ALK inhibitors like crizotinib and ceritinib which show relevant activity at several unrelated kinases. (CH5424802=alectinib, LDK378=ceretinib)

CH5424802 Crizotinib LDK378 IC50 (nmol/L) 1000 100 10 1 1000 100 10 1 1000 100 10 1 ALK ALK ALK RET RET RET INSR INSR INSR KDR KDR KDR ROS1 ROS1 ROS1 ABL ABL ABL EGFR EGFR EGFR FGFR2 FGFR2 FGFR2 HER2 HER2 HER2 IGF1R IGF1R IGF1R JAK1 JAK1 JAK1 KIT KIT KIT MET MET MET PDGFRβ PDGFRβ PDGFRβ RON RON RON SRC SRC SRC Akt1 Akt1 Akt1 Aurora A Aurora A Aurora A CDK1 CDK1 CDK1 CDK2 CDK2 CDK2 MEK1 MEK1 MEK1 PKA PKA PKA PKCα PKCα PKCα PKCβ1 PKCβ1 PKCβ1 PKCβ2 PKCβ2 PKCβ2 Raf-1 Raf-1 Raf-1

Table 1: Kinase Selectivity of Alectinib compared to Crizotinib and LDK378 The antiproliferative effect of alectinib was investigated in a number of cell lines, including NSCLC, lymphoma, and neuroblastoma cell lines. These experiments included cell lines with wild-type ALK, ALK fusion proteins, ALK amplification, ALK mutation, and ALK deletion.

Table 2: Growth Inhibitory Activity of Alectinib on Human NSCLC, Lymphoma and Neuroblastoma Cell Lines with Gene Alterations of ALK

Tumor type Cell line ALK status IC50 (nM) NSCLC NCI-H2228 EML4-ALK 12 PC-9 WT 400 NCI-H1993 WT >1,000 A549 WT >1,000 Lymphoma KARPAS-299 NPM-ALK 14 SR NPM-ALK 14 U-937 ALK Negative >1,000 Neuroblastoma NB-1 Gene amplification 4.5 KELLY F1174L 62 SK-N-FI WT >10,000

Assessment report EMA/197343/2017 Page 21/122 Cells were seeded and cultured overnight, and then treated with alectinib for four or five days. Cell growth was measured using the WST-8 or CellTiter-Glo assays, and the IC50 values were calculated.

Alectinib also showed activity against ALK point mutations associated with acquired crizotinib resistance. The

IC50 for these kinases varied between 2-fold lower to 20-fold higher than for native ALK. Alectinib inhibited the growth of tumor cell lines where ALK activation is due to the formation of a fusion protein (EML4-ALK). A comparison of the in vitro anti-tumor activity between alectinib and crizotinb was performed with number of murine Ba/F3-derived cell lines expressing different EML4-ALK mutants (see the following figure and table).

A B CH5424802 Crizotinib

Ba/F3 parental cell line and Ba/F3 cell lines expressing native or mutated EML4-ALK were treated with alectinib (CH5424802) (A) or crizotinib (B) for 2 days. Cell viability was analyzed by CellTiter-Glo® Luminescent Cell Viability Assay. Data are shown as mean ± SD (n = 3). Table 3: Anti-Tumour activity of Alectinib or Crizotinib in Ba/F3 Cell Lines Expressing Native and Mutated EML4-ALK A comparison of the inhibitory activity of alectinib vs crizotinib for unmutated and mutated ALK was conducted in vitro (Study 1061439). The following table shows the IC50 values for the decrease in cell viability. In this table are also shown the ratios where the IC50 value with the mutant ALK is divided with the

IC50 value for the WT ALK – thus representing how much each mutant affects the activity of the two ALK inhibitors.

Table 4: Inhibitory activity of alectinib and crizotinb against Ba/F3 cell lines expressing native and mutated AML4-ALK Error! Objects cannot be created from editing field codes.

The in vivo anti-tumor effect of alectinib was studied in mouse xenograft models. Alectinib strongly inhibited the growth of the EML4-ALK expressing cell line NCI-H228 with tumour regression observed from 6 mg/kg/day.

Table 5: Anti-Tumour Effect of Alectinib in NCI-H2228 and A549 Xenografts

Assessment report EMA/197343/2017 Page 22/122 Cell line Dose (mg/kg/day) TGI (%) P value

NCI-H2228 0.2 34 0.4078

0.6 47 0.0947

2 84 0.0004

6 147 < 0.0001

20 168 < 0.0001

A549 2 < 1 1.0000

6 22 0.5448

20 31 0.2874 Mice bearing NCI-H2228 (A, human NSCLC cell line harboring EML4-ALK) and A549 (B, human NSCLC cell line with wild- type ALK) were administered alectinib (free base) orally once daily for 11 days at the indicated doses. TGI = Tumor growth inhibition In a further set of studies, crizotinib-resistant EML4-ALK mutants, expressed in the murine Ba/F3 cell line were implanted subcutaneously in mice, and the anti-tumor effect of alectinib and crizotinib was studied. Alectinib at 60 mg/kg inhibited the growth of many of these mutants, where crizotinib at 100 mg/kg was without effect.

G1269A G1202R 1000

1000 ) 3 800 N.S. 800 N.S. N.S. 600 600 Vehicle CH5424802 400 400 Crizotinib

200 200 Tumor volume (mm volume Tumor 0 *** 0 10 17 10 17 Days after tumor implantation Days after tumor implantation 1151Tins F1174L S1206Y

1000 1000 800 )

3 800 800 600 N.S.

600 N.S. 600 400 400 400 ** 200 200 200 *** *** Tumor volume (mm volume Tumor *** 0 0 0 10 17 7 14 10 17 Days after tumor implantation Days after tumor implantation Days after tumor implantation

Table 6: Anti-tumour activity of Alectinib against EML4-ALK mutant-driven tumours in a subcutaneous mouse model In a CNS tumour model, with NCI-H228 cells injected intracranially in mice, alectinib significantly prolonged survival compared to crizotinib.

Assessment report EMA/197343/2017 Page 23/122 100

80 Vehicle Crizotinib CH5424802 60 N=5 N=5 N=4 40 Treatment 20 Survival rate (%) rate Survival Day20-45 0 0 20 40 60 80 100 120 Days after implantation

NCI-H2228 cells were injected intracranially into mice at 4 x 105 cells/mouse. Starting 20 days after implantation, mice were treated with control vehicle (closed circle), alectinib (60 mg/kg/day, free base; open square), or crizotinib (100 mg/kg/day; open triangle) once daily from Day 20 to 45 or until death.

Table 7: Survival Rate of Mice Treated with either Crizotinib or Alectinib in an Intracerebral NCI-H2228 Implantation Model In a second model using tumour cells expressing the luciferase gene and bioluminescence as readout, Alectinib showed anti-tumor activity in the brain against these tumors, while crizotinib did not.

The main human metabolite, M4 was determined to have an IC50 of 1.2 nM against human recombinant ALK, equivalent to that of alectinib (IC50 of 1.9 nM). Moreover, M4 was also tested in enzyme assays against ALK mutations associated with crizotinib resistance, and found to have similar inhibition to the parental molecule, alectinib. The kinase selectivity of M4 was also explored in an enzyme assay against various protein kinases, and found to be very similar to alectinib. The growth inhibiting activity of M4 was also evaluated using EML4- ALK fusion-positive cell line NCI-H2228. M4 inhibited growth of the NCI-H2228 cell line with an IC50 of 37 nM, which is in the same range as that for alectinib (IC50 of 12 nM)..

Secondary pharmacodynamic studies

In vitro screening assays investigated the effects of a single concentration of alectinib (10 μM, 4826 ng/mL) on ligand binding to 109 types of receptors, ion channels and transporters, and on the activity of 42 types of enzymes. The results of these assays showed that at concentrations far exceeding the clinically relevant free exposure (by ~690 fold), alectinib had an inhibitory effect of >50% on ligand binding (peripheral benzodiazepine receptors, dopamine receptors, muscarinic receptors, neurokinin receptors, serotonin receptors, sigma [σ] receptors, glucocorticoid receptors, urotensin receptors, L-type Ca2+ channels, Cl - channels, serotonin transporters, norepinephrine transporters, dopamine transporters) and enzyme activity (CaMK2α).

Functional assays were then performed on the basis of the outcomes of these binding assays. The functions of 20 receptors expressed in cells and the synaptosomal uptake of 2 types of monoamines (norepinephrine and dopamine) in rats were evaluated at 2 concentrations (1 and 10 μM, 482.6 and 4826 ng/mL). The receptors evaluated included those with ligand binding inhibited by alectinib, namely dopamine, muscarine, neurokinin, serotonin, and urotensin receptors that could be tested with functional assays. > 50% inhibition of synaptosomal uptake of norepinephrine and dopamine occurred at a concentration of 1 μM (53% and 83%,

Assessment report EMA/197343/2017 Page 24/122 respectively). At 10 μM, % inhibition of uptake was 49% for dopamine and 43% for norepinephrine. For serotonin, >50% inhibition of serotonin receptor 5-HT2B activity was observed at 10 μM (82%). No other receptor functions were affected.

To further investigate the inhibitory action seen in these functional assays, inhibition of synaptosomal uptake of serotonin, norepinephrine, and dopamine was examined at concentrations of 10-3 to 1 μM. The IC50s of serotonin, norepinephrine, and dopamine uptake were determined to be 0.10, 0.33, and 0.26 μM (48.3, 159.3, and 125.5 ng/mL), respectively, which corresponds to 7- to 22-fold the human relevant free exposure of approximately 7 ng/mL.

Safety pharmacology programme

Cardiovascular

Alectinib blocked hERG currents with an IC20 of 0.12 μM (58 ng/mL) and an IC50 of 0.45 μM (217 ng/mL).

These concentrations are approximately 8- to 30-fold the anticipated free clinical Cmax of ~7 ng/mL at 600 mg BID. Alectinib has the potential to block the voltage-dependent Cav1.2 channel with an IC20 of 0.203 μM

(98 ng/mL) and an IC50 of 0.461 μM (222 ng/mL)

An exploratory cardiovascular safety pharmacology study was performed in monkeys. At 20 and 60 mg/kg

(group mean Cmax on first dosing day: 719 and 695 ng/mL, respectively; data taken from Day 1 of the 2- week repeat-dose monkey study), a slight hypotensive effect (approximately -10 mm Hg) was seen, but there were no effects on the ECG or heart rate. The hypotension occurring after the 20 mg/kg dose resolved approximately 22 hours after dosing, whereas after the 60 mg/kg dose, it persisted at least until the end of the recording period (24 hours), showing dose dependence mainly in duration of effect.

A GLP cardiovascular study was performed in monkeys with ascending doses of 1.7, 5, and 15 mg/kg alectinib. No effects on the cardiovascular system or body temperature were seen at any dose up to 15 mg/kg (4-hour postdose mean plasma drug concentration: 279 ng/mL). Based on these findings, the NOEL for the hypotension seen in the exploratory telemetry study was deemed to be 15 mg/kg.

To elucidate the mechanism responsible for the alectinib-induced hypotension seen in the exploratory study of telemetered monkeys, the effects of alectinib (free base) on vasoconstriction of isolated rat aorta induced by a high-concentration potassium solution were evaluated using Magnus' method. Alectinib inhibited the vasoconstriction induced by the high-concentration potassium solution with an IC20 of 0.0153 μM (7.38 ng/mL, reaching the free clinical exposure range and the free concentration in the exploratory telemetry study) and an IC50 of 0.168 μM (81.1 ng/mL). This vasodilatory effect was considered the cause of the hypotension seen in cynomolgus monkeys

CNS

To investigate the effects of alectinib on the CNS, a single dose of the vehicle (10% w/v HPCD, 0.01 M MSA, and 0.001 M HCL solution) or 3, 30, or 300 mg/kg alectinib was orally administered to male rats (N = 6 in each group), and behavior and general condition were observed using the modified Irwin test. A separate pharmacokinetic group consisting of 3 male rats was used to confirm exposure. No effects on the CNS were seen at any of the doses up to 300 mg/kg (mean Cmax: 1770 ng/mL).

Respiratory

To investigate the effects of alectinib on the respiratory system, a single dose of the vehicle (10% w/v HPCD, 0.01 M MSA, and 0.001 M HCL solution) or 3, 30, or 300 mg/kg alectinib was orally administered to conscious

Assessment report EMA/197343/2017 Page 25/122 male rats (N = 8 in each group), and respiratory rate and ventilation volume (tidal volume and minute volume) were measured (whole-body plethysmography). No effects on the respiratory system were seen at any of the doses up to 300 mg/kg (for comparison, mean Cmax at 300 mg/kg in the modified Irwin's test is 1770 ng/mL).

Gastrointestinal

The potential effect of alectinib on gastrointestinal motor function was investigated in male rats using a phenol red method. Alectinib was orally administered as a single dose to 6 animals per group at 0, 3, 30 and 300 mg/kg (vehicle: 10% w/v HPCD, 0.01 M MSA and 0.001 M HCL solution). Gastric emptying and small intestinal transit of phenol red, which was orally administrated by gavage 4 hours after test article administration, were evaluated. No test article-related effects on gastrointestinal motor function were observed.

Pharmacodynamic drug interactions

No pharmacodynamic drug interaction studies have been conducted.

2.3.3. Pharmacokinetics

The measurement of plasma concentrations of alectinib in rats, rabbit, and monkeys was validated using a liquid chromatography — tandem mass spectrometry (LC - MS/MS) method with a supernatant obtained after plasma protein precipitation.

The measurement of plasma concentrations of the metabolite M4 in monkeys was validated using LC - MS/MS after plasma protein precipitation.

Absorption

Single dose pharmacokinetic studies were performed in rats and monkeys with intravenous and oral administration. Pharmacokinetic parameters are presented in the following tables:

Table 8: Mean pharmacokinetic parameters after a single intravenous dose of alectinib to rats and cynomolgus monkeys

Species, Sex Dose CLtot Vss t1/2 (Number of animals) (mg/kg) (mL/min/kg) (L/kg) (h) SD rat, male (3) 1 7.79 ± 0.556 9.33 ± 1.04 17.8 ± 2.4 [1054082] HccHan:WIST rat, male (3) [1056248] 1 11.0 ± 2.8 13.3 ± 4.3 24.4 ± 10.9 Monkey, male and female (2) 0.5 6.04 5.28 10.4 [1054083]

Table 9: Mean pharmacokinetic parameters after a single oral dose of alectinib to rats and cynomolgus monkeys

Species, Sex Dose Cmax Tmax AUCinf t1/2 F (Number of animals) (mg/kg) (ng/mL) (h) (ng∙h/mL) (h) (%) SD rat, male [1054082] 1 60.8 ± 11.7 7.0 ± 0.0 1400 ± 492 12.6 ± 2.5 65.2 HccHan:WIST rat, male 1 60.9 ± 14.0 8.0 ± 0.0 1400 ± 320 32.1 ± 4.9 88.6 [1056248] Monkey, male and female 0.5 55.8 2.0 696 8.38 50.4 [1054083]

Distribution

Assessment report EMA/197343/2017 Page 26/122 The tissue distribution studies following a single oral dose of [14C]-alectinib to albino, pigmented and pregnant rats demonstrated that drug-related material:

1) distributed well to all tissues with highest tissue-to-plasma concentration ratios in harderian glands, adrenal glands, lungs, brown adipose tissue and liver;

2) penetrated into the CNS tissues with brain tissue-to-plasma concentration ratio of approximately 1;

3) had high accumulation in the melanin-containing tissue uveal tract of the eyes;

4) penetrated the foetus with a distribution pattern to that in maternal rats

In vitro plasma protein binding

14 The plasma protein binding ratios (fb) of [ C]-alectinib at concentrations of 100, 1000, and 10,000 ng/mL (1000 ng/mL in mice) were all high, > 99%, regardless of species or concentration. The [14C]-alectinib–HSA binding ratio at concentrations of 100, 300, and 1000 ng/mL was 97%, but the ratio of binding to α1–AAG was low, ≤ 4.9%, at all concentrations. These findings suggest that albumin is the main binding protein in human plasma. M4 was highly protein bound, like the parent molecule, with fb>99% in all species and concentrations tested.

In vitro blood cell distribution

At concentrations of 100 to 1000 ng/mL, the blood/plasma concentration ratio (Rb) values of rats (2.06), cynomolgus monkeys (3.95 - 3.52), and humans (2.92 - 2.64) were similar. At a concentration of 10,000 ng/mL, the Rb values decreased to 1.40 in rats, 1.68 in cynomolgus monkeys, and 1.29 in humans.

The blood cell distribution of M4 in rat, monkey, and human blood was evaluated in vitro at concentrations of

0.3, 1 and 2 µg/mL. The quantification of M4 was determined by a LC - MS/MS-based method. The Rb values were 2.70 - 2.45 in rats, 2.26 - 2.18 in monkeys, and 2.58 - 2.50 in humans.

Metabolism

The metabolic pathway of alectinib in humans is proposed based on the results of in vitro metabolism studies using human hepatocytes, human liver microsomes, microsomes expressing human CYP isoforms (Figure 1) and the results from the radiolabelled human ADME study.

Assessment report EMA/197343/2017 Page 27/122

Figure 1: Proposed Metabolic Pathway for Alectinib in Humans The enzymes involved in the metabolism of alectinib in human hepatocytes were evaluated by reacting [14C]- alectinib (2 µM) with a human hepatocyte suspension (prepared from cryopreserved hepatocytes) for 60 minutes at 37˚C in the presence of benzylimidazole or ketoconazole. This study suggests that CYP3A is the primary enzyme responsible formation of M4 and M6 and contributes to 40% - 50% of total hepatic metabolism.

M4 and M6 exposure levels were measured in cryopreserved plasma samples from the 13-week oral repeat- dose toxicity study of alectinib in rats from the high dose group (27 mg/kg/day) using a qualified LC -

MS/MS-based method. Exposures (Cmax and AUC0-24) on Days 1 and 91 are shown in In rats, the exposure to M4 and M6 was higher in male rats, ranging from 1.4 to 2.0 times the levels in females. There was no pronounced increase in plasma trough level from the Day 28 onward, and Cmax and AUC0-24 on Day 91 were 1.7 - 2.0 times higher (M4) and 1.1 - 1.7 times higher (M6) than on Day 1. M6 had much lower exposure than M4 (6% - 11% of M4) and is considered a minor/trace metabolite. Based on the mean AUC0-24 (35,300 and 41,100 ng∙h/mL in male and females, respectively) of alectinib on Day 91, the M4-to-parent AUC ratios were 3 - 5%.

In monkeys, there were no clear sex differences in exposure to metabolites M4 or M6. There was no pronounced increase in plasma trough level from Day 28 dose onward, and Cmax and AUC0-24 on Day 91 were 1.9 to 2.4 times higher (M4) and 2.1 to 3.7 times higher (M6) than on Day 1. M6 had much lower exposure than M4 (6% to 10% of M4) and is considered a minor/trace metabolite. Based on the mean AUC0-24 (6990 ng∙h/mL, gender combined) of alectinib on Day 91, the M4-to-parent AUC ratios were 24%.

Excretion

Assessment report EMA/197343/2017 Page 28/122 The excretion of radioactivity in urine and faeces was evaluated up to 168 hours after a single oral dose of 1 mg/kg [14C]-alectinib in male RccHan™: WIST rats.

The main route of excretion for drug-related substances was via faeces, and cumulative faecal excretion at 24, 48, 72, 96, and 168 hours postdose was 51.9%, 78.5%, 87.2%, 91.3%, and 95.7% of the dose, respectively. Excretion in urine was minimal and nearly complete within 96 hours of dosing. The total amount of radioactivity excreted in urine by 168 hours postdose was 0.5% of the dose.

2.3.4. Toxicology

The performed studies include repeat-dose toxicity, genotoxicity, preliminary embryo-foetal development toxicity, phototoxicity and toxicity evaluation of the clinical formulation in rats. In addition, the major human metabolite M4 was investigated in genotoxicity studies in vitro. The rat and cynomolgus monkey were the selected rodent and non-rodent species, respectively, for general toxicity studies based on their suitable pharmacokinetic profiles and representation of the major metabolism pathways in humans.

Table 10: Summary of main toxicity studies of alectinib. Route or Dose (mg/kg/day) GLP Type of study Species/cell line test method /concentrationa compliance Repeat-dose toxicity 4-week dose finding Rat Oral 0, 3, 10, 30 Non-GLP 4 week Rat Oral 0, 6, 20, 60 GLP 13-week Rat Oral 0, 3, 9, 27 GLP Preliminary 2-week Monkey Oral 0, 6, 20, 60 Non-GLP 4-week dose finding Monkey Oral 0, 1, 3, 10 Non-GLP 4-week Monkey Oral 0, 1.7, 5, 15 GLP 13-week Monkey Oral 0, 1.3, 4, 12 GLP Genotoxicity Bacterial reverse mutation S. typhimurium; E. coli In vitro 0, 31.3–1000 µg/plate GLP Chromosomal aberration CHO cells In vitro 0, 3–10 µg/mL GLP Micronucleus (mechanism) TK6 cells In vitro 0, 0.4-6.3 µg/mL Non-GLP Micronucleus Rat Oral 0, 6-2000 GLP Micronucleus (mechanism) Rat Oral 0, 100-1000 GLP Reproductive & developmental toxicity Embryo-fetal development DF Rat Oral 0, 3, 9, 27 GLP Embryo-fetal development DF Rabbit Oral 0, 3, 9, 27 GLP Other toxicity Screening Ames test (M4) S. typhimurium; E. coli In vitro 0, 1.58–500 µg/plate Non-GLP Micronucleus (M4) TK6 cells In vitro 0, 0.1–2.8 µg/mL Non-GLP Phototoxicity 3T3 cells In vitro 0, 0.02–50 µg/mL Non-GLP 4-week (±SLS)b Rat Oral 0, 20 GLP a Free-base alectinib was used as the test article in the photosafety study, but alectinib hydrochloride was used in all other studies. All of the doses/concentrations are shown converted to the free base. b A formulation with ± 10 mg/kg/day SLS for each group has been used.

Single dose toxicity

No dedicated single-dose studies with alectinib have been conducted. However acute toxicity was evaluated on the basis of general condition, body weight, and food consumption until a day after the second dose in the

Assessment report EMA/197343/2017 Page 29/122 micronucleus test study in rats and on the basis of general condition the day after the first dose in the preliminary 2-week toxicity study in monkeys. There were no deaths or deteriorations in general condition at ≤ 2000 mg/kg/day in rats or at ≤ 60 mg/kg/day in monkeys.

Repeat dose toxicity

The repeat-dose toxicity of alectinib after oral administration was investigated in studies up to 13 weeks in rats and monkeys. Identified target organs in both rats and monkeys were the erythroid system, GI tract, hepatobiliary system and adrenal cortex. In rats only, additional effects were observed in the respiratory tract, bone, incisor teeth, and prolongation of clotting time accompanied by haemorrhagic changes in the ileum. The changes seen in the 4- and 13-week repeat-dose toxicity studies had recovered or had shown a tendency to recover by the end of the respective 4- and 8-week recovery periods.

Morbidity and mortality

In the preliminary 2-week toxicity study conducted in monkeys, one of two animals in the high dose group was sacrificed in moribund condition on Day 13. The major cause of the moribund condition was thought to be GI tract abnormalities including retention of digestive tract contents and digestive tract dilatation.

Therefore, the dose limiting toxicity at 60 mg/kg (AUC0-24 on Day 1, gender combined: 12050 ng·h/mL) in monkeys was attributed to alectinib’s effect on the GI tract. No deaths or deteriorations in condition due to toxicity were seen in either species in the 4- and 13-week repeat-dose toxicity studies.

Effects on the erythroid system (rats and monkeys)

Abnormal red blood cell (RBC) morphology (poikilocytosis, i.e., mainly burr cells and fragmented erythrocytes) was one of the changes seen at a low exposures. In most cases, abnormal RBC morphology was seen at low dose, and then as the exposure level increased, a tendency to slight anaemia was seen in monkeys. In rats, changes suggestive of a compensatory response to anaemic stimuli (e.g., increased reticulocytes, increased extramedullary hematopoiesis in the spleen and mild increase in spleen weight) were seen. In the 13-week study in rats, in contrast to monkeys, increased extramedullary hematopoiesis in the spleen was seen at all doses, while abnormal RBC morphology was only seen at higher exposure levels. The changes in erythroid system parameters were generally mild and no exacerbation was noted with continued treatment. Similar findings were also seen in rabbits. The findings were considered reversible or showed a trend towards reversibility after the recovery period.

Effects on the gastrointestinal tract (rats and monkeys)

Effects on GI epithelia (proliferative zone extension in GI mucosa) were seen in both species. In addition, in rats, degeneration and mucinous hypertrophy of glandular stomach mucosal epithelium, mixed cell infiltration in mucosa, disarrangement/desquamation of mucosal epithelium in the small intestine, and increased serum intestinal ALP were noted. The multinucleated giant cells seen in the GI mucosa were similar to macrophages phagocytizing foreign material. The content in these cells was most likely the test article dosing formulation (suspension). In monkeys, dilatation of the large intestine was also seen, and, in the preliminary 2-week study, decreased stool and retention of GI contents were seen in both animals at 60 mg/kg/day and were accompanied by very slight erosions/ulcers in the stomach, duodenum, caecum and colon at histopathology.

Assessment report EMA/197343/2017 Page 30/122 One animal in that group was sacrificed moribund due to worsening of GI symptoms. In rats, no exacerbation of GI changes was seen after 13 weeks treatment when compared to the 4-week study.

Effects on the hepatobiliary system (rats and monkeys)

In rats, increases in liver weight were noted with vacuolation and degeneration/necrosis of bile ductal epithelium, bile duct proliferation, hepatocellular hypertrophy and necrosis, and, hepatic sinusoidal swelling/yellow-brown pigmentation. The hepatobiliary changes were associated with increases in serum hepatic ALP, -GT and total cholesterol. The changes were mostly minimal to mild; the only moderate change was focal hepatocellular necrosis. In monkeys, increases in serum hepatic ALP, direct bilirubin, inflammatory cell infiltration in Glisson’s sheath, hypertrophy of hepatocytes or Kupffer cells, and increases in liver weight were seen in the 4-week repeat-dose toxicity study. In addition, bile duct proliferation and increased serum -GT were seen in the 13-week study. The histopathological changes were minimal, and the changes in laboratory values were mild. Hepatocellular hypertrophy was also observed in the 2-week preliminary study at 20 mg/kg/day accompanied by very slight bile duct proliferation in one animal. Extension of the dosing period from 4 to 13 weeks did not lead to a marked increase in the severity of hepatobiliary findings, but findings occurred at lower exposure. The findings showed complete reversibility during the recovery period.

Effects on the adrenal gland (rats and monkeys)

In rats, findings included increased adrenal gland weight and cortical hypertrophy (zona fascicularis and reticularis) and increased large lipid droplets or decreased lipid droplets in fascicular cells. In monkeys, cortical hypertrophy of zona fascicularis/reticularis with decreased lipid droplets in fascicular cells was seen. The histopathological findings were of minimal to mild severity and were reversible. Many of these histopathological changes are similar to those usually induced by stress. Accordingly, slight increases of blood glucose, cholesterol, neutrophils and decreases in lymphocytes (rat 4-week repeat-dose study) might have been responses to adrenal cortical hypertrophy.

Effects on the respiratory system (lung and trachea, rats)

Minimal to mild foamy macrophage infiltration in alveoli was seen in the 4- and 13-week repeat-dose toxicity studies in rats. There was no clear exacerbation with longer treatment at comparable exposures, but an increased severity and incidence at higher doses compared to the control groups. Additionally, mixed cellular infiltration was seen in the tracheal mucosa (macrophages/multinucleated giant cells and/or inflammatory cells); and in some animals in the 13-week study in rats, minimal or mild disarrangement in tracheal mucosal epithelium was also seen.

Effects on bone (rats)

Minimal or mild changes in bone were seen in the high dose groups of the 4- and 13-week repeat-dose toxicity studies in rats (group mean AUCs ≥38000 ng·h/mL) and these changes were reversible. Findings included increases in activated osteoclasts and decreased trabecular bone in femur and sternum associated with mild elevations of bone ALP and inorganic phosphorus. The exposure levels in the high dose groups in the 4- and 13-week studies were approximately the same, but the longer dosing period led to an increased severity and incidence of histopathological changes.

Effects on teeth (rats)

In rats, effects on incisor teeth were seen in the high dose group (27 mg/kg/day; mean AUC0-24, gender combined: 38200 ng·h/mL) following 13-weeks treatment. The changes were reversible following the 8-week recovery period. Findings included white discoloration and shortening of incisor teeth, as well as

Assessment report EMA/197343/2017 Page 31/122 disarrangement, degeneration, and/or necrosis of ameloblasts, and dilatation of capillaries in the papillary/odontoblast layers, and disarrangement of odontoblasts. These changes were seen mainly in the middle region of the incisor, which is where enamel is produced.

Prolongation of clotting time and ileal hemorrhage (rats)

Prolongation of clotting time (PT and APTT) and ileal hemorrhage were seen at 27 mg/kg/day in the 13-week repeat-dose toxicity study in rats (group mean AUC; 38200 ng·h/mL) and were considered reversible. The ileal haemorrhage is considered likely secondary to the prolonged clotting time. Other potential consequences included blackish feces, dark-red/gelatinous content in the large intestine, and blood loss anemia (decreased RBC parameters, discoloration of eyeballs considered due to anemia).

Other findings (rats)

Mild effects on the thrombopoietic system, pituitary gland, and blood glucose were seen in repeat-dose toxicity studies in rats but are considered of limited toxicological significance.

Genotoxicity

Alectinib was not mutagenic in vitro in the bacterial reverse mutation (Ames) assay but induced a slight increase in numerical aberrations in the in vitro cytogenetic assay using Chinese Hamster Lung (CHL) cells with metabolic activation, and micronuclei in a rat bone marrow micronucleus test. The mechanism of micronucleus induction was abnormal chromosome segregation (aneugenicity), and not a clastogenic effect on chromosomes.

Carcinogenicity

No carcinogenicity studies with alectinib have been conducted.

Reproduction Toxicity

Fertility and early embryonic development studies were not conducted with alectinib.

Preliminary embryo-foetal developmental studies were performed in rats and rabbits. In rats, maternal toxicity was evident at ≥9 mg/kg/day as shown by decreased body weight gain or body weight and decreased food consumption. As for the reproductive function of dams, increases in embryonic death ratio, early foetal death ratio and total death ratio were observed resulting in a total litter loss in all dams at 27 mg/kg/day. In foetal examination, decreased foetal weight, decrease in the number of sacral and caudal vertebrae, and increases in visceral anomalies ratio, including dilation of ureter, thymic cord, small heart ventricle and thinning of ventricular wall, were observed at 9 mg/kg/day. The findings in heart were seen at a low incidence, in 2 foetuses from 1 dam. Due to the total litter loss at the highest dose, a potential dose- response relationship may be masked by the lethality and cannot be evaluated. Based on the absence of similar observations in the control and low dose groups, a relation to treatment cannot be excluded. In conclusion, the NOAEL for maternal toxicity and embryo-foetal toxicity in rats was 3 mg/kg/day (AUC0-24h 13900 ng∙h/mL on GD17).

In rabbits, maternal toxicity was observed at 27 mg/kg/day including low food consumption and body weight, and changes considered to be associated with anaemic changes, deterioration of nutritional condition, inflammation and/or stress in haematology and blood chemistry. In addition, erythrocyte morphological

Assessment report EMA/197343/2017 Page 32/122 abnormalities, in which were based mainly on poikilocyte (burr cell) were observed at ≥9 mg/kg/day. At 27 mg/kg/day, one dam aborted and two others had total litter loss. Increased post-implantation loss, reduced foetal and placental weights, reduced litter size and an increased incidence of foetuses with additional ribs were also noted in the dams given 27 mg/kg/day. The NOAEL for maternal toxicity in rabbits was 3 mg/kg/day (AUC0-24h, 2950 ng∙h/mL on GD18), and 9 mg/kg/day for embryo-foetal development.

No studies on pre-and post-natal development have been performed

Toxicokinetic data

Toxicokinetic (TK) assessment was included in all toxicology studies conducted in rats, monkeys and rabbits. In addition, the exposure of metabolites M4 and M6 was measured in cryopreserved plasma samples from the 13-week oral repeat-dose toxicity studies using a qualified LC MS/MS-based method. Plasma samples from the high dose groups only were evaluated, i.e. 27 mg/kg/day and 12 mg/kg/day, in rats and monkeys, respectively.

Table 11: Mean alectinib exposure levels after repeat dosing in the pivotal in vivo toxicity studies and in patients. a Type of study Dose Cmax AUC0-24 (mg/kg/day) (ng/mL), M/F (ng·h/mL), M/F

Humana 1200 mg/day 665 14900 Rat, 4-week 6 673/841 9340/13900 20 2190/1770 38100/33200

60 MTD 2200/1990 39800/40300

Rat, 13-week 3 370/524 5100/8450 9 1250/1520 18600/25800

27 MTD 1840/2170 35300/41100

Monkey, 4-week 1.7b 102/92.5 1160/1300 5 205/302 2400/3980

15 456/464 6530/6540

Monkey, 13-week 1.3 83.4/94.3 894/1030 4 243/189 3610/2810

12 461/463 7060/6920

Rat, embryo-fetal development 3 –/827 –/13900 (pregnant rats) 9 –/2140 –/40600 27 –/3590 –/66400

Rabbit, embryo-fetal development 3 –/163 –/2950 (pregnant rabbits) 9 –/385 –/6650 27 –/2090 –/43200

Rat, micronucleus 6c 412/– 6480/– 20c 1560/– 24200/–

60c 1770/– 28400/–

200 1850/– 36700/–

500 3030/– 55800/–

1000 3020/– 59400/–

2000 2810/– 53700/–

Rat, micronucleus 100 2100/– 37000/– (mechanism elucidation) 200 3210/– 58700/– 500 4140/– 74900/–

1000 4440/– 84200/–

–: Not applicable a Human AUCSS,24: The value used was 2 times the geometric mean AUCSS,12 from population PK analysis for Phase I/II studies NP28761 and NP28673 on patients who received repeated oral administration of the recommended clinical dose, 1200 mg/day (600 mg administered twice a day). bClose to NOAEL based on only one animal with poikilocytosis. c Cmax and AUC0-24 were extrapolated from the 4-week oral repeat-dose study in rats.

Assessment report EMA/197343/2017 Page 33/122 Local Tolerance

No dedicated local tolerance studies were performed. The local tolerance after oral administration is considered adequately assessed in performed oral repeat-dose toxicology studies. Local effects in the GI tract were seen in both rats and monkeys and are considered dose-limiting.

Other toxicity studies

Phototoxicity

Alectinib shows affinity with eyes and skin and absorbs light in the 290-700 nm range. In the 3T3 NRU assay in vitro, alectinib was positive suggesting a potential risk of phototoxicity.

4-week rat study with alectinib in the SLS-containing clinical formulation

The clinical formulation contains sodium lauryl sulfate (SLS) at a concentration of 50% (w/w SLS to alectinib). This formulation was not used in the repeat-dose toxicity studies and therefore, a dedicated 4- week rat study was performed to assess the tolerability of the clinical formulation and to explore potential toxicity effects of SLS combined with alectinib. The study showed that there were no remarkable differences in the alectinib toxicity and plasma concentration time profiles between alectinib in the clinical SLS formulation compared to alectinib in the formulation used in the toxicology studies. In the group treated with the clinical formulation (without alectinib but with SLS), no effects were noted. Based on the acceptable non- clinical and clinical tolerability, the amount of SLS in the formulation is not considered to pose a significant toxicological concern.

Metabolites

Metabolite M4 was identified as a major metabolite of alectinib in humans representing ~15% of the total plasma radioactivity in the mass balance study. M4 was not mutagenic in the screening Ames test and nor was micronucleus induction observed in the in vitro micronucleus test in TK6 cells. M4 was present in both rats and monkeys and the AUC exposure to M4 in the 13-week repeat-dose studies represents ~33 to 59% of the clinical M4 exposure.

Impurities

The Applicants proposes to apply the less-than-lifetime (LTL) approach to determine the acceptable daily intakes for mutagenic impurities as described in Note 7 in the ICH M7 guideline. Based on the daily dose of alectinib, these intakes correspond to LTL concentration limits of 8 ppm (individual mutagenic impurities) and 23 ppm (total of mutagenic impurities). Considering alectinib’s indication in advanced cancer and longest expected treatment duration based on patient life expectancy (less than 10 years), this approach is considered acceptable. The proposed specifications for the Class 1 compounds are considered acceptable and in agreement with ICH Q3C R(5) and ICH M7, respectively. For the remaining Class 2 or 3 compounds, they are all considered as likely below the LTL limit for individual mutagenic impurities of 8 ppm. The sum of these Class 2 and 3 compounds, assuming a worst-case scenario is also below the LTL limit for total mutagenic impurities of 23 ppm. Although it is unclear what toxicological data was used by the Applicant to calculate the ADI of the non-classified fluorinated solvent used during drug substance manufacture, robust data on genotoxicity with negative results are available in the ECHA database.

Assessment report EMA/197343/2017 Page 34/122 Ecotoxicity/environmental risk assessment Table 1. Summary of main study results Substance (INN/Invented Name): Alectinib CAS-number: 1256589–74–8 (Alectinib HCl), 1256580–46–7 (Alectinib free base) PBT screening Result Conclusion Bioaccumulation potential- log OECD117 3.09 (pH 5) Potential PBT (N) Kow 3.86 (pH 7) 3.85 (pH 9) PBT-assessment Parameter Result relevant Conclusion for conclusion

Bioaccumulation log Kow 3.86 not B BCF steady-state 148 (see OECD 305) not B

Persistence DT50 or ready Not ready biodegradable (see P biodegradability OECD 301F) Toxicity CMR T PBT-statement : The compound is not considered as PBT nor vPvB. However, a ready biodegradability test showed that it is not biodegradable, and hence should be considered as vP. Phase I Calculation Value Unit Conclusion

PEC surfacewater , default 6 g/L > 0.01 threshold (Y) Other concerns (e.g. chemical N/A N/A N/A class) Phase II Physical-chemical properties and fate Study type Test protocol Results Remarks

Adsorption-Desorption OECD 106 Koc = Koc > 10000 L/kg Soil 1 (pH 5.5)= 21575 L/kg triggers terrestrial Soil 2 (pH 5.0)= 18928 L/kg testing Soil 3 (pH 7.4)= 14921 L/kg Sludge 1 (pH 6.1)= 7508 L/kg Sludge 2 (pH 6.1)= 5707 L/kg Ready Biodegradability Test OECD 301F Not ready biodegradable Aerobic and Anaerobic OECD 308 Sediment 1 Anaerobic Transformation in Aquatic DT50, water = 2.3 days conditions not Sediment Systems DT50, whole system = 75 days tested Sediment 2 Shifting to sediment DT50, water = 1.5 days triggers sediment DT50, whole system = 273 days testing >10% shifting to sediment Phase IIa Effect Studies Study type Test protocol Endpoint value Unit Remarks

Algae, Growth Inhibition Test OECD 201 NOECgrowth rate 11.7 mg/L Pseudokirchneriella NOECbiomass subcapitata Daphnia sp. Reproduction Test OECD 211 NOEC 0.133 mg/L Daphnia magna

Fish, Early Life Stage Toxicity OECD 210 NOEC 1.04 µg/L Danio rerio Test

Activated Sludge, Respiration OECD 209 EC50 >1000 mg/L Inhibition Test EC10 1000 Phase IIb Studies

Bioaccumulation in fish OECD 305 BCFsteady state, 434 L/kg Danio rerio parent normalized to 5% lipids

Aerobic and Anaerobic OECD 307 DT50, soil1 4.1 days Soil 1: Kenslow Transformation in Soil DT50, soil2 14.9 loam DT50, soil3 (276) Soil 2: Lufa Speyer %CO2 in all 2.3 loam soils was Soil 3: Lufa Speyer negligable 6S clay Soil Micro-Organisms: Nitrogen OECD 216 NOEC 1000 mg/kg Transformation Test dry weight

Assessment report EMA/197343/2017 Page 35/122 Terrestrial Plants, Growth Test OECD 208 NOEC 31.6 mg/kg Allium cepa, Avena dry sativa, Brassicus weight napus, cucumis sativus, Lycopersicon esculentum, Pisum sativum Earthworm, Chronic Toxicity OECD222 NOEC 1000 mg/kg Eisenia fetida Tests dry weight

Toxicity to Terrestrial Springtails OECD232 NOECreproduction 316 mg/kg Folsomia candida (Collembola) dry weight Chronic Toxicity to Sediment mg/kg Chironomus riparius OECD218 NOEC 1060 Dwelling Organisms dry weight Alectinib is persistent (P) or very persistent (vP) in the environment but not bioaccumulative (not B) nor very bioaccumulative (not vB). It is toxic (T) due to mammalian embryotoxicity at dose levels causing maternal toxicity and potentially T due to chronic aquatic ecotoxicity. As all the three criteria must be fulfilled for a PBT substance, the conclusion is that alectinib is not a PBT substance.

2.3.5. Discussion on non-clinical aspects

Pharmacology

In an enzyme assay, alectinib inhibits ALK with an IC50 of 1.9 nM. In addition RET is inhibited with an IC50 of 4.8 nM. No kinase inhibitory effect was seen with a large panel of other tyrosine kinases. This is in contrast to other ALK inhibitors like crizotinib and ceritinib which show relevant activity at several unrelated kinases. Alectinib also showed activity against ALK point mutations associated with acquired crizotinib resistance. The

IC50 for these kinases varied between 2-fold stronger to 20-fold lower than for native ALK.

The biology of ALK and RET is discussed and it is agreed that current understanding propose a limited role of these kinases in the adult, whereas they may play critical roles during development; in particular for the neural development. For RET, there seems to be some role for neural functions also in adults. No findings related to the functions identified in knockout models have been observed in animal studies or in humans.

Alectinib inhibited the growth of tumor cell lines where ALK activation is due to the formation of a fusion protein (EML4-ALK). In one study, alectinib was compared to crizotinib for growth inhibition. In this assay alectinib was about 10-fold more potent than crozitinib in inhibiting the growth of cells expressing the native EML4-ALK enzyme.

When comparing the activity of alectinib and crizotinib with a series of Ba/F3 cell transfected with EML4-ALK as well as a number of EML4-ALK mutants there was no difference in the decrease of inhibitory activity for ALK mutants between alectinib and crizotinb. Thus, the ability of alectinib to show efficacy with crizotinib- resistant tumors is not due to a qualitative difference in the inhibition of ALK and ALK mutants. An alternative explanation is proposed: The activity at the parental Ba/F3 cell line, which is independent of ALK for its growth, represents off target activity and is likely a good surrogate for dose-limiting toxicity. For crizotinib the ratio between the IC50 for growth inhibition of Ba/F3 and the IC50 for the Ba/F3 EML-ALK (WT ALK sequence) is only 6.2 while for alectinib the corresponding figure is 48. For crizotinib, dose-limiting toxicity will limit the dose resulting in an exposure that will be sufficient to inhibit the growth of tumour cells expressing the WT ALK sequence but not most of the mutant ALK sequences. For alectinib, a higher relative exposure can be reached, sufficient to act at many of the mutants. For one of the mutants, ALK G12002R,

Assessment report EMA/197343/2017 Page 36/122 alectinib actually shows a stronger loss of activity than crizotinib (13x vs 2.6x), which is well in line with the fact that alecctinib lacks activity on tumors harboring this mutation.

These conclusions on the mechanistic background for the activity of alectinib in patients which have failed on crizotinb, while differing from the view presented by the applicant, will not have any further consequences from the non-clinical perspective.

Alectinib was active in models with intrancranially injected tumors. These data, and the substantial alectinib exposure in CNS suggest that alectinib has the potential to be effective against brain metastases.

The main human metabolite, M4 shows a similar or even higher activity to ALK in enzyme assays, and a similar selectivity. The activity in the cellular assay was somewhat lower (3-fold) which may be due to a lower cell penetration. This difference may be of importance for clinical PK-PD considerations, where substantial differences in relations between alectinib and M4 exposures have been observed (see Clinical section).

In the screening for secondary pharmacology, some secondary targets with an IC50 < 1 μM were observed. While an in vivo activity on serotonin, norepinephrine and/or dopamine uptake cannot be fully excluded based on the relatively limited multiples to clinical exposure, no CNS finding were seen in the safety pharmacology study in rats.

Alectinib has inhibitory activity on the hERG channel as well as on the Cav1.2 channel. At IC20 the multiple to the free clinical exposure is 8-fold for hERG and 14-fold for the Cav1.2 channel. In an ex-vivo aorta model, alectinib inhibited vasoconstriction with an IC20 value similar to the free clinical exposure. A mild hypotensive effect has been observed in monkeys at around clinically relevant exposures. No cardiovascular effects were observed in a GLP cardiovascular study in monkeys, but in this study the exposure did not reach clinical exposure (ca 3-fold below free clinical exposure).

Animal data suggest a risk for cardiovascular effects (hypotension). There are no preclinical findings suggesting a QT liability. However, no or limited exposure multiples to clinical exposure have been obtained in these studies and cardiovascular effects need to be carefully addressed in the clinical setting.

Pharmacokinetics

The pharmacokinetics of alectinib was adequately evaluated in the two toxicology species rats and cynomolgus monkeys.

The distribution studies show widespread distribution. Exposure in CNS suggests a potential for treatment of CNS metastases (as shown in pharmacology studies). There is retention of drug-related material in the eyes from pigmented animals, but not in skin.

Drug-related material crosses the placenta and the exposure in the foetal tissues exceeded exposure in maternal plasma. There is a potential for direct interference with the growing foetus.

The metabolism of alectinib was adequately evaluated in humans, rat and monkeys.

Toxicology

The toxicological profile of alectinib has been evaluated in agreement with recommendations in ICH S9. The rat and cynomolgus monkey were the selected rodent and non-rodent species, respectively, for general toxicity studies based on their suitable pharmacokinetic profiles and representation of the major metabolism pathways in humans.

Assessment report EMA/197343/2017 Page 37/122 The repeat-dose toxicity of alectinib after oral administration was investigated in studies up to 13 weeks in rats and monkeys. In the repeat-dose studies in rats, the AUC exposure multiples at the highest dose levels compared to the clinical AUC were 2.4- to 2.8-fold, while in monkeys, the AUC exposure were less than the clinical exposure in all studies.

Target organs in both rat and monkey at clinically relevant exposures in the repeat-dose toxicology studies included, but were not limited to the erythroid system, gastrointestinal tract, and hepatobiliary system. These effects were observed at exposures below that of the recommended clinical dose. However, most changes were minimal to mild, and changes of moderate or greater severity were only seen in rats at exposures approximately three times that of the recommended clinical dose. Furthermore, all findings were considered reversible changes that either recovered or showed a tendency to recover during the recovery period. It is also possible to monitor the functioning of these organs through clinical testing (haematology and blood chemistry tests) and clinical symptoms.

Alectinib is a lipophilic drug, with a >1 blood-to-plasma distribution, and this may have contributed to a reduced membrane integrity and life span of erythrocytes.

Abnormal erythrocyte morphology was observed at exposures equal or greater than 10-60% the human exposure by AUC at the recommended dose. Erythrocytes are considered particularly sensitive because of their limited capacity to renew their cytoskeleton due to lack of a nucleus. In the proposed SmPC section 4.8, anaemia is included as a very common adverse reaction in clinical trials with alectinib.

The effects on the gastrointestinal tract were considered dose-limiting in monkeys as one animal was sacrificed moribund due to worsening of GI symptoms. Proliferative zone extension in GI mucosa in both species was observed at exposures equal to or greater than 20-120% of the human AUC exposure at the recommended dose.

The severity was graded as minimal to mild in all affected animals with exception of one female high dose rat in the 1-month study (60 mg/kg/day) which was graded as moderate. There was no apparent progression of the finding by extending the daily dosing period from 4 to 13 weeks. A complete recovery was noted in the 1- month rat study and in the 1- and 3-month monkey studies while a tendency for reversibility was noted in the rat 3-month study. In addition to the proliferative zone extension, desquamation of surface epithelia and inflammatory cell infiltration was observed in the 1- and 3-month rat studies but not in monkeys. In the preliminary 14-day monkey study, very slight erosions ulcers were noted in the stomach, duodenum, cecum and colon in both high dose animals (60 mg/kg/day). These findings indicate that alectinib is a slight GI irritant. In addition, the non-clinical vehicle also seems to have irritative properties as proliferative zone extension of minimal degree was observed in 18 out of 20 vehicle control rats in the 3-month rat study. Therefore, it is agreed with the Applicant that the observed proliferative zone extension is considered a secondary response to irritation. In clinical trials, the GI tolerability is considered as acceptable. Adverse GI effects are reported as very common and included constipation (36%), nausea (22%), diarrhoea (18%) and vomiting (13%). Most of these events were of mild or moderate severity and declined in frequency after the first month of treatment.

Alectinib’s high liver:plasma distribution, its biliary excretion (including metabolite M4) and inhibition of BSEP

(IC50 = 0.9 µM alectinib), may contribute to hepatobiliary toxicity. Increased hepatic alkaline phosphatase (ALP) and direct bilirubin as well as vacuolation/degeneration/necrosis of bile duct epithelium and enlargement/focal necrosis of hepatocytes was observed in rats and/or monkeys at exposures equal to or greater than 20-30% of the human exposure by AUC at the recommended dose.

Assessment report EMA/197343/2017 Page 38/122 Adverse hepatobiliary reactions are listed as very common in the SmPC section 4.8 and adequate warnings are included in section 4.4. In addition, bilirubin and hepatic transaminase elevations are included as important identified risks in the RMP.

The reversible effects on the adrenal glands of minimal to mild severity are similar to those usually induced by stress. However, as quantitative whole-body autoradiography of rats revealed a high distribution of radioactivity in the adrenal gland, an effect on the adrenal gland in addition to stress cannot be fully ruled out. In rats only, additional effects were observed in the respiratory tract, bone, incisor teeth, and prolongation of clotting time accompanied by haemorrhagic changes in the ileum. The effects on the respiratory system are considered likely caused by the oral gavage administration, because similar changes are commonly seen in oral gavage studies in rats even in vehicle groups and these were not observed in monkeys. The effects were not associated with any findings seen in pneumonia or bronchitis and were reversible, so they were not considered to suggest that alectinib causes a clinically significant respiratory disorder. The effects seen in the rat long bones and incisor teeth are considered of limited relevance for adult humans. However, potential effects of alectinib on bone and tooth development in children cannot be ruled out. A prolonged of clotting time (PT and APTT) was seen at a dose higher than the dose causing hepatotoxicity. Since it is known that blood coagulation factors are produced in hepatocytes and bleeding tendency is seen in hepatic diseases, the prolongation of clotting time may be secondary to the hepatotoxicity in rats. The ileal haemorrhage is considered likely secondary to the prolonged clotting time.

Alectinib was tested in a complete package of genotoxicity studies in agreement with ICH S2(R1) guidance, and in addition, mechanistic in vitro and in vivo micronucleus tests combined with centromere analysis was conducted. The obtained data indicate that alectinib induces abnormal chromosomal segregation (aneugenicity). Such abnormal chromosome segregations are considered to have a threshold dose, below which no increase in DNA damage is expected. Accordingly, the threshold for the induction of micronuclei by alectinib using the no observed effect level (NOEL) for micronucleus induction in rats was 200 mg/kg/day

(Cmax 1850 ng/mL and AUC0-24h 36700 ng·h/mL) corresponding to 2.5-fold clinical exposure based on AUC.

No carcinogenicity studies were performed which in view of the applied indication is acceptable according to the guideline ICH S9.

Fertility and early embryonic development studies were not conducted with alectinib in agreement with ICHS9. No adverse effects on male and female reproductive organs were observed in general toxicology studies. These studies were conducted in rats and monkeys at exposures equal to or greater than 2.6- and 0.5-fold, respectively, of the human exposure, measured by AUC, at the recommended dose of 600 mg twice daily.

Alectinib caused embryo-foetal toxicity in pregnant rats and rabbits. In pregnant rats, alectinib caused total embryo-foetal loss (miscarriage) at exposures 4.5-fold of the human AUC exposure and small foetuses with retarded ossification and minor abnormalities of the organs at exposures 2.7-fold of the human AUC exposure. In pregnant rabbits, alectinib caused embryo-foetal loss, small foetuses and increased incidence of skeletal variations at exposures 2.9-fold of the human AUC exposure at the recommended dose. Embryo-foetal toxicity is included as an important potential risk in the RMP.

Alectinib absorbs UV light between 200 and 400 nm and demonstrated a phototoxic potential in an in vitro photosafety test in cultured murine fibroblasts after UVA irradiation. Photosensitivity is listed as a common

Assessment report EMA/197343/2017 Page 39/122 adverse reaction in the SmPC section 4.8 and adequate warnings are included in section 4.4. Photosensitivity is also considered an important identified risk for alectinib in the RMP.

The clinical formulation contains amounts of SLS that are than that previously approved within EU. The daily dose of SLS at the maximum recommended clinical dose of alectinib (1200 mg/day) is 600 mg/day i.e. ~10 mg/kg in a 60 kg patient. This is similar to the SLS dose tested in the 4-week rat study where no SLS-related effects in any of the examined parameters were observed. Additionally, SLS did not alter the toxicity of alectinib.

From a toxicological perspective, SLS has protein denaturing properties and may cause damage to cell membranes. However, no SLS-related histopathological effects were seen in the 4-week rat study. Data from oral administration of SLS in rats are also described in the literature. Results from 4- and 13-week rat studies showed the primary effect of SLS to be local irritation of the GI tract, with NOAELs at 100 mg/kg/day. Based on the available data, the amount of SLS in the formulation is not considered to pose a significant concern although local toxicity in the GI tract is a potential concern. As adverse GI effects have been noted after administration of alectinib in both non-clinical studies (without SLS) and in clinical studies (with SLS), the actual contribution of SLS to the clinical reactions is unclear. Therefore, a reduction in the SLS amount may be beneficial for the clinical GI tolerability (see also discussion on clinical pharmacology). The applicant is recommended to introduce an alternative capsule formulation for oral administration containing a lower amount of SLS post-approval.

Metabolite M4 was present in both rats and monkeys and the AUC exposure to M4 in the 13-week repeat- dose studies represents ~33 to 59% of the clinical M4 exposure. Although, M4 cannot be considered as fully qualified from a non-clinical perspective, further non-clinical testing is not warranted based on the intended target population with advanced cancer. Overall, the characterisation of M4 is in agreement with recommendations given by CHMP (EMEA/H/SA/2659/1/2013/III).

Based on environmental fate and toxicity data as well as on maximal epidemiology-based predicted amounts of alectinib on the European market, no significant risks to the environmental compartments sewage treatment, surface water, groundwater, sediment and soils have been identified. Overall, therefore, no significant risk has been identified for the intended medical use of alectinib in ALK+ NSCLC cancer patients in Europe.

2.3.6. Conclusion on the non-clinical aspects

The pharmacologic, pharmacokinetic, and toxicological characteristics of alectinib are well characterized.

In order to reduce the potential SLS-related effect on GI tract, the applicant is recommended to introduce an alternative capsule formulation for oral administration containing a lower amount of SLS post-approval.

Assessment report EMA/197343/2017 Page 40/122 2.4. Clinical aspects

2.4.1. Introduction

GCP

The Clinical trials were performed in accordance with GCP as claimed by the applicant.

The applicant has provided a statement to the effect that clinical trials conducted outside the community were carried out in accordance with the ethical standards of Directive 2001/20/EC.

 Tabular overview of clinical studies

Study Objective Populatio Dosing regimen n n NP29040 BE (fasted/fed) between Healthy 600 mg SD n=49 four candidate 150 mg capsules with different SLS (fasted) formulations for alectinib content n=48 (fed) NP28989 Absolute bioavailability Healthy Period 1: SD IV 50 µg and PO 600 n=6 and mass balance mg Period 2: 600 mg SD NP28990 Posaconazole DDI Healthy Cohort A: Alectinib SD 40 Cohort A mg±posaconazole 400 mg BID n=6 Cohort B: Alectinib SD 300 mg± Cohort B posaconazole 400 mg BID n=17 NP29042 Rifampin DDI Healthy Period 1: Alectinib SD 600 mg n=24 Period 2: Alectinib SD 600 mg + RIF 600-mg QD NP28991 Food-effect and Healthy FE: Alectinib SD 600 mg fasted or FE n=18 esomeprazole DDI high fat meal DDI n=24 PPI DDI: Alectinib SD 600 mg +esomeprazole 40 mg QD NP28673 Safety, tolerability, ALK- 600 mg oral BID fed n=130 efficacy, PK positive patients Midazolam DDI ALK- Midazolam SD 2 mg± alectinib: 600 n=15 substudy in study positive mg BID NP28673 patients NP28761 Safety, tolerability, ALK- Phase I: 240 mg SD fasted, 300 Phase I efficacy, PK positive (fasted), 460, 600, 760, 900 mg BID n=42 patients fed Phase II Phase II: 600mg BID fed n=85 AF-001JP Safety, tolerability, ALK- 20, 40, 80, 160 (fasted) mg oral BID n = 55-75 efficacy, PK positive 240 and 300 (fed/fasted) mg oral japanese BID patients

2.4.2. Pharmacokinetics

The pharmacokinetics (PK) of Alecensa has been assessed in subjects with NSCLC who have progressed on or are intolerant to crizotinib as well as in healthy subjects in five Phase I and two Phase I/II studies (see table above). Also, the disposition of alectinib has been evaluated in a number of in vitro studies, including metabolism studies, CYP inhibition and induction studies and studies to evaluate transport, transporter inhibition, protein binding and blood-plasma ratio.

Assessment report EMA/197343/2017 Page 41/122 An active metabolite (M4) of alectinib, with an estimated relative contribution to the activity of Alecensa of 30-50% based on target binding, has been quantified in laboratory animals and humans. The PK of M4 was investigated in most of the submitted in vivo and in vitro studies.

Pharmacokinetic data analysis

Standard statistical methods and non-compartmental methods were used to characterize the PK of alectinib and M4. The population PK of alectinib was characterised by non-linear mixed effects modelling including data from the clinical studies in patients (study NP28673 and NP28761). A physiologically based pharmacokinetic (PBPK) model was built using physiological, in vitro and in vivo data and verified based on available clinical PK and DDI data for prediction alectinib drug interactions.

Analytical methods

In all the completed clinical pharmacology studies included in the submission, plasma concentrations of alectinib and the active metabolite M4 were quantified using validated LC-MS/MS bioanalytical methods. The two applied bioanalysis methods were adequately validated and showed acceptable performance during analysis of study samples. However, there was a 21% bias observed between the methods in a cross- validation study. However, as the poor performing method was only used in study NP28761 this is not considered relevant for the PK assessment and will not affect the efficacy/safety evaluation of alectinib.

Formulations

The majority of the studies have been performed with the intended commercial formulation, 150 mg hard capsule. In addition, an intravenous solution and an oral suspension formulation containing radiolabeled alectinib and 20/40 mg hard capsules were used during drug development.

Due to the high SLS content (50%) of the formulation used during drug development, bioequivalence to the 50% SLS formulation was investigated for formulations with lower SLS content (25, 12.5 and 3%).

Absorption

The time to maximum observed plasma concentration (tmax) was reached after approximately 4 to 6 hours. The in vitro permeability of alectinib was studied in Caco-2 cells and was low at 1.810-6 cm/s. The aqueous solubility of alectinib was low across the pH range.

• Bioavailability

Absolute bioavailability was investigated as part of study NP28989 in six healthy volunteers.

Assessment report EMA/197343/2017 Page 42/122

Table 12: Mean (± SD) alectinib concentration versus time profiles following a 600 mg single oral dose of alectinib and following a 50 μg single IV dose of [14C]-alectinib in period 1

The geometric mean exposure ratio (PO/IV) for dose normalized alectinib AUC0-∞ as a measure of absolute bioavailability was 36.9% (90% CI: 33.9 to 40.3) under fed conditions in healthy subjects.

Alectinib steady-state is reached within 7 days with continuous 600 mg twice daily dosing. The accumulation ratio for the twice-daily 600 mg regimen was approximately 6-fold.

• Bioequivalence (Study NP29040)

The bioequivalence of the different alectinib formulations following oral administration in healthy subjects was investigated in a randomized, open-label, single dose, cross-over study. The bioequivalence of treatments A (reference, 50% SLS), B (25% SLS), C (12.5% SLS) and D (3% SLS) was investigated both in a fasted and fed state.

Assessment report EMA/197343/2017 Page 43/122 Table 13: Statistical estimates of effect of formulation on the PK parameters of alectinib and M4 in part 1 (Fasted)

Table 14: Statistical estimates of effect of formulation on the PK parameters of alectinib and M4 in part 2 (Fed)

Bioequivalence (based on Cmax and AUC) between formulations was also evaluated based on the pharmacokinetic parameters of the total molar sum of alectinib and M4 since both contributed to the effect of Alecensa. A similar pattern as presented above for alectinib was seen, only with formulation B (25% SLS) being bioequivalent to formulation A (50% SLS).

Assessment report EMA/197343/2017 Page 44/122 • Influence of food

The effect of a high-fat, high-calorie meal on the single oral dose PK of alectinib and M4 in healthy subjects was investigated as part of Study NP28991.

Table 15: Mean plasma concentration vs time profiles of alectinib (RO5424802) and M4 (RO5468924)

Table 16: Statistical estimates of effect of high-fat meal on the PK parameters of alectinib (RO5424802) and M4 (RO5468924)

The alectinib median Tmax was prolonged from 4 hours in the fasting state to 8 hours in the fed state. For M4 the median Tmax was prolonged from 8 hours in the fasting state to 10 hours in the fed state.

The molecular weight-adjusted M/P ratios of AUC0-∞ were similar for fasted and fed conditions.

Distribution

Human plasma protein binding and binding proteins of alectinib were evaluated in vitro by equilibrium dialysis in studies 1054086 and 1056250. The plasma protein binding ratio of alectinib at concentrations of 100-

Assessment report EMA/197343/2017 Page 45/122 10000 ng/mL were all >99%. The alectinib–HSA binding ratio was 97% at concentrations of 100-1000 ng/mL, but was only 4.9% for a1–AAG, at all concentrations.

The human plasma protein binding of metabolite M4, evaluated by equilibrium dialysis assay, was >99% at concentrations of 0.3-2 mg/mL in study 1057255.

The mean in vitro human blood-to-plasma concentration ratios of alectinib and M4 are 2.64 and 2.50, respectively, at clinically relevant concentrations.

The geometric mean volume of distribution at steady state (Vss) of alectinib following IV administration was 475 L, indicating extensive distribution into tissues.

Alectinib demonstrated penetration into the CNS. CSF samples were collected in 8 patients receiving alectinib in study NP28761. The alectinib concentration in CSF was 0.2-0.5% of the concentration in plasma, which is similar to unbound fraction in plasma based on in vitro protein binding (0.3%). Penetration into CNS of M4 has not been demonstrated.

Elimination

Alectinib is suggested to be eliminated by two main primary metabolic clearance pathways, the M4/M6 pathway by CYP3A4 metabolism (estimated to contribute 40-50% of alectinib metabolism) and the M1a/M1b pathway likely to be catalyzed by a combination of CYP isozymes (including isozymes other than CYP3A) and aldehyde dehydrogenase (ALDH) enzymes.

Results from the human mass balance study demonstrated that alectinib and M4 were the main circulating moieties in plasma with 76% of the total radioactivity in plasma. The geometric mean Metabolite/Parent ratio at steady state is 0.399.

Table 17: Proposed metabolic pathway of alectinib in human

Assessment report EMA/197343/2017 Page 46/122 Based on the final popPK analysis, the geometric mean of the individual elimination half-life estimates after single and multiple doses of alectinib is 20 hours and 32.5 hours respectively. The individual elimination half- life estimates after multiple doses of M4 is 30.7 hours.

The apparent clearance (CL/F) was 81.9 L/hour and 217 L/hour for alectinib and M4 respectively.

The molecular weight adjusted geometric mean M/P (M4/alectinib) ratios for AUC0-t and Cmax were 0.368- 0.515 and 0.324-0.494, respectively, following 600 mg BID repeated dosing of alectinib in phase II studies.

Following administration of a 600 mg single oral dose of [14C]-alectinib in study NP28989, administered orally to healthy subjects the majority of radioactivity was excreted in faeces (mean recovery 97.8%) with minimal excretion in urine (mean recovery 0.46%). In faeces, 84% of the dose was excreted as unchanged alectinib.

In faeces and urine 5.8, 0.2 and 7.7% of the alectinib dose was excreted as M4, M6 and M1a/M1b respectively. The remaining identified radioactivity in excreta was made up by alectinib.

Dose proportionality and time dependencies

Dose proportionality analysis of alectinib and M4 Cmax and AUCinf after single and multiple dose (460-900 mg) was performed as part of study NP28761.

After single dose administration, Cmax and AUCinf increased with increasing doses in the investigated dose range. A linear regression was performed and alectinib Cmax and AUCinf had a slope of 0.750 (90% CI: 0.188, 1.313) and 0.752 (90% CI: 0.060, 1.440) across the 460 to 900 mg dose range respectively.

After multiple dose administration, Cmax and AUCinf increased with increasing doses in the investigated dose range. Linear regression of alectinib Cmax and AUClast resulted in slopes of 0.908 (90% CI: 0.526, 1.290) and 1.110 (90% CI: 0.652, 1.573) across the 460 to 900 mg BID dose range.

In addition, dose proportionality was assessed using popPK modelling. The comparisons of the post hoc estimate for the PK parameters (Table 18), results confirmed that alectinib is dose proportional from 300 mg to 900 mg.

Table 18: Post hoc (CL/F and V/F) estimates of the PK parameters for alectinib by dose category The mean accumulation ratio for alectinib and M4 ranges 4.55-8.87 and 5.23-11.5 respectively across the 300-900 mg dose range without no clear trend in study NP28761. The accumulation ratios in study NP28673, based on AUC0-10 ratios day 21/day 1 was similar at 6.95 and 8.68 for alectinib and M4 respectively. The accumulation ratio is higher than than expected based on estimations using alectinib and M4 single-dose half- lifes (around 3-fold accumulation).

Assessment report EMA/197343/2017 Page 47/122 Population PK model

The objectives of the population pharmacokinetic (popPK) analysis was to develop a model to describe the PK of alectinib and M4, explore covariates that contribute to the between-patient variability in PK parameters of alectinib and M4 and derive individual estimates of PK parameters.

Only data from study NP28673 was used to characterise the PK of alectinib and M4 in this popPK model. Population PK analyses were performed using the software NONMEM version 7.2.0.

The population PK of alectinib and M4 was described by a one-compartment with first-order elimination. Body weight statistically influenced both CL/F and V/F of alectinib and M4 and was therefore included in the base structural PK model for alectinib and M4.

The geometric mean (coefficient of variation %) steady-state Cmax, Cmin and AUC0-12hr for alectinib were approximately 665 ng/mL (44.3%), 572 ng/mL (47.8%) and 7430 ng*h/mL (45.7%), respectively. The geometric mean steady-state Cmax, Cmin and AUC0-12hr for M4 were approximately 246 ng/mL (45.4%), 222 ng/mL (46.6%) and 2810 ng*h/mL (45.9%), respectively.

The NONMEM parameter estimates for the final popPK model for alectinib are summarized in Table 19.

Table 19: NONMEM parameter estimates for the final PK model for alectinib

The NONMEM parameter estimates for the final popPK model for M4 are summarized in Table 20.

Assessment report EMA/197343/2017 Page 48/122 Table 20: NONMEM parameter estimates for the final PK model for M4

An adequate predictive performance of the proposed model was demonstrated by prediction corrected visual predictive checks.

Special populations

No dedicated renal or hepatic impairment study has been submitted by the applicant.

Negligible amounts of alectinib and the active metabolite M4 are excreted unchanged in urine (< 0.2% of the dose). Based on a population pharmacokinetic analysis alectinib and M4 exposures were similar in patients with mild and moderate renal impairment and normal renal function. The pharmacokinetics of alectinib has not been studied in patients with severe renal impairment.

Table 21: PK parameters (Mean ± SD) in patients with severe, moderate and mild renal impairment and with normal creatinine clearance based on individual estimates from the popPK model for alectinib

Assessment report EMA/197343/2017 Page 49/122 Table 22: PK parameters (Mean ± SD) in patients with severe, moderate and mild renal impairment and with normal creatinine clearance based on individual estimates from the popPK model for M4

As elimination of alectinib is predominantly through metabolism in the liver, hepatic impairment may increase the plasma concentration of alectinib and/or its major metabolite M4. Based on a population pharmacokinetic analysis, alectinib and M4 exposures were similar in patients with mild hepatic impairment and normal hepatic function.

Table 23: PK Parameters in patients with hepatic impairment and with normal hepatic function (NCI Criteria) based on individual estimates from the final popPK model for alectinib

Table 24: PK Parameters in patients with hepatic impairment and with normal hepatic function (NCI Criteria) based on individual estimates from the final popPK model for M4

The pharmacokinetics of alectinib and M4 has not been studied in patients with moderate to severe hepatic impairment.

An exploratory analysis of the effect of race, white (n=6) vs. asian (n=20) populations, on alectinib PK was included in study NP28673. No statistical significant influence of race on the PK parameters was identified during covariate model building in the popPK model for either alectinib or M4. However, the clearance was lower in asian patients as compared to white patients both for alectinib (66±25.7 vs. 91.4±37.8 L/h) and M4 (202.5±72.9 vs. 226.6±111.5).

Assessment report EMA/197343/2017 Page 50/122 No effect of gender was seen in the popPK model.

Body weight was identified as a significant covariate on both CL/F and V/F for both alectinib and M4 and the alectinib and M4 exposure decreased with increasing weight (24-30% decrease in exposure for patients 90 kg as compared to patients 60-90 kg). Alectinib has not been dosed in patients > 128 kg.

No studies have been conducted to investigate the pharmacokinetics of alectinib in paediatric patients.

No effect of age was observed in the popPK model.

Age 65-74 Age 75-84 Age 85+ (Older subjects number (Older subjects number (Older subjects number /total number) /total number) /total number)

PK Trials including 26 9 0 NP29761 and NP28673

Pharmacokinetic interaction studies

The drug-drug interaction potential of alectinib was investigated in a number of in vitro and in vivo studies.

Based on in vitro data, CYP3A4 is the primary enzyme mediating the metabolism of both alectinib and its major active metabolite M4, and CYP3A contributes to 40%  50% of total hepatic metabolism. M4 has shown similar in vitro potency and activity against ALK.

In vitro, neither alectinib nor M4 are considered clinically relevant systemic direct inhibitors of CYP-enzymes. Alectinib is likely to be a time-dependent and intestinal inhibitor of CYP3A4. A signal of CYP1A2, 2B6 and 3A in vitro induction was observed.

Alectinib did not inhibit OATP1B1/OATP1B3, OAT1, OAT3 or OCT2 at clinically relevant concentrations in vitro. Based on in vitro data, alectinib is not a substrate of P-gp. Alectinib and M4 are not substrates of BCRP or organic anion-transporting polypeptide (OATP) 1B1/B3. M4 was identified as a P-gp substrate.

Co-administration of multiple oral doses of 600 mg rifampicin once daily, a strong CYP3A inducer, with a single oral dose of 600 mg alectinib reduced alectinib Cmax,and AUCinf by 51% and 73% respectively and increased M4 Cmax and AUCinf 2.20 and 1.79-fold respectively. The effect on the combined exposure of alectinib and M4 was minor, reducing Cmax and AUCinf by 4% and 18%, respectively.

Co-administration of multiple oral doses of 400 mg posaconazole twice daily, a strong CYP3A inhibitor, with a single oral dose of 300 mg alectinib increased alectinib exposure Cmax and AUCinf by 1.18 and 1.75-fold respectively, and reduced M4 Cmax and AUCinf by 71% and 25% respectively .The effect on the combined exposure of alectinib and M4 was minor, reducing Cmax by 7% and increasing AUCinf 1.36-fold.

The metabolic ratio (M/P) of M4 and alectinib is 0.3-0.5 at normal conditions. However, when Alectinib is coadministrated with the potent CYP3A4 inhibitor posaconazole and CYP3A inducer rifampin the M/P of M4 is altered to 0.22 and 3.74 respectively.

Alectinib and M4 were identified as P-gp and BCRP inhibitors in vitro.

Assessment report EMA/197343/2017 Page 51/122 As alectinib exhibits pH dependent solubility, a potential interaction with esomeprazole that change gastric pH has been evaluated. Multiple doses of esomeprazole, a proton pump inhibitor, 40 mg once daily, demonstrated no clinically relevant effect on the combined exposure of alectinib and M4.

In vitro, alectinib and its major active metabolite M4 are inhibitors of the efflux transporters P-glycoprotein (P-gp) and BCRP.

In vitro, alectinib and M4 show weak time-dependent inhibition of CYP3A4, and alectinib exhibits a weak induction potential of CYP3A4 and CYP2B6 at clinical concentrations.

Multiple doses of 600 mg alectinib had no influence on the exposure of midazolam (2 mg), a sensitive CYP3A substrate.

2.4.3. Pharmacodynamics

Mechanism of action

No mechanism of action studies have been conducted.

Primary and Secondary pharmacology

Primary pharmacology

The EML4-ALK fusion gene product is a chimeric oncoprotein with constitutive tyrosine kinase (TK) activity. A large number of ALK EML4 fusion variants have been described and there are also additional fusion partners such as TGF.

ALK positive (ALK+) tumours treated with the first licensed ALK inhibitor (ALKi) invariably develop secondary resistance that might be related to secondary mutations, fusion gene amplification and activation of alternative signalling pathways. Alectinib shows in essence full inhibitory activity in the most predictable secondary mutation (L1196M) associated with resistance to treatment with crizotinib, but less activity in case of other mutations such as G1202R. Resistance associated mutations related to alectinib treatment have also been described, in case of the resistance mutation I1171N ceritinib has been shown to be active. Of note in terms of fold change to the wild type mutation alectinib appears as sensitive to resistance mutations as crizotinib, the main difference being that higher exposure intensity is tolerated for alectinib (for details see non-clinical section).

RET (rearranged during transfection), the other TK targeted by alectinib, encodes a receptor tyrosine kinase of importance for proliferation and differentiation. Mutations in RET have been identified e.g. in NSCLC and alectinib has shown promising inhibitory potential.

Secondary pharmacology

QTc: Based on non-clinical studies the risk for QT prolongation is considered low, about 8-30 fold margin based on free Css and Cmax (please refer to the non-clinical section). A total of 221 patients who received alectinib 600 mg BID were analysed with respect ECG changes without conspicuous findings.

There was no observed exposure – QTcF relationship.

Assessment report EMA/197343/2017 Page 52/122

Heart rate: The heart rate (HR) decreased with about 10 to 13 beats per minute (see safety). This effect is considered ALK inhibitor class related. A clear exposure – HR relationship was shown. This was not seen in non-clinical studies, but might have been masked by hypotension induced increase in HR. Hypotension was not reported in clinical studies.

Across the dose range of 300 to 900 mg BID in phase studies, graphical analyses showed an exposure- response relationship for the change in tumour size over time. Lower exposure was associated with less decrease in tumour size over time. At the 600 mg dose level, the Cox proportional-hazards analysis showed that the variability in exposure was not associated with differences in PFS.

Assessment report EMA/197343/2017 Page 53/122 The explorative exposure-response analyses indicate that lower alectinib doses could potentially results in reduced efficacy, putatively related also to pattern of secondary mutations. If an increase in exposure from doses higher than 600 BID will result in higher efficacy is difficult to determine from the rather limited data available. However, an indication that the effect appears to plateau at 900 mg BID is visible in the graphical analysis of the phase I data.

2.4.4. Discussion on clinical pharmacology

M4, but not alectinib, is a substrate of P-glycoprotein (P-gp). Alectinib, but not the active metabolite M4, demonstrated penetration into the central nervous system (CNS) with CSF concentrations corresponding to the unbound concentration in plasma. Even so, Alectinib has shown to be effective also against brain metastases and therefore the uncertainty of M4 CNS distribution is considered acceptable.

The combined data on absolute bioavailability and mass balance of alectinib in study NP28989 is crucial to elucidate the elimination pathways of alectinib. Metabolism is the major elimination pathway with some possible contribution of biliary/intestinal secretion.The major drug interaction risks have been identified and handled.

The main elimination pathway of alectinib is via metabolism/excretion in the liver. Therefore, the ongoing clinical study on the effect of hepatic impairment on alectinib and M4 PK is important and planned to be submitted post-approval. Thus, this study has been included in the risk management plan as a post-approval commitment.

Population PK analysis supports dose proportionality for alectinib across the dose range of 300 to 900 mg under fed conditions.

Based on available data, no dose adjustment is required in patients with mild hepatic impairment. Alectinib has not been studied in patients with moderate to severe hepatic impairment. Therefore, Alectinib is not recommended in patients with moderate to severe hepatic impairment (see sections 4.2 and 5.2 of the SmPC).

The effect of hepatic impairment on the pharmacokinetics of alectinib will be further investigated in a multiple-center, open-label study following single oral dosing of alectinib (study NP29783) in subjects with hepatic impairment and in matched healthy subjects. The final study report will be submitted by 30 April 2018 (see RMP).

No dose adjustment is required in patients with mild or moderate renal impairment. Alectinib has not been studied in patients with severe renal impairment. However, since alectinib elimination via the kidney is negligible, no dose adjustment is required in patients with severe renal impairment (see sections 4.2 and 5.2 of the SmPC).

Age, race and gender had no clinically meaningful effect on the systemic exposure of alectinib and M4.

Exposure of alectinib and M4 decreased with increasing weight after dosing to steady-state. However, the exposure seems to reach a plateau and no dose adjustments seems warranted. Therefore, it is stated in the SmPC that although PK simulations do not indicate a low exposure in patients with extreme body weight (i.e. >130 kg), alectinib is widely distributed and clinical studies for alectinib enrolled patients within a range of body weights of 36.9-123 kg. There are no available data on patients with body weight above 130 kg.

Assessment report EMA/197343/2017 Page 54/122 The limited data on the safety and efficacy of Alectinib in patients aged 65 years and older do not suggest that a dose adjustment is required in elderly patients (see sections 4.2 and 5.2 of the SmPC). There are no available data on patients over 80 years of age.

Based on the effects on the combined exposure of alectinib and M4, no dose adjustments are required when Alectinib is co-administered with CYP3A inducers. Appropriate monitoring is recommended for patients taking concomitant strong CYP3A inducers (including, but not limited to, carbamazepine, phenobarbital, phenytoin, rifabutin, rifampicin and St. John’s Wort (Hypericum perforatum)).

Based on the effects on the combined exposure of alectinib and M4, no dose adjustments are required when Alectinib is co-administered with CYP3A inhibitors. Appropriate monitoring is recommended for patients taking concomitant strong CYP3A inhibitors (including, but not limited to, ritonavir, saquinavir, telithromycin, ketoconazole, itraconazole, voriconazole, posaconazole nefazodone, grapefruit or Seville oranges).

No dose adjustments are required when Alectinib is co-administered with proton pump inhibitors or other medicinal products which raise gastric pH (e.g. H2 receptor antagonists or antacids).

A risk for induction of CYP2B6 and PXR regulated enzymes apart from CYP3A4 cannot be completely excluded based on the submitted in vitro data. However, given the totality of data regarding CYP3A induction and TDI (e.g. midazolam study, hepatocyte induction study) the risk for either a clinically relevant PXR/CAR induction or a clinically relevant TDI of CYP3A4 is considered limited and no dose adjustment is required for co- administered CYP3A substrates. The effectiveness of concomitant administration of oral contraceptives may be reduced.

Since alectinib and its major active metabolite M4 are inhibitors of the efflux transporters P-gp and BCRP, they may have the potential to increase plasma concentrations of co-administered substrates of P-gp or BCRP. When Alectinib is co-administered with P-gp substrates (e.g., digoxin, dabigatran etexilate, topotecan, sirolimus, everolimus, nilotinib and lapatinib), or BCRP substrates with narrow therapeutic index (e.g., methotrexate, mitoxantrone, topotecan and lapatinib), appropriate monitoring is recommended (see section 4.5 of the SmPC).

Alectinib has shown reproductive toxicity in animal studies and conducting an interaction study between alectinib and oral contraceptives is not considered feasible. Therefore, the issue of appropriate contraception during and after alectinib treatment has been solved by clear recommendations in the SmPC.

As part of a scientific advice procedure, the Applicant was recommended to introduce an alternative capsule formulation for oral administration containing a lower amount of SLS post-approval.

2.4.5. Conclusions on clinical pharmacology

The pharmacokinetic characteristics of alectinib and its active metabolite M4 have been sufficiently characterized.

The CHMP considers the following measures necessary to address the issues related to pharmacology:

- In order to reduce the potential SLS-related effect on GI tract, the applicant is recommended to introduce an alternative capsule formulation for oral administration containing a lower amount of SLS post-approval.

- Alecensa has not been studied in patients with moderate to severe hepatic impairment. The applicant will conduct a study to assess the PK in subjects with hepatic impairment and submit the results by 30 April 2018.

Assessment report EMA/197343/2017 Page 55/122 2.5. Clinical efficacy

The first in human Phase I/II study, AF-001JP, in ALK-positive crizotinib-naïve NSCLC patients was conducted by Chugai. Based on data from this study (ORR about 90+%) alectinib was granted marketing access in Japan on 4 July 2014 with the indication: ALK-fusion gene-positive, unresectable, recurrent / advanced NSCLC. The dose was 300 mg BID. Further dose escalation was stopped due to regulatory concerns related to the high dose of the excipient SLS (see quality and non-clinical).

In July 2014, Roche initiated a Phase III global study to evaluate alectinib versus crizotinib in treatment- naïve patients with advanced ALK-positive NSCLC (BO28984, ALEX study). This trial is currently ongoing and no data are available. A separate Phase III Chugai-sponsored study (JO28928) with a design similar to the BO28984 (ALEX) study in patients who were ALK-inhibitor naïve and may have received up to one prior chemotherapy regimen for ALK–positive NSCLC is ongoing in Japan.

This submission is based on two ongoing Phase I/II studies evaluating alectinib in ALK-positive NSCLC patients who have progressed on crizotinib treatment (studies NP28673 and NP28761).

2.5.1. Dose response study

Study NP28761 – Phase I

A total of 47 patients with ALK positive NSCLC after failure on crizotinib, were enrolled and treated in cohorts in a staggered manner at an initial dose level of 600 mg/day (300 mg BID). Patients in Cohort 1 received alectinib under fasting conditions whereas all other cohorts received alectinib under fed conditions.

Figure 2: Phase I Study Design :

Assessment report EMA/197343/2017 Page 56/122

Figure 3: Phase I Dosing Schedule :

Figure 4: Phase I Dose Escalation Process:

Table 25: Phase I Cohort Dose Levels

DLTs

DLT was defined as:

1. Grade 4 thrombocytopenia

2. Grade 3 thrombocytopenia with bleeding

3. Grade 4 neutropenia continuing for ≥ 7 consecutive days

4. Non-hematological toxicity of Grade 3 or higher

5. Adverse events that required suspension of treatment for a total of ≥ 7 days

Assessment report EMA/197343/2017 Page 57/122 Two DLTs were reported in the alectinib 900 mg BID PK bridging cohort: AE of Grade 3 headache in one patient (progressive but asymptomatic brain metastases) and an AE of Grade 3 neutrophil count decreased in one patient requiring study drug interruption for a week.

2.5.2. Main studies

Due to the similarity in study designs and patient eligibility criteria in the two pivotal studies NP28761 and NP28673, data will be presented in parallel.

Study NP28761: A Phase I/II Study of the ALK Inhibitor alectinib (CH5424802/RO5424802) in patients with ALK-rearranged Non-Small Cell Lung Cancer previously treated with Crizotinib.

Study NP28761 is an on-going multi-centre, single arm, open-label, dose escalation study to determine the safety, tolerability and activity of alectinib as a single agent in patients with either locally advanced (AJCC Stage IIIB) and not amenable to curative therapy or metastatic (AJCC Stage IV) ALK-positive NSCLC who had experienced disease progression on crizotinib with or without prior chemotherapy.

Radiologic assessments in study NP28761 were performed every six weeks from Cycle 1 Day 1 at Cycles 2, 4, and 6 between Days 14 to 21 and every 9 weeks thereafter.

Methods

• Study participants

Inclusion criteria

1. Provided informed consent

2. ≥ 18 years

3. ECOG PS ≤ 2

4. Patients with locally advanced (AJCC Stage IIIB) not amenable to curative therapy or metastatic (AJCC Stage IV) NSCLC

5. Histologically-confirmed NSCLC

6. ALK-rearrangement confirmed by a FDA-approved test

7. Prior treatment with crizotinib. A washout period of at least 1 week was required from last crizotinib dose and dosing with alectinib

7. Measurable disease defined by RECIST 1.1

8. Adequate hematologic function

9. Adequate hepatic function

10. Adequate renal function

11. For women who were not postmenopausal (12 months of amenorrhea) or surgically sterile (absence of ovaries and/or uterus): agreement to use two adequate methods of contraception, including at least one method with a failure rate of < 1% per year (e.g., hormonal implants, combined oral contraceptives,

Assessment report EMA/197343/2017 Page 58/122 vasectomized partner), during the treatment period and for at least 3 months after the last dose of study drug

12. For men: agreement to use a barrier method of contraception during the treatment period and for at least 80 days after the last dose of study drug

13. Women of childbearing potential had to have a negative serum or urine pregnancy test result within 7 days of starting study drug

14. Willingness and ability to comply with scheduled visits, treatment plans, laboratory tests, and other study procedures

Exclusion Criteria

1. History of hypersensitivity to any of the additives in the alectinib formulation (lactose monohydrate, microcrystalline cellulose, sodium starch glycolate, hydroxypropyl cellulose, sodium lauryl sulfate (SLS), magnesium stearate)

2. Cytotoxic chemotherapy within 4 weeks or 5 half-lives of previous therapy prior to starting the study drug, whichever was shorter

3. Radiotherapy within 14 days prior to starting the study drug

4. Prior therapy with an ALK inhibitor other than crizotinib.

5. Not recovered from adverse events (AEs) or hematologic and/or biochemical toxicities due to previous treatments to a Grade 1 or less specified in NCI − CTCAE v4.0 except for alopecia

6. Brain or leptomeningeal metastases that were symptomatic and/or requiring treatment. Patients were allowed on study only if the following criteria were met:

● Patients who had previously been treated with whole-brain radiotherapy (WBRT) or gamma-knife radiosurgery had to have completed treatment and discontinued the use of corticosteroids for this indication for ≥ 2 weeks and any signs and/or symptoms of brain metastases had to be stable for at least 2 weeks prior to the first dose of alectinib

● Patients who had not previously been treated with WBRT or gamma knife radiosurgery had to be asymptomatic without neurological signs and clinically stable for ≥ 2 weeks without steroid treatment for CNS metastases prior to the first dose of alectinib [Patients who had not previously been treated with WBRT or gamma knife radiosurgery and who were not asymptomatic were not eligible for the study. However, they could be re-screened after completing WBRT or gamma-knife treatment. They had to have discontinued the use of corticosteroids for this indication for ≥ 2 weeks and any symptoms had to be stable without deterioration for at least 2 weeks prior to the first dose of study drug]

7. History of myocardial infarction or stroke within 6 months, congestive heart failure greater than the New York Heart Association (NYHA) class II, unstable angina pectoris, cardiac arrhythmia requiring treatment or family history of sudden death from cardiac-related causes

8. History of or current active infection with HIV, hepatitis B virus, or hepatitis C virus

9. Major surgery within 4 weeks of starting study drug. Surgery for brain metastases within 2 weeks of starting study drug. Intravenous port placement was not considered as a major surgery

Assessment report EMA/197343/2017 Page 59/122 10. Clinically significant GI abnormality that would affect the absorption of drug such as malabsorption syndrome or major resection of the small bowel or stomach

11. Uncontrolled intercurrent illness or psychiatric illness/social situations that would limit compliance with study requirements

12. Currently under alcohol or drug abuse rehabilitation or treatment program

13. History of another malignancy, unless patient had been disease-free for 3 years. Individuals with the following cancers were eligible if diagnosed and treated within the past 3 years: any carcinoma in situ, and basal cell or squamous cell carcinoma of the skin

14. Pregnant or lactating women

15. Baseline corrected QT interval (QTc) > 470 ms, or baseline symptomatic bradycardia < 45 beats per minute

16. Administration of strong/ potent cytochrome P450 3A (CYP3A) inhibitors or inducers within 14 days or 5 half-lives (whichever was longer) prior to first administration of study drug(s) and while on treatment except for oral corticosteroids up to 20 mg prednisolone equivalent per day, or agents with potential QT prolonging effects within 14 days prior to first administration of study drug(s) and while on treatment.

• Treatments

Alectinib was administered to the patient orally BID starting on Day 1 through Day 21 of each 21-day treatment cycle. The RP2D determined in Phase I was 600 mg BID. Patients were treated continuously until disease progression, death, or withdrawal for any other reason.

Figure 5: Phase II dosing schedule (NP28761)

• Objectives

Primary objective

To evaluate the efficacy of alectinib based on ORR (according to RECIST 1.1) as per independent review committee (IRC) in patients ALK-positive NSCLC who have failed crizotinib.

Secondary objectives

• ORR according to RECIST 1.1 based on investigator review of radiographs

• DCR, DOR and PFS according to RECIST v1.1 by IRC and investigator review of radiographs

• OS

• safety of alectinib

Assessment report EMA/197343/2017 Page 60/122 • to characterize the PK of alectinib and metabolite(s)

• to assess QoL using the EORTC QLQ − C30 and − LC13

• CNS ORR (CORR) in patients with CNS metastases who have measurable disease in the CNS at baseline, based on IRC review of radiographs by RECIST 1.1 and Response Assessment in Neuro-Oncology (RANO) criteria

• CNS duration of response (CDOR) in patients who have a CNS objective response based on IRC review of radiographs by RECIST 1.1 and RANO criteria

• CNS progression rates (CPR) at 3, 6, 9 and 12 months based on cumulative incidence by IRC review of radiographs by RECIST v1.1 and RANO criteria.

Exploratory objectives

• To assess blood based methods to detect point mutations in plasma

• To identify molecular determinants of clinical resistance to ALK inhibitors

• To identify potential genetic determinants of PK variability and safety parameters (pharmacogenomics research)

• To assess alectinib CNS penetration by measuring the cerebrospinal fluid (CSF): plasma concentration ratio

• Outcomes/endpoints

Primary endpoint

ORR (IRC assessed) - Proportion of patients achieving confirmed CR or PR

Secondary endpoints

ORR (investigator assessed) - Proportion of patients achieving confirmed CR or PR

DCR - Proportion of patients achieving CR, PR or SD lasting ≥ 12 weeks

DOR – duration of response

PFS – progression free survival

OS – Overall survival

CORR (IRC assessed) – CNS ORR

CDOR (IRC assessed) – CNS DOR

CPR (IRC assessed) – CNS progression rate

• Sample size

Assessment report EMA/197343/2017 Page 61/122 A sample size of 85 was chosen such that the lower limit of the 2-sided 95% confidence interval (CI) (using a Clopper-Pearson CI) around the point estimate of the ORR allowed identification of a clinically relevant response in order to reject the null hypothesis (H0) that ORR = 35%. With 85 patients, an observed response rate of 46% (39/85 responses) would have a lower limit of the two-sided 95% CI of 35% which is considered to be clinically relevant, and thus the null hypothesis could be rejected.

The primary analysis was to occur once all patients enrolled in Phase II were followed for a minimum of 12 weeks, essentially after two tumour assessments had been performed in order that any observed CR or PR could be confirmed, unless patients progressed, died or withdrew sooner.

A non-binding interim analysis for futility was planned after 30 patients had been dosed and followed at least until their first post-baseline tumour assessment (6 weeks post-treatment). The analysis was based on the investigator-assessed ORR in RE population. If ≥ 9 responders were observed, the Phase II portion would continue to enrol a planned total of approximately 85 patients. If ≤ 8 responders were observed, accrual could be discontinued. The futility analysis was performed by an IMC.

• Randomisation

N/A

• Blinding (masking)

N/A

Statistical methods

The following null (H0) and alternative (Ha) hypotheses were tested at a two-sided 5% significance level to compare ORR according to RECIST per IRC review in the response evaluable (RE) population (i.e., the primary efficacy endpoint):

H0: ORR = 35% versus Ha: ORR ≠ 35%

A sample size of 85 was chosen such that the lower limit of the 2-sided 95% confidence interval (CI) (using a Clopper-Pearson CI) around the point estimate of the ORR allowed identification of a clinically relevant response in order to reject the null hypothesis (H0) that ORR = 35%. With 85 patients, an observed response rate of 46% (39/85 responses) would have a lower limit of the two-sided 95% CI of 35%.

The primary analysis, reported in this CSR, was to occur once all patients enrolled in Phase II were followed for a minimum of 12 weeks, i.e., after two tumor assessments had been performed in order that any observed CR or PR could be confirmed, unless patients progressed, died or withdrew sooner.

A non-binding interim analysis for futility was planned for Phase II after 30 patients had been dosed and followed at least until their first post-baseline tumor assessment (6 weeks post-treatment). The analysis was based on the investigator-assessed ORR using RECIST 1.1 in RE population. If ≥ 9 responders were observed, the Phase II portion would continue to enroll a planned total of approximately 85 patients. If ≤ 8 responders were observed, accrual could be discontinued. The futility analysis was performed by an IMC.

Assessment report EMA/197343/2017 Page 62/122 Results

Participant flow

Disposition of Patients in Phase II:

Recruitment

Patients were enrolled in 26 centres in the US and one site in Canada. The first patient was enrolled on 03 May 2012 and the last patient was enrolled on 04 August 2014.

The data cut-off date for the primary analysis was 24 October 2014 and the database lock date was 10 December 2014.

Conduct of the study

Six protocol amendments were made and a summary of the major changes relating to study design, objectives or entry criteria is provided below.

Protocol Version 2: 18 November 2011

In Version 2 “NSCLC that has failed crizotinib treatment” was added to the inclusion criteria for Phase I.

Protocol Version 3: 14 May 2012

Key changes in Version 3 were modifications to the dose-escalation design of Phase I.

Protocol Version 4: 10 August 2012

Version 4 added surgery for brain metastases within two weeks of starting treatment to the list of exclusion criteria and to allow concurrent radiation therapy for palliation or symptom relief of brain or bone metastases within 24 hours of last dose of alectinib and re-start of study drug upon resolution of any radiation toxicity to ≤ Grade 1.

Assessment report EMA/197343/2017 Page 63/122 Version 5: 12 March 2013

Phase II consisted of a single population of patients that had failed crizotinib Phase I (instead of two sub- populations A [crizotinib-failed] and B [crizotinib-naïve]); the study design was modified to conduct Phase II under fed conditions only; and measures of efficacy in the CNS (CNS response rate in patients with brain lesions that were not irradiated and CNS relapse rate) were added to the secondary objectives for Phase II.

Version 6: 24 June 2013

The wash-out period for crizotinib was decreased from 2 weeks to 1 week in order to avoid potential disease flare ups.

Version 7: 17 December 2013

The study population definition was changed to “Patients with locally advanced (AJCC Stage IIIB) not amenable to curative therapy or metastatic (AJCC Stage IV) NSCLC” (previously “locally advanced or metastatic NSCLC”); CORR, CDOR, CPR, and QoL using EORTC QLQ-C30 and QLQ-LC13 were added as secondary efficacy endpoints for Phase II; measures of efficacy in CNS (CNS response rate in patients with brain lesions that were not irradiated, and CNS relapse rate), were changed to CNS ORR, CNS DOR, and CNS progression rate (CPR) assessed using RECIST v1.1 and RANO criteria; the statistical hypothesis for efficacy and statistical methods (null hypothesis that ORR = 35% instead of ORR = 50%) were clarified.

Protocol violations

One patient had a major protocol violation at baseline. The patient tested ALK-positive with a non FDA- approved FISH test. In the absence of any safety concerns noted, the patient continued to receive alectinib and was not excluded from the study or analyses.

During the study ten major protocol violations occurred in eight patients (9%) for the following reasons:

• Use of prohibited medications/procedures (7 violations in five patients). These violations pertained to the administration of corticosteroids at doses higher than allowed per protocol, potent CYP3A4 inhibitor and radiosurgery/ radiotherapy. • Study drug not taken according to protocol (2 violations in 2 patients)

• Missing tumour assessment at a scheduled visit (1 violation in 1 patient).

At the time of discovering these violations, no safety concerns were identified. No patient who violated the study protocol was excluded from any analysis and all patients continued on study drug as planned.

Baseline data

Table 26: Patient demographics and baseline characteristics in the safety population for Phase II

Assessment report EMA/197343/2017 Page 64/122

Assessment report EMA/197343/2017 Page 65/122 Table 27: Disease history of NSCLC in safety population (Phase II)

Table 28: Target lesions at baseline (IRC) in response evaluable population (Phase II)

Assessment report EMA/197343/2017 Page 66/122 Numbers analysed

Table 29: Summary of analysis populations in Phase II

Outcomes and estimation

Primary Endpoint

ORR (IRC Assessed in RE Population)

Table 30: Objective Response Rate (IRC) in Phase II (RE Population)

Assessment report EMA/197343/2017 Page 67/122

Figure 6: Waterfall Plot of Target Lesions Sum of Longest Diameter Best Change from Baseline (IRC) in RE population Secondary endpoints

ORR (Investigator Assessed)

Table 31: ORR and DCR in Phase II (Investigator) (RE Population):

Assessment report EMA/197343/2017 Page 68/122

Figure 7: Waterfall Plot of Target Lesions Sum of Longest Diameter Best Change from Baseline (Investigator) in Phase II (RE Population) Concordance Analysis between IRC and Investigator Assessments of ORR

Table 32: Concordance Analysis between IRC-Determined and Investigator-Determined Best Overall Response in Phase II (Investigator RE Population)

Table 33: Concordance Analysis between IRC-Determined and Investigator-Determined Best Overall Response in Phase II (IRC RE Population)

Duration of Response (DOR)

Of the 33 patients achieving PR in the IRC assessment of the RE population, six patients experienced progression of their disease. The median DOR was 7.5 months.

Progression - Free Survival (PFS)

At the time of the data cut off, 35 of 87 patients (40%) had progressed or died based on IRC assessments. The majority due to PD. The estimated median duration of PFS was about 6 months.

Overall Survival (OS)

At the time of data cut-off, 14% enrolled in Phase II had died. Due to the low number of deaths, the median OS could not be estimated.

Assessment report EMA/197343/2017 Page 69/122 Ancillary analyses

Health-Related Quality of Life (HRQoL),

Improvements (changes from baseline of ≥ 10%; Osoba et al 1998, 2005) were seen in the European Organization for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire (QLQ-C30) items Global Health Status, Emotional Functioning and Social Functioning and also the symptom scales Fatigue and Pain, indicating improvement of these symptoms.

Summary of main study(ies)

The following tables summarise the efficacy results from the main studies supporting the present application. These summaries should be read in conjunction with the discussion on clinical efficacy as well as the benefit risk assessment (see later sections).

Table 34: Summary of Efficacy for trial NP28761 Title: A Phase I/II Study of the ALK Inhibitor alectinib (CH5424802/RO5424802) in patients with ALK-rearranged Non-Small Cell Lung Cancer previously treated with Crizotinib. Study identifier NP28761 Design multi-centre, single arm, open-label, dose escalation study

Duration of main phase: 03 May 2012 to 24 October 2014 (first patient screened to data cut-off date for primary analysis)

Hypothesis Exploratory Treatments groups Alectinib Alectinib was administered to the patient orally BID starting on Day 1 through Day 21 of each 21-day treatment cycle. The RP2D determined in Phase I was 600 mg BID. Patients were treated continuously until disease progression, death, or withdrawal for any other reason.

Endpoints and Primary IRC assessed Proportion of patients achieving confirmed* CR or definitions endpoint ORR PR, calculated using best overall response

Secondary Investigator- Proportion of patients achieving confirmed CR endpoint assessed or PR, calculated using best overall response ORR Secondary DOR Interval between 1st documented response (CR or endpoint PR) and 1st documented PD or death Secondary PFS Interval between 1st dose of alectinib and date of endpoint 1st documented PD (IRC or investigator) or death Secondary OS Interval between 1st dose of alectinib and endpoint date of death

Results and Analysis

Analysis description Primary Analysis Database lock: 24 October 2014 Analysis population and Response evaluable population: all patients with measurable disease at baseline time point description who had a baseline tumour assessment and received at least one dose of alectinib. Descriptive statistics and Treatment group ALK-positive NSCLC previously treated with crizotinib estimate variability ORR (IRC) % 47.8 N= 69

95% CI [35.6, 60.2]

Assessment report EMA/197343/2017 Page 70/122 ORR (investigator) % 46.0 N= 87

95% CI [35.2, 57.0]

DOR (months) 7.5 Median N=33

95% CI [4.9, NE]

PFS (months) 6.3 Median N=87

95% CI [5.5, NE]

OS (months) NE Median N=87

95% CI [10.1, NE]

Analysis description Secondary analysis Database lock: 22 January 2016

Descriptive statistics and Treatment group ALK-positive NSCLC previously treated with crizotinib estimate variability

ORR (IRC) % 52.2 N= 67

95% CI [39.7, 64.6]

ORR (investigator) % 52.9 N=87

95% CI [41.9, 63.7]

DOR (months) 14.9 Median N=35

95% CI [6.9, NE]

PFS (months) 8.2 Median N=87

95% CI [6.3, 12.6]

OS (months) 22.7 Median N=87

95% CI [17.2, NE]

Assessment report EMA/197343/2017 Page 71/122 Study NP28673: An open-label, non-randomized, multicentre Phase I/II trial of alectinib (RO5424802) given orally to Non-Small Cell Lung Cancer patients who have ALK mutation and who have failed crizotinib treatment.

This study was conducted in three parts:

● Phase I, Part 1: A dose-escalation phase with the objective of assessing the safety, tolerability, and PK of alectinib 600 mg BID and 900 mg BID dose regimens. During the conduct of Part 1 for this study, the RP2D (600 mg BID) was confirmed in study NP28761 and Part 1 and Part 2 were combined.

● Phase II, Part 2: A safety and efficacy evaluation phase

● Phase II, Part 3: A post-progression treatment phase

Radiologic assessments were performed at screening and every 8 weeks in the first year, every 12 weeks in the second year and every 16 weeks subsequently until disease progression.

Patients in Part 2 that were considered by the investigator to still benefit from treatment, could be included in the treatment beyond progression in Part 3, where they could continue to receive alectinib alone or, if their tumour harboured an EGFR activating mutation, in combination with erlotinib. However, all patients in Part 3 have continued with alectinib monotherapy only. No patients have received erlotinib in combination with alectinib at this point.

Figure 8: Study design – Part 2 and 3 (Study NP28673)

Methods

Study Participants

See inclusion/exclusion criteria for Study NP28671

Assessment report EMA/197343/2017 Page 72/122 Treatments

Alectinib as a single agent was administered orally at a dose of 600 mg BID within 30 minutes after meals in the morning and evening. Patients received alectinib at the stated dose for 5 cycles (28 days) continuously, starting on Cycle 1, Day 1. When a patient missed a dose of alectinib, it could be taken within 4 hours of the scheduled time. If the time was greater than 4 hours, or the patient vomited the dose, the patient was instructed to wait and take the next scheduled dose.

Objectives/endpoints

Co-Primary Efficacy Objectives/endpoints

• ORR as per central IRC using RECIST v1.1 in the overall population (with and without prior exposure of cytotoxic chemotherapy treatments).

• ORR as per central IRC using RECIST v1.1 in patients with prior exposure of cytotoxic chemotherapy treatments.

Secondary Efficacy Objectives/endpoints

• ORR as per central IRC using RECIST criteria v1.1 in patients without prior exposure of cytotoxic chemotherapy treatments

• ORR per investigator review of radiographs using RECIST v1.1

• DCR based on IRC and investigator review of radiographs

• DOR based on IRC and investigator review of radiographs

• PFS based on IRC and investigator review of radiographs

• CNS Objective Response Rate (CORR) in patients with CNS metastases who had measurable disease in the CNS at baseline, based on IRC review of radiographs

• CNS Duration of Response (CDOR) in patients who have a CNS Objective Response based on IRC review of radiographs

• CNS Progression Rates (CPRs) at 3, 6, 9 and 12 months based on cumulative incidence by IRC review of radiographs

• To assess OS

Safety Objectives

• To evaluate the safety profile of alectinib using the NCI-CTCAE v 4.03

• To assess the effect of alectinib on cardiac repolarization (corrected QT interval [QTc])

Exploratory Objectives

• To evaluate ORR in patients with EGFR mutation who had experienced disease progression on alectinib alone and were subsequently treated with a combination of alectinib and erlotinib (Part 3)

• To evaluate the PK of alectinib and metabolite(s), and erlotinib in patients with EGFR mutation who have experienced disease progression on alectinib alone and who received the combination of alectinib and erlotinib (Part 3)

Assessment report EMA/197343/2017 Page 73/122 • To evaluate the safety profile of alectinib in combination with erlotinib using the NCI-CTCAE v 4.03 (Part 3)

• To evaluate the ALK mutations in tumor and blood samples and correlate it with response to alectinib, where possible

• To evaluate the co-expression of EGFR mutation with ALK translocation and ALK mutation status

• To investigate potential bypass mechanisms of ALK inhibition, such as cMET, KRAS and cKIT

Sample size

A sample size of 85 was chosen such that the lower limit of the two-sided 95% confidence interval (CI; using an exact Clopper-Pearson CI) around the point estimate of the ORR allowed identifying a clinically relevant response in order to reject the null hypothesis that ORR = 35%. With 85 patients, an observed response rate of 46% (39/85 responses) would have a lower limit of the two-sided 95% CI of 35.02% which is considered to be clinically relevant and the null hypothesis would be rejected. Therefore, at least 85 patients treated with prior chemotherapy would need to be enrolled. A total enrolment of 130 patients was planned, with a maximum of 45 chemo-naïve patients to ensure the minimum number requirement of 85 prior-chemotherapy patients was met.

With 130 patients, an observed response rate of 44% (57/130 responses) would have a lower limit of the two-sided 95% CI of 35.2% which is considered to be clinically relevant and the null hypothesis would be rejected. With 130 patients, there was 93% power, based on an exact test for single proportion, to detect a 15% increase in ORR from 35% to 50% at the 5% two-sided significance level.

Randomisation

N/A

Blinding (masking)

N/A

Statistical methods

Hierarchical testing was used. The primary analysis tested the null hypothesis of the best ORR for alectinib equal to 35%, against the alternative hypothesis that the best ORR for alectinib was not equal to 35%, with two-sided alpha of 0.05. This primary analysis was tested in the all patients group (patients with and without prior exposure to cytotoxic chemotherapy treatment).

If the primary test was significant, then the subsequent test also tested the null hypothesis of the best ORR for alectinib equal to 35%, against the alternative hypothesis that the best ORR for alectinib was not equal to 35%, with two-sided alpha of 0.05. However this subsequent analysis was tested only in the group of patients with prior exposure to cytotoxic chemotherapy treatment.

Assessment report EMA/197343/2017 Page 74/122 Results

Participant flow

Recruitment

A total of 138 patients were enrolled and treated in Phase II of the study at 56 sites in 16 countries globally (France (11), the US (10), Italy (8), Spain (6), the Republic of Korea (4), Taiwan (3), Germany (3), the Netherlands (2), Australia (2), UK (1), Luxemburg (1), Belgium (1), Sweden (1), Singapore (1), Denmark (1) and the Russian Federation(1).

Data cutoff date was 18 August 2014.

Conduct of the study

The initial protocol, dated 12 December 2012, was amended four times. A summary of the main changes made to the protocol are summarised below.

Protocol Amendment 1 (Version 2): 21 December 2012

Protocol Version 1 was amended to remove exclusion criterion regarding use of anticoagulation or thrombolytic agent within 2 weeks prior to Day 1.

Protocol Amendments 3 and 4 (Versions 4 and 5): 19 November 2013 (v4) and 30 January 2014 (v5)

During submission of protocol Version 4, additional changes were identified that required the creation of protocol amendment Version 5. The following changes include those made to Versions 4 and 5:

1. To incorporate a MDZ DDI sub-study to be conducted in approximately 14 additional patients with ALK- positive NSCLC.

Assessment report EMA/197343/2017 Page 75/122 2. To clarify the Stage IIIb NSCLC study patients to be included in the study, as per FDA request: Only Stage IIIb NSCLC patients whose disease was not amenable to curative therapy would be eligible for the study, as other Stage IIIb patients may be eligible to other multimodality treatments.

3. To remove the restriction for the last dose of crizotinib to be within 60 days from the first dose of alectinib. This restriction was first put in place to avoid possible resensitization to crizotinib and ensure that all patients are true crizotinib-failure. However, this limitation was thought to potentially affect enrollment as it may not have allowed enough time for patients to progress on subsequent chemotherapy treatments.

4. To update the permitted and prohibited medications based on the available DDI information for alectinib

Violations at Baseline

Table 35: Summary of Major Protocol Violations at Baseline (NP28673)

Violations during study

In total, 26 major protocol violations occurred in 23 patients (17%) during the study for the following reasons:

• Use of prohibited medications/procedures (19 violations [18 medications, 1 procedure] in 16 patients). Of note, of the 18 violations due to use of prohibited medications, nine were due to use of antacids and P-gp substrates. While these were considered prohibited concomitant medications at the time these violations occurred, their use was permitted in the latest version (Version E) of the study protocol.

• Study drug not taken according to protocol (6 violations in 6 patients)

• Missing tumour assessment at a scheduled visit (1 violation in 1 patient).

No patient who violated the study protocol was excluded from any analysis (efficacy or safety) and all patients continued on study drug as planned.

Assessment report EMA/197343/2017 Page 76/122 Baseline data

Table 36: Demographics and Baseline Characteristics in the Safety Population (Data Cut-off Date: 18 August 2014) – study NP28673

Assessment report EMA/197343/2017 Page 77/122 Table 37: Disease History of NSCLC in Safety Population (Data Cutoff Date: 18 August 2014) – study NP28673

Assessment report EMA/197343/2017 Page 78/122 Table 38: Number of Prior Regimens for Metastatic NSCLC and Prior Crizotinib Treatment (Safety Population)

Numbers analysed

Table 39: Analysis Populations (NP28673)

Assessment report EMA/197343/2017 Page 79/122 Outcomes and estimation

Primary endpoint

ORR – IRC-Assessed (Overall RE Population)

Table 40: ORR and DCR (IRC) in the RE Population (NP28673)

Figure 9: Waterfall Plot of Target Lesions Sum of Longest Diameter Best Change from Baseline (IRC) in RE Population (NP28673)

Assessment report EMA/197343/2017 Page 80/122 Updated ORR data were submitted during the procedure.

Table 41: ORR in Response Evaluable Population as Assessed by the IRC (NP28673)

Assessment report EMA/197343/2017 Page 81/122 ORR – IRC-Assessed (Chemotherapy-Pretreated Patients)

Table 42: ORR and DCR (IRC) in Patients with Prior Chemotherapy in the RE Population (NP28673)

Assessment report EMA/197343/2017 Page 82/122 Secondary endpoints

ORR (Investigator Assessment)

Table 43: ORR and DCR (Investigator) in the RE Population (NP28673)

Assessment report EMA/197343/2017 Page 83/122 Updated DCR data were submitted during the procedure.

Table 44: Summary of DCR in Response Evaluable Population as Assessed by the IRC and Investigator (NP28673)

Figure 10: Waterfall Plot of Target Lesions Sum of Longest Diameter Best Change from Baseline (Investigator) in RE Population (NP28673)

Assessment report EMA/197343/2017 Page 84/122 Table 45: Concordance Analysis Between the IRC-Determined and the Investigator-Determined Best Overall Response (NP28673)

DOR

Table 46: Summary of Duration of Response (NP28673)

Assessment report EMA/197343/2017 Page 85/122 CNS ORR (CORR, by IRC)

Table 47: CORR Based on RECIST (IRC) in Patients with Measurable CNS Lesions at Baseline (NP28673)

Figure 11: Waterfall Plot of CNS Target Lesions Sum of Longest Diameters Best Change from Baseline for Patients with Measurable CNS Lesions at Baseline Based on RECIST (IRC) – Study NP28673

Assessment report EMA/197343/2017 Page 86/122 CNS DOR (CDOR, by IRC)

Table 48: CDOR for Patients with Measurable CNS Lesions at Baseline Based on RECIST (IRC) – study NP28673

CNS Progression Rate (CPR, by IRC)

The analysis of CPR was based on the CNS RECIST v1.1 and the full RECIST v1.1 IRC review of scans. Progressions were identified in target, non-target and new lesions from CNS RECIST v1.1 and from progressions in the CNS in non-target and new lesions in the full RECIST v1.1 IRC review.

Table 49: CPRs for All Patients Based on CNS RECIST (IRC) in the Safety Population (NP28673)

Assessment report EMA/197343/2017 Page 87/122 Ancillary analyses

Subgroup analyses

Table 50: – ORR by subgroup (INV) RE population (NP28673)

Summary of main study(ies)

The following tables summarise the efficacy results from the main studies supporting the present application. These summaries should be read in conjunction with the discussion on clinical efficacy as well as the benefit risk assessment (see later sections).

Table 51: Summary of Efficacy for trial NP28673 Title: An open-label, non-randomized, multicentre Phase I/II trial of alectinib (RO5424802) given orally to non- small cell lung cancer patients who have ALK mutation and who have failed crizotinib treatment. Study identifier NP28673 Design An open-label, non-randomized, multicentre Phase I/II trial Duration of main phase: 20 June 2013 − 18 August 2014 (first patient in until data cut-off date for primary analysis).

Hypothesis Exploratory Treatments groups Alectinib 600 mg BID alectinib, orally for 5 cycles (28 days) continuously starting on Cycle 1, Day 1 until disease progression, death, or withdrawal for any other reasons.

Endpoints and Co-Primary ORR by IRC in A best overall response (BOR) of complete definitions endpoint the response response (CR) or partial response (PR) in all evaluable patients with measurable disease at baseline population

Assessment report EMA/197343/2017 Page 88/122 Co-Primary ORR by IRC in BOR of CR or PR in the subgroup of endpoint the subgroup patients who received previous chemotherapy of patients who received previous chemotherapy Secondary ORR by Proportion of patients achieving best endpoint Investigator response of CR or PR Secondary CORR (IRC- Proportion of patients achieving CR or PR of endpoint assessed) BL CNS lesions Secondary DOR Interval between 1st documented response endpoint (CR or PR) and 1st documented PD or death Secondary PFS Interval between 1st dose of alectinib and endpoint date of 1st documented PD (IRC or investigator) or death Secondary OS Interval between 1st dose of alectinib and endpoint date of death

Results and Analysis

Analysis description Primary Analysis Database lock: 8 January 2015 Analysis population and Response evaluable population: all patients with measurable disease at baseline who time point description had a baseline tumour assessment and received at least one dose of alectinib. Descriptive statistics and Treatment group ALK-positive NSCLC previously treated with crizotinib estimate variability ORR (IRC) % 50.0 N= 122

95% CI [40.8, 59.2]

ORR (IRC) in 44.8 patients pre- treated with chemotherapy % N= 96

95% CI [34.6, 55.3]

CORR (IRC- 57.1 assessed) % N=35

95% CI [39.4, 73.7]

DOR (months) 11.2 Median N=61

95% CI [9.6, NE]

PFS (months) 8.9 Median N=122

95% CI [5.6, 11.3]

OS (months) NE Median N=138

95% CI [14.6, NE]

Analysis description Secondary analysis Database lock: 1 February 2016

Assessment report EMA/197343/2017 Page 89/122 Descriptive statistics and Treatment group ALK-positive NSCLC previously treated with crizotinib estimate variability ORR (IRC) % 50.8 N= 122

95% CI [41.6, 59.9]

ORR (IRC) in 44.8 patients pre-treated with chemotherapy % N= 96

95% CI [34.6, 55.3]

CORR (IRC- 58.8 assessed) % N=34

95% CI [40.7, 75.4]

DOR (months) 15.2 Median N=62

95% CI [11.2, 24.9]

PFS (months) 8.9 Median N=138

95% CI [5.6, 12.8]

OS (months) 26.0 Median N=138

95% CI [21.5, NE]

Updated efficacy data

The updated data cut-off date for the NP28761 study was 22 January 2016 providing 17 months of follow-up. The updated data cut-off date for the NP28673 study was 1 February 2016 providing 21 months of follow-up.

Assessment report EMA/197343/2017 Page 90/122 Table 52: Overview of efficacy results in pivotal studies

The event rate in regard to PFS is 67 % and 71 % for NP28761 and NP28673 respectively. For OS, the event rate is 41 % and about 44 % respectively. The level of data maturity is considered acceptable.

The median PFS in the NP28761 study has increased with about 2 months (6→8 months) over time whilst remaining unchanged in study NP28673 (about 9 months).

Assessment report EMA/197343/2017 Page 91/122 Analysis performed across trials (pooled analyses and meta-analysis)

CNS ORR and DCR

Table 53: Summary of the pooled analysis of CNS endpoints from studies NP28673 and NP28761 CNS Parameters (NP28673 and NP28761) Alectinib 600 mg twice daily Patients with measurable CNS lesions at baseline N = 50 CNS ORR (IRC)

Responders (%) 32 (64.0%) [95% CI] [49.2%, 77.1%] Complete response 11 (22.0%) Partial response 21 (42.0%)

CNS DOR (IRC) N=32 Number of patients with events (%) 18 (56.3%) Median (months) 11.1 [95%CI] [7.6, NE] CI  confidence interval; DOR  duration of response; IRC  independent review committee; ORR  objective response rate; NE = not estimable

Clinical studies in special populations

Table 54: Elderly patients included in the pivotal studies (NP28761 and NP28673)

Supportive study

Study AF-001JP is a multicentre, open label Phase I/II Study of CH5424802 in Patients with Non-small Cell Lung Cancer Harbouring the ALK Fusion Gene.

The study was conducted in 13 sites in Japan. The study period was 24 August 2010 to 18 April 2013.

A total of 24 patients were enrolled in Phase I and were assigned to one of the seven alectinib dose cohorts (20, 40, 80, 160, 240, and 300 [fed/fasted] mg BID). In Phase II, 46 patients were enrolled and all received treatment with alectinib 300 mg BID. In phase II at least one prior line of chemotherapy was required. Prior treatment with an ALK inhibitor was an exclusion criterion.

Assessment report EMA/197343/2017 Page 92/122 In Phase II, mean age of the population was 49.5 years (range: 26 to 75 years), with 42/46 patients being < 65 years old. There were 52% women and 48% men. The majority of patients (59%) had never smoked (or hardly smoked). ECOG PS was 0 in 20 patients (44%) and 1 in 26 patients (57%). The histological type was adenocarcinoma in all 46 patients. Disease stage was IIIB in 2 patients (4%), IV in 31 patients (67%), and postoperative recurrence in 13 patients (28%).

In Phase I, alectinib 300 mg BID was chosen as the RP2D in Japanese patients on the basis of safety, PK, and efficacy data. Of the 20 Phase I patients, 18 patients had an investigator-assessed BOR of PR and 2 patients had SD.

In Phase II, the IRC-assessed response rate in the 46 crizotinib-naïve patients enrolled was 93.5% (95% CI: 82.1% to 98.6%), with 7 of the patients achieving a CR and 36 achieving a PR. Due to the low number of events, the median values for PFS and OS could not be estimated. One-year PFS rate was 83% (95% CI: 68% to 92%) and 1-year survival rate was 93% (95% CI: 81% to 98%).

Study JO28928 (J-ALEX)

During the procedure, the applicant submitted high-level data from J-ALEX, an open-label, randomized (1:1) Phase III study comparing alectinib head to head with crizotinib in patients who were ALK-inhibitor naïve and may have received up to one prior chemotherapy regimen for ALK–positive NSCLC.

This study is conducted in Japanese patients and the dosing is alectinib 300 BID as opposed to the dosing in the two studies pertinent to this application i.e. 600 BID.

The data presented are from a pre-planned interim analysis with 50% of the targeted PFS events. The median duration of follow-up at the time of this analysis was 12 months (range: 1 to 23 months) for the alectinib arm and 12 months (range: 0 to 20 months) for the crizotinib arm.

2.5.3. Discussion on clinical efficacy

Design and conduct of clinical studies

The benefit assessment is based on two Phase I/II studies, NP28761 (n=87, US/ Canada) and NP28673 (n=138, global) that enrolled patients with ALK positive NSCLC previously treated with crizotinib (Xalkori).

The amendments and protocol violations in the two studies are deemed unlikely to have a relevant impact on the integrity of the study.

In further support of the application the AF-001JP study has been submitted which is a Phase I/II study performed in Japanese patients with ALK positive NSCLC that had not received prior crizotinib. In this study conducted in chemotherapy pre-treated but ALK inhibitor naïve patients, the ORR was very high about 94 % and the 1-year PFS rate was 83%. This study did not report any CORR and CNS DCR data.

The design of the dose finding study (phase I of study NP28761) is considered standard for an oncology study. The recommended phase 2 dose (RP2D) determined i.e. 600 mg BID, is considered reasonably well justified. Study NP28673 also had a dose escalation portion (Part 1) but as the RP2D was defined in study NP28761, patients in the NP28673 Phase I part were included in the Phase II part (n=6).

Both pivotal studies were single arm studies and very similar in study design and patient eligibility criteria. Patient demographics were consistent with that of a NSCLC ALK positive population. The demographic

Assessment report EMA/197343/2017 Page 93/122 characteristics of the overall study population were 67% Caucasian, 26% Asian, 56% females, and the median age was 52 years. The majority of patients had no history of smoking (70%).

The ECOG (Eastern Cooperative Oncology Group) performance status at baseline was 0 or 1 in 90.6% of patients and 2 in 9.4% of patients. At the time of entry in the study, 99% of patients had stage IV disease, 61% had brain metastases and in 96% of patients tumours were classified as adenocarcinoma. Among patients included in the study, 20% of the patients had previously progressed on crizotinib treatment only, and 80% had previously progressed on crizotinib and at least one chemotherapy treatment.

Furthermore the patients enrolled were rather heavily pre-treated with 54% having received ≥ 3 lines of therapy prior to enrolment (including crizotinib). The majority (74%) had received a previous platinum-based chemotherapy regimen and 64% had received prior radiotherapy for NSCLC with main targets brain metastases (47 %), bone and lung. About half of the population had undergone at least one previous surgical procedure for NSCLC.

As required by the study protocols, all 225 patients in the overall population had previously received crizotinib. The median time on crizotinib was about 12 months (range 1 day to 4.5 years) and time since last dose of crizotinib was about a fortnight (range 7 days to 2 years).

The Applicant has provided best response achieved while on crizotinib for the respective studies. For the 87 patients in NP28761: one patient had CR (1.1 %), 28 patients had PR (32.2 %), 22 patients with SD (25.3 %) and 27 patients with PD (31.0 %). A total of nine patients were deemed NE/NA/Unknown (10.3 %). The corresponding numbers for study NP28673 were: PR 75 patients (54.3 %), SD 30 patients (21.7 %), PD 27 patients (19.6 %) and 6 (4.3 %) patients NA/Unknown. No patient had CR as best reported response on crizotinib.

Efficacy data and additional analyses

While primary endpoint in NP28761 and NP28673 was ORR by IRC assessment (RECIST 1.1 with 4 week confirmation), NP28673 also had a co-primary endpoint of ORR by IRC assessment in patients pre-treated with chemotherapy.

The primary objective was met in both studies and ORR by IRC was 48 % and 49 % in NP28761 and NP 28673 respectively. ORR by IRC for the patients pre-treated with chemotherapy was 44 %.

A partial response was obtained in 48 % and 49 % in in NP28761 and NP 28673 respectively and stable disease was achieved in 32 % and 30 % respectively. About 16 % had progressive disease, similar in both studies.

Data cut-off for NP28761 was 24 October 2014. For NP28673 data cut-off for the primary analysis was 18 August 2014 with an update provided with data cut-off date 8 January 2015. Only limited number of additional events had occurred at the latter.

Among secondary endpoints was ORR by Investigator with 46 % responders in NP28761 and 48 % in NP28673. ORR (Inv) in patients pre-treated with chemotherapy in NP28673 was 46 %.

Results on ORR based on subgroups (by age, gender, race, ECOG PS, CNS metastasis, smoking status, baseline body weight, and prior chemotherapy use), were generally consistent with the ORR in the overall population for both pivotal studies.

Assessment report EMA/197343/2017 Page 94/122 DCR as assessed by IRC was about 64 % in both studies although the definition of SD varied between the two studies with duration for at least 12 weeks in NP28761 and 16 weeks in NP28673.

In terms of time-related secondary endpoints such as DOR, PFS and OS, too few events had occurred which precluded any firm conclusions to be drawn. The Applicant has as requested (former MO Q 22), provided updates with data cut-off dates of 22 January 2016 for NP28761 and 1 February 2016 for NP28673:

For PFS, with 67% of the patients in study NP28761 experiencing an event, a mature estimate of the median PFS can be expected. The median PFS as assessed by IRC is about 8 months. In study NP28673, the median PFS by IRC (71% data maturity) is about 9 months. The pooled analysis shows a median PFS of approximately 8 months (69% data maturity). These results are considered clinically meaningful compared with the results obtained with ceritinib in a similar patient population (Study A: 6.9 months and Study B: 5.7 months).

For OS, the pooled updated median OS is 26 months based on 43% data maturity. These results are also clinically meaningful, and indirect comparison with ceritinib seems to show that alectinib fares at least as good (Study A: 16.7 months, Study B: 14.0 months). Study 1007 (crizotinib vs chemotherapy, second-line) showed a median PFS of 7.7 months and a median OS of 20.3 months.

In regard to CNS DOR a total 50 patients had measurable CNS lesions at baseline. The updated analyses show that the median CNS DOR for these patients is 11 months (which is slightly better than the primary pooled analysis of 9 months). This is considered of high clinical relevance. In the pivotal trial in the approval of ceritinib, patients with symptomatic CNS metastases were excluded.

Overall, the updated results from the two pivotal studies continue to demonstrate a clinically relevant effect of alectinib in the proposed patient population.

The Applicant has also provided data from a pre-planned interim analysis from the alectinib vs. crizotinib head to head, Japanese J-ALEX study. The comparative data form the J-ALEX study further supports the results from the two main studies [data not shown].

The Applicant was also requested to provide further data on the NP28673 Part 3. However, only 11 consented to an optional biopsy for EGFR testing and none of these were found to harbour an EGFR activating mutation. Hence no data on the combination of alectinib and erlotinib is available. As regards to the 44 patients in part 3 that continued on alectinib monotherapy beyond progression, 33 patients had discontinued alectinib at the time of the updated 01 February 2016 cut-off (main reason lack of further clinical benefit [26 patients]). The median treatment duration was 3.4 months (range 0 to 22 months, censored).

Additional efficacy data needed in the context of a conditional MA

The updated information from the phase II studies supporting the ALKi pretreated indication confirm the efficacy results initially observed, and provide reassurance on the robustness of these results. The high-level comparative data form the Japanese J-ALEX study further supports the results from the two main studies.

Although a positive benefit-risk profile can be concluded, the data have limitations inherent to the non- comparative nature of the pivotal studies supporting the recommendation for a conditional MA.

In order to confirm the positive benefit-risk profile, the applicant will submit the results of the ALEX study, a phase 3, two-arm, randomized study, comparing alectinib and crizotinib in the first line. Although this study is not in the exact same setting where alectinib is authorised, it will provide comparative efficacy and safety

Assessment report EMA/197343/2017 Page 95/122 data for alectinib versus crizotinib. The approval of alectinib in the EU is not expected to impact completion of the ALEX study as it is not conducted in the indication being applied for.

2.5.4. Conclusions on the clinical efficacy

The updated results from the two pivotal studies with now a reasonable level of data maturity continue to demonstrate a clinically relevant effect of alectinib in the proposed patient population.

The CHMP considers the following measures necessary to address the missing efficacy data in the context of a conditional MA:

In order to further confirm the efficacy and safety of alectinib in the treatment of patients with ALK-positive NSCLC, the MAH should submit the clinical study report of the phase III study ALEX comparing alectinib versus crizotinib in treatment naïve patients with ALK-positive NSCLC. The clinical study report will be submitted by 30 April 2018.

2.6. Clinical safety

The primary cut-off dates for the MAA were 24 October 2014 and 18 August 2014 for Studies NP28761 and NP28673, respectively. This safety update assessment provides cumulative safety data with data cut-off dates of 22 January 2016 and 01 February 2016 for Studies NP28761 and NP28673, respectively.

1. Group 1 includes patients from Phase II of pivotal Study NP28761.

2. Group 2 includes patients from Phase II of pivotal Study NP28673.

3. Group 3 includes patients from Phase I and II of Study NP28761 (only patients in Phase I who received alectinib at 600 mg twice daily [BID] were included), and Phase II and the midazolam (MDZ) drug-drug interaction (DDI) substudy of Study NP28673.

The safety of Alectinib has been evaluated in 253 patients in pivotal phase II clinical trials (NP28761, NP28673) with ALK-positive non-small cell lung cancer (NSCLC) treated with the recommended dose of 600 mg twice daily. The median duration of exposure to Alectinib was 11 months.

Assessment report EMA/197343/2017 Page 96/122 Patient exposure

Table 55: Patient Status at Time of Data Cut-off Date for the Safety Update

About 1/3 of the patients were still on therapy and slightly more than 40% of the patients were dead at time of this update meaning that safety data are considered reasonably stable and robust.

Table 56: Study Drug Exposure in Studies NP28761 (Group 1) and NP28673(Group 2) and the Integrated Safety Population (Group 3) (Safety Population)

Mean and median duration of therapy were roughly 1 year at close to 100% dose intensity.

Assessment report EMA/197343/2017 Page 97/122 Adverse events

Table 57: Overview of Adverse Events in Studies NP28761 (Group 1) and NP28673 (Group 2) and the Integrated Safety Population (Group 3) (Safety Population)

AEs (related or not) leading to withdrawal of treatment is overall about 6%

Table 58: Adverse Events (Grade 3−5) with an Incidence of At Least 2% in Studies NP28761 (Group 1), NP28673 (Group 2), or the Integrated Safety Population (Group 3) (Safety Population)

Assessment report EMA/197343/2017 Page 98/122

Significant adverse events of special interest

Interstitial lung disease (ILD) / pneumonitis

Severe ILD/pneumonitis occurred in patients treated with Alectinib. In the pivotal phase II clinical trials (NP28761, NP28673), 1 out of 253 patients treated with Alectinib (0.4%) had a Grade 3 ILD. This event led to withdrawal from Alectinib treatment. There were no fatal cases of ILD.

Hepatocellular or Cholestatic Adverse Events and Abnormal Liver Function Tests

Hepatotoxicity is considered a class effect of ALK inhibitors and has been classified as an important identified risk for alectinib.

In the pivotal phase II clinical trials (NP28761, NP28673) two patients with Grade 3-4 AST/ALT elevations had documented drug induced liver injury by liver biopsy. One of these cases led to withdrawal from Alecensa treatment. Adverse reactions of increased AST and ALT levels (16% and 14% respectively) were reported in patients treated with Alectinib in pivotal phase II clinical trials (NP28761, NP28673). The majority of these events were of Grade 1 and 2 intensity, and events of Grade ≥ 3 were reported in 2.8% and 3.2% of the patients, respectively. The events generally occurred during the first 3 months of treatment, were usually transient and resolved upon temporary interruption of Alectinib treatment (reported for 1.2% and 3.2% of the patients, respectively) or dose reduction (1.6% and 0.8%, respectively). In 1.2% and 1.6% of the patients, AST and ALT elevations, respectively, led to withdrawal from Alectinib treatment.

Adverse reactions of bilirubin elevations were reported in 17% of the patients treated with Alectinib in pivotal phase II clinical trials (NP28761, NP28673). The majority of the events were of Grade 1 and 2 intensity; Grade 3 events were reported in 3.2% of the patients. The events generally occurred during the first 3 months of treatment, were usually transient and resolved upon temporary interruption of Alectinib treatment (4.7% of the patients) or dose reduction (2.8%). In 4 patients (1.6%), bilirubin elevations led to withdrawal from Alectinib treatment.

Concurrent elevations in ALT or AST greater than or equal to three times the ULN and total bilirubin greater than or equal to two times the ULN, with normal alkaline phosphatase, occurred in one patient (0.2%) treated in Alectinib clinical trials.

Assessment report EMA/197343/2017 Page 99/122 Table 59: Hepatocellular or Cholestatic Damage Adverse Events and Abnormal Liver Function Tests by Preferred Term in the Integrated Safety Population in more than one patient (Safety Population)

Two patients with Grade 3-4 AST/ALT elevations had documented drug-induced liver injury (DILI) by liver biopsy in Group 3:

 Patient 1 in Study NP28761: Liver biopsy showed morphological features of autoimmune hepatitis, but there was no clear diagnostic evidence of a separate autoimmune hepatitis. Since a similar pattern may also be seen secondary to drug, and in the patient's clinical setting, a drug-induced hepatitis was favoured as diagnosis. The event let to permanent discontinuation of study drug and was considered resolved by Day 53.

 Patient 2 in Study NP28673: Ultrasound guided biopsy of the liver performed on Day 222 showed centrilobular hepatic injury deemed compliant with a toxic effect and thus indicative for DILI. The events of Grade 3 ALT and AST elevations led to study drug interruption on Day 208, and ALT elevation finally led to permanent discontinuation on Day 215. The events were considered resolved by Day 268.

In addition

 Patient 3 in Study BO28984 (ongoing phase III) was hospitalized with Grade 4 drug induced liver injury DILI (considered serious). ALP was within normal range. The patient already had an event of increased AST, ALT and bilirubin (up to a maximum of 17 fold, 14 fold and 4.7 fold the ULN, respectively), along with subclinical icterus, abdominal pain, nausea, vomiting, loss of appetite and increased C-reactive protein approximately 2 weeks before study start. The event of DILI led to permanent discontinuation of study drug and was considered resolved by Day120.

No cases of hepatic failure have been reported.

Bradycardia

Cases of bradycardia (7.9%) of Grade 1 or 2 have been reported in patients treated with Alectinib in pivotal phase II clinical trials (NP28761, NP28673). There were 44 of 221 patients (20%) treated with Alectinib who had post-dose heart rate values below 50 beats per minutes. No case of bradycardia led to withdrawal from Alectinib treatment.

Severe myalgia and CPK elevations

Cases of myalgia (31%) including myalgia events (25%) and musculoskeletal pain (7.5%) have been reported in patients treated with Alectinib in pivotal phase II clinical trials (NP28761, NP28673). The majority of events were Grades 1 or 2 and three patients (1.2%) had a Grade 3 event. Dose modifications of Alectinib treatment due to these adverse events were only required for two patients (0.8%). Alectinib treatment was not withdrawn due to these events of myalgia. Elevations of CPK occurred in 46% of 219 patients with CPK

Assessment report EMA/197343/2017 Page 100/122 laboratory data available in pivotal phase II clinical trials (NP28761, NP28673) with Alectinib. The incidence of Grade 3 elevations of CPK was 5.0%. Median time to Grade 3 CPK elevation was 14 days. Dose modifications for elevation of CPK occurred in 4.0% of patients; withdrawal from Alectinib treatment did not occur due to CPK elevations..

Gastrointestinal effects

Constipation (36%), nausea (22%), diarrhoea (18%) and vomiting (13%) were the most commonly reported gastrointestinal (GI) reactions. Most of these events were of mild or moderate severity; Grade 3 events were reported for diarrhea (1.2%), nausea (0.4%), and vomiting (0.4%). These events did not lead to withdrawal from Alecensa treatment. Median time to onset for constipation, nausea, diarrhea, and/or vomiting events was 18 days. The events declined in frequency after the first month of treatment.

Adverse drug reactions

The safety database related to the dose proposed for marketing represents 253 patients with a median duration of therapy of approximately 1 year.

Assessment report EMA/197343/2017 Page 101/122 Table 60: Adverse Reactions in ALK-Positive Patients Treated with Alectinib

Assessment report EMA/197343/2017 Page 102/122 Serious adverse event/deaths/other significant events

Table 61: Serious Adverse Events Occurring in at Least 2 Patients (Safety Population)

Group 1

Fifteen patients (17%) experienced 19 SAEs. All events were reported in single patients.

SAEs lead to dose reduction (1 patient), dose interruption (8 patients), and study drug withdrawal (1 patient).

The event was considered related in one patient who reported a Grade 3 SAE of tumour necrosis beginning on Day 465. Study drug therapy was not changed.

Group 2

Thirty-one patients (23%) experienced 48 SAEs. All events were reported in single patients, with the exceptions of hyperbilirubinemia (3 patients [2%]), pulmonary embolism (3 patients [2%]), dyspnoea (3 patients [2%]), ALT increased (2 patients [1%]), and AST increased (2 patients [1%]).

SAEs leading to dose reduction (5 patients), dose interruption (11 patients), and study drug withdrawal (8 patients).

Assessment report EMA/197343/2017 Page 103/122 Table 62: Death and Cause of Death in Studies NP28761 (Group 1) and NP28673 (Group 2) and the Integrated Safety Population (Group 3) (Safety Population)

Laboratory findings

The following table displays treatment-emergent laboratory abnormalities occurring in >10% (all grades) of patients treated with Alectinib.

Table 63: Key Laboratory Abnormalities Parameter Alectinib N= 250* All Grades (%) Grade 3 -4 (%)# Chemistry Increased Blood Creatinine 99 1.2 Increased AST 61 3.2 Increased ALT 45 4.0 Increased Blood Creatine Phosphokinase 40 4.4 Increased Bilirubin (total) 35 1.6 Haematology Decreased Hemoglobin 85 2.0 AST - Aspartate Aminotransferase, ALT - Alanine Aminotransferase, * n=203 for Creatine Phosphokinase. #Treatment-Emergent only.

Assessment report EMA/197343/2017 Page 104/122 Safety in special populations

Table 64: Adverse Events by Age Group in the Integrated Safety Population (Group 3) (Safety Population)

Table 65: Adverse Events by Sex in the Integrated Safety Population (Group 3) (Safety Population)

Assessment report EMA/197343/2017 Page 105/122 Table 66: Adverse Events by Race in the Integrated Safety Population (Group 3) (Safety Population)

There are no or limited amount of data from the use of Alectinib in pregnant women.

Safety related to drug-drug interactions and other interactions

See clinical pharmacology.

Dose modifications or interruptions

Group 1

In total, 16 patients (18%) experienced AEs that led to a dose reduction.

Since the data cutoff date for the SCS, 4 additional patients reported a total of 5 new AEs leading to dose reduction of study drug.

The most common reasons for dose reduction were blood CPK increased (5%), AST increased (3%), and ALT increased and blood bilirubin increased (2% each).

Compared with the SCS, the following new AEs were reported leading to dose reduction: AST increased (1 patient), ALT increased (1 patient), constipation (1 patient), nausea (1 patient), and oedema peripheral (1 patient).

At the time of the data cutoff for the safety update (22 January 2016), the median time to dose reduction was 46 days (range: 4−644 days).

Group 2

In total, 15 patients (11%) experienced AEs that led to a dose reduction. Since the data cutoff date for the SCS, 3 additional patients reported a total of 9 new AEs leading to dose reduction of study drug.

The most common reasons for dose reduction were blood bilirubin increased (3%), abdominal pain upper (1%), and vomiting (1%). Compared with the SCS, the following new AEs were reported leading to dose

Assessment report EMA/197343/2017 Page 106/122 reduction: abdominal pain upper (2 patients [1%]), blood bilirubin increased (2 patients [1%]), rash maculo- papular (1 patient [1%]), and urticaria (1 patient [1%]).

At the time of the data cutoff for the safety update (01 February 2016), the median time to dose reduction was 102.5 days (range: 2−577 days).

Discontinuation due to adverse events

Altogether 15 patients (6%) discontinued due to AEs, hepatic laboratory events dominated, but there were also single cases of pneumonia, ILD, ligament rupture.

Data from supportive study

J-ALEX was a randomised crizotinib comparative trial conducted in Japan. The dose of alectinib was 300 bid instead of 600 mg bid. The estimated systemic exposure in this study was about 35% lower than in the pivotal trials. This is not likely to have a more than minor impact to safety.

Table 67: Summary of AEs

There was no case of death related to AEs.

Assessment report EMA/197343/2017 Page 107/122 Table 68: Adverse Events Leading to Discontinuation

2.6.1. Discussion on clinical safety

The safety database related to the dose proposed for marketing is small (253) but median duration of therapy is approximately 1 year. Still the long-term safety of alectinib is not sufficiently characterised and will be monitored as part of routine pharmacovigilance. The crizotinib comparative trial conducted at a lower dose of alectinib (J-ALEX) is considered informative and contextualises the safety profile.

Almost every patient in the two pivotal studies experienced at least one AE. There are comparable numbers of deaths (all cause) and SAEs in the two studies.

The most common adverse drug reactions (ADRs) (≥ 20%) were constipation (36%), oedema (34%, including oedema peripheral, oedema, generalised oedema, eyelid oedema, periorbital oedema), myalgia (31%, including myalgia and musculoskeletal pain) and nausea (22%).

Cases of ILD/pneumonitis have been reported in clinical trials with Alectinib (see sections 4.4 and 4.8 of the SmPC). Patients should be monitored for pulmonary symptoms indicative of pneumonitis. Alectinib should be immediately interrupted in patients diagnosed with ILD/pneumonitis and should be permanently discontinued if no other potential causes of ILD/pneumonitis have been identified (see section 4.2 of the SmPC).

Gastrointestinal AEs seem to be common. This may be due to the high concentration of SLS in the tablets. SLS is a known gastrointestinal irritant that can cause the above mentioned common GI AEs, and these same GI AEs are also observed in patients treated with crizotinib. They are clinically manageable and do not pose any major concerns. In Group 4 liver related AEs are more common than GI related AEs. Patients in Group 4

Assessment report EMA/197343/2017 Page 108/122 received only 300 mg BID and thus had a lower exposure to SLS could explain the lower incidence of GI related AEs.

The applicant is recommended to introduce an alternative capsule formulation for oral administration containing a lower amount of SLS post-approval.

Hepatotoxicity is a class effect of ALK-inhibitors, and is also observed in patients treated with crizotinib and ceritinib.

Elevations in alanine aminotransferase (ALT) and aspartate aminotransferase (AST) greater than 5 times the ULN as well as bilirubin elevations of more than 3 times the ULN occurred in patients in pivotal phase II clinical trials (NP28761, NP28673) with Alectinib (see sections 4.4 and 4.8 of the SmPC). The majority (76% of the patients with hepatic aminotransferase elevations and 60% of the patients with bilirubin elevations) of these events occurred during the first 3 months of treatment. In the pivotal phase II clinical trials (NP28761 and NP28673), two patients with Grade 3-4 AST/ALT elevations had documented drug induced liver injury by liver biopsy. Concurrent elevations in ALT or AST greater than or equal 3 times the ULN and total bilirubin greater than or equal 2 times the ULN, with normal alkaline phosphatase, occurred in one patient treated in Alectinib clinical trials.

Liver function, including ALT, AST, and total bilirubin should be monitored at baseline and then every 2 weeks during the first 3 months of treatment. Thereafter, monitoring should be performed periodically, since events may occur later than 3 months, with more frequent testing in patients who develop aminotransferase and bilirubin elevations. Based on the severity of the adverse drug reaction, Alectinib should be withheld and resumed at a reduced dose, or permanently discontinued (see section 4.2 of the SmPC).

Vision disorders were reported in both pivotal and supportive studies. Most common vision disorders are blurred vision and visual impairment. All events in Group 3 were of Grade 1 or 2, except for one serious Grade 3 event of retinal detachment.

Myalgia or musculoskeletal pain occurred in 31% of patients in pivotal phase II trials (NP28761, NP28673) with Alectinib. The incidence of Grade 3 myalgia/musculoskeletal pain was 1.2%. Dose modifications for myalgia/musculoskeletal pain were required in 0.8% of patients.

Elevations of CPK occurred in 46% of 219 patients with CPK laboratory data available in pivotal phase II trials (NP28761, NP28673) with Alectinib. The incidence of Grade 3 elevations of CPK was 5.0%. Median time to Grade 3 CPK elevation was 14 days. Dose modifications for elevation of CPK occurred in 4.0 % of patients (see sections 4.4 and 4.8 of the SmPC).

Patients should be advised to report any unexplained muscle pain, tenderness, or weakness. CPK levels should be assessed every two weeks for the first month of treatment and as clinically indicated in patients reporting symptoms. Based on the severity of the CPK elevation, Alectinib should be withheld, then resumed or dose reduced (see section 4.2 of the SmPC).

Symptomatic bradycardia can occur with Alectinib (see sections 4.4 and 4.8 of the SmPC). Heart rate and blood pressure should be monitored as clinically indicated. Dose modification is not required in case of asymptomatic bradycardia (see section 4.2 of the SmPC). If patients experience symptomatic bradycardia or life-threatening events, concomitant medicinal products known to cause bradycardia, as well as anti- hypertensive medicinal products should be evaluated and Alectinib treatment should be adjusted as described in sections 4.2 and 4.5 of the SmPC (P-gp substrates and BCRP substrates).

Assessment report EMA/197343/2017 Page 109/122 Photosensitivity to sunlight has been reported with Alectinib administration (see sections 4.4 and 4.8 of the SmPC). Patients should be advised to avoid prolonged sun exposure while taking Alectinib, and for at least 7 days after discontinuation of treatment. Patients should also be advised to use a broad-spectrum Ultraviolet A (UVA)/ Ultraviolet B (UVB) sun screen and lip balm (SPF ≥50) to help protect against potential sunburn.

Alectinib may cause foetal harm when administered to a pregnant woman. Women of childbearing potential must be advised to avoid pregnancy while on Alectinib. Female patients of child-bearing potential receiving Alectinib, must use highly effective contraceptive methods during treatment and for at least 3 months following the last dose of Alectinib (see sections 4.4, 4.6 and 5.3 of the SmPC). Female patients, who become pregnant while taking Alectinib or during the 3 months following the last dose of Alectinib must contact their doctor and should be advised of the potential harm to the foetus.

It is unknown whether alectinib and its metabolites are excreted in human milk. A risk to the newborn/infant cannot be excluded. Mothers should be advised against breast-feeding while receiving Alectinib.

Alectinib has minor influence on the ability to drive and use machines. Caution should be exercised when driving or operating machines as patients may experience symptomatic bradycardia (e.g., syncope, dizziness, hypotension) or vision disorders while taking Alectinib (see sections 4.7 and 4.8 of the SmPC).

This medicinal product contains lactose. Patients with rare hereditary problems of galactose intolerance, a congenital lactase deficiency or glucose-galactose malabsorption should not take this medicinal product.

The recommended daily dose (1200 mg) of Alectinib contains 2.1 mmol (or 48 mg) sodium. To be taken into consideration by patients on a controlled sodium diet.

There is no specific antidote for overdose with Alectinib. Patients who experience overdose should be closely supervised and general supportive care instituted.

From the safety database all the adverse reactions reported in clinical trials have been included in the Summary of Product Characteristics.

Additional safety data needed in the context of a conditional MA

The submission of the final CSR for the global Phase III comparative alectinib versus crizotinib ALEX study will provide more mature safety data in order to further confirm the safety profile of alectinib.

2.6.2. Conclusions on the clinical safety

The safety profile of alectinib is as expected for ALK inhibitors. Hepatotoxicity, anaemia, bradycardia, ILD and photosensitivity are observed. However, QT interval prolongation is not observed in patients being treated with alectinib, while it is observed in patients being treated with other ALK inhibitors. The discontinuation rate is only 5% and considered relatively low.

Overall, the safety profile of alectinib is acceptable and in line with other ALK inhibitors.

The CHMP considers the following measures necessary to address the missing safety data in the context of a conditional MA:

In order to further confirm the efficacy and safety of alectinib in the treatment of patients with ALK-positive NSCLC, the MAH should submit the clinical study report of the phase III study ALEX comparing alectinib versus crizotinib in treatment naïve patients with ALK-positive NSCLC. The clinical study report will be

Assessment report EMA/197343/2017 Page 110/122 submitted by 30 April 2018.

2.7. Risk Management Plan

Safety concerns

Table 69: Summary of the Safety Concerns as proposed by the applicant: Summary of safety concerns

Important identified risks  Interstitial lung disease/pneumonitis

 Hepatotoxicity

 Photosensitivity

 Bradycardia

 Severe myalgia and CPK elevations

Important potential risks  Embryo-foetal toxicity

Missing information  Long-term safety

 Treatment in lactating women

 Treatment in patients with moderate or severe hepatic impairment

 Treatment in patients with severe renal impairment

Pharmacovigilance plan

Table 70: Ongoing and planned additional pharmacovigilance studies/activities in the Pharmacovigilance Plan Study/activity Objectives Safety Status Date for submission Type, title and concerns of interim or final category (1-3) addressed reports

Assessment report EMA/197343/2017 Page 111/122 BO28984 Primary Objective Does not address Ongoing Study start date: a specific safety August 2014. Randomized, To evaluate and compare the efficacy concern. multicenter, Phase III, of alectinib compared to crizotinib in Provision of open-label study of patients with treatment-naive ALK- mature safety Primary CSR and alectinib versus positive advanced NSCLC, as data requested related documents: crizotinib in measured by investigator-assessed by the EMA. April 2018. treatment-naïve PFS. anaplastic lymphoma Secondary Objectives kinase-positive advanced non-small To evaluate and compare ORR, DOR, cell lung cancer (2) IRC-assessed PFS, OS.

To evaluate and compare the time to progression in the CNS on the basis of IRC review of radiographs by RECIST v1.1 and RANO criteria, as well as: To evaluate C-ORR in patients with CNS metastases who have measurable disease in the CNS at baseline; To assess C-DOR in patients who have a CNS Objective Response; To assess C-PR at 6, 12, 18, and 24 months on the basis of cumulative incidence.

To evaluate the safety and tolerability of alectinib compared to crizotinib.

To characterize the pharmacokinetics of alectinib and metabolite(s).

To evaluate and compare TTD in patient-reported lung cancer symptoms of cough, dyspnea (single item and multi-item subscales), chest pain, arm and shoulder pain, and fatigue as measured by the EORTC Quality of Life Questionnaire−Core (QLQ-C30) and the supplemental lung cancer module (QLQ-LC13) as well as a composite of three symptoms (cough, dyspnea, chest pain).

To evaluate and compare PROs of HRQoL, patient functioning, and side effects of treatment as measured by the EORTC QLQ-C30 and EORTC QLQ- LC13.

Assessment report EMA/197343/2017 Page 112/122 NP29783 Primary Objective Treatment of Ongoing Study start date: patients with December 2015. The effect of hepatic To assess the PK of alectinib in moderate and impairment on the subjects with hepatic impairment and Final CSR: Estimated severe hepatic pharmacokinetics of in matched healthy subjects after a December 2017 impairment alectinib: a multiple- single oral dose. (submission of report: center, open-label April 2018). Secondary Objective study following single oral dosing of alectinib To assess the PK of the major active to subjects with metabolite of alectinib, M4, and the hepatic impairment combined exposure of alectinib and and matched healthy M4 in subjects with hepatic subjects with normal impairment and in matched healthy hepatic function (3) subjects after a single oral dose.

To investigate safety and tolerability of alectinib in subjects with hepatic impairment and in matched healthy subjects.

Exploratory Objective

To evaluate the relationship, if any, between measures of hepatic impairment (e.g., Child-Pugh Scores, NCI Criteria for hepatic impairment categories, albumin concentration, etc.) and PK parameters for alectinib and/or M4, as appropriate.

ALK  Anaplastic Lymphoma Kinase; C-DOR  CNS Duration Of Response; CSR Clinical Study Report; C-ORR  CNS Objective Response

Rate; C-PR  CNS Progression Rates; DOR  Duration Of Response; EMA  European Medicines Agency; EORTC  European Organization for the Research and Treatment of Cancer; HRQoL  Health-Related Quality of Life; IRC  Independent Review Committee; NCI National

Cancer Institute; NSCLC  Non-Small Cell Lung Cancer; ORR  Objective Response Rate; OS  Overall Survival; PFS  Progression-Free

Survival; PK Pharmacokinetics; PROs  Patient Reported Outcomes; RANO  Revised Assessment in Neuro Oncology; RECIST  Response

Evaluation Criteria In Solid Tumors; TTD  Time to Deterioration.

Risk minimisation measures

The Table from the RMP section V.3 has been omitted as there is no required additional risk minimisation measures proposed.

Safety concern Routine risk Additional risk

minimization measures minimization measures Interstitial Lung Disease None SmPC sections: 4.2; 4.4; 4.8 (ILD)/Pneumonitis

Assessment report EMA/197343/2017 Page 113/122 Safety concern Routine risk Additional risk

minimization measures minimization measures

Hepatotoxicity SmPC sections: 4.2; 4.4; 4.8 None

Photosensitivity SmPC section: 4.4 None

Bradycardia SmPC sections: 4.2; 4.4; 4.8 None

Severe myalgia and SmPC sections: 4.2; 4.4; 4.8 None CPK elevations

Embryo-fetal toxicity SmPC sections: 4.4; 4.6 None Long-term safety No data available. None

Treatment in lactating SmPC section: 4.6 None women

Treatment in patients SmPC section: 4.2; 5.2 None with moderate or severe hepatic impairment

Treatment in patients SmPC section: 4.2; 5.2 None with severe renal impairment

Conclusion

The CHMP and PRAC considered that the risk management plan version 1.3 dated 15th December 2016 is acceptable.

2.8. Pharmacovigilance

Pharmacovigilance system

The CHMP considered that the pharmacovigilance system summary submitted by the applicant fulfils the requirements of Article 8(3) of Directive 2001/83/EC.

2.9. New Active Substance

The applicant compared the structure of alectinib with active substances contained in authorised medicinal products in the European Union and declared that it is not a salt, ester, ether, isomer, mixture of isomers, complex or derivative of any of them.

The CHMP, based on the available data, considers alectinib to be a new active substance as it is not a constituent of a medicinal product previously authorised within the European Union.

Assessment report EMA/197343/2017 Page 114/122 2.10. Product information

2.10.1. User consultation

The results of the user consultation with target patient groups on the package leaflet submitted by the applicant show that the package leaflet meets the criteria for readability as set out in the Guideline on the readability of the label and package leaflet of medicinal products for human use.

2.10.2. Additional monitoring

Pursuant to Article 23(1) of Regulation No (EU) 726/2004, Alecensa (alectinib) is included in the additional monitoring list as:

- It contains a new active substance which, on 1 January 2011, was not contained in any medicinal product authorised in the EU

- It is approved under a conditional marketing authorisation [REG Art 14(7)]

Therefore the summary of product characteristics and the package leaflet includes a statement that this medicinal product is subject to additional monitoring and that this will allow quick identification of new safety information. The statement is preceded by an inverted equilateral black triangle.

3. Benefit-Risk Balance

3.1. Therapeutic Context

The original proposed indication for alectinib was “adult patients with anaplastic lymphoma kinase (ALK)- positive, locally advanced or metastatic non-small cell lung cancer (NSCLC) who have progressed on or are intolerant to crizotinib”.

Intolerance to crizotinib as part of the label was questioned by the CHMP and it has been shown that only a very limited number of patients were in fact included in the studies based on crizotinib-intolerance (five in NP28761, none in NP28673). Consequently, the Applicant has proposed a revision of the indication wording to state “Alectinib as monotherapy is indicated for the treatment of adult patients with ALK-positive, advanced NSCLC previously treated with crizotinib”.

For a new ALKi to be used after failure on prior ALKi treatment it is clinically relevant to define whether resistance is related to secondary ALK mutations or other mechanisms underlying resistance; a priori activity e.g., in case of resistance due to activation of alternative signalling pathways bypassing ALK appears remote. Similarly, to what degree the secondary mutation affects the activity of the new compound is of importance. Such data are presently available to a very limited and non-informative extent. The applicant is recommended to provide final analysis on secondary ALK mutation positive/negative samples and correlation with clinical outcome.

Assessment report EMA/197343/2017 Page 115/122 3.1.1. Disease or condition

NSCLC is the leading cause of cancer-related mortality worldwide and represents a major health problem.

Approximately 5% of patients with NSCLC have tumours that contain an inversion in chromosome 2 that juxtaposes the 5' end of the echinoderm microtubule-associated protein-like 4 (EML4) gene with the 3' end of the ALK gene, resulting in the novel fusion oncogene EML4-ALK. EML4-ALK fusion, and ALK fusions with other genes (Kinesin Family member 5B [KIF5B]), KLC1 and others define a distinct clinicopathologic subset of NSCLC.

Patients with ‘ALK-positive’ tumours (tumours harboring a rearranged ALK gene/fusion protein) tend to have specific clinical features, including never or light smoking history, younger age, NSCLC with adenocarcinoma histology, and are highly sensitive to therapy with ALK inhibitors.

ALK positive tumours are addicted to ALK signalling meaning that inhibition leads to rapid cell death and early tumour responses. Loss of activity invariably follows and may be related to secondary ALK resistance mutations, but also fusion gene amplification and activation of alternative signalling pathways.

3.1.2. Available therapies and unmet medical need

After the licensure of crizotinib and ceritinib, alectinib is the third ALK inhibitor (ALKi) submitted for marketing authorisation for the treatment of ALK+ NSCLC.

The kinase inhibitory profiles of these compounds are clearly different; alectinib inhibits ALK and RET, crizotinib has a broader profile: ALK, ROS1, MET, etc. and ceritinib ALK, ROS1, IGF1R (for details please refer to the non-clinical pharmacology section). Of potential clinical relevance for the activity in different tumour types/subsets, apart from ALK is inhibition of RET, ROS1, MET and IGF1R, but a broader inhibition profile is also of importance from a safety perspective.

3.1.3. Main clinical studies

The Applicant initially applied for full approval for alectinib based on two single arm Phase I/II studies, NP28761 (n=87, US/ Canada) and NP28673 (n=138, global) supported by data from a pre-planned interim analysis from the 1st line, Japanese, crizotinib comparative study (J-ALEX).

3.2. Favourable effects

Whilst primary endpoint in both studies was ORR by IRC assessment (RECIST 1.1 with confirmation), NP28673 also had a co-primary endpoint of ORR by IRC assessment in patients pre-treated with chemotherapy.

The studies met their primary endpoints with ORR, in the updated analyses; IRC 52 % and 51 % in NP28761 and NP 28673, respectively. The co-primary endpoint ORR by IRC for the patients pre-treated with chemotherapy in NP28673 was 45 %. About 16 % had progressive disease as best response, similar in both studies.

For both pivotal studies, ORR by subgroups (age, gender, race, ECOG PS, CNS metastasis, smoking status, baseline body weight, and prior chemotherapy use), were generally consistent with the ORR in the overall population.

Assessment report EMA/197343/2017 Page 116/122 Updated median PFS was 8.9 months in both studies, at event rates of 67 and 71%. Data are thus highly likely to be stable after prolonged follow up.

Out of the 50 patients with measurable CNS lesions in the overall efficacy population, the overall response rate was 64%, thereof CR in 22% (IRC).

In the supportive Japanese study J-ALEX in ALKi naïve patients alectinib 300 mg BID was compared with crizotinib in the EU licensed dose. In the interim anaysis at event rates of 24% vs. 56% the PFS HR was 0.34 (p<0.0001)(IRC) and response rates were 92% vs. 79%.

3.3. Uncertainties and limitations about favourable effects

The major uncertainty in the assessment of optimised benefit of alectinib is absence of informative data on mechanisms underlying resistance to crizotinib; secondary ALK resistance mutations, fusion gene amplification and activation of alternative signalling pathways as such data may guide the proper use of alectinib in the now targeted indication.

Out of 225 patients enrolled in the studies, baseline plasma samples were available from 178 individuals and now data are presented from 49 samples.

Based on the existence of alternative mechanisms of resistance unlikely to be responsive to alectinib therapy, a higher ORR would be expected in case of tumours with secondary ALK mutation positive samples, but ORR was similar (1/3) in patients with ALK mutation positive and negative liquid samples and apparently lower than in the full study population (1/2). Whether the analysed 49 samples are representative for the full study population could obviously be questioned and the sensitivity of liquid biopsies to detect tumour tissue mutations is unknown and circulating tumour DNA may relate to tumour burden/invasiveness.

Of some interest, PD as best response was more commonly reported in patients with ALK mutation negative samples than ALK mutation positive samples (12/35 vs. 2/21).

Patients with CNS progression only should have been excluded from the group of patients with mutation negative liquid biopsies due to the postulated poorer uptake of crizotinib in CNS metastases. This will not be further pursued at this stage.

Available data cannot be used to guide the use of alectinib, second line to crizotinib.

The final study report on secondary ALK-mutation positive/negative samples and correlation with clinical outcome should be submitted (please refer to Section 7 Proposed list of recommendation).

3.4. Unfavourable effects

Tolerability is considered possible to estimate with reasonable certainty at the dose proposed for marketing, even though the number of patients is low, roughly 250 individuals treated for a for median about 1 year.

The most common side effects are constipation, oedema (including peripheral oedema, generalised oedema, eyelid oedema, periorbital oedema), myalgia (including myalgia and musculoskeletal pain) and nausea.

Tolerability in terms of discontinuations (6%, related or not) and dose reductions (about 15%, mainly laboratory events among likely related) are considered favourable when contextualised. This refers not only to ALKi but to anti-cancer therapy in general.

Assessment report EMA/197343/2017 Page 117/122 Hepatic events and CPK/myalgia events need to monitored in clinical practice and proper guidance is provided in the SmPC.

3.5. Uncertainties and limitations about unfavourable effects

Events such as pulmonary embolism and haemorrhages always occur in patients with advanced malignancies, including ALK+NSCLC. Whether treatment with alectinib increases or decreases the risk cannot be assessed, but good control of the underlying disease should decrease the disease related event rate.

As only single arm or active, reference controlled trials are possible, the absolute risk increase or decrease cannot be assessed. The Japanese, first-line, crizotinib comparative trial is considered informative and indicates a favourable safety profile, with the caveat that the dose of alectinib was lower (estimated exposure reduced about 35%). The only signal of increased toxicity related to alectinib was myalgia/CPK increase.

There is another ongoing first-line, crizotinib comparative trial at the “EU dose”.

3.6. Effects Table

Table 71: Effects Table for Alectinib in ALK+ NSCLC Effect Short data cut-off: Updated data cut- Uncertainties/ Description off Strength of evidence Favourable Effects NP28761 24 October 2014 22 January 2016 ORR by IRC Proportion of 48 % 52 % Overall, the updated patients achieving results confirm the [95% CI] best response of CR findings in the primary or PR (%) (35.6%, 60.2%) (39.7%, 64.6%) analyses. CNS ORR by Proportion of 69 % 75 % IRC patients achieving [95% CI] CR or PR of BL CNS (41.3%, 89.0%) (47.6%, 92.7%) lesions CNS DOR NE 11 (months) [95% CI] (NE, NE) (5.8, NE) PFS IRC 6 8 (months) (5.5, NE) (6, 13) [95% CI] OS (months) NE 23 [95% CI] (10, NE) (17, NE) NP28673 08 January 2015 01 February2016 ORR by IRC 50 % 51 % Overall, the updated [95% CI] (40.8%, 59.2%) (41.6%, 59.9%) results confirm the findings in the primary analyses. CNS ORR by 57 % 59 % IRC (41%,75%) [95% CI] (39%, 74%) CNS DOR 9 11 (months) (5.8, NE) (7.1, NE) [95% CI] PFS IRC 9 9 (months) (5.6, 11.3) (5.6, 13) [95% CI] OS (months) NE 26 [95% CI] (14.6, NE) (21.5,NE)

Assessment report EMA/197343/2017 Page 118/122 Unfavourable Effects: Group 3 (pooled safety dataset: NP28761 (Phase I/II), NP28673 (Phase II) and MDZ N=253 SCS Update Comments At least 1 AE 97% 99 % AE grade 3-5 28% 39.5 SAE 16% 22 % “Related” SAE 5% 6 % AE leading to dose 26% 33 % reduction/interruption AE leading to withdrawal 5% 6 % Number of deaths 17% 41.5 % (all causes) Number of AE with fatal outcome 2% 2.8 % Hepatocellular AEs Grade >=3 6% Visions disorders 12% Photosensitivity 32% GI AEs 55% 31% had GI events considered by the Grade 3 3 % investigator to be related to study drug. The relevance of the high dose of SLS is unknown. Kidney AEs 15% Bradycardia 5.9% 7.9 %

Abbr: NE – Not estimated, SCS – Summary of Clinical Safety

3.7. Benefit-risk assessment and discussion

3.7.1. Importance of favourable and unfavourable effects

Available clinical data illustrate the likely benefit of targeting the driver mutation of the malignancy only. Very high activity in terms of ORR and superior PFS compared with crizotinib has thus been reported in the first- line J-ALEX trial. This is interpreted as indicative of anti-tumour activity not being restricted by dose limiting side effects not related to inhibition of ALK (and RET). It is easily understood why the Applicant views the currently proposed indication as a first step only, with an ongoing comparative study versus crizotinib at the EU recommended dose. In this context it is of interest to notice that fold change in relation to the wild type translocation is similar to fold change for crizotinib for crizotinib resistance related mutations.

The reported benefit in terms of ORR and PFS after failure on crizotinib is considered clinically meaningful. Actually, it has been shown that tumour responses at much lower rates than 50% are associated with measurable symptomatic benefit in clinical studies. Similarly, delayed progression by imaging is associated with delayed symptomatic progression of similar magnitude. This means that treatment with alectinib is highly likely to provide symptomatic benefit on a group level. At the reported high frequency of tumour responses in case of brain metastases, symptomatic benefit is expected also in this group of patients, very commonly encountered in ALK+ NSCLC.

Reported adverse reactions are considered manageable and reversible and unlikely to affect the tolerability of the product in more than a small proportion of patients. There are obvious caveats in relation to hepatic and muscular events and, by convention, ILD. Proper guidance is provided in the SmPC.

To contextualise, the ORR is of similar magnitude to that reported for ceritinib after crizotinib failure. Despite brain metastases in a high proportion of patients in the ceritinib studies, only 11 patients had brain metastasis assigned as target lesions resulting in poor estimates of ORR: 45.5% (95% CI: 16.7, 76.6)

Assessment report EMA/197343/2017 Page 119/122 (EPAR). Selection “bias” should also be taken into account when small proportions of lesions are assigned as target lesions.

It should also be noticed that tumours addicted to a driver mutation tend to show higher response rates to conventional chemotherapy. In the confirmatory crizotinib study, second-line to chemotherapy, the response rate in patients treated with pemetrexed or docetaxel was 20% (vs. 65% in the crizotinib arm).

Altogether benefit clearly outweighs the risk in the target population. As activity is unlikely in case of resistance to crizotinib if related to signalling bypassing ALK, there is likely room to improve B/R by proper patient selection. This issue may not be possible to resolve within this procedure, and should be a post approval commitment.

3.7.2. Balance of benefits and risks

Benefit-risk is considered positive.

3.7.3. Additional considerations on the benefit-risk balance

Alectinib has shown a clinically relevant ORR rate in patients that have previously been treated with crizotinib.

Conditional marketing authorisation

As comprehensive data on the product are not available, a conditional marketing authorisation was proposed by the CHMP during the assessment, after having consulted the applicant.

The CHMP considered that ALECENSA falls under the scope of Article 2 of Commission Regulation (EC) No. 507/2006 as eligible for a Conditional Marketing Authorisation as it belongs to:

b) Medicinal products which aim at the treatment, the prevention or the medical diagnosis of seriously debilitating diseases or life-threatening diseases.

Furthermore, the CHMP considers that the product fulfils the requirements for a conditional marketing authorisation:

 The benefit-risk balance is positive. The ORR of 51-52% obtained with alectinib in pivotal studies is significant in patients with relapsed and refractory ALK+ NSCLC, whose prior therapy included crizotinib. Together with an acceptable safety-profile in patients in the proposed indication, the benefit-risk balance is considered positive.

 It is likely that the applicant will be able to provide comprehensive data. The applicant will provide further comprehensive clinical data to confirm efficacy and safety of alectinib in the proposed indication. More specifically the applicant will provide:

 Study ALEX, a phase 3, two-arm, randomized study, comparing alectinib and crizotinib in the first line.

This phase 3 study is not in the exact same setting where alectinib is authorised, but it will provide comparative efficacy and safety data for alectinib versus crizotinib.

 Unmet medical needs will be addressed. In view of the positive benefit-risk balance, alectinib fulfils an

Assessment report EMA/197343/2017 Page 120/122 unmet medical need for patients with poor long-term prognosis and limited treatment alternatives. Although alectinib has not a new mechanism of action, the safety profile is manageable, and treatment with alectinib was associated with high ORR and durable responses in patients with CNS metastases. In this context it should be noted that ceritinib (Zykadia) licensed in the same indication as currently applied for by the Applicant (i.e. after prior crizotinib treatment), is conditionally approved at the time of the CHMP opinion.

 The benefits to public health of the immediate availability outweigh the risks inherent in the fact that additional data are still required.

In view of the favourable benefit-risk profile, the high ORR, especially in patients with CNS metastases, the immediate availability of Alecensa on the market outweighs the risk inherent in the fact that additional data are still required.

3.8. Conclusions

The overall B/R of Alecensa is positive.

4. Recommendations

Outcome

Based on the CHMP review of data on quality, safety and efficacy, the CHMP considers by consensus that the risk-benefit balance of Alecensa is favourable in the following indication:

Alecensa as monotherapy is indicated for the treatment of adult patients with anaplastic lymphoma kinase (ALK) positive advanced non-small cell lung cancer (NSCLC) previously treated with crizotinib.

The CHMP therefore recommends the granting of the conditional marketing authorisation subject to the following conditions:

Other conditions or restrictions regarding supply and use

Medicinal product subject to restricted medical prescription (see Annex I: Summary of Product Characteristics, section 4.2).

Conditions and requirements of the marketing authorisation

Periodic Safety Update Reports

The requirements for submission of periodic safety update reports for this medicinal product are set out in the list of Union reference dates (EURD list) provided for under Article 107c(7) of Directive 2001/83/EC and any subsequent updates published on the European medicines web-portal.

The marketing authorisation holder shall submit the first periodic safety update report for this product within 6 months following authorisation.

Assessment report EMA/197343/2017 Page 121/122 Conditions or restrictions with regard to the safe and effective use of the medicinal product

Risk Management Plan (RMP)

The MAH shall perform the required pharmacovigilance activities and interventions detailed in the agreed RMP presented in Module 1.8.2 of the marketing authorisation and any agreed subsequent updates of the RMP.

An updated RMP should be submitted:

 At the request of the European Medicines Agency;

 Whenever the risk management system is modified, especially as the result of new information being received that may lead to a significant change to the benefit/risk profile or as the result of an important (pharmacovigilance or risk minimisation) milestone being reached.

Specific Obligation to complete post-authorisation measures for the conditional marketing authorisation

This being a conditional marketing authorisation and pursuant to Article 14(7) of Regulation (EC) No 726/2004, the MAH shall complete, within the stated timeframe, the following measures:

Description Due date

In order to further confirm the efficacy and safety of alectinib in the treatment of 30 April 2018 patients with ALK-positive NSCLC, the MAH should submit the clinical study report of the phase III study ALEX comparing alectinib versus crizotinib in treatment naïve patients with ALK-positive NSCLC.

Conditions or restrictions with regard to the safe and effective use of the medicinal product to be implemented by the Member States

Not applicable.

New Active Substance Status

Based on the CHMP review of the available data, the CHMP considers that alectinib is considered to be a new active substance as it is not a constituent of a medicinal product previously authorised within the European Union.

Assessment report EMA/197343/2017 Page 122/122