208692Orig1s000

CENTER FOR DRUG EVALUATION AND RESEARCH

APPLICATION NUMBER:

208692Orig1s000

MEDICAL REVIEW(S) CLINICAL REVIEW

Application Type Supplemental NDA Application Number(s) 208692 Priority or Standard Priority

Submit Date(s) Dec 22, 2015 Received Date(s) Dec 22, 2015 PDUFA Goal Date June 22, 2016 Division / Office Office of Hematology & Oncology Products

Reviewer Name(s) Michael Brave (Efficacy) Harpreet Singh (Safety) Julia A. Beaver (TL) Review Completion Date April 15, 2016

Established Name Cabozantinib (Proposed) Trade Name Cabometyx Therapeutic Class Kinase inhibitor Applicant Exelixis, Inc.

Formulation(s) Tablet Dosing Regimen 60 mg once daily Indication(s) Treatment of patients with advanced renal cell carcinoma who have received prior anti- angiogenic therapy Intended Population(s) \HDUV RI DJH

Reference ID: 3917720 Template Version: March 6, 2009

APPEARS THIS WAY ON ORIGINAL

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table of Contents

1 RECOMMENDATIONS/RISK BENEFIT ASSESSMENT...... 10 1.1 Recommendation on Regulatory Action ...... 10 1.2 Risk Benefit Assessment...... 10 1.3 Recommendations for Postmarket Risk Evaluation and Mitigation Strategies . 12 1.4 Recommendations for Postmarket Requirements and Commitments ...... 12 2 INTRODUCTION AND REGULATORY BACKGROUND ...... 12 2.1 Product Information ...... 13 2.2 Tables of Currently Available Treatments for Proposed Indications ...... 13 2.3 Availability of Proposed Active Ingredient in the United States ...... 14 2.4 Important Safety Issues with Consideration to Related Drugs...... 15 2.5 Summary of Presubmission Regulatory Activity Related to Submission ...... 16 2.6 Other Relevant Background Information ...... 18 3 ETHICS AND GOOD CLINICAL PRACTICES...... 18 3.1 Submission Quality and Integrity ...... 18 3.2 Compliance with Good Clinical Practices ...... 18 3.3 Financial Disclosures...... 22 4 SIGNIFICANT EFFICACY/SAFETY ISSUES RELATED TO OTHER REVIEW DISCIPLINES ...... 22 4.1 Chemistry Manufacturing and Controls ...... 22 4.2 Clinical Microbiology...... 26 4.3 Preclinical Pharmacology/Toxicology ...... 26 4.4 Clinical Pharmacology...... 26 4.4.1 Mechanism of Action...... 26 4.4.2 Pharmacodynamics...... 26 4.4.3 Pharmacokinetics...... 26 5 SOURCES OF CLINICAL DATA...... 27 5.1 Tables of Studies/Clinical Trials ...... 27 5.2 Review Strategy ...... 29 5.3 Discussion of Individual Studies/Clinical Trials...... 30 6 REVIEW OF EFFICACY...... 44 Efficacy Summary...... 44 6.1 Indication ...... 45 6.1.1 Methods ...... 45 6.1.2 Demographics...... 45 6.1.3 Subject Disposition ...... 50 6.1.4 Analysis of Primary Endpoint(s) ...... 54

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Patients in the primary analysis were censored for PFS if they were determined to have progressive disease by ICR or if they died. Table 19 shows the reasons for censoring of patients...... 55 6.1.5 Analysis of Secondary Endpoints(s)...... 56 6.1.6 Other Endpoints ...... 59 6.1.7 Subpopulations ...... 62 6.1.8 Analysis of Clinical Information Relevant to Dosing Recommendations .... 66 6.1.9 Discussion of Persistence of Efficacy and/or Tolerance Effects...... 67 6.1.10 Additional Efficacy Issues/Analyses...... 68 7 REVIEW OF SAFETY...... 73 7.1 Methods...... 75 7.1.1 Studies/Clinical Trials Used to Evaluate Safety ...... 75 7.1.2 Categorization of Adverse Events...... 75 7.1.3 Pooling of Data Across Studies/Clinical Trials to Estimate and Compare Incidence...... 75 7.2 Adequacy of Safety Assessments ...... 75 7.2.1 Overall Exposure at Appropriate Doses/Durations and Demographics of Target Populations...... 75 7.2.2 Explorations for Dose Response...... 76 7.2.3 Special Animal and/or In Vitro Testing ...... 76 7.2.4 Routine Clinical Testing ...... 76 7.2.5 Metabolic, Clearance, and Interaction Workup ...... 77 7.2.6 Evaluation for Potential Adverse Events for Similar Drugs in Drug Class .. 77 7.3 Major Safety Results ...... 77 7.3.1 Deaths...... 77 7.3.2 Nonfatal Serious Adverse Events ...... 85 7.3.3 Dropouts and/or Discontinuations ...... 86 7.3.4 Significant Adverse Events ...... 93 7.3.5 Submission Specific Primary Safety Concerns ...... 95 7.4 Supportive Safety Results ...... 117 7.4.1 Common Adverse Events ...... 117 7.4.2 Laboratory Findings ...... 122 7.4.3 Vital Signs...... 125 7.4.4 Electrocardiograms (ECGs) ...... 126 7.4.5 Special Safety Studies/Clinical Trials...... 126 7.4.6 Immunogenicity...... 126 7.5 Other Safety Explorations...... 126 7.5.1 Dose Dependency for Adverse Events ...... 126 7.5.2 Time Dependency for Adverse Events...... 127 7.5.3 Drug-Demographic Interactions ...... 127 Reviewer comment: ...... 132 7.5.4 Drug-Disease Interactions...... 132 7.5.5 Drug-Drug Interactions...... 132

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

7.6 Additional Safety Evaluations ...... 132 7.6.1 Human Carcinogenicity...... 132 7.6.2 Human Reproduction and Pregnancy Data...... 133 7.6.3 Pediatrics and Assessment of Effects on Growth ...... 133 7.6.4 Overdose, Drug Abuse Potential, Withdrawal and Rebound...... 133 7.7 Additional Submissions / Safety Issues...... 133 8 POSTMARKET EXPERIENCE...... 134 9 APPENDICES ...... 135 9.1 Literature Review/References ...... 135 9.2 Labeling Recommendations...... 135 9.3 Advisory Committee Meeting...... 136

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table of Tables

Table 1: FDA approved Therapies Indicated for Second-Line Treatment of Patients with Advanced Renal-Cell Carcinoma...... 14 Table 2: Warnings and Precautions and Select Adverse Reactions of Kinase Inhibitors Approved for Patients with Renal Cell Carcinoma ...... 16 Table 3. Overview of Important Protocol Deviations Excluding Eligibility Criteria...... 19 Table 4: Concordance between eCRF and IVRS stratification data...... 20 Table 5: Summary of OSI Findings ...... 21 Table 6: Summary of Clinical Safety Studies of Cabozantinib Included in the Safety Analysis ...... 28 Table 7 XL184-308 Allowable Dose Reductions of Cabozantinib ...... 35 Table 8. XL184-308 Analysis Populations...... 37 Table 9. XL184-308 Key Demographic Characteristics...... 46 Table 10. XL184-308 Cancer History and Baseline Status ...... 46 Table 11. XL184-308 Prior Anti-Cancer Therapies ...... 48 Table 12. XL184-308 Prior Chemotherapies...... 49 Table 13. XL184-308 Subject Disposition ...... 50 Table 14. XL184-308 Grade 3-4 AEs Reported by Patients who Withdrew (ITT pop.).. 51 Table 15. XL184-308 Subject Disposition by Age ...... 52 Table 16. XL184-308 Subject Disposition by Sex ...... 52 Table 17. XL184-308 Concomitant and Subsequent Chemotherapy ...... 53 Table 18. XL184-308 Progression-Free Survival per IRC (PITT Population) ...... 54 Table 19. XL184-308 Patient Censoring in PFS Analysis (PITT Population) ...... 55 Table 20. XL184-308 Overall Survival (ITT Population) ...... 57 Table 21. XL184-308 Tumor Response per RECIST 1.1 (ITT Population)...... 59 Table 22. XL184-308 Duration of Response (IRC-Determined; ITT Population)...... 59 Table 23. XL184-308 Time to Tumor Response per RECIST 1.1 (ITT Population)...... 60 Table 24. XL184-308 Bone Scan Response (IRC-Determined; ITT Population)...... 61 Table 25. XL184-308 Post-Randomization SREs (IRC-Determined; ITT Pop.)...... 61 Table 26. XL184-308 Patients with IRC-Determined rPD...... 67 Table 27. XL184-308 Outcome of Treatment Beyond IRC-Determined rPD...... 68 Table 28. XL184-308 Patients with Investigator-Determined rPD ...... 68 Table 29. XL184-308 Outcome of Treatment Beyond Investigator-Determined rPD..... 68 Table 30. XL184-308 Sensitivity Analyses of PFS (PITT Population) ...... 69 Table 31. XL184-308 Sensitivity Analyses of PFS (ITT Population)...... 69 Table 32. XL184-308 PFS (IRC-Determined; ITT Population) ...... 70 Table 33. XL184-308 Investigator-Determined PFS (PITT Population)...... 71 Table 34. XL184-308 Comparison of Progression Events (PITT population)...... 72 Table 35. XL184-308 PFS (IRC-Determined, First 375 Patients As-Treated) ...... 73 Table 36: XL184-308 Summary of Adverse Events (Safety Population) ...... 74 Table 37: XL184-308 Summary of Exposure (All Treated Population)...... 76 Table 38: XL184-308 Summary of Deaths (All Treated Population)...... 77

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 39 : XL184-308 Deaths for Other than Disease Progression Cabozantinib Group Up to 30 Days of Last Dose (All Treated Population)...... 79 Table 40: XL184- 6HULRXV $GYHUVH (YHQWV 5HSRUWHG LQ &DER]DQWLQE *URXS (All Treated Population) ...... 86 Table 41: XL184-308 Cabozantinib Dose Reductions...... 87 Table 42: XL184- $(V 7KDW /HG WR 6WXG\ 7UHDWPHQW 'LVFRQWLQXDWLRQ LQ RI Subjects in Either Arm (All Treated Population)...... 89 Table 43: XL184- $(V WKDW /HG WR 'RVH 5HGXFWLRQ LQ RI 6XEMHFWV LQ (LWKHU $UP (All Treated Population) ...... 91 Table 44: XL184- $GYHUVH (YHQWV WKDW /HG WR 'RVH ,QWHUUXSWLRQV LQ RI 6XEMHFWV in Either Treatment Arm (All Treated Population) ...... 91 Table 45: XL184- $GYHUVH (YHQWV 7KDW /HG WR 'RVH 0RGLILFDWLRQ LQ RI Subjects in Cabozantinib Arm (All Treated Population) ...... 92 Table 46: XL184- *UDGH DQG $GYHUVH (YHQWV LQ &DER]DQWLQLE *URXS $OO Treated Population) ...... 94 Table 47 Safety for Medically Important Cabozantinib-Related Adverse Events across Tumor Types...... 97 Table 48: XL184-308 Incidence of Toxicities Associated with VEGFR-TKI/ VEGF Inhibition (All Treated Population)...... 99 Table 49: XL184-308 Incidence of Hypertension Adverse Events (All Treated Population)...... 100 Table 50: XL184-308 Incidence of Diarrhea (All Treated Population) ...... 102 Table 51: XL184-308 Patients with Diarrhea Serious Adverse Events (All Treated Population)...... 103 Table 52 XL184-308 Incidence of Palmar-Plantar Erythrodysaesthesia Syndrome (All Treated Population) ...... 104 Table 53 XL184-308 Incidence of Dermatologic Toxicities (All Treated Population)... 105 Table 54 : XL184-308 Incidence of Selected Renal Failure and Proteinuria Adverse Events (All Treated Population) ...... 106 Table 55: XL184-308 Incidence of Adverse Events in the Hepatobiliary Disorders SOC (All Treated Population) ...... 108 Table 56: XL184-308 Incidence of Hemorrhage (All Treated Population) ...... 110 Table 57: XL184-308 Incidence of Arterial Thrombotic Events (All Treated Population) ...... 111 Table 58: XL184-308 Patients with Arterial Thrombotic Events SAE’s (Cabozantinib) 112 Table 59: XL184-308 Incidence of Venous and Mixed Thrombotic Events (All Treated Population)...... 113 Table 60: XL184-308 Patients with Venous and Mixed Thrombotic Events SAE’s (Cabozantinib)...... 114 Table 61: XL184-308 Summary of System Organ Class (All Treated Population) ...... 118 Table 62: XL184-308 Treatment Emergent Adverse Events Reactions occurring in &DER]DQWLQLE *URXS $OO 7UHDWHG 3RSXODWLRQ ...... 119 Table 63: XL184-308 7UHDWPHQW (PHUJHQW $GYHUVH (YHQWV E\ +LJK /HYHO 7HUP (All Treated Population) ...... 121

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 64: XL184-308 Treatment Emergent $GYHUVH (YHQWV E\ +LJK /HYHO *URXS Term (All Treated Population)...... 122 Table 65 XL184-308 Treatment Emergent Laboratory AbnorPDOLWLHV 2FFXUULQJ LQ of Patients (Cabozantinib)...... 123 Table 66: XL184-308 Potential DILI Cases by Laboratory Screening Criteria (All Treated Population)...... 124 Table 67 XL184- ,QFLGHQFH RI :HLJKW /RVV RI %DVHOLQH %RG\ :HLJKW $OO Treated Population) ...... 126 Table 68 XL184-308 Time to First Occurrence of Submission Specific Adverse Events ...... 127 Table 69 XL184-308 Treatment Emergent Adverse Events by Subgroup (All Treated Population)...... 128 Table 70 XL184-308 Treatment Emergent Events Leading to Dose Reduction by Subgroup (All Treated Population)...... 130 Table 71 XL184-308 Treatment Emergent Serious Adverse Events by Subgroup...... 131

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table of Figures

Figure 1. Trial XL184-308 Progression Free Survival per IRC (PITT Population) ...... 55 Figure 2. XL184-308 Overall Survival (2nd Interim Analysis; ITT Population)...... 58 Figure 3. XL184-308 Duration of Response (IRC-Determined; ITT Population)...... 60 Figure 4. XL184-308 Subgroup Analyses of PFS (IRC Determined, PITT Pop.)...... 63 Figure 5. XL184-308 Subgroup Analyses of PFS (IRC-Determined, ITT Pop.)...... 64 Figure 6. XL184-308 Subgroup Analyses of OS (Unplanned 2nd Interim, ITT Pop.)...... 65 Figure 7. XL184-308 PFS, IRC-Determined (ITT Pop.)...... 70 Figure 8. XL184-301 Inv.-Determined PFS, Clinical or Radiographic (PITT Pop.)...... 71 Figure 9. XL184-301 Inv.-Determined PFS, Clinical or Radiographic (PITT Pop.)...... 72 Figure 10 XL184-308 Treatment Emergent Adverse Events by Subgroup (All Treated Population)...... 129 Figure 11 XL184-308 Treatment Emergent Events Leading to Dose Reduction by Subgroup (All Treated Population)...... 130 Figure 12 XL184-308 Treatment Emergent Serious Adverse Events by Subgroup .... 132

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

1 Recommendations/Risk Benefit Assessment

1.1 Recommendation on Regulatory Action

The reviewers recommend approval of cabozantinib for the treatment of adult patients with advanced renal cell carcinoma who have received prior anti-angiogenic therapy at a dose of 60 mg orally daily.

The reviewers do not recommend any clinical postmarketing commitments (PMC) or postmarketing requirements (PMR).

The clinical data reviewed in support of this NDA are from a multicenter, randomized, open-label trial (XL184-308) in patients with advanced renal carcinoma who have had at least one prior VEGF therapy and demonstrate a statistically significant and clinically meaningful benefit in progression free survival (PFS) and overall survival (OS) compared to everolimus.

1.2 Risk Benefit Assessment

Renal cell carcinoma is a serious and life-threatening condition. In 2015, it is estimated that there will be 61,560 new cases and 14,080 deaths from RCC in the U.S.1

The current standard of care for advanced RCC patients whose disease has progressed on or who are resistant to VEGFR-TKI therapy is treatment with everolimus or axitinib.2,3 Sorafenib is also recommended in this setting based on a study in patients who had progressed on a prior systemic (mainly cytokine-based) therapy.4 Everolimus is the most frequently used second-line therapy following a VEGFR-TKI in patients with RCC. However, these approved second-line options for patients (everolimus and axitinib) are limited by modest PFS benefit and a lack of improvement in OS. In November 2015, nivolumab was approved based on improved OS but no PFS benefit as a second-line treatment.5

Cabozantinib is an orally bioavailable multi-targeted kinase inhibitor with activity against MET (hepatocyte growth factor receptor protein) and VEGFR2 (vascular endothelial growth factor receptor), as well as other receptors that have been implicated in tumor pathobiology, including RET (glial cell derived neurotropic factor rearranged during transfection), KIT (stem cell factor receptor), AXL (GAS6 receptor), and FLT (Fms-like tyrosine kinase-3). Cabozantinib suppresses MET and VEGFR2 signaling, inducing apoptosis of endothelial and tumor cells.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

The recommendation for approval of cabozantinib is based on the results of Trial XL- 184308, a multicenter, open-label, international, randomized (1:1) trial of cabozantinib versus everolimus in adult patients with advanced renal cell carcinoma who have had one prior VEGF therapy. Patients were stratified by the number of prior of VEGFR- targeting therapies and number of risk factors per Memorial Sloan-Kettering Cancer Center prognostic criteria for previously treated patients with RCC.

Patients (n=658) were randomized and treated with either cabozantinib 60 mg orally daily (n=330) or everolimus 10 mg orally daily (n=328) until disease progression or unacceptable toxicity. Patients on both arms who had disease progression could continue treatment at the discretion of the investigator. The primary endpoint was PFS among the first 375 randomized patients (PITT population), by IRC-determined RECIST criteria. There was a statistically significant improvement in PFS for patients in the cabozantinib arm compared with the everolimus arm, with a 3.6 month difference in median PFS (7.4 vs. 3.8 months). The hazard ratio adjusted for stratification factors was 0.59 (95% CI: 0.46, 0.76; stratified log-rank p-value < 0.0001) per eCRF and 0.58 (95% CI: 0.45, 0.74; stratified log-rank p-value < 0.0001) per IVRS. Key secondary endpoints were OS and objective response rate (ORR) in the Intent-to-Treat population (ITT). Median OS in the ITT population was 21.4 and 16.5 months in the cabozantinib and everolimus arms, respectively [HR 0.66 (95% CI: 0.53, 0.83); p=0.0003]. Confirmed ORR was 17% (95% CI: 13, 22) in the cabozantinib arm and 3% (95% CI: 2, 6) in the everolimus arm.

Safety was evaluated in 331 patients treated with cabozantinib. The most common (greater than or equal to 20%) adverse reactions included diarrhea, fatigue, nausea, vomiting, constipation, abdominal pain, decreased appetite, hypertension, rash, palmar- plantar erythrodysesthesia, weight decreased, and dygeusia. Serious adverse events were reported in 40% of patients. The most common serious adverse events (greater than or equal to 2%) were abdominal pain, pleural effusion, diarrhea, nausea, and anemia. Fifteen deaths were reported within 30 days of the last cabozantinib dose. Eight of the deaths were attributed to progressive disease, five deaths were due to cognitive deterioration, pneumonia, hemorrhage, cardiac failure, and urosepsis respectively, and two deaths were due to sudden cardiac death. Sixty percent of patients treated with cabozantinib had at least one dose reduction while on study. Ten percent of patients in the cabozantinib group required treatment discontinuation and seventy percent required dose modification, including dose interruption or dose reduction. The most frequent AEs leading to dose reduction of cabozantinib were diarrhea (16%), palmar-plantar erythrodysesthesia syndrome (11%), fatigue (10%), and hypertension (7.6%).

Given the rate of dose reductions and modifications in Trial XL184-308, the requirement of a dose finding study was considered by the review team. However, modeling and simulations indicated that lower starting doses such as 40 mg or 20 mg daily could possibly compromise efficacy though incidences of dose reductions may be lower than

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

at the 60 mg dose. In addition, despite the rate of dose reductions in the cabozantinib group, the rate of treatment discontinuations due to adverse events was only 10%, suggesting that most AE’s were appropriately managed with dose interruptions and supportive measures. The dose selection of 60 mg daily is adequate based on exposure-response analyses, and a safety profile which is acceptable for the patient population.

Cabozantinib demonstrated superiority to everolimus with a statistically significant improvement in PFS, OS and ORR. The improvement in median PFS and OS is considered clinically meaningful. Cabozantinib has a favorable benefit-risk profile for treatment of adult patients with advanced renal cell carcinoma who have received prior anti-angiogenic therapy.

1.3 Recommendations for Postmarket Risk Evaluation and Mitigation Strategies

There are no safety issues identified at this time requiring Risk Evaluation and Mitigation Strategies (REMS). No new safety signals were identified in the evaluation of this trial.

1.4 Recommendations for Postmarket Requirements and Commitments

There are no clinical recommendations for PMR or PMC.

The following Clinical Pharmacology PMC is recommended:

Combine all available PK data from different patient populations and healthy subjects in an integrated PK model to evaluate the potential impact of tumor types on the PK of cabozantinib.

2 Introduction and Regulatory Background

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

2.1 Product Information

2.2 Tables of Currently Available Treatments for Proposed Indications

Anti-angiogenic therapy has been shown to have clinical benefit in patients with advanced renal cell carcinoma (RCC). During the past decade, the FDA has granted regular approval to seven targeted drugs for patients with RCC. Six of these drugs were approved based on improvement in progression-free survival (PFS), whereas temsirolimus was approved in poor-prognosis patients based on improved OS. In December 2015, nivolumab, an immunotherapeutic agent that blocks the T-cell checkpoint programmed death receptor-1 (PD-1) was granted regular approval for the treatment of advanced RCC patients who have received prior antiangiogenic therapy based on an improved OS. Nivolumab is currently approved for the treatment of advanced melanoma, non-small cell lung cancer, and renal cell carcinoma.

Table 1 lists the FDA-approved therapies for previously treated advanced renal-cell cancer and the clinical basis for approval.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 1: FDA approved Therapies Indicated for Second-Line Treatment of Patients with Advanced Renal-Cell Carcinoma FDA Approved Approval Trial Design/ Endpoint(s) Clinical Benefit/Effect Drug Year Population Affinitor 2009 Multicenter, PFS (by Everolimus vs. placebo (everolimus) randomized, double- central PFS: HR:0.33 (95% CI:0.25- blinded, two arm review) 0.43) median 4.9 vs 1.9 m

Prior sorafenib or sunitinib Inlyta 2012 Multicenter, PFS (by Axitinib vs. Sorafenib (axitinib) randomized, open- central PFS: HR:0.67 (95% CI:0.54- label, two-arm review) 0.81) median 6.7 vs. 4.7 m

Prior anti-angiogenic or cytokine therapy Opdivo 2015 Prior anti-angiogenic OS Nivolumab vs. Everolimus (nivolumab) therapy OS: HR:0.73 (95% CI:0.60- 0.89) median 25.0 vs 19.6 m PFS: HR:0.88 (95% CI:0.75- 1.03) median 4.6 vs 4.4 m Source: Affinitor USPI; Inlyta USPI; Opdivo USPI. Abbreviations in Table: m, months; HR, hazard ratio; PFS, progression-free survival; OS, overall survival

Reviewer Comments:

1. All approved therapies for advanced renal cell carcinoma are included in this table, given the clinical use of tyrosine kinase inhibitors as a class of drugs.

2. Nivolumab, everolimus and axitinib are approved for patients with advanced renal cell carcinoma after anti-angiogenic therapy. Everolimus was able to demonstrate an improvement in PFS compared to placebo. Nivolumab demonstrated an OS benefit compared to everolimus.

3. The landscape of metastatic RCC is evolving with recent approvals, and it is unclear how this will impact the therapeutic sequence. Given the overall survival advantage demonstrated by nivolumab, it will be interesting to observe how providers select the optimal second line therapy for metastatic RCC.

2.3 Availability of Proposed Active Ingredient in the United States

Cabozantinib 20 mg and 80 mg capsules (under the proprietary name COMETRIQ®) was approved for use in the US on November 29, 2012 for the treatment of patients with progressive, metastatic medullary thyroid cancer (MTC). The major safety concerns of the capsule formulation are included in the COMETRIQ® label and include:

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

x Gastrointestinal Perforations and Fistulas: Monitor for symptoms of perforations and fistulas. (b) (4) discontinue in patients who experience a perforation or fistula. x Hemorrhage: Do not administer to patients with a recent history of hemorrhage or hemoptysis. x Thrombotic events: (b) (4) discontinue in patients who develop an acute myocardial infarction or any other clinically significant arterial thromboembolic complication. x Wound Complications: Stop treatment at least 28 days prior to scheduled surgery. Resume therapy after surgery based on clinical judgment of adequate wound healing. x Hypertension: Monitor blood pressure prior to initiation and regularly during treatment. Withhold for hypertension that is not adequately controlled with medical management. Discontinue for severe hypertension that cannot be controlled with anti-hypertensive therapy. x Osteonecrosis of the Jaw: Withhold for at least 28 days prior to scheduled (b) (4) . Perform an oral examination prior to initiating and periodically during therapy. x Palmar-plantar Erythrodysesthesia Syndrome: Withhold in patients who develop intolerable Grade 2 or Grade 3-4 PPES until improvement to Grade 1; resume at a reduced dose. x Proteinuria: Monitor urine protein regularly during treatment. (b) (4) (b) discontinue (4) nephrotic syndrome. x Reversible Posterior Leukoencephalopathy Syndrome (RPLS): (b) (4) (b) discontinue (4) RPLS. (b) (b) (4) (b) (4) x Embryo-fetal Toxicity: Can cause fetal harm. potential to (4) fetus (b) (4)

2.4 Important Safety Issues with Consideration to Related Drugs

Several multi-targeted kinase inhibitors have been approved in the US for the treatment of hematologic and oncologic malignancies. While the targets and relative levels of target inhibition for these compounds vary, many toxicities of these multi-targeted TKIs are overlapping. However each compound does retain a distinctive toxicity profile. In addition to other TKIs, the safety experience with selective monoclonal antibodies such as bevacizumab, which targets VEGFR, is also relevant.

Toxicities reported with agents used in the treatment of renal cell carcinoma in the Warnings and Precautions or Adverse Reactions: Postmarketing Experience sections of the current or proposed US Package Insert (USPI) are listed below in Table 2. Any of these toxicities might be anticipated in post-marketing surveillance of cabozantinib.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 2: Warnings and Precautions and Select Adverse Reactions of Kinase Inhibitors Approved for Patients with Renal Cell Carcinoma Sorafenib Sunitinib Pazopanib Axitinib Bevacizumab Hepatic toxicity/ XXXX hepatic impairment Congestive heart XXX failure Hemorrhagic events X XX XX Ischemic (b) (4) cerebrovascular events QT prolongation X XX Dermatologic X toxicities/ Stevens- Johnson syndrome Hypothyroidism XX X Hypertension X XX XX Reversible posterior XX(b) (4) leukoencephalopathy syndrome (RPLS) Arterial thrombotic XXX events Venous thrombotic XXX events Gastrointestinal and XXX non-gastrointestinal perforations and fistulas Impaired wound XX X healing Proteinuria/ renal XXXX failure Infection X Mucositis/stomatitis X Osteonecrosis X X

2.5 Summary of Presubmission Regulatory Activity Related to Submission

The following summarizes the regulatory activity for cabozantinib and is modified from the original NDA cabozantinib review.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

On June 10, 2005, the Applicant opened the initial Investigational New Drug application for cabozantinib (XL-184; IND 72,596).

On November 29, 2012, cabozantinib (Cometriq) was approved for the treatment of patients for patients with progressive, metastatic, medullary thyroid cancer (MTC).

On January 17, 2013, FDA and the Applicant discussed the Applicant’s proposal regarding a planned registrational study in advanced RCC. FDA agreed on the design of Trial XL184-308, including the choice of comparator arm and the use of Progression Free Survival (PFS) as the primary endpoint with a robust evaluation of Overall Survival (OS) as a secondary end point.

On February 27, 2013, Protocol XL184-308, A Phase III, Randomized, Controlled Study of Cabozantinib (XL184) vs Everolimus in Patients with Metastatic Renal Cell Carcinoma that is Refractory to or has Progressed after prior VEGFR Tyrosine Kinase Inhibitor Therapy was submitted to the IND.

On September 13, 2013, the Applicant discussed with FDA their plans for a tablet formulation of cabozantinib.

On April 8, 2015, the Applicant was granted Fast Track Designation for treatment of patients with advanced renal cell carcinoma who have received one prior therapy.

On July 27, 2015, the Applicant submitted a request for Breakthrough Therapy Designation which was granted on August 21, 2015 based on preliminary evidence from Trial XL184-308 which indicated that the treatment effect of cabozantinib may represent a substantial improvement over existing therapies for the treatment of advanced renal cell carcinoma after prior VEGFR-TKI therapy.

On September 2, 2015, FDA addressed the general submission plan for the NDA. FDA stated that the Applicant should focus on the pivotal Phase 3 study in in their filing. FDA also stated that a review of additional OS analysis would occur during the NDA review. The Applicant states that additional PK reports would be submitted after the initial filing. FDA also stated that a 120-Day Safety update was not required, and recommended a rolling NDA submission.

On October 13, 2015, the Applicant initiated a rolling NDA submission for the treatment of patients with metastatic RCC after prior VEGFR TKI therapy.

On December 22, 2015, the Applicant submitted the final component (Clinical) of the rolling NDA submission.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

2.6 Other Relevant Background Information

Cabozantinib, 20 mg and 80 mg capsules, (under the proprietary name COMETRIQ®) was approved for use in the US on November 29, 2012 for the treatment of patients with progressive, metastatic MTC. The recommended daily dose is 140 mg.

On January 11, 2016, the applicant submitted cabozantinib for the treatment of patients with advanced renal cell carcinoma who have received one prior therapy to the European Medicines Agency for accelerated assessment.

3 Ethics and Good Clinical Practices

3.1 Submission Quality and Integrity

The submission contains all required components of the eCTD. The overall quality and integrity of the application appear acceptable.

3.2 Compliance with Good Clinical Practices

The Applicant states that the trial was conducted in accordance with Good Clinical Practice (GCP), as defined by the International Conference on Harmonization (ICH) and in accordance with the ethical principles underlying European Union Directive 2001/20/EC and the United States Code of Federal Regulations, Title 21, Part 50 (21CFR50). The laws and regulatory requirements of all countries that had sites participating in this trial were adhered to. The trial was conducted in accordance with the ethical principles that have their origin in the Declaration of Helsinki. The protocol, amendments, and patient informed consent received appropriate approval by the IRB/IEC prior to initiation of trial at the site. No breaches of the conditions and principles of GCP in connection with the trial or protocol were reported.

Protocol deviations were identified by the applicant via on-site monitoring and programmed checks based on data collected in the case report form (CRF) as well as review of data listings.

The Applicant defined the following protocol deviations as “important”: x Any eligibility violation that, in the judgment of the medical monitor, could put an individual subject’s safety at risk or result in enrollment of a subject who is not within the intended study population. x Any other deviation that, in the judgment of the medical monitor, could put an individual subject’s safety at risk or meaningfully impact the interpretation and/or validity of study results.

The number of subject eligibility violations was similar between arms (11% in both arms). The most common inclusion criteria deviations were not meeting required

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

screening laboratory values within the required time periods, and receipt of the last dose of the most recently received VEGFR-TKI more than 6 months before the date of randomization. The most common exclusion criteria deviations were not meeting the required criteria for QTcF and receipt of any type of small molecule kinase inhibitor within two weeks before randomization.

Important protocol deviations that did not involve eligibility criteria were also well balanced between treated arms and are summarized in Table 3. The most frequently observed category was randomization irregularities that potentially impacted efficacy. Deviations observed in this category included stratification errors by incorrect determination of the MSKCC risk group or the number of prior VEGFR-TKIs. The primary PFS analysis by the Applicant was stratified by data collected on the eCRFs. A concordance analysis of eCRF- and IVRS-based stratification values revealed some discrepancies, more so for the MSKCC risk factors than the prior VEGFR-TKIs (See Table 4). Discrepancies were consistent between the ITT and PITT populations.

Table 3. Overview of Important Protocol Deviations Excluding Eligibility Criteria Cabozantinib Everolimus (N=330) (N=328) n (%) n (%) Potential Impact Potential Impact Deviation Safety Efficacy Both Other Safety Efficacy Both Other Prohibited 2 (0.6) 1 (0.3) 1 (0.3) 0 1 (0.3) 3 (0.9) 2 (0.6) 0 medication Treatment 9 (2.7) 0 4 (1.2) 0 9 (2.7) 0 0 0 deviation Withdrawal 0000001 (0.3)0 deviation Randomization 1 (0.3) 23 (7.0) 0 0 0 24 (7.3) 0 0 irregularity Source: CSR

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 4: Concordance between eCRF and IVRS stratification data

PITT Population ITT Population Cabozantinib Everolimus Cabozantinib Everolimus N=187 N=188 N = 330 N=328 Number of Prior VEGFR-TKIs Concordance between eCRF 182 (97%) 184 (98%) 322 (98%) 322 (98%) and IVRS Discordance between eCRF 5 (3%) 4 (2%) 8 (2%) 6 (2%) and IVRS Number of risk factors per MSKCC prognostic criteria for previously treated patients with RCC Concordance between eCRF 167 (89%) 168 (89%) 300 (91%) 293 (89%) and IVRS Discordance between eCRF 20 (11%) 20 (11%) 30 (9%) 35 (11%) and IVRS Source: statistical review

A high incidence of protocol deviations categorized as ‘other’ that could have potentially impacted safety or efficacy was also observed (51% in the cabozantinib arm and 44% in the everolimus arm for those potentially impacting safety and 29% and 28% in the cabozantinib and everolimus arms respectively for those potentially impacting efficacy. Deviations in this category that potentially impact safety included ECGs and vital signs not obtained per protocol, and revised informed consent form not signed by subject in a timely manner. Deviations in this category that potentially impacted efficacy included tumor imaging not obtained per protocol and subject Health Related Quality of Life questionnaires not completed per protocol.

Reviewer comments: 1. The nature and distribution of the eligibility or important protocol violations are unlikely to have substantially affected the efficacy outcomes. Although some patients were stratified incorrectly, we performed sensitivity analyses of PFS using IVRS data and the PFS analysis remained in favor of the cabozantinib arm. To follow the intent-to-treat principle, IVRS data was used in the stratified analysis as well as in the product label (See Section 6.1.4 and 6.1.5) 2. Protocol violations were balanced evenly between treatment arms. 3. While there are a substantial number of protocol deviations categorized as ‘other,’ it is unlikely that they had a substantial impact on the safety or efficacy of the study. As the primary endpoint was PFS, it is unlikely that these protocol deviations could have impacted the assessment of progression. For safety, missing data such as ECG and vital signs were unlikely to impact the overall assessment. Many of these assessments were captured throughout the study and did not reveal any new safety signals.

Reference ID: 3917720

Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Long term stability data including up to 36 months at 25 °C/60%RH for the drug product batches manufactured with a process representative of the proposed commercial manufacturing process and at the intended commercial site (Patheon, Mississauga, Ontario, Canada) were submitted. Based on stability results from these batches, an expiry of 36 months was granted for the drug product stored in the proposed commercial packaging system at 25 °C/60% RH.

Summary of Drug Product Intended Use

Proprietary Name of the Drug Product CABOMETYX Non Proprietary Name of the Drug Product Cabozantinib tablets Non Proprietary Name of the Drug cabozantinib Substance Proposed Indication(s) including Intended CABOMETYX is a kinase inhibitor Patient Population indicated for the treatment of advanced renal cell carcinoma (RCC) in patients who have received (b) (4) prior therapy. Duration of Treatment ? (b) Maximum Daily Dose (4) mg per day Alternative Methods of Administration N/A

Biopharmaceutics Considerations Cabozantinib is classified as a BCS Class 2 (low solubility, high permeability) that demonstrates a pH-dependent solubility profile. Cabozantinib (XL184 free base) was determined to be practically insoluble in water (<0.0001 mg/mL) and polymorphic. The applicant did not request a BE waiver.

Besides the approved capsule formulation (COMETRIQ®, NDA 203756), a tablet formulation (20 mg, 40 mg and 60 mg) was later developed and has been used in the other Phase 3 efficacy and/or safety studies, including biopharmaceutical studies XL184-308. All of the drug product batches were manufactured at the intended commercial site (Patheon, Mississauga, Ontario, Canada). The tablet formulation did not change between that used in the Phase 3 study and the proposed commercial drug product.

The biopharmaceutics review evaluates the data provided to support; 1) proposed in vitro dissolution testing method and 2) the release/stability in vitro dissolution acceptance criteria. The proposed dissolution method and the revised acceptance criteria ((b) (4) % at 15 min.) for the dissolution test are found to be acceptable.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

4.2 Clinical Microbiology

Not applicable.

4.3 Preclinical Pharmacology/Toxicology

Refer to the FDA Nonclinical Review for details.

4.4 Clinical Pharmacology

Refer to the FDA Clinical Pharmacology Review for details.

4.4.1 Mechanism of Action

Cabozantinib is an inhibitor of multiple intracellular kinases involved in a range of pathologic processes such as oncogenesis, tumor angiogenesis, and maintenance of the tumor environment. In vitro biochemical or cellular assays have shown that cabozantinib inhibits the tyrosine kinase activity of RET, mesenchymal epithelial transition factor (MET), vascular endothelial cell growth factor (VEGFR) receptors, KIT, TRKB, FLT-3, AXL, and TIE-2 receptors.

4.4.2 Pharmacodynamics

Refer to Clinical Pharmacology Review for details.

4.4.3 Pharmacokinetics

A population PK analysis of cabozantinib was performed using data from 318 patients with RCC and 63 normal healthy volunteers following oral administration of doses of 60 mg, 40 mg, and 20 mg. The predicted terminal half-life is approximately 99 hours, the oral volume of distribution (Vz/F) is approximately 319 L, and the apparent clearance (CL/F) at steady-state is estimated to be 2.2 L/hr. It is noted that the estimated PK parameters are different from MTC population, where the half-life at steady state is approximately 55 hours, the oral volume of distribution is approximately 349 L, and the clearance (CL/F) at steady-state was estimated to be 4.4 L/hr. Steady-state exposure (Cmin,ss) were similar in RCC (1260 ng/mL, 44.4% CV) and in MTC (1380 ng/mL, 53% CV) following different formulations at different doses of cabozantinib in RCC (60 mg once daily tablet) and MTC (140 mg once daily capsule) patient populations.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

5 Sources of Clinical Data

5.1 Tables of Studies/Clinical Trials

In support of the NDA application for the treatment of RCC, the applicant provided full or abbreviated study reports for three clinical trials (Table 6)

The primary evidence to support this application is derived from data from Trial XL184- 308. Additional safety data were analyzed from two randomized, double-blind, active comparator-controlled Phase 3 studies in patients with metastatic castration-resistant prostate cancer. In Study XL184-307, 1023 patients received study drug (681 cabozantinib; 342 prednisone),and OS was the primary endpoint. In Study XL184-306, 117 patients received either cabozantinib (n=60) or mitoxantrone plus prednisone (n=57), and the primary endpoint was the proportion of patients showing a pain response at Week 6 confirmed at Week 12. Finally, safety data from Trial XL184-301 in patients with MTC was also analyzed with this NDA.

See Section 5.3 for discussion of each protocol included in support of this application.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 6: Summary of Clinical Safety Studies of Cabozantinib Included in the Safety Analysis

Study Phase Study Design Cabozantinib Study Population No. Patients No. Patients Study Status: (XL184) Doses Treated in the Safety Type of Analysis Set Report XL184-308 3 Open-label, randomized, Treatment: 60 mg Advanced RCC 331 cabo 331 cabo Active; CSR controlled tablet (post-VEGFR 322 evero 322 evero Control: TKI) everolimus 10 mg XL184-307 3 Double-blinded, Treatment: 60 mg Metastatic CRPC 681 cabo 681 Active; CSR randomized, controlled tablet (post-docetaxel 342 pred Control: pred 5 mg and abiraterone or enzalutamide XL184-306 3 Double-blind, Treatment: 60 mg Metastatic CRPC 60 cabo 60 Completed; randomized, controlled tablet (post docetaxel 57 mito+pred CSR Control: mito + and abiraterone pred or enzalutamide XL184-301 3 Double-blind, Treatment: 140 mg Progressive, 214 cabo 214 Completed; randomized, controlled capsule metastatic MTC 109 placebo CSR Control: placebo Cabo, cabozantinib; CRPC, castration-resistant prostate cancer; CSR, clinical study report; evero, everolimus; mito, mitoxantrone; pred, prednisone; RCC, renal cell carcinoma, VEGF-TKI, vascular endothelial growth factor tyrosine kinase inhibitor

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

5.2 Review Strategy

The FDA clinical sNDA review was conducted by two primary clinical reviewers. Dr. Michael Brave conducted the efficacy review and Dr. Harpreet Singh conducted the safety review. A statistical review was conducted by D. Joyce Chen.

The NDA submission contained Trial XL184-308, a randomized, open-label, controlled Phase 3 multi-center trial of cabozantinib vs. everolimus in advanced renal cell carcinoma that has progressed after prior VEGF tyrosine kinase inhibitor therapy. The clinical review of efficacy focuses on the intent-to-treat population of 375 patients, 187 randomized to the cabozantinib arm and 188 randomized to the everolimus arm. Additional OS analyses focused on the total enrolled population after the second unplanned interim analysis. The review and analysis of data included the clinical study report (CSR), CRF’s, and datasets in consultation with the FDA statistical review team. The following efficacy endpoints were analyzed: PFS, OS, and ORR.

The clinical review of safety focused on 653 treated patients in Trial XL184-308 (331 on the cabozantinib arm and 322 on the everolimus arm) with additional analyses focused on select adverse events (AEs) with cabozantinib (e.g. deaths, non-fatal serious adverse events (SAEs), AEs leading to treatment modifications), including a detailed review and analysis of data including the CSR, CRFs, narratives, and datasets.

The clinical review included the following: x Review of published literature on RCC epidemiology, and treatment, including other kinase inhibitors x Review of Applicant-submitted Trial XL184-308 including CSR, protocols, and protocol amendments x Review and assessment of Applicant analysis of cabozantinib efficacy and safety, for evaluation of Applicant’s claims x Review of datasets submitted as xpt.files x Review of patient narratives of SAEs and deaths x Review of meeting minutes conducted during drug development x Assessment of the Module 2 summaries including the Summary of Clinical Safety x Review of reviews conducted by other FDA disciplines including Clinical Pharmacology and Biostatistics x Review of consultation reports of Office of Scientific Investigations x Requests for additional information from the Applicant and review of Applicant responses x Formulation of the benefit-risk analysis and recommendations x Review and evaluation of proposed labeling

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

5.3 Discussion of Individual Studies/Clinical Trials

Protocol XL184-308

Clinical Trial Title A Phase 3, Randomized, Controlled Study of Cabozantinib (XL184) Versus Everolimus in Subjects with Metastatic Renal Cell Carcinoma that has Progressed after Prior VEGFR Tyrosine Kinase Inhibitor Therapy

Trial Sites 658 patients were treated at approximately 182 sites in 24 countries (Australia, Austria, Belgium, Canada, Chile, Czech Republic, Denmark, Finland, France, Germany, Hungary, Ireland, Republic of Korea, Netherlands, Poland, Portugal, Russian Federation, Slovakia, Spain, Sweden, Province of China Taiwan, Turkey, United Kingdom, and United States).

Objectives Primary Objective x To compare the clinical benefit, as measured by duration of progression-free survival (PFS), provided by cabozantinib vs everolimus in patients with advanced renal cell cancer that has progressed after prior VEGFR tyrosine kinase inhibitor therapy.

Secondary Objectives x To compare the duration of (investigator-assessed) overall survival (OS) of cabozantinib vs everolimus in patients with advanced renal cell cancer that has progressed after prior VEGFR tyrosine kinase inhibitor therapy. x To compare the response rate (ORR) per independent radiology committee (IRC) of cabozantinib vs everolimus in patients with advanced renal cell cancer that has progressed after prior VEGFR tyrosine kinase inhibitor therapy.

Key Exploratory Objectives x To evaluate duration of radiographic response of cabozantinib and everolimus. x To assess the overall safety and tolerability of cabozantinib vs everolimus. x To evaluate changes in kidney-cancer related symptoms as assessed by the Functional Assessment of Cancer Therapy-Kidney Cancer Symposium Index (FKSI-19). x To evaluate health-related quality of life (HRQoL) as assessed by EuroQol Health questionnaire instrument (EQ-5D-5L). x To characterize the pharmacokinetics of cabozantinib.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Enrollment began on August 22, 2013. The 375th patient was randomized on April 15, 2014, and the final patient was enrolled on July 14, 2014. The data cutoff date for the primary analysis was May 22, 2015.

Patient population Inclusion Criteria 1. Documented histological or cytological diagnosis of renal cell cancer with a clear cell component. 2. Measurable disease per the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 as determined by the investigator. 3. Must have received at least one VEGFR-targeting TKI (e.g., sorafenib, sunitinib, axitinib, pazopanib or tivozanib).

Note: Prior treatment with other anticancer therapies including cytokines (e.g., interleukin-2, interferon-alfa), monoclonal antibodies, (e.g., bevacizumab), and cytotoxic chemotherapy was allowed (except Exclusion Criterion #1).

4. For the most recently received VEGFR-targeting TKI the following criteria applied: a. Must have radiographically progressed during treatment, or been treated for at least 4 weeks and radiographically progressed within 6 months after the last dose. Radiographic progression was defined as unequivocal progression of existing tumor lesions or developing new tumor lesions as assessed by the investigator on CT or MRI scans. b. The last dose must have been within 6 months before the date of randomization. 5. 5HFRYHU\ WR EDVHOLQH RU *UDGH 1DWLRQDO &DQFHU ,QVWLWXWH 1&, &RPPRQ Terminology Criteria for Adverse Events (CTCAE v4) from toxicities related to any prior treatments, unless AE(s) were clinically nonsignificant and/or stable on supportive therapy. 6. Age eighteen years or older on the day of consent. 7. Karnofsky Performance Status VFRUH RI 8. Adequate organ and marrow function, based upon all of the following laboratory criteria within 10 days before randomization: a. Absolute neutrophil count PL b. 3ODWHOHWV PL c. +HPRJORELQ JG/ d. Alanine aminotransferase (ALT) and aspartate aminotransferase (AST) < 3.0 × upper limit of normal e. 7RWDO ELOLUXELQ î WKH XSSHU OLPLW RI QRUPDO )RU VXEMHFWV ZLWK *LOEHUW¶V disease PJG/ ȝPRO/ f. )DVWLQJ VHUXP WULJO\FHULGHV î XSSHU OLPLW RI QRUPDO $1' WRWDO FKROHVWHURO PJG/ PPRO/ /LSLG-lowering medication was allowed.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

g. +E$F )RU VXEMHFWV ZLWK D FRQGLWLRQ H.g., hemoglobin variant) that affected the LQWHUSUHWDWLRQ RI +E$F UHVXOWV D IDVWLQJ JOXFRVH PJG/ PPRO/ h. 6HUXP FUHDWLQLQH î XSSHU OLPLW RI QRUPDO RU FDOFXODWHG FUHDWLQLQH clearance P/PLQ P/VHF XVLQJ WKH &RFNURIW-Gault equation i. Urine protein-to-creatinine rDWLR 83&5 PJPJ PJPPRO creatinine or 24-hour urine protein < 1 g 9. Capable of understanding and complying with the protocol requirements and must have signed the informed consent document. 10.Sexually active fertile subjects and their partners must have agreed to use medically accepted methods of contraception (e.g., barrier methods, including male condom, female condom, or diaphragm with spermicidal gel) during the course of the study and for 4 months after the last dose of study treatment. 11.Female subjects of childbearing potential must not have been pregnant at screening. Females of childbearing potential were defined as premenopausal females capable of becoming pregnant (i.e., females who had any evidence of menses in the past 12 months, with the exception of those who had prior hysterectomy). However, women who had been amenorrheic for 12 or more months were still considered to be of childbearing potential if the amenorrhea was possibly due to prior chemotherapy, antiestrogens, low body weight, ovarian suppression, or other reasons.

Exclusion Criteria 1. Prior treatment with everolimus, or any other specific or selective TORC1/PI3K/AKT inhibitor (e.g., temsirolimus), or cabozantinib. 2. Receipt of any type of small molecule kinase inhibitor (including investigational kinase inhibitor) within 2 weeks before randomization. 3. Receipt of any type of anticancer antibody (including investigational antibody) within 4 weeks before randomization. 4. Radiation therapy for bone metastasis within 2 weeks, any other external radiation therapy within 4 weeks before randomization. Systemic treatment with radionuclides within 6 weeks before randomization. Subjects with clinically relevant ongoing complications from prior radiation therapy were not eligible. 5. Known brain metastases or cranial epidural disease unless adequately treated with radiotherapy and/or surgery (including radiosurgery) and stable for at least 3 months before randomization. Eligible subjects must have been neurologically asymptomatic and without corticosteroid treatment at the time of randomization. 6. Concomitant anticoagulation at therapeutic doses, with oral anticoagulants (eg, warfarin, direct thrombin, and Factor Xa inhibitors) or platelet inhibitors (eg, clopidogrel).

Note: Low-dose aspirin for cardioprotection (per local applicable guidelines), low- dose warfarin (< 1 mg/day), and low dose, low molecular weight heparins (LMWH) were permitted. Anticoagulation with therapeutic doses of LMWH were

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

allowed in subjects without radiographic evidence of brain metastasis, who were on a stable dose of LMWH for at least 12 weeks before randomization, and who had experienced no complications from a thromboembolic event or the anticoagulation regimen.

7. Chronic treatment with corticosteroids or other immunosuppressive agents (with the exception of inhaled or topical corticosteroids or corticosteroids with a daily dosage HTXLYDOHQW PJ SUHGQLVRQH LI JLYHQ IRU GLVRUGHUV RWKHU WKDQ UHQDO FHOO cancer). Subjects with brain metastases requiring systemic corticosteroid were not eligible. The subject had uncontrolled, significant intercurrent or recent illness including, but not limited to, the following conditions: a. Cardiovascular disorders: i. Congestive heart failure New York Heart Association class 3 or 4, unstable angina pectoris, serious cardiac arrhythmias ii. Uncontrolled hypertension defined as sustained blood pressure > 150 mm Hg systolic or > 100 mm Hg diastolic despite optimal antihypertensive treatment iii. Stroke (including transient ischemic attack), myocardial infarction, or other ischemic event, or thromboembolic event (e.g., deep venous thrombosis, pulmonary embolism) within 6 months before randomization b. Gastrointestinal disorders (GI) including those associated with a high risk of perforation or fistula formation: i. Tumors invading the GI-tract, active peptic ulcer disease, inflammatory bowel disease, diverticulitis, cholecystitis, symptomatic cholangitis or appendicitis, acute or acute obstruction of the pancreatic or biliary duct, or gastric outlet obstruction ii. Abdominal fistula, gastrointestinal perforation, bowel obstruction, or intra-abdominal abscess within 6 months before randomization

Note: Complete healing of an intra-abdominal abscess must have been confirmed before randomization.

c. Clinically significant hematuria, hematemesis, or hemoptysis of > 0.5 teaspoon (2.5 mL) of red blood, or other history of significant bleeding (eg, pulmonary hemorrhage) within 3 months before randomization d. Cavitating pulmonary lesion(s) or known endobronchial disease manifestation e. Lesions invading major pulmonary blood vessels f. Other clinically significant disorders such as: i. Active infection requiring systemic treatment, infection with human immunodeficiency virus (HIV) or acquired immunodeficiency syndrome (AIDS)- related illness, or chronic hepatitis B or C infection

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

ii. Serious nonhealing wound/ulcer/bone fracture iii. Malabsorption syndrome iv. Uncompensated/symptomatic hypothyroidism v.Moderate to severe hepatic impairment (Child-Pugh B or C) vi. Requirement for hemodialysis or peritoneal dialysis vii. History of solid organ transplantation 8. Major surgery (e.g., GI surgery, removal or biopsy of brain metastasis) within 2 months before randomization. Complete wound healing from major surgery must have occurred 1 month before randomization and from minor surgery (e.g., simple excision, tooth extraction) at least 10 days before randomization. Subjects with clinically relevant ongoing complications from prior surgery were not eligible. 9. Corrected QT interval calculated by the Fridericia formula (QTcF) > 500 ms within 10 days before randomization.

Three ECGs were performed. If the average of these three consecutive results for QTcF was PV WKH VXEMHFW PHW WKLV HOLJLELOLW\ FULWHULRQ

10.Pregnant or lactating females 11.Inability to swallow tablets or capsules 12.Previously identified allergy or hypersensitivity to components of the study treatment formulations 13.Diagnosis of another malignancy within 2 years before randomization, except for superficial skin cancers, or localized, low grade tumors deemed cured and not treated with systemic therapy

Treatment Plan

Patients on the cabozantinib arm were to receive oral cabozantinib 60 mg daily, while those on the everolimus arm were to be treated with oral everolimus 10 mg daily. There were to be allowances for dose interruptions for up to 6 weeks for study treatment related AE’s. Treated patients were to be evaluated for response according to RECIST 1.1 guidelines beginning 8 weeks after randomization and continuing every 8 weeks for the first 12 months, and then every 12 weeks until disease progression or treatment discontinuation, which occurs later.

Treatment beyond initial radiographic progression per RECIST 1.1 could be considered in patients experiencing investigator-assessed clinical benefit and tolerating study therapy.

Concomitant medications

The following medications were prohibited: immunosuppressive agents, systemic corticosteroid > 10 mg daily prednisone equivalent, anticoagulation at therapeutic doses with oral anticoagulants or platelet inhibitors, and any concurrent antineoplastic therapy.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Dose Modification and Management Algorithms The assigned dose for cabozantinib is 60 mg/day. Two dose reduction levels of cabozantinib were permitted (Table 7) Table 7 XL184-308 Allowable Dose Reductions of Cabozantinib Starting Dose First Dose-Level Second Dose-Level Reduction Reduction 60-mg cabozantinib 40-mg cabozantinib 20-mg cabozantinib

The dose of cabozantinib was to be reduced or interrupted at the discretion of the investigator for the following:

x Grade 2 AEs which were intolerable and could not be adequately managed x Grade 3 AEs except clinically nonrelevant laboratory abnormalities x Grade 4 AEs except clinically nonrelevant laboratory abnormalities

Interruption of cabozantinib for treatment of AEs could occur at any time per investigator discretion. If treatment was interrupted due to AEs for more than 6 weeks, cabozantinib was to be discontinued.

Discontinuation Treatment of cabozantinib was to be permanently discontinued for the following: x Any Grade 4 drug-related adverse event, except clinically non-relevant laboratory abnormalities x Hypertensive emergency x Nephrotic syndrome x Severe bleeding complication x GI or non-GI perforation or fistula x Any dosing interruption lasting >6 weeks, unless continuation of treatment is approved by the Sponsor for interruption s which are not due to AEs. x Patient no longer experiences clinical benefit as determined by the investigator x Unacceptable side effects the investigator feels may be due to study treatment x The investigator feels it is not in the best interest of the patient to continue on study x Patient participation in another clinical study using an investigational agent or investigational medical device x Necessity for treatment with other systemic anticancer therapy. x Pregnancy x Termination of the trial by Exelixis, Inc. x Patient’s request to stop study treatment x Additional protocol specific reasons for discontinuation

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Trial Assessments

There are four phases of the trial:

Screening Phase: x Potential subjects screened to determine if they meet the required eligibility criteria x Qualifying screen assessments must be performed within 28 days before randomization unless otherwise specified

Treatment Phase: x Randomization to either the cabozantinib or everolimus arm x Randomization is stratified by the number of prior of VEGFR-targeting therapies and number of risk factors per Memorial Sloan-Kettering Cancer Center prognostic criteria for previously treated patients with RCC

Patients receive treatment as long as they continue to experience clinical benefit in the opinion of the investigator or until there is unacceptable toxicity or the need for subsequent anticancer treatment. Treatment may continue after radiographic RCC progression per RECIST 1.1 as long as the investigator believes that the subject is still receiving study treatment outweighs potential risks. Crossover between treatment arms will not be allowed.

Maintenance Phase (added as part of Amendment 1):

When sufficient data have been collected to adequately evaluate all study endpoints, and upon site notification by the Sponsor, patients who continue study treatment will enter the study Maintenance Phase. Upon initiation of the Maintenance Phase, the Sponsor considers the safety and efficacy profile of the drug within this study to have been sufficiently established for regulatory purposes.

In the Maintenance Phase, patients continue to receive study treatment until they meet the protocol-required criteria for treatment discontinuation. Patients undergo periodic safety assessments and tumor assessments. Reporting of SAEs, certain AEs, and other reportable events is to continue per protocol.

Follow-Up Phase:

x Begins when a patient is discontinued from treatment. x Patients will continue radiographic tumor assessments and HRQOL assessments x Patients will be contact every 8 weeks after the Post-Treatment Follow-up Visit to assess survival status and document receipt of subsequent anticancer therapy

Analysis Populations

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

The ITT population contained 658 patients: 330 randomized to cabozantinib and 328 randomized to everolimus. The PITT population (the first 375 patients randomized) included 187 patients in the cabozantinib arm and 188 patients in the everolimus arm. The safety population consisted of all randomized patients who received at least one dose of study treatment (n = 653): 331 in the cabozantinib arm and 322 subjects in the everolimus arm. Five patients randomized to receive everolimus did not receive study treatment, and one patient (Patient 1417-3624) was randomized to the everolimus arm but received cabozantinib as study treatment and therefore was evaluated in the cabozantinib arm for the Safety population (Table 8). Table 8. XL184-308 Analysis Populations Population Cabozantinib Everolimus All Patients ITT 330 328 Safety 331 322 First 375 Patients Randomized As randomized (“PITT”) 187 188 As treated 187 185 Source: dataset adtte; variables: arm, actarm, pittfl, and trta

Statistical Considerations

Efficacy

The primary efficacy endpoint was PFS, using RECIST 1.1 criteria, as determined by an independent radiology committee (IRC). Secondary efficacy endpoints were OS and objective response rate (ORR).

It was estimated that 375 patients were sufficient to assess the primary PFS endpoint. However, 650 patients were needed to provide sufficient power to evaluate OS. Therefore, to allow longer PFS follow-up for fewer patients and to avoid over- representing patients with early progression, the primary analysis was to be conducted after 259 PFS events had been observed among the first 375 randomized patients. Thus, only the first 375 patients randomized were included in the primary endpoint intent to treat population (“PITT population”), whereas all 650 randomized patients were included in analyses of OS (ITT population).

For the primary endpoint, it was estimated that 259 PFS events in the first 375 randomized patients would provide 90% power to detect a HR of 0.667, using the log- rank test and a 2-sided significance level of 5%. This corresponds to a 33.3% reduction in the risk of progression or death, or a 50% improvement in median PFS from 5 months to 7.5 months. Under this design, the minimum observed effect that would result in

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

statistical significance is a 27.8% improvement (HR = 0.783) in PFS from 5 to 6.39 months.

When the required number of events was reached for analysis of the primary endpoint, PFS was to be tested at a 2-sided 0.05 significance level. If the null hypothesis was rejected, two secondary endpoints were then to be tested in parallel: ORR was to be tested at a 2-sided 0.01 significance level, and an interim analysis of OS at a 2-sided 0.0001 level, per a Lan-DeMets O’Brien-Fleming alpha spending function. The PFS analysis could not be performed until all 650 patients in the ITT population were enrolled.

For the secondary efficacy endpoint of OS, assuming a single interim analysis at the 33% information fraction (at the time of the primary analysis of PFS) and a subsequent primary analysis, 408 deaths among 650 patients were estimated to provide 80% power to detect a HR of 0.75, using the log-rank test and a 2-sided significance level of 4%. This corresponds to a 25% reduction in the risk of death, or a 33% improvement in median OS from 15 to 20 months. Under this design, the minimum observed effect that would result in statistical significance is a 22.5% improvement (HR=0.816) in OS from 15 to 18.38 months.

The ORR was defined as the proportion of patients whose best response was complete response (CR) or partial response (PR), as assessed by the IRC per RECIST 1.1, FRQILUPHG E\ D VXEVHTXHQW YLVLW GD\V ODWHU 7KH DQDO\VLV RI 255 ZDV WR EH performed at the time of the primary PFS analysis, using a chi-squared test at a 2-sided significance level of 1%, based on the ITT population available at that time. Patients with no post-baseline tumor assessments were to be considered nonresponders.

Reviewer’s comment:

This hierarchical testing plan adequately controls the overall type 1 error.

Additional endpoints

Each additional endpoint will be analyzed using an appropriate two-sided statistical test without adjustment for multiplicity. Statistical results for exploratory endpoints will be considered descriptive. Exploratory analyses will be performed using the ITT and PITT populations unless specified otherwise.

Exploratory efficacy assessments included the following:

x Duration of response (DOR), defined as the time from the first tumor assessment that documented PR or CR that was subsequently confirmed at least 28 days later until the date of documented progression by IRC, per RECIST 1.1. The median duration of the objective response and the associated 95% confidence

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

interval for each treatment arm will be estimated using the Kaplan-Meier method. The hazard ratio and the 95% CI will be estimated using a Cox proportional- hazard model with the treatment group as the independent variable and stratified by the same randomization stratification factors as were used for the log-rank test. Supportive analysis will include DOR per investigator’s assessment.

x Changes in bone scans assessments, which is defined as the proportion of the subset of patients in the ITT population who have bone lesions per bone scan at baseline for whom the best bone scan response at the time of data cutoff is responder as assessed by the IRC. Point estimates of bone scan response, the difference in response rates between the two treatment arms, and associated confidence intervals will be provided. Confidence intervals will be calculated using exact methods. A waterfall plot of best percentage change in bone scan lesion area per IRC will be presented for each treatment group using the ITT and PITT population.

x Serum bone markers: Baseline and changes in bone markers from baseline will be summarized. Descriptive statistics will be presented by treatment group using all available data from protocol-defined time-points. Appropriate transformations may be applied to normalize the data for presentation or analysis. A waterfall plot of best percentage change from baseline will be presented for each treatment group using the ITT population.

x Plasma biomarkers: Baseline values of AXL, IL-6, osteopontin and TIMP 1 will be summarized. Descriptive statistics will be presented by treatment group using all available data from protocol-defined time-points.

x Skeletal-related events: Frequency counts and percentages will be presented by overall, SRE type and treatment arm for patients who experience an SRE post randomization, and for those who had a prior SRE before randomization and experience a SRE post randomization.

x Characterization of the pharmacokinetics of cabozantinib

x Patient reported outcomes, as assessed by the Functional Assessment of Cancer Therapy (FKSI-19) and the EuroQol Health questionnaire instrument (EQ-5D-5L). Repeated measure mixed-effects models will include the outcome variable of the QOL score change from baseline. The predictors will be the baseline scores, treatment arms, visit, and randomization strata. The individual subject nested within the planned treatment arm will be the random effect. All available data will be included for the analysis. The estimated mean change from baseline for each treatment arm as well as the difference between the 2 arms will be calculated. The p-values comparing the 2 treatment arms and the effect size will be presented. No adjustment will be made for multiple comparisons. An effect

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

size of differences in the 0.3-0.5 range will be considered potentially clinically meaningful; greater than 0.5 will be considered clinically meaningful.

Safety

The safety profile was to be measured by incidences of treatment emergent-adverse events (TEAEs), changes in laboratory parameters and vital signs from baseline, and ECOG PS. Adverse event assessments and laboratory tests were assessed at baseline and continuously throughout the trial at the beginning of each subsequent cycle. The Applicant also submitted narrative summaries of deaths, SAEs, AEs leading to discontinuation, overall AEs, select AEs, and laboratory abnormalities. All on-study AEs were summarized for the entire treatment period from first dosing to the last dosing date and the primary analyses of safety occurring within a 30-day window after the last dosing.

Protocol Amendments

The Applicant submitted one amendment Protocol XL184-308 (Amendment 1.0; 17 April 2014) in addition to several country-specific amendments. Key changes are summarized here: x A Maintenance Phase was added to the Treatment Period to be implemented when sufficient data had been collected to adequately evaluate all study endpoints. This is further described above. x The study population was limited to including approximately 10% of subjects (maximum of approximately 65 subjects) who had been previously treated with antibodies targeting the programmed cell death immune receptor, PD-1, or its ligands, PD-L1/L2. x Two exploratory study endpoints were added: changes in bone scans and changes in serum calcium from baseline. x Clarifications on imaging including time points of assessments were added. Within the summary of the schedule of bone scan assessments, the required interval for patients who had received 12 months of study treatment was expanded from 16 weeks to 24 weeks. There were no changes regarding the method of evaluation, imaging guidelines, or instructions for investigator assessments.

Reviewer comments:

1. None of the changes introduced by Amendment 1.0 significantly altered the patient population, primary efficacy analysis, or management of toxicity.

2. The study population was limited to include approximately 10% of patients who had been previously treated with antibodies targeting PD-1 or PD-L1/L2, however only about

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

5% of patients received prior immunotherapy, most commonly nivolumab. See Section 6.1.2 for further description of prior systemic therapies.

Phase 3 Study XL184-307

Title: A Phase 3, randomized, double-blind, controlled study of cabozantinib (XL184) vs. prednisone in mCRPC patients who have received prior docetaxel and prior abiraterone and enzalutamide. Patients were enrolled in 14 countries at 216 sites.

The primary objective of Study XL184-307 was to evaluate the effect of cabozantinib compared to prednisone on OS. The secondary objective was bone scan response (BSR) at the end of Week 12 of treatment with cabozantinib compared with prednisone.

3DWLHQWV LQ WKLV VWXG\ ZHUH PDOH DQG \HDUV Rf age with documented histological or cytological diagnosis of prostate cancer and evidence of its metastasis to bone (as determined by a bone scan). Patients were to have received prior docetaxel and either abiraterone or enzalutamide with evidence of investigator-assessed prostate cancer progression on each agent independently. If a patient had an AE related to a prior WUHDWPHQW WKH $( ZDV WR KDYH UHVROYHG WR EDVHOLQH RU &7&$( Y *UDGH 3DWLHQWV without prior orchiectomy must have been taking luteinizing hormone-releasing hormone (LHRH) analogue therapy at baseline and concomitantly throughout the study. In addition, patients were to have an ECOG performance status of 0-2, a serum WHVWRVWHURQH OHYHO RI ޒ QJG/ DQG DGHTXDWH RUJDQ DQG PDUURZ IXQFWLRn as determined by clinical laboratory tests.

Eligible patients were randomized 2:1 to the cabozantinib and prednisone treatment arms, respectively. Randomization was stratified by the following factors: prior treatment with cabazitaxel, baseline pain severity, and ECOG performance status.

Based on treatment assignment, patients received one of the following regimens:

x Cabozantinib treatment arm: oral cabozantinib (60 mg, qd) plus prednisone- matched placebo (twice daily [bid]) x Prednisone treatment arm: oral prednisone (5 mg, bid) plus cabozantinib- matched placebo (qd)

Patients maintained on daily doses of prednisone prior to randomization were allowed to continue this treatment during the study. For these patients, prednisone/prednisolone was considered a concomitant medication following randomization.

Patients received study treatment as long as they continued to experience clinical benefit, as determined by the investigator. Reasons for discontinuation included, among others, an unacceptable toxicity or the need for non-protocol systemic anticancer

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

therapy (including the use of bone-targeted radiopharmaceuticals). Crossover between treatment arms was not allowed.

Clinic visits occurred at minimum every 3 weeks after treatment was initiated through Week 12 and then every 6 weeks thereafter. Routine safety evaluations included assessments of AEs, vital signs, laboratory tests, and concomitant medications. New or worsening AEs were collected at study visits, over the phone, or by spontaneous subject report from informed consent through 30 days after the date of the decision to discontinue study treatment.

Study XL184-307 is complete; the study did not meet its primary endpoint of OS. Active patients were required to discontinue study treatment and revert to standard of care thereby limiting the overall exposure to treatment on the study.

Phase 3 Study XL184-306

Title: A Phase 3, randomized, double-blind, controlled study of cabozantinib (XL184) vs. prednisone plus mitoxantrone plus prednisone in men with previously treated symptomatic CRPC. Patients were enrolled in 5 countries at 42 sites.

The primary efficacy endpoint of Study XL184-306 was to evaluate the effect of cabozantinib compared with comparator (mitoxantrone + prednisone) on the proportion of patients with pain response at Week 6 confirmed at Week 12. The study was conducted in men with previously treated mCRPC with bone-dominant disease who had experienced disease progression while on docetaxel-containing chemotherapy and either abiraterone or enzalutamide. The secondary efficacy endpoints were BSR at the end of Week 12. The study was conducted in men with previously treated mCRPC with bone-dominant disease who had experienced disease progression while on docetaxel- containing chemotherapy and either abiraterone or enzalutamide. The secondary efficacy endpoints were BSR at the end of Week 12 of treatment and OS.

Patients had to have documented pain from bone metastases requiring opioid narcotics. During a 7-day Run-In Stage, patients had to meet stringent pain (average daily worst SDLQ VFRUH DQG Rn the BPI scale of increasing pain 0 to 10) and narcotic use criteria to be eligible for study.

2WKHU VWXG\ FULWHULD 0DOH SDWLHQWV LQ WKLV VWXG\ ZHUH UHTXLUHG WR EH \HDUV RI DJH with a documented histological or cytological diagnosis of prostate cancer and evidence of metastasis to the bone. 3DWLHQWV ZHUH WR KDYH UHFHLYHG SULRU GRFHWD[HO F\FOHV RU have progressed after any amount of docetaxel) and either abiraterone or enzalutamide with evidence of investigator-assessed prostate cancer progression. If a patient had an $( UHODWHG WR SULRU WUHDWPHQW WKH $( ZDV WR KDYH UHVROYHG WR EDVHOLQH RU &7&$( Y Grade 1. Patients without prior orchiectomy must have been taking LHRH analogue

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

therapy at baseline and concomitantly throughout the study. In addition, patients were to KDYH D VHUXP WHVWRVWHURQH OHYHO RI ޒ QJG/ DQG DGHTXDWH RUJDQ DQG PDUURZ IXQFWLRQ as determined by clinical laboratory tests.

Eligible subjects were randomized 1:1 to the cabozantinib or mitoxantrone plus prednisone treatment arms. Randomization was stratified by the following factors: prior treatment with cabazitaxel and ECOG performance status.

Based on treatment assignment, patients received one of the following regimens:

x Cabozantinib treatment arm: oral cabozantinib qd plus mitoxantrone-matched placebo infusion every 3 weeks plus oral prednisone-matched placebo bid x Mitoxantrone-prednisone treatment arm: mitoxantrone infusion 12 mg/m2 q3w plus oral prednisone-matched placebo qd

Patients received study treatment as long as they continued to experience clinical benefit, as determined by the investigator. Reasons for discontinuation of treatment included, among others, an unacceptable toxicity or the need for non-protocol systemic anticancer therapy. Crossover between protocol treatment arms was not allowed.

Clinic visits occurred at minimum every 3 weeks through treatment discontinuation with extended follow-up to assess survival status and to document receipt of subsequent anticancer therapy. Routine safety evaluations included assessments of AEs, vital signs, laboratory test, and concomitant medications. New or worsening AEs were collected at study visits, over the phone, or by spontaneous subject report from informed consent through 30 days after the date of the decision to discontinue study treatment.

Study XL184-306 was stopped early due to lack of a survival benefit observed in Study XL184-307; active patients were required to discontinue study treatment and revert to standard of care thereby limiting the overall exposure to treatment on the study. Based on the available data, the study did not meet its primary endpoint of the proportion of patients with a pain response at Week 6 confirmed at Week 12.

Phase 3 Study XL184-301

Study title: An International, Randomized, Double-Blinded, Phase 3 Efficacy Study of XL184 Versus Placebo in Subjects with Unresectable, Locally Advanced, or Metastatic Medullary Thyroid Cancer

Details of this study protocol can be found in Dr. Ruthann Giusti’s review of NDA 203756 at the link provided:

http://www.accessdata.fda.gov/drugsatfda docs/nda/2012/203756Orig1s000MedR.pdf

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

6 Review of Efficacy

Efficacy Summary This sNDA was supported by data from Trial XL184-308, entitled A Phase 3, Randomized, Controlled Study of Cabozantinib (XL184) versus Everolimus in Subjects with Metastatic Renal Cell Carcinoma that has Progressed after Prior VEGFR Tyrosine Kinase Inhibitor Therapy. This open-label international trial randomized patients with previously untreated, advanced RCC to receive cabozantinib 60 mg orally once daily or everolimus 10 mg orally once daily until disease progression or unacceptable toxicity. Randomization was stratified by number of prior VEGFR-targeting TKI therapies (1 vs. 2 or more) and Memorial Sloan-Kettering Cancer Center (MKSCC) prognostic score for previously treated patients with RCC (0 vs. 1 vs. 2 or 3 risk factors). The primary efficacy endpoint was PFS, with OS and ORR as key secondary endpoints.

The intent to treat (ITT) population contained 658 patients randomized to receive cabozantinib (n=330) or everolimus (n=328). Demographics and baseline disease characteristics appeared generally well-balanced between treatment arms. The majority of patients were male (75%) with a median age of 62 years. Seventy-one percent (71%) received only one prior VEGFR-directed therapy, 69% received only one prior anti- angiogenic therapy, and 41% received sunitinib as their only prior therapy. Forty-six percent of patients were in the MSKCC favorable risk category, 42% were in the intermediate risk category, and 13% were in the poor risk category.

The results of XL184-308 demonstrated a statistically significant and clinically meaningful improvement in PFS, as determined by the IRC, in the first 375 patients randomized (PITT population) to cabozantinib compared to those treated with everolimus with a 3.6 month improvement in median PFS with stratification data from the IVRS at the time of randomization (HR=0.58 [95% CI: 0.45, 0.74]; stratified log rank p-value <0.0001). The ORR (as determined by the IRC) was 17% (95% CI: 13, 22) versus 3% (95% CI: 2, 6) in the cabozantinib and everolimus arms, respectively (unstratified p-value <0.0001). The Kaplan-Meier estimate of median duration of response (DOR) had not been reached (95% CI: 7.2, NE) in the cabozantinib arm and was 7.4 months (95% CI: 1.9, NE) in the everolimus arm. The treatment effect of cabozantinib appeared consistent across all patient subgroups analyzed.

At the time of the primary PFS analysis OS data was not mature. On 11 February 2016, the Applicant submitted the results of the second interim OS analysis with a cutoff date of 31 December 2015. The analysis demonstrated a statistically significant difference in median OS for patients in the cabozantinib arm compared with the everolimus arm with a 4.9 month improvement in median OS with stratification data from the IVRS at the time of randomization (HR 0.66 [95% CI: 0.53, 0.83] stratified log-rank p-value = 0.0003). Kaplan-Meier estimates of median duration of OS were 21.4 months in the cabozantinib arm and 16.5 months in the everolimus arm. A number of sensitivity

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

analyses performed by the applicant as well as the FDA support the robustness of the primary PFS findings.

With the recent approval of nivolumab, patients with previously treated metastatic RCC in the US may no longer receive everolimus as second-line therapy. However, there is no comparative efficacy requirement to support regular approval, everolimus is an approved and active comparator, and the results of XL184-308 show statistically significant and clinically meaningful improvements in PFS and OS compared to everolimus.

6.1 Indication

The Applicant proposed the following indication for the cabozantinib label:

CABOMETYX is indicated for the treatment of advanced renal cell carcinoma (RCC) in patients who have received (b) (4) prior therapy.

This reviewer recommends the following indication:

CABOMETYX is indicated for the treatment of advanced renal cell carcinoma (RCC) in patients who have received prior anti-angiogenic therapy.

Reviewer’s Comment: Labeling negotiations over the indication are ongoing at the time of this review. Similar language was used in the labeling for nivolumab and axitinib.

6.1.1 Methods

The Applicant submitted one clinical trial to support the efficacy of cabozantinib for the NDA review. Trial XL184-308 is described in detail in Section 5.3. The review strategy is described in detail in Section 5.2.

6.1.2 Demographics

The key demographics are shown in Table 9. Approximately 30% of patients in each WUHDWPHQW JURXS ZHUH DJH DQG DSSUR[LPDWHO\ LQ HDFK JURXS ZHUH 75. Approximately 75% of patients in each group were male, and approximately 80% were white.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 9. XL184-308 Key Demographic Characteristics ITT Population PITT Population Cabozantinib Everolimus Cabozantinib Everolimus (N = 330) (N = 328) (N = 187) (N = 188) Median Age (years) 62.5 62 62 61 Male Gender 77% 73% 76% 69% White Race 82% 80% 84% 78% Non-Hispanic Ethnicity 84% 83% 86% 85% Europe Region 51% 47% 44% 45% North America Region 36% 37% 41% 34% Prior VEGFR-TKI = 1 71% 70% 73% 72% Prior VEGFR-7., 29% 30% 27% 28% MSKCC risk factors = 0 45% 46% 43% 44% MSKCC risk factors = 1 42% 41% 42% 40% MSKCC risk factors = 2 or 3 13% 13% 16% 16% Baseline KPS 70 29 (9%) 22 (7%) 15 (8%) 16 (9%) 80 75 (23%) 90 (27%) 43 (23%) 56 (30%) 90 127 (38%) 142 (43%) 72 (39%) 81 (43%) 100 99 (30%) 74 (23%) 57 (30%) 35 (19%) Source: dataset adtte; variables: trt01p, age, sex, race, region, prtkcgr1, and mskccgr1

Reviewer’s comment: Treatment arms appeared well balanced with respect to age, gender, race, ethnicity, enrollment region, number of prior VEGF TKI therapies, and MSKCC risk factors.

Table 10 summarizes the cancer history and baseline disease status of the population enrolled in XL184-308. Table 10. XL184-308 Cancer History and Baseline Status ITT Population PITT Population Cabozantinib Everolimus Cabozantinib Everolimus (N = 330) (N = 328) (N = 187) (N = 188) Dx of RCC with a clear cell component by histol or cytol, n (%) 330 (100%) 327 (100%) 187 (100%) 187 (100%) Time since initial histol/cytol diagnosis to randomization <1 year 59 (18%) 76 (23%) 34 (18%) 44 (23%) 1 year 271 (82%) 251 (77%) 153 (82) 143 (76%) Median (range) (years) 2.8 (0-30) 2.5 (0-33) 2.6 (0-33) 2.4 (0-33)

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Disease stage, n (%) IV 272 (82%) 287 (88%) 153 (82%) 166 (88%) III 34 (10%) 24 (7%) 20 (11%) 13 (7%) Unknown 24 (7) 16 (5%) 14 (8%) 8 (4%) Extent of disease by IRC, n (%) Visceral 241 (73%) 245 (75%) 139 (74%) 142 (76%) Lung 204 (62%) 212 (65%) 115 (61%) 126 (67%) Liver 88 (27%) 103 (31%) 52 (28%) 58 (31%) Brain 2 (1%) 1 (<1%) 2 (1%) 1 (1%) Lymph Node 206 (62%) 199 (61%) 124 (66%) 110 (59%) Kidney 70 (21%) 66 (20%) 46 (25%) 36 (19%) Bone (CT or MRI) 77 (23%) 65 (20%) 39 (21%) 32 (17%) Other 23 (7%) 21 (6%) 16 (9%) 10 (5%) Involved Organs by IRC, n (%) 1 59 (18%) 56 (17%) 31 (17%) 31 (16%) 2 101 (31%) 77 (23%) 57 (30%) 48 (26%) 168 (51%) 190 (58%) 98 (52%) 105 (56%) Missing 2 (1%) 5 (2%) 1 (1%) 4 (2%) Sum of Lesion Diameters (mm) N 330 327 187 187 Mean (SD) 77 (55) 81 (54) 81 (57) 88 (56) Median (range) 65 (0-291) 65 (0-258) 70 (0-291) 77 (0-231) Source: dataset adtte; variables: trt01p, bkps, diagtgr1, cdstage, metlngfl, metlivfl, metbrnfl, metbonfl, metlymfl, metbonfl, metothfl, numorg, and bsod

Reviewer’s comment: The population enrolled in XL184-308 appeared well balanced between treatment arms with respect to cancer history and baseline disease status.

All patients received prior systemic anticancer therapy. About 70% of patients had received one prior VEGFR-TKI and about 25% received two prior VEGFR-TKIs. Sixty nine percent of patients had received one prior anti-angiogenic therapy. The most common prior therapy was sunitinib (64% cabozantinib, 63% everolimus; ITT population). Most subjects were on prior VEGFR-TKI treatment for > 6 months. In addition, approximately 85% of patients had undergone prior nephrectomy, and approximately one third had received prior radiation (Table 11).

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 11. XL184-308 Prior Anti-Cancer Therapies ITT Population PITT Population Cabozantinib Everolimus Cabozantinib Everolimus (N = 330) (N = 328) (N = 187) (N = 188) Prior Radiation Therapies N 110 (33%) 108 (33%) 56 (30%) 61 (32%) Mean (SD) 1.2 ± 0.5 1.2 ± 0.4 1.2 ± 0.5 1.2 ± 0.4 Median (range) 1.0 (1.4) 1.0 (1.3) 1.0 (1.4) 1.0 (1.3) Prior Systemic Therapies N 330 (100%) 328 (100%) 187 (100%) 188 (100%) Mean 1.6 ± 0.8 1.7 ± 0.9 1.5 ± 0.8 1.6 ± 0.9 Median 1.0 (1-6) 1.0 (1-7) 1.0 (1-6) 1.0 (1-7) Prior Nephrectomy 283 (86%) 278 (85%) 157 (84%) 153 (81%) Source: dataset adtte; variables: trt01p, numprra, numprac, and prnsgfl

All patients had received at least one prior kinase inhibitor, per the protocol eligibility criteria. The most frequently used prior kinase inhibitors in both treatment arms were sunitinib, pazopanib, and axitinib. Only three patients had received prior lenvatinib, and none had received a prior anti-PD1 agent (Table 12).

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 12. XL184-308 Prior Chemotherapies ITT Population PITT Population Cabozantinib Everolimus Cabozantinib Everolimus (N = 330) (N = 328) (N = 187) (N = 188) Sunitinib 210 (64%) 205 (63%) 114 (61%) 113 (60%) Pazopanib 144 (44%) 136 (41%) 87 (47%) 78 (41%) Axitinib 52 (16%) 55 (17%) 28 (15%) 28 (15%) Sorafenib 21 (6%) 31 (9%) 11 (6%) 19 (10%) Interleukins 20 (6%) 29 (9%) 11 (6%) 13 (7%) Interferon 19 (6%) 24 (7%) 6 (3%) 13 (7%) Monoclonal antibodies 17 (5%) 14 (4%) 9 (5%) 11 (6%) Investigational drug 11 (3%) 15 (5%) 5 (3%) 9 (5%) Tivozanib 5 (15%) 5 (15%) 2 (1%) 4 (2%) Bevacizumab 5 (15%) 11 (3%) 1 (<1%) 7 (4%) Lenvatinib 3 (1%) 0 (0%) 2 (1%) 0 (0%) Erlotinib 2 (1%) 0 (0%) 1 (1%) 0 (0%) Ipilimumab 1 (<1%) 0 (0%) 1 (<1%) 0 (0%) Hydroxyprogesterone 1 (<1%) 0 (0%) 1 (1%) 0 (0%) Tasquinimod 1 (<1%) 1 (<1%) 0 (0%) 1 (1%) Paclitaxel with carboplatin 1 (<1%) 0 (0%) 0 (0%) 0 (0%) Vinblastine 1 (<1%) 2 (1%) 0 (0%) 1 (1%) Zoledronic acid 1 (<1%) 0 (0%) 0 (0%) 0 (0%) Regorafenib 0 (0%) 2 (1%) 0 (0%) 0 (0%) Actinimycines 0 (0%) 1 (<1%) 0 (0%) 1 (1%) Cyclophosphamide 0 (0%) 3 (1%) 0 (0%) 1 (1%) Fluorouracil 0 (0%) 2 (1%) 0 (0%) 2 (1%) FOLFOX 0 (0%) 1 (<1%) 0 (0%) 1 (1%) Gemcitabine 0 (0%) 2 (1%) 0 (0%) 0 (%) Hyrdoxychloroquine 0 (0%) 1 (<1%) 0 (0%) 1 (1%) Methotrexate 0 (0%) 1 (<1%) 0 (0%) 1 (1%) Temsirolimus 0 (0%) 2 (1%) 0 (0%) 0 (0%) Source: dataset adcm; variables: trtp, pittfl, cmcat; cmdecod, and cmclas

Reviewer’s comments:

1. The treatment arms of XL184-308 appeared well balanced with respect to prior anti-cancer therapy.

2. Prior systemic therapies other than VEGFR-TKIs were observed infrequently with no individual therapies having been used in > 10% of patients. Only about 5% of patients received prior anti-PDL1 immunotherapy, most commonly nivolumab.

3. The treatment history of the population studied supports this reviewer’s recommendation for the indication granted (the recommended wording of the indication is discussed in Section 6.1 of this review).

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

6.1.3 Subject Disposition

More patients in the everolimus arm discontinued treatment compared to the cabozantinib arm (80% vs 60%). This difference was driven by discontinuations attributed to progressive disease (48% vs. 37%) and clinical deterioration (16% vs. 8%). Rates of discontinuation due to adverse events were similar in both groups. Table 13. XL184-308 Subject Disposition ITT Population PITT Population Cabozantinib Everolimus Cabozantinib Everolimus (N = 330) (N = 328) (N = 187) (N = 188) Treatment ongoing 133 (40%) 67 (20%) 56 (30%) 33 (18%) Discontinued treatment 197 (60%) 261 (80%) 131 (70%) 155 (82%) Progressive disease 122 (37%) 158 (48%) 82 (44%) 92 (49%) Adverse event 32 (10%) 31 (9%) 21 (11%) 20 (11%) Clinical deterioration 28 (8%) 51 (16%) 18 (10%) 29 (16%) Withdrawal by subject 6 (2%) 11 (3%) 3 (2%) 7 (4%) Physician decision 5 (2%) 2 (1%) 4 (2%) 2 (1%) Lack of efficacy 3 (1%) 0 (0%) 2 (1%) 0 (0%) Protocol violation 1 (<1%) 1 (<1%) 1 (1%) 1 (1%) No study treatment given 0 (%) 5 (2%) 0 (0%) 3 (2%) Sponsor decision 0 (0%) 1 (<1%) 0 (0%) 0 (0%) Other 0 (0%) 1 (<%) 0 (0%) 1 (1%) Source: adsl.xpt; variables TRT01P, PITTFL, and EOTREAS

Reviewer’s Comment: Rates of trial discontinuation which the Applicant attributed to clinical deterioration in the cabozantinib group of the overall as-treated population were numerically slightly higher among patients age 65 or older. Rates of trial discontinuation attributed to progressive disease in the cabozantinib group of the overall as-treated population were numerically slightly higher among patients younger than age 65.

For 17 patients (6 randomized to cabozantinib and 11 randomized to everolimus), the reason for discontinuing XL184-308 was “withdrawal by subject”. The Applicant submitted narrative reports for two of those 17 patients (Patients 36073575 and 54023044) x Patient 36073575 was a 66 year old male with a baseline KPS of 100% and prior nephrectomy was hospitalized on Day 32 for Grade 1 bullous dermatitis. Cabozantinib was interrupted. On Day 37, cabozantinib was resumed at the same dose. On Day 142, bullous dermatitis recurred (Grade 2). On Day 175, cabozantinib was interrupted due to bullous dermatitis. On Day 230, the patient requested to stop treatment. As of the data cutoff (Day 298), the patient was alive. x Patient 54023044 was a 71-year-old white female with a baseline KPS of 80% and prior nephrectomy was diagnosed on Day 169 with Grade 3 anemia and was hospitalized on Day 175. On Day 187, the patient was hospitalized for Grade 3

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

fatigue. On Day 200, the patient withdrew consent for treatment. On 23 Day 250, 47 days after the last dose of cabozantinib, the patient died of progressive RCC.

Eight of the 17 patients whose reason for discontinuing was “withdrawal by subject” reported a total of 15 Grade 3-4 AEs (Table 14). Table 14. XL184-308 Grade 3-4 AEs Reported by Patients who Withdrew (ITT pop.) Study Days Study Day of Patient Event Max Grade of AE Withdrawal Cabozantinib 33573783 Asthenia 3 102-140 111 Decreased appetite 3 102-140 36073575 Blood pressure increased 3 32-32 229 Hypokalemia 3 169-197 Hypomagnesemia 3 141-197 36073841 Hypertensive crisis 3 32-33 70 54023044 Anemia 3 169-176 199 Fatigue 3 187-201 Hypokalemia 3 169-197 Everolimus 14133111 Nausea 3 391-401 456 14173706 Hyperglycemia 3 13- 63 Hypertriglyceridemia 4 27- 44443365 Cholecystitis 3 55-66 77 61193079 Pneumonia 3 229- 253 Source: datasets: adae and adsl where EOTREAS = “withdrawal by subject”; variables: subjid, trtp, aedecod, and aetoxgr

Reviewer’s comment: The number of who discontinued treatment because of “withdrawal by subject” and had Grade 3-4 AE(s) prior to withdrawing appeared balanced between treatment arms.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 15. XL184-308 Subject Disposition by Age ITT Population PITT Population Cabozantinib Everolimus Cabozantinib Everolimus (N = 330) (N = 328) (N = 187) (N = 188) Adverse Event <65 13 (4%) 14 (4%) 11 (6%) 11 (6%) 19 (6%) 17 (5%) 10 (5%) 9 (5%) No study treatment <65 0 4 0 (0%) 2 (1%) 010(0%)1(1%) Clinical Deterioration <65 17 (5%) 29 (9%) 9 (5%) 18 (10%) 12 (4%) 21 (6%) 9 (5%) 11 (6%) Lack Of Efficacy <65 2 (1%) 0 (0%) 2 (1%) 0 (0%) 1 (<1%) 0 (0%) 0 (0%) 0 (0%) Other <65 0 (0%) 1 (<1%) 0 (0%) 1 (1%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) Physician Decision <65 3 (1%) 2 (1%) 2 (1%) 2 (1%) 2 (1%) 0 (0%) 2 (1%) 0 (0%) Progressive Disease <65 85 (26%) 99 (30%) 57 (3%) 58 (31%) 37 (11%) 59 (20%) 25 (13%) 34 (18%) Protocol Violation <65 1 (<1%) 0 (%) 1 (1%) 0 (0%) 0 (0%) 1 (<1%) 0 (0%) 1 (1%) Sponsor Decision <65 0 (0%) 0 (%) 0 (0%) 0 (0%) 0 (0%) 1 (<1%) 0 (0%) 0 (0%) Withdrawal By Subject <65 3 (1%) 8 (2%) 2 (1%) 5 (3%) 3 (1%) 3 (1%) 1 (1%) 2 (1%) Source: adsl.xpt; variables TRT01P, AGE, PITTFL, and EOTREAS

Reviewer's Comment: Patients under age 65 appeared at slightly higher risk of clinical deterioration and progressive disease. This finding could reflect an association between younger age and more aggressive underlying disease; thus, no firm conclusions can be drawn from this observation.

Table 16. XL184-308 Subject Disposition by Sex ITT Population PITT Population Cabozantinib Everolimus Cabozantinib Everolimus (N = 330) (N = 328) (N = 187) (N = 188) Adverse Event M 21 (6%) 21 (3%) 13 (7%) 12 (6%) F 11 (3%) 10 (30%) 8 (4%) 8 (4%) Missing 0 (0%) 0 (0%) 0 (0%) 0 (0%) Clinical Deterioration M 20 (%) 40 (%) 13 (7%) 23 (13%) F 8 (2%) 11 (%) 5 (3%) 6 (3%) Missing 0 (0%) 0 (0%) 0 (0%) 0 (0%) Lack Of Efficacy M 3 (1%) 0 (0%) 2 (1%) 0 (0%) F 0 (0%) 0 (0%) 0 (0%) 0 (0%) Missing 0 (0%) 0 (0%) 0 (0%) 0 (0%) Other M 0 (0%) 1 (<1%) 0 (0%) 1 (<1%) F 0 (0%) 0 (0%) 0 (0%) 0 (0%) Missing 0 (0%) 0 (0%) 0 (0%) 0 (0%) No treatment given M 0 (0%) 3 (1%) 0 (0%) 2 (1%) F 0 (0%) 2 (1%) 0 (0%) 1 (1%)

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Missing 0 (0%) 0 (0%) 0 (0%) 0 (0%) Physician Decision M 5 (2%) 1 (<1%) 4 (2%) 1 (<1%) F 0 (0%) 0 (0%) 0 (0%) 0 (0%) Missing 0 (0%) 1 (<1%) 0 (0%) 1 (1%) Progressive Disease M 96 (29%) 115 (35%) 64 (34%) 64 (34%) F 26 (8%) 43 (13%) 18 (10%) 28 (15%) Missing 0 (0%) 0 0%) 0 (0%) 0 (0%) Protocol Violation M 0 (0%) 0 (0%) 0 (0%) 0 (0%) F 1 (<1%) 1 (1<%) 1 (1%) 1 (1%) Missing 0 (0%) 0 (0%) 0 (0%) 0 (0%) Sponsor Decision M 0 (0%) 1 (<1%) 0 (0%) 0 (0%) F 0 (0%) 0 (0%) 0 (0%) 0 (0%) Missing 0 (0%) 0 (0%) 0 (0%) 0 (0%) Withdrawal By Subject M 4 (1%) 10 (3%) 2 (1%) 6 (3%) F 2 (1%) 1 (<1%) 1 (1%) 1 (1%) Missing 0 (0%) 0 (0%) 0 (0%) 0 (0%) Source: adsl.xpt; variables TRT01P, SEX, PITTFL, and EOTREAS

Table 17 summarizes post-cabozantinib systemic anti-cancer treatments received by patients in each arm. Table 17. XL184-308 Concomitant and Subsequent Chemotherapy ITT Population PITT Population Cabozantinib Everolimus Cabozantinib Everolimus (N = 330) (N = 328) (N = 187) (N = 188) Everolimus 75 (23%) 13 (4%) 52 (28%) 7 (4%) Kinase inhibitors 54 (16%) 133 (41%) 36 (20%) 84 (45%) Monoclonal antibodies 6 (2%) 8 (2%) 5 (3%) 4 (2%) Pyrimidine analogues 5 (2%) 3 (1%) 4 (2%) 2 (1%) Interferons 3 (1%) 6 (2%) 3 (2%) 3 (2%) Nitrogen mustard analogues 1 (<1%) 1 (<1%) 1 (1%) 1 (1%) Ixabepalone 1 (<1%) 0 1 (1%) 0 Platinum compounds 1 (<1%) 1 (1%) 0 1 (1%) Vinblastine 1 (<1%) 1 (<1%) 1 (1%) 1 (1%) Interleukins 0 4 (1%) 0 3 (2%) Source: subjid, trtp, cmcat, cmclas, and cmdecod

Reviewer’s Comments: 1. Patients on the everolimus arm received more concomitant and subsequent kinase inhibitors, whereas patients on the cabozantinib arm received more everolimus. Few patients in either arm received subsequent chemotherapeutic agents other than kinase inhibitors or everolimus. 2. All concomitant/subsequent monoclonal antibodies were bevacizumab except one which was not otherwise specified. 3. No patients received concomitant or subsequent immunotherapies. 4. There did not appear to be a notable imbalance of subsequent therapies between arms.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

6.1.4 Analysis of Primary Endpoint(s)

Trial XL184-308 met its primary endpoint of demonstrating improved PFS in the first 375 randomized patients (PITT population), by IRC-determined RECIST criteria. The results of the analysis demonstrated a statistically significant improvement in PFS for patients in the cabozantinib arm compared with the everolimus arm. There was a statistically significant improvement in PFS for subjects in the cabozantinib arm compared with the everolimus arm, with a 3.6 month difference in median PFS (7.4 vs. 3.8 months). The hazard ratio adjusted for stratification factors was 0.59 (95% CI: 0.46, 0.76; stratified log-rank p-value < 0.0001) per eCRF and 0.58 (95% CI: 0.45, 0.74; stratified log-rank p- value < 0.0001) per IVRS.

Reviewer’s comment: The sponsor pre-specified their primary analysis to be stratified by eCRF values. The FDA recommended that the results from the IVRS analysis be used in the label to more closely follow the ITT principle; however, analyses performed using CRF data and IVRS showed consistent results (see Table 18). Table 18. XL184-308 Progression-Free Survival per IRC (PITT Population) Cabozantinib Everolimus (N = 187) (N = 188) No. of Events 121 (65%) 126 (67%) Median (months) 7.4 (5.6, 9.1) 3.8 (3.7, 5.4) p-value (stratified log-rank per eCRF) < 0.0001 Hazard ratio (95% CI; stratified per eCRF) 0.59 (0.46, 0.76) p-value (stratified log-rank per IVRS) < 0.0001 Hazard ratio (95% CI; stratified per IVRS) 0.58 (0.45, 0.74) Source: dataset adtte where pittfl = “Y” and param = “PFS1 per IRC by trtp, aval, cnsr (censor where cnsr 0)

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Figure 1. Trial XL184-308 Progression Free Survival per IRC (PITT Population)

CABOZANTINIB EVEROLIMUS 0.8

0.6

0.4

0.2

Progression-Free Survival (%) 0.0 0 5 10 15 20 Months

Source: dataset adtte where pittfl = “Y” and param = “PFS1 per IRC (months) by trtp, aval, cnsr (censor where cnsr 0)

Patients in the primary analysis were censored for PFS if they were determined to have progressive disease by ICR or if they died. Table 19 shows the reasons for censoring of patients.

Table 19. XL184-308 Patient Censoring in PFS Analysis (PITT Population) Cabozantinib Everolimus (N = 187) (N = 188) No. of patients censored 66 (35%) 62 (33%) PLVVHG DVVHVVPHQWV 1 (<1%) 5 (3%) Anti-cancer therapy 24 (13%) 31 (16%) No event by last assessment 39 (21%) 23 (12%) No post-baseline assessment 0 3 (2%) Surgery 2 (1%) 0 Event 121 (65%) 126 (67%) Death 8 (4%) 13 (7%) Progression 113 (60%) 113 (60%) Source: dataset adtte where pittfl = “Y” and param = “PFS1 per IRC (months) by trtp, aval, cnsr (censor where cnsr 0)

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Reviewer’s Comments: 1. Reviewer’s comment: The datasets submitted corroborate the efficacy findings reported in the Clinical Study Report. 2. This statistically significant and magnitude of median PFS improvement representing a clinically meaningful result supports this reviewers recommendation for approval of cabozantinib in advanced RCC patients who have received prior anti-angiogenic therapy.

6.1.5 Analysis of Secondary Endpoints(s)

Overall Survival

At the time of the primary PFS analysis of the PITT population (data cutoff May 22, 2015), a first planned interim analysis of OS was performed in the ITT population. At that time, 202 deaths had been observed: 89 (27%) deaths on the cabozantinib arm and 113 (34%) deaths on the everolimus arm. Kaplan-Meier estimates of median duration of OS had not been reached in either treatment arm; the HR was 0.69 (95% CI: 0.53, 0.92; p = 0.01). This difference was not statistically significant, as the p-YDOXH RI 0.0019 required to achieve statistical significance at the time of the first interim analysis was not reached.

Reviewer comment: The OS data was not mature at the time of the primary analysis, as only 202 (49%) of the 408 required for the prespecified primary analysis of OS had been observed.

On 11 February 2016, The Applicant submitted a Clinical Study Report Addendum containing results of a second interim OS analysis in the ITT population, with a cutoff date of 31 December 2015. This analysis included 320 total deaths, representing 78% of the total 408 deaths required for the planned final OS analysis. The minimum time of follow-up from randomization of the 658th patient through 31 December 2015 was 13 months. Survival status as of 31 December 2015 was determined for 642 (98%) of the 658 randomized patients. For the 16 patients for which it could not be established (6 in the cabozantinib arm and 10 in the everolimus arm), 13 withdrew consent and 3 were lost to follow-up. These patients were censored at the date of most recent contact prior to the data cutoff.

The analysis demonstrated a statistically significant difference in duration of OS for subjects in the cabozantinib arm compared with the everolimus arm: There was a 4.9 month improvement in median OS (21.4 vs. 16.5 months) for subjects in the cabozantinib arm compared with the everolimus arm. The hazard ratio adjusted for stratification factors was 0.67 (95% CI: 0.53, 0.83; stratified log-rank p-value = 0.0003) per eCRF and 0.66 (95% CI: 0.53, 0.83; stratified log-rank p-value = 0.0003) per IVRS (Table 19 and Figure 2). The p-YDOXH RI 0.0163 (78% information fraction) required to achieve statistical significance at the time of this unplanned second interim analysis was

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

reached. Consistent results were see OS analyses using either the IVRS data or CRF data (see Table 20). Table 20. XL184-308 Overall Survival (ITT Population) 1st Interim 2nd Interim (22 May 2015 cutoff) (31 December 2015 cutoff) Cabozantinib Everolimus Cabozantinib Everolimus N = 330 N = 328 N = 330 N = 328 No. of patients censored 241 (73%) 215 (66%) 190 (58%) 148 (45%) No. of Deaths 89 (27%) 113 (34%) 140 (42%) 180 (55%) Median OS (months, 18.2 (16.1, NE) NE (13.9, NE) 21.4 (18.7, NE) 16.5 (14.7, 95% CI) 18.8) p-value (stratified log- 0.006 0.0003 rank per eCRF) Hazard ratio (95% CI; 0.68 (0.51, 0.90) 0.67 (0.53, 0.83) stratified per eCRF) p-value (stratified log- 0.007 0.0003 rank per IVRS) Hazard ratio (95% CI; 0.68 (0.52, 0.90) 0.66 (0.53, 0.83) stratified per IVRS) p-value (unstratified log- 0.010 0.0004 rank test) Hazard ratio (95% CI; 0.69 (0.53, 0.92) 0.67 (0.54, 0.84) unstratified) Source: dataset adtt3 by trtp, aval, cnsr (censor if cnsr 0), and evntdesc

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Figure 2. XL184-308 Overall Survival (2nd Interim Analysis; ITT Population)

CABOZANTINIB 0.8 EVEROLIMUS

0.6

0.4

0.2

Progression-Free Survival (%) 0.0 0 5 10 15 20 25 30

Months

Source: dataset adtte where param = Time to death (months) by trtp, aval, cnsr (censor if cnsr 0).

Reviewer’s comment: This statistically significant and clinically meaningful improvement in median OS supports this reviewer’s recommendation for approval of cabozantinib in patients with advanced RCC who have received prior anti-angiogenic therapy.

Overall Response Rate The primary analysis of ORR (as assessed by the IRC per RECIST 1.1) was conducted in the ITT population at the time of the primary analysis of PFS (cutoff date 22 May 2015). The ORR was 17% (95% CI: 13, 22) and 3% (95% CI: 2, 6) for patients in the cabozantinib and everolimus arms, respectively (stratified chi-squared test p-value < 0.001). All responses were PRs. The median time to objective response was 1.91 months (range 1.6, 11.0) in the cabozantinib arm and 2.14 months (range 1.9, 9.2) in the everolimus arm.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 21. XL184-308 Tumor Response per RECIST 1.1 (ITT Population) Cabozantinib (N = 330) Everolimus (N = 328) Overall response rate N 57 (17%) 11 (3%) 95% CI 13%, 22% 2%, 6% Stratified CMH test p-value < 0.0001 Confirmed complete response 0 (0%) 0 (0%) 95% CI 0%, 0% 0%, 0% Confirmed partial response 57 (17%) 11 (3%) 95% CI 13%, 22% 2%, 6% Source: adrs.xpt where param = Best overall tumor response by arm and avalc

Reviewer’s Comments: 1. The primary datasets submitted corroborate the response rates presented in the Clinical Study Report. 2. The ORR results support the conclusion that cabozantinib has activity in patients with advanced RCC after prior anti-angiogenic therapy and support the recommendation for approval.

6.1.6 Other Endpoints

Duration of Response Duration of response was defined as the time from the first tumor assessment that documented PR or CR and was subsequently confirmed at least 28 days later until the date of IRC-determined disease progression (RECIST 1.1). Only patients with confirmed CR or PR were included in this analysis. The censoring rules used for the primary PFS analysis were applied. At the time of the analysis, the median duration of response had not been reached for the cabozantinib arm and was 7.4 months for the everolimus arm. Table 22. XL184-308 Duration of Response (IRC-Determined; ITT Population) Cabozantinib (N = 57) Everolimus (N = 11) No. of patients censored 44 (77%) 7 (63%) PLVVHG DVVHVVPHQWV 01(10%) Anti-cancer therapy 7 (12%) 0 No event by last ATA 36 (63%) 6 (55%) Surgery 1 (2%) 0 Event Death 1 (2%) 0 Progression 12 (21%) 4 (36%) Kaplan-Meier estimate (mo) Median (95% CI) NR (7.2, NR) 7.4 (1.9, NR) Range 0-14.8+ 0-10.1+ Source: dataset adtte where param = DOR per IRC (months) by trtp, censor if cnsr 0 +: ongoing response

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Figure 3. XL184-308 Duration of Response (IRC-Determined; ITT Population)

CABOZANTINIB EVEROLIMUS 0.8

0.6

0.4

0.2 Maintained Response (%) 0.0 0 2 4 6810 12 14 Months

Source: dataset adtte where param = DOR per IRC by trtp, censor if cnsr 0

Reviewer Comment: The data regarding DOR remain immature at the time of this review. However, these preliminary findings are consistent with the other efficacy measures reported and support the overall conclusion that cabozantinib has activity in this clinical setting.

Time to Response The median time to objective response was 1.91 months (range 1.6, 11.0) in the cabozantinib arm and 2.14 months (range 1.9, 9.2 months) in the everolimus arm. Table 23. XL184-308 Time to Tumor Response per RECIST 1.1 (ITT Population) Cabozantinib (N = 330) Everolimus (N = 328) N 57 (17%) 11 (3%) Mean ± SD 2.96 ± 2.04 3.55 ± 2.50 Median (range) 1.91 (1.6 – 11.0) 2.14 (1.9-9.2) Source: dataset adrs where PARAM = “Time to First IRC Objective Tumor Response” by trtp and aval

Reviewer’s comment: The time to tumor response was similar in the cabozantinib and everolimus arms.

Changes in Bone Scans and Skeletal Related Events

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

The bone scan response (BSR) rate was defined as the proportion of patients in the ITT population who had bone lesions on bone scan at baseline and for whom the best bone scan response at the time of data cutoff was a response, responder (R) as assessed per CAD by the IRC. The bone scan response rate was 18% (95% CI: 11, 27) in the cabozantinib arm and 10% (95% CI: 4, 19) in the everolimus arm. Table 24. XL184-308 Bone Scan Response (IRC-Determined; ITT Population) Cabozantinib (N = 330) Everolimus (N = 328) Bone lesions at baseline 105 (32%) 73 (22%) Responder 9 (18%) 7 (10%) Stable disease 69 (66%) 47 (64%) Progressive disease 17 (16%) 17 (23%) Source: dataset adtte where param = DOR per IRC (months) by trtp, censor if cnsr 0

The protocol defined pathologic fractures, spinal cord compression, external radiation therapy to bone, and surgery to bone as SREs. A total of 12% of patients in the cabozantinib arm and 14% in the everolimus arm had an SRE post-randomization. Table 25. XL184-308 Post-Randomization SREs (IRC-Determined; ITT Pop.) Cabozantinib (N = 330) Everolimus (N = 328) Any post-randomization SRE 38 (12%) 46 (14%) External radiotherapy to bone 25 (8%) 35 (11%) Pathologic fracture 16 (5%) 11 (4%) Spinal cord compression 4 (1%) 8 (2%) Surgery to bone 11 (3%) 10 (3%) Source: dataset adae; variables subjid, trtp, radbonea, aepfayn, aescayn, and surbonea

Reviewer Comment: Rates of post-randomization SREs appeared similar between treatment arms, although the trial was not powered to detect differences with respect to SREs.

Patient Reported Outcomes Heath-related quality of life (HRQOL) data were collected through the FKSI-19 and EQ- 5D-5L questionnaires. The HRQOL analysis was presented for the ITT population. The FKSI-19 questionnaire scores each symptom was scored on a scale of 1 to 5: “not at all” vs “very much.” Total scores and four disease-related symptoms (DRS) subscales (FKSI-DRS-Physical, FKSIDRS-Emotional, FKSI-Treatment Side Effects, and FKSI- Function/Well-Being) were derived. No treatment differences were apparent for the FKSI-total score and three of the four subscales, although the score for diarrhea was higher in the cabozantinib arm (data not shown). See the statistical review for further information.

Reviewer’s Comments: 1. Patient reported outcome data were exploratory, without statistical power allocated.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

2. The observed mean values for patient reported outcomes represent patients who remained on therapy, and not those who discontinued treatment, and may be confounded by the open label design of the trial. 3. The higher FKSI score for diarrhea in the cabozantinib arm is consistent with other safety data collected for XL184-308.

6.1.7 Subpopulations

The review team analyzed PFS and OS in in several prespecified patient subgroups according to age, sex, geographic region, race, and disease history (Figures 4, 5 and 6).

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Figure 4. XL184-308 Subgroup Analyses of PFS (IRC Determined, PITT Pop.)

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Figure 5. XL184-308 Subgroup Analyses of PFS (IRC-Determined, ITT Pop.)

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Figure 6. XL184-308 Subgroup Analyses of OS (Unplanned 2nd Interim, ITT Pop.)

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Reviewer’s Comment: 1. The treatment effect on PFS appeared reasonably consistent across a variety of prespecified patient subgroups. 2. In the PITT population, the treatment effect on PFS appeared limited to men. However, in the larger ITT population, an effect on PFS was apparent in both women and men. Given the sample sizes and the benefit in the full PITT population no conclusions can be drawn from this.

6.1.8 Analysis of Clinical Information Relevant to Dosing Recommendations

The following is excerpted from the clinical pharmacology review. Cabozantinib was first approved for the treatment of medullary thyroid cancer in 2012 (as CometriqTM) under NDA 203756. The approved dosing regimen for Cometriq was 140 mg (capsules) by mouth once daily, taken at least 1 hour before or 2 hours after a meal. A post-marketing requirement was to evaluate a lower dose (b) (4) (b) (4)

In the current submission, the Applicant seeks the approval of a tablet formulation of cabozantinib (CabometyxTM) for the treatment of advanced RCC in patients who have received prior anti-angiogenic therapy. The proposed dosing regimen for Cabometyx tablets is 60 mg by mouth once daily, taken at least 1 hour before or 2 hours after a meal. In Trial XL184-308, Cabometyx (60 mg tablet QD) improved median PFS and OS; however, the dose was reduced in 60% of patients, with 20% receiving 20 mg as their lowest dose. The most frequent adverse reactions leading to dose reduction were diarrhea, palmar-plantar erythrodysesthesia syndrome, fatigue, and hypertension. The Clinical Pharmacology review team considers the Cabometyx dose of 60 mg QD to be acceptable for the RCC indication and the clinical review team agrees. A higher dose is not considered appropriate, given dose reductions occurred in 60% patients at 60 mg QD. Modeling and simulations indicated that lower starting doses such as 40 mg or 20 mg QD likely compromises efficacy though the incidences of dose reductions may be lower than 60 mg QD dose. Taking the cabozantinib concentration of 1125 ng/mL (median concentration at 60 mg QD dose) as the reference, the hazard ratio would be 1.1 and 1.39 if the exposure was reduced to 67% and 33% of the reference concentration, respectively. Reduction in tumor size in the 20 mg QD (median percent change from baseline = -4.45%) and 40 mg QD (-9.1%) starting dose groups were predicted to be smaller than that in the 60 mg QD dose group (-11.9%). Objective response rates at 40 mg QD (15.6%) and 20 mg QD (8.70%) were predicted lower than that in the 60 mg QD (19.1%) dose groups. Cross-study PK analysis indicated similar steady-state exposures at different doses across patient populations of medullary thyroid cancer (140 mg capsules), advanced RCC (60 mg tablet), and castration-resistant prostate cancer (60 mg tablet).

Reference ID: 3917720

Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 27. XL184-308 Outcome of Treatment Beyond IRC-Determined rPD Cabozantinib Everolimus (N = 106) (N = 108) RECIST SD or PR 4 (4%) 2 (2%) SOD < baseline SOD 43 (41%) 30 (28%) <20% increase in SOD vs. SOD at date of rPD 70 (66%) 76 (70%) Source: Applicant’s response to FDA Information Request dated 10 March 2016

Exploratory analyses of treatment beyond progression in XL184-308 were repeated for patients with Investigator-determined rPD. Table 28. XL184-308 Patients with Investigator-Determined rPD Cabozantinib Everolimus (N = 193) (N = 226) Continued study treatment beyond rPD 142 (74%) 188 (83%) Continued VWXG\ WUHDWPHQW GD\V beyond rPD 45 (23%) 49 (22%) Still on study treatment after rPD as of data cutoff date 29 (15%) 22 (10%) Continued study treatment GD\V beyond rPD or still 74 (38%) 71 (31%) on study treatment after rPD as of data cutoff date Source: Applicant’s response to FDA Information Request dated 10 March 2016

Table 29. XL184-308 Outcome of Treatment Beyond Investigator-Determined rPD Cabozantinib Everolimus (N = 74) (N = 71) RECIST SD or PR 5 (7%) 6 (8%) SOD < baseline SOD 34 (46%) 15 (21%) <20% increase in SOD vs. SOD at date of rPD 51 (69%) 49 (69%) Source: Applicant’s response to FDA Information Request dated 10 March 2016

Reviewer’s Comment: Similar proportions of patients were treated beyond rPD, whether determined by the IRC or investigators, in both arms. Substantial fractions of patients in both arms experienced continued tumor control (post-PD SOD < baseline SOD) after the first date of progression. The labeling will state that the protocol allowed for patients to receive treatment until disease progression or experiencing unacceptable toxicity and that patients on both arms who had disease progression could continue treatment at the discretion of the investigator.

6.1.10 Additional Efficacy Issues/Analyses

This reviewer performed the following exploratory sensitivity analyses of the results of XL184-308:

Sensitivity Analyses of Primary PFS Endpoint

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

The review team calculated the difference between treatment arms in terms of median PFS using several alternative patient groupings and censoring schemes (Table 30). Table 30. XL184-308 Sensitivity Analyses of PFS (PITT Population) Cabozantinib Everolimus Stratified HR Sensitivity Analysis (95% CI) Median PFS (months) Uniform dates 7.4 3.9 0.59 (0.45, 0.76) Investigator claims 7.3 4.0 0.59 (0.47, 0.74) Investigator-determined Radiographic PD 7.4 5.3 0.61 (0.48, 0.77) Censoring Scheme 1 5.7 3.8 0.73 (0.57, 0.92) Censoring Scheme 2 5.6 3.7 0.70 (0.56, 0.89) Censoring Scheme 3 5.7 3.7 0.58 (0.46, 0.73) Censoring Scheme 4 5.6 3.8 0.75 (0.59, 0.95) Safety Population 7.4 3.8 0.59 (0.46, 0.76) All IRC Radiographic PDs as Events 6.9 3.8 0.59 (0.46, 0.75) Combination of IRC and Investigator 5.6 3.7 0.59 (0.47, 0.74) Source: FDA statistical review Table 31. XL184-308 Sensitivity Analyses of PFS (ITT Population) Cabozantinib Everolimus Stratified HR Sensitivity Analysis (95% CI) Median PFS (months) Uniform dates 7.4 3.9 0.52 (0.42, 0.64) Investigator claims 7.3 3.9 0.53 (0.45, 0.64) Investigator-determined Radiographic PD 7.4 5.1 0.54 (0.45, 0.66) Censoring Scheme 1 6.0 3.9 0.66 (0.54, 0.79) Censoring Scheme 2 5.8 3.7 0.65 (0.54, 0.78) Censoring Scheme 3 6.0 3.7 0.53 (0.45, 0.64) Censoring Scheme 4 5.8 3.9 0.69 (0.57, 0.84) Safety Population 7.4 3.9 0.52 (0.43, 0.64) All IRC Radiographic PDs as Events 7.4 3.9 0.53 (0.44, 0.64) Combination of IRC and Investigator 5.8 3.7 0.53 (0.44, 0.63) First 263 Events 3.7 6.0 0.49 (0.38, 0.63) Source: FDA statistical review

Reviewer’s comments: 1. The hazard ratios for PFS were similar using all patient subgroups and censoring schemes analyzed. 2. Although the Applicant’s Clinical Study Report reported the primary efficacy analysis stratified by data entered on Case Reports Form, the FDA clinical and statistical review teams felt that stratification by the Interactive Voice Recording System was more appropriate for labeling purposes.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 32. XL184-308 PFS (IRC-Determined; ITT Population) Cabozantinib (N = 330) Everolimus (N = 328) No. of Deaths 180 (55%) 214 (65%) Median (months) 7.4 (6.6, 9.1) 3.8 (3.7, 5.1) Hazard ratio (95% CI) 0.52 (0.43, 0.64) Source: dataset adtte where param = PFS1 by trtp, aval, cnsr (censor if cnsr 0)

Figure 7. XL184-308 PFS, IRC-Determined (ITT Pop.) 1.0 m

CABOZANTINIB EVEROLIMUS 0.8

0.6

0.4

0.2 Progression-Free Survival (IRC-Deter 0.0 0 5 10 15 20

Months

Source: dataset adtte where param = PFS1 by trtp, aval, cnsr (censor if cnsr 0)

Reviewer’s Comment: The hazard ratio for PFS in the ITT population is similar to that in the pre-specified primary analysis for PFS in the PITT population.

Progression-Free Survival (Investigator-Determined, PITT Population) This reviewer repeated the PFS analysis using Investigator-determined progression dates for the PITT population. When both clinical and radiographic progression events were counted, a total of 301 events had occurred as of the 22 May 2015 data cutoff. Kaplan-Meier estimates for median duration of PFS were 7.3 months in the cabozantinib arm vs 4.0 months in the everolimus arm (HR 0.58 [95% CI: 0.46, 0.73].

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 33. XL184-308 Investigator-Determined PFS (PITT Population) Clinical or Radiographic Radiographic-Only Total censored 44 (24%) 30 (16%) 54 (29%) 47 (25%) Total events 143 (76%) 158 (84%) 133 (71%) 141 (75%) Median (months) 7.3 4.0 7.4 5.3 Stratified HR (95% CI) 0.59 (0.47, 0.74) 0.61 (0.48, 0.77) Source: dataset adtte where param = PFS3 or PFS4 by trtp, aval, cnsr (censor if cnsr 0) Figure 8. XL184-301 Inv.-Determined PFS, Clinical or Radiographic (PITT Pop.)

1.0 CABOZANTINIB EVEROLIMUS 0.8

0.6

0.4

0.2

Inv.-Determined PFS, Clinical or Radiog 0.0 0 5 10 15 20 Months

Source: dataset adtte where param = PFS3 by trtp, aval, cnsr (censor if cnsr 0)

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Figure 9. XL184-301 Inv.-Determined PFS, Clinical or Radiographic (PITT Pop.)

CABOZANTINIB EVEROLIMUS 0.8

0.6

0.4

0.2

Inv.-Determined PFS, Radiographic-On 0.0 0 5 10 15 20

Months

Source: dataset adtte where param = PFS4 by trtp, aval, cnsr (censor if cnsr 0)

Reviewer’s Comment: The hazard ratios for PFS in the PITT population based on Investigator-determined response dates were similar to those in the primary analysis for PFS using IRC-determined response dates. This was true when considering RECIST- only progression or both clinical and radiographic progression.

In the PITT population, 10 progression events in the cabozantinib arm and 17 progression events in the everolimus arm were detected clinically without radiographic progression. Table 34. XL184-308 Comparison of Progression Events (PITT population) Clinical or Radiographic Prog. Radiographic-Only Prog. Cens Cabo Evero Cens Cabo Evero ? (N = 187) (N = 188) ? (N = 187) (N = 188) Anti-cancer therapy No 6 (3%) 14 (7%) Yes 9 (5%) 17 (9%) Clinical deterioration No 10 (5%) 15 (8%) No 0 0 Death No 4 (2%) 6 (3%) No 10 (5%) 14 (7%) Documented progression No 123 (66%) 123 (66%) No 123 (66%) 127 (68%) Surgery No 0 0 Yes 0 0 No post baseline ATA Yes 0 3 (2%) Yes 0 3 (2%) PLVVHG $7$ prior to event Yes 4 (2%) 4 (2%) Yes 5 (3%) 4 (2%) No event by last ATA Yes 40 (21%) 23 (12%) Yes 40 (21%) 23 (12%) Source: dataset adtte; where PITT = “Y” and PARAM = “PFS3” or “PFS4” by TRTP, EVNTDESC, and CNSR

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Progression-Free Survival (IRC-Determined, First 375 Patients As-Treated) The protocol-specified primary efficacy analysis was based upon the first 375 patients as randomized (187 to cabozantinib and 188 to everolimus). However, three of the first 188 patients randomized to everolimus did not receive treatment. Including those three untreated patients in the primary efficacy analysis reduces the median PFS (because those three patients are assigned PFS times of zero). Table 35 presents a sensitivity analysis excluding those three patients (i.e., based on the first 375 patients as treated instead of as randomized). Table 35. XL184-308 PFS (IRC-Determined, First 375 Patients As-Treated) Cabozantinib (N = 187) Everolimus (N = 185) No. of Events 121 (65%) 126 (67%) Median (months) 7.4 (5.6, 9.1) 3.8 (3.7, 5.4) Hazard ratio (95% CI) 0.59 (0.46, 0.76) Source: dataset adtte where pittfl = “Y” and param = “PFS1 per IRC by trta, aval, cnsr (censor where cnsr 0)

7 Review of Safety

Safety Summary

The safety of cabozantinib for this supplement was evaluated in Trial XL184-308, a randomized, open-label, international trial in patients with metastatic renal cell carcinoma that has progressed after prior anti-angiogenic kinase inhibitor therapy. The clinical review of safety focused on the 653 treated patients, 331 in the cabozantinib group and 322 patients in the everolimus group. The safety review also analyzed data from three other trials; two metastatic prostate cancer (Trial XL184-307 and Trial XL184-306) and medullary thyroid cancer (Trial XL184-301) resulting in a total safety database of 1286 cabozantinib treated patients. The overall safety profile of cabozantinib in Trial XL184-308 was similar to that of the described safety of cabozantinib in the label for the indication of medullary thyroid cancer. Cabozantinib exposure was adequate to assess safety with a median duration of 32 weeks in the cabozantinib group and 19 weeks in the everolimus group. However, the median relative dose intensity was 76% in the cabozantinib group compared to 92% in the everolimus group.

Overall, 40% of patients in the cabozantinib group and 43% of patients in the everolimus group experienced serious adverse events (SAEs). The most common SAEs observed with cabozantinib were abdominal pain, pleural effusion, diarrhea, nausea, and anemia.

Adverse events were managed with dose reductions; doses were reduced in 60% of patients who received cabozantinib and in 25% of those who received everolimus.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

A higher proportion of patients died in the everolimus group (34%) compared to the cabozantinib group (27%). The majority of deaths were attributed to disease progression (22.4% in the cabozantinib group and 28.4% in the everolimus group).

AEs, unrelated to disease progression, leading to treatment discontinuation were similar between groups (9.1% in cabozantinib group and 10% in everolimus group).

The safety profile of cabozantinib is acceptable for patients with advanced renal cell carcinoma who have progressed after prior anti-angiogenic therapy. The size of the safety database and duration of cabozantinib exposure were sufficient to characterize the safety of cabozantinib in the studied patient population. The safety profile is similar to the previously characterized profile from thyroid cancer studies. Patients treated with the 60 mg tablet formulation used in the renal cell cancer studies experienced fewer events such as GI perforations and fistulas, than those treated with the 140 mg capsule formulation in the MTC trial. However, the rates of other adverse events, such as diarrhea, hypertension, and PPES were similar across tumor types.

The reviewer does not recommend a risk evaluation and mitigation strategy (REMS) given the current safety profile of cabozantinib and the experience of the medical community in managing toxicities related to tyrosine kinase inhibitor therapy. Recommendations for safe and effective use of cabozantinib will continue to be reflected in labeling and patient medication guide.

A brief summary of adverse events is summarized in Table 36. Table 36: XL184-308 Summary of Adverse Events (Safety Population) Cabozantinib Everolimus N=331 N=322 n (%) n (%) AE 331 (100) 321 (100) Related AE 322 (97) 293 (91) Serious AE 131 (40) 139 (43) Serious and related AE at any time 50 (15) 41 (13) Worst AE, Grade 3 or 4 226 (68) 186 (58) Worst AE, Grade 4 26 (7.9) 26 (8.1) Grade 5 AE at any time 23 (6.9) 28 (8.7) *UDGH $( GD\V DIWHU VWXG\ WUHDWPHQW 15 (4.5) 23 (7.1) Grade 5 AE > 30 days after last dose of 8 (2.4) 5 (1.6) study treatment Related Grade 5 AE at any time 1 (0.3) 2 (0.6) Source: ADSL, ADAE

Reviewer comment:

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

The rates of AE’s were similar in both treatment groups. There were higher rates of Grade 3 or 4 AEs in the cabozantinib group (68%) compared to the everolimus group (58%).

7.1 Methods

The safety review of cabozantinib in patients with metastatic renal cell carcinoma who have progressed after prior VEGFR tyrosine kinase inhibitor therapy was evaluated in Trial XL184-308 as described in Section 5.3. The Applicant mapped and coded verbatim adverse events (AE) terms for Trial XL184-308 using MedDRA version 17.0.

7.1.1 Studies/Clinical Trials Used to Evaluate Safety

The safety review is analyzed from data from Trial XL184-308 which is described in Section 5.3.

7.1.2 Categorization of Adverse Events

The Applicant graded the severity of AEs observed on Trial XL184-308 using NCI CTCAE version 4.0. The Applicant mapped and coded verbatim AE terms for Trial XL184-308 using MedDRA version 17.0.

7.1.3 Pooling of Data Across Studies/Clinical Trials to Estimate and Compare Incidence

Clinical safety data from trials of cabozantinib in mCRPC and MTC were used to review events of rare adverse events associated with cabozantinib. These trials are described in Section 5 of this review.

7.2 Adequacy of Safety Assessments

7.2.1 Overall Exposure at Appropriate Doses/Durations and Demographics of Target Populations

In the All Treated Population, 331 patients received at least one dose of cabozantinib and 322 patients received at least one dose of everolimus. The demographics of the treated patients and randomized patients were similar as only five patients were not treated. Please see Section 6.1.2 for demographics.

The median duration of therapy was 32.1 weeks in the cabozantinib group compared to 18.9 weeks in the everolimus group. The median average daily dose was 45.3 mg cabozantinib and 9.2 mg everolimus. The relative dose intensity was 75.5% in patients treated with cabozantinib versus 91.5% in patients treated with everolimus.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 37 shows the summary of exposure to the study drugs. Table 37: XL184-308 Summary of Exposure (All Treated Population) Cabozantinib Everolimus N=331 N=322 Duration of exposure (weeks) Mean 33.1 23.9 Median 32.1 18.9 Min, max 1.1, 89.3 0.9, 82.1 Average daily dose (mg/day) Mean (SD) 45.2 (13.1) 8.4 (2.0) Median 45.3 9.2 Min, max 14.2, 82.7 2.8, 14.2 Relative dose intensity Mean (SD) 75.25 (21.8) 83.9 (19.8) Median 75.5 91.5 Min, Max 23.7, 137.8 28.3, 142.3 Source: ADEX.xpt and CSR Table 46

Reviewer comment: The size of the safety database and duration of cabozantinib exposure are sufficient to characterize the safety of cabozantinib for this sNDA. While the median duration of therapy is longer in the cabozantinib group (32.1 weeks) compared to the everolimus group (18.9 weeks), dose intensity was lower in the cabozantinib group. This is likely due to the higher rate of dose interruptions and dose reductions in the cabozantinib group, which are described further in Section 7.3.3.

7.2.2 Explorations for Dose Response

Refer to Clinical Pharmacology review by Dr. Peingfei Song and Dr. Chao Liu.

7.2.3 Special Animal and/or In Vitro Testing

Not applicable to this supplement.

7.2.4 Routine Clinical Testing

The routine clinical testing of patients enrolled in the clinical trial, including efforts to examine adverse event data by monitoring laboratory tests and vital signs, appear to have been adequate.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

7.2.5 Metabolic, Clearance, and Interaction Workup

See clinical pharmacology review.

7.2.6 Evaluation for Potential Adverse Events for Similar Drugs in Drug Class

Evaluations for adverse events related to drugs similar to cabozantinib were adequate (see Section 2.4).

7.3 Major Safety Results

7.3.1 Deaths

Deaths listed included deaths during treatment, deaths occurring up to 30-days of last dose of study drug, and deaths more than 30 days after the last dose of study treatment, as of the clinical cut-off data for the report on May 22, 2015.

A higher proportion of patients died in the everolimus group (34%, 110 patients) compared to the patients in the cabozantinib group (27%, 90 patients). The majority of deaths in both groups were due to disease progression.

Table 38 summarizes the primary causes of death for patients in the all treated safety population in Trial XL184-308. Table 38: XL184-308 Summary of Deaths (All Treated Population) Primary Cause of Death Cabozantinib Everolimus N=331 N=322 n (%) n (%) Patients who Died 90 (27.2) 110 (34.2) Disease Progression 73 (22.1) 91 (28.3) Other 17 (5.1) 19 (5.9) 30 Days of Last Dose Disease Progression 8 (2.4) 11 (3.4) Other 7 (2.1) 12 (3.7) After 30 days of Last Dose Disease Progression 65 (19.6) 80 (24.8) Other 10 (3.0) 7 (2.2) Source: ADSL.xpt

In Trial XL184-308, thirty-six patients died for reasons other than disease progression. The investigator classified the primary reason for death as “other” for seventeen (5.1%) patients in the cabozantinib group and nineteen (5.9%) patients in the everolimus group. Nineteen of these deaths occurred within 30-days of study treatment. Seven were in the cabozantinib group and twelve in the everolimus group.

Reference ID: 3917720

Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Patient ID Summary of Case Reviewer Comment Group Cabozantinib was interrupted on Day 32 possibly related to the due to abdominal pain. Treatment included post-operative dexamethasone, omeprazole, and complications. amoxicillin/clavulanate. His abdominal pain resolved and Cabozantinib was resumed on Day 37 at unchanged dose. On (b) (6) (b) (6) (Study Day 41), he was hospitalized with Grade 3 acute ischemic leg. Treatment with cabozantinib was discontinued due to clinical deterioration and acute ischemic leg. The patient underwent thrombectomy of the groin and fasciotomy of the legs, as well as heparin for blood clot prophylaxis. On (b) (6) (Study Day 52), the patient died due to post-operative hemorrhage. The investigator listed the death as “other”. 184308- 79 year old male with RCC with metastasis The death is unlikely due 5401-3537 to the retroperitoneum and lymph nodes. to cabozantinib. The Cabozantinib The patient had a history of ischemic heart patient had multiple risk disease, hyperlipidemia, diabetes, and factors for cardiac failure, anemia. He received his first dose of including ischemic heart cabozantinib on (b) (6) . On (b) (6) disease, diabetes, and (b) (6) (Study Day 15), he experienced hyperlipidemia. His Grade 1 abdominal pain, and a Grade 1 abdominal pain was due to decrease in both serum sodium and progression of disease. magnesium. On (b) (6) (Study Day 18), the patient was hospitalized for Grade 3 abdominal pain. Treatment included a saline infusion, ciprofloxacin, piracetam, and allopurinol. An abdominal ultrasound performed on (b) (6) showed a 10.6 x 8.1 cm tumor causing pressure of the surrounding structures with compression of the gall bladder. Another 5.6 cm tumorous lesion was located in the area of the right adrenal gland. Surgery was recommended, however the patient was considered a high risk surgical candidate due to atrial fibrillation and underlying heart disease. On (b) (6) (b) (6) , a CT confirmed significant

Reference ID: 3917720

Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Patient ID Summary of Case Reviewer Comment Group (Study Day 189), a fleet enema was initiated as treatment for fecal loading. On (b) (6) (Study Day 190), the patient experienced Grade 2 fecal loading and was hospitalized on (b) (6) . An abdominal x-ray showed a moderate amount of fecal loading throughout the colon. Treatment included docusate and intravenous hydration. On an unknown day in (b) (6) , the patient experienced Grade 2 urinary tract disorder. On (b) (6) (Study Day 227), he experienced Grade 4 clinical deterioration. Cabozantinib was discontinued on the same due to clinical deterioration. On (b) (6) , the patient was hospitalized due to urosepsis. On (b) (6) (Study Day 239), 13 days after the last dose of cabozantinib, the patient died due to urosepsis. No autopsy was performed. The investigator listed the death as “other.” Source: Narratives, CRFs, ADAE.xpt, ADSL.xpt.

Reviewer Comments: 1. The sponsor categorized deaths as due to disease progression or other. They further categorized the deaths in the “other” category into deaths causally associated with renal cell carcinoma. Deaths were classified as “yes”, “no”, or “unknown” if causally associated with renal cell carcinoma. This reviewer did not take these categorizations into consideration and independently performed a review of the deaths.

2. There were Grade 5 AE’s of hemorrhage (Patient ID 4444-3161 and Patient ID 4501- 3191) which the sponsor did not attribute to cabozantinib. After reviewing the case reports, the reviewer attributed these two deaths to cabozantinib. The proposed labeling includes hemorrhage as a Warning and Precaution, given the fatal adverse events of hemorrhage which occurred not only in Trial XL184-308 but across the clinical program.

3. While the number of deaths within the safety database is limited, they were comparable in number to everolimus control arm, and most appear to be confounded by underlying disease factors, reports to date suggest a pattern of toxicity consistent with

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

the drug’s known toxicity profile, such as hypertension, hemorrhage, thrombotic events, fistulas and perforations.

4. In general the reviewer agreed with the assessments of death. Deaths that occurred with ongoing SAEs at the time of death after 30 days post-treatment were not captured, thus the reviewer could not determine if any of the later deaths attributed to renal cell carcinoma disease progression were also associated with ongoing SAEs.

5. Patients 184308-3603-3606 and 184308-3410-3021 were described in cases narratives as having unexplained deaths. Upon further review, the medical reviewer classified these two cases as sudden cardiac deaths.

7.3.2 Nonfatal Serious Adverse Events

Serious adverse events of any grade up to 30-days after last dose of study therapy, regardless of causality, were similar in both treatment groups (39.6% cabozantinib, 43.2% everolimus). 6HULRXV $(V UHSRUWHG IRU RI VXEMHFWV LQ WKH FDER]DQWLQLE group were renal cell carcinoma, abdominal pain, pleural effusion, diarrhea, nausea, anemia, back pain, dyspnea, fatigue, pneumonia, pulmonary embolism, vomiting, and pain. Table 40 VXPPDUL]HV IUHTXHQWO\ UHSRUWHG 6$(V WKRVH UHSRUWHG IRU RI subjects in the cabozantinib group).

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 40: XL184-308 6HULRXV $GYHUVH (YHQWV 5HSRUWHG LQ &DER]DQWLQE *URXS (All Treated Population) Preferred Term Cabozantinib Everolimus N=331 N=322 n(%) n(%) Total patients with event 131 (39.6) 139 (43.2) Renal cell carcinoma 11 (3.3) 11 (3.4) Abdominal pain 10 (3.0) 3 (0.9) Pleural effusion 10 (3.0) 6 (1.9) Diarrhea 7 (2.1) 2 (0.6) Nausea 7 (2.1) 2 (0.6) Anemia 6 (1.8) 12 (3.7) Back pain 6 (1.8) 4 (1.2) Dyspnea 6 (1.8) 13 (4.0) Fatigue 6 (1.8) 5 (1.6) Pneumonia 6 (1.8) 13 (4.0) Pulmonary embolism 6 (1.8) 1 (0.3) Vomiting 6 (1.8) 4 (1.2) Pain 5 (1.5) 4 (1.2) Source: ADAE.xpt, ADSL.xpt, analysis generated by MAED

Reviewer comment:

The rates of serious adverse events were similar in the cabozantinib and everolimus treated groups. Outside of renal cell carcinoma, many SAE’s were GI related, such as abdominal pain, diarrhea, nausea, and vomiting.

7.3.3 Dropouts and/or Discontinuations

A summary of dose reductions for cabozantinib is presented in Table 41. A total of 59.8% of patients in the cabozantinib arm had a dose reduction due to AE; the median time to first dose reduction was 55.0 days. A second dose reduction due to an AE occurred in 19.3% of patients. The median time from initiation of drug to second dose reduction was 93.0 days.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 41: XL184-308 Cabozantinib Dose Reductions Cabozantinib N=331 Patients treated, n (%) 331 (100.0) Patients with any dose reduction resulting from AE, n (%) 198 (59.8) Received dose level, n(%) Assigned dose level (60 mg) 331 (100.0) First dose-level reduction (40 mg) resulting from AE 192 (58.0) Second dose-level reduction (20 mg) resulting from AE 64 (19.3) Lowest dose level received (excluding dose interruptions), n(%) Assigned dose level (60 mg) 133 (40.2) First-level dose reduction (40 mg) resulting from AE 132 (39.9) Second-level dose reduction (20 mg) resulting from AE 65 (19.6) Last dose level received (excluding dose interruptions), n (%) 60 mg 142 (42.9) 40 mg 132 (39.9) 20 mg 56 (16.9) Last dose level received (including dose interruptions), n (%) 60 mg 98 (29.6) 40 mg 97 (29.3) 20 mg 45 (13.6) 0 mg 91 (27.5) Median (range) time (days) on treatment at: Assigned dose level (60 mg) 73.0 (3, 560) First dose-level reduction (40 mg), resulting from AE 83.5 (1, 472) Second dose-level reduction (20 mg), resulting from AE 117.0 (2, 426) Median (range) time to first dose reduction resulting from AE (days) 55.0 (10, 355) Median (range) time to second dose reduction resulting from AE (days) 93.0 (29, 317)

Reviewer comment:

There were more dose reductions in the cabozantinib treated group than the everolimus treated group. Any dose reduction to do an AE occurred in 59.8% of 24.2% of patients in the cabozantinib and everolimus arms, respectively. However as of the data cutoff, median study drug exposure was approximately twice as long in the cabozantinib arm (32 weeks in the cabozantinib arm and 19 weeks in the everolimus arm). Dose reductions and interruptions were used (in both arms) to titrate the dose to individual subject tolerability.

Adverse events that led to discontinuation of study drug (excluding AEs of disease SURJUHVVLRQ UHSRUWHG IRU RI SDWLHQWV LQ HLWKHU WUHDWPHQW DUP DUH VXPPDUL]HG LQ Table 42. The incidence of AEs that led to discontinuation of study drug was similar between treatment arms (10% cabozantinib, 9.5% everolimus).

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

The most common AEs which led to treatment discontinuation in the cabozantinib group were decreased appetite, fatigue, asthenia, diarrhea, and proteinuria.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 42: XL184-308 AEs That Led to Study Treatment Discontinuation LQ %of Subjects in Either Arm (All Treated Population) Cabozantinib Everolimus N=331 N=322 System Organ Class All grades Grade 3/4 All grades Grade 3/4 Preferred Term n (%) n (%) n (%) n (%) Patients with an AE 34 (10) 19 (5.7) 31 (9.6) 19 (5.9) General Disorders And Administration Site Conditions Fatigue 4 (1.2) 0 3 (0.9) 1 (0.3) Asthenia 3 (0.9) 2 (0.6) 1 (0.3) 0 Death 2 (0.6) 0 0 0 Gastrointestinal Disorders Diarrhea 3 (0.9) 0 3 (0.9) 1 (0.3) Vomiting 2 (0.6) 0 0 0 Nausea 1 (0.3) 0 0 0 Constipation 1 (0.3) 0 0 0 Anal fistula 1 (0.3) 1 (0.3) 0 0 Dry mouth 1 (0.3) 0 0 0 Dysphagia 1 (0.3) 0 0 0 Gastrointestinal perforation 1 (0.3) 1 (0.3) 0 0 Pancreatitis 1 (0.3) 1 (0.3) 0 0 Skin And Subcutaneous Tissue Disorders PPE 1 (0.3) 0 0 0 Musculoskeletal And Connective Tissue Disorders Pain in extremity 2 (0.6) 1 (0.3) 1 (0.3) 0 Renal and Urinary Disorders Proteinuria 3 (0.9) 3 (0.9) 0 0 Renal failure 1 (0.3) 1 (0.3) 0 0 Metabolism and Nutrition Disorders Decreased appetite 6 (1.8) 3 (0.9) 3 (0.9) 0 Dehydration 1 (0.3) 0 2 (0.6) 1 (0.3) Hyperkalemia 1 (0.3) 1 (0.3) 0 0 Neoplasms benign, malignant and unspecified Adenocarcinoma 1 (0.3) 0 0 0 Leiomyoma 1 (0.3) 0 0 0 Infections and Infestations Anal abscess 1 (0.3) 1 (0.3) 0 0 Anorectal cellulitis 1 (0.3) 1 (0.3) 0 0 Device related infection 1 (0.3) 1 (0.3) 0 0 Investigations AST increased 1 (0.3) 0 0 0 Weight decreased 1 (0.3) 1 (0.3) 0 0 Cardiac disorders Cardiac failure 1 (0.3) 0 0 0 Ventricular arrhythmia 1 (0.3) 0 0 0 Hepatobiliary disorders Hepatitis cholestatic 1 (0.3) 1 (0.3) 0 0 Nervous System Disorders Muscle spasticity 1 (0.3) 0 0 0 Respiratory, Thoracic And Mediastinal Disorders Pneumothorax 1 (0.3) 1 (0.3) 0 0

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Pulmonary embolism 1 (0.3) 1 (0.3) 0 0 Source: ADAE, Summary of Clinical Safety Table 26

Reviewer comment:

While hypertension, diarrhea, and PPES were frequent events in the cabozantinib group, the low incidence of hypertension, diarrhea, or PPES leading to treatment discontinuation suggests that these events were managed effectively with dose modifications and supportive care in most cases.

As discussed in Section 6.1.3, seventeen patients were listed as discontinuing treatment due to “withdrawal by subject.” Of those, eight patients reported a total of 15 Grade 3-4 AEs, which may have contributed to why these patients withdrew from treatment. Four patients on the cabozantinib arm who withdrew reported the following Grade 3-4 AEs: asthenia, decreased appetite, blood pressure increased, hypokalemia, hypomagnesemia, hypertensive crisis, anemia, fatigue, and hypokalemia. Four patients on the everolimus arm discontinued treatment due to “withdrawal by subject” and had a total of five Grade 3-4 AEs.

Interestingly, asthenia, decreased appetite, fatigue and hypertension were the most common reasons for treatment discontinuation, and were also experienced by the patients in the cabozantinib group who discontinued treatment due to “withdrawal by subject.” If these four patients were to be included in the group of patients who discontinued therapy due to AE, the rate of discontinuation for cabozantinib treated patients would only increase slightly, from 10% to 11.5%.

Dose Reductions

The dose of study drug was reduced due to an AE for 60% of patients in the cabozantinib group and 24% of patients in the everolimus group. The most frequently reported AEs that led to dose reduction are summarized in Table 43.

The most frequently reported AEs leading to dose reduction corresponded to the frequently reported AEs overall.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 43: XL184-308 AEV WKDW /HG WR 'RVH 5HGXFWLRQ LQ of Subjects in Either Arm (All Treated Population) Cabozantinib Everolimus N=331 N=322 Preferred Term All grades Grade 3/4 All grades Grade 3/4 n(%) n (%) n (%) n (%) Number of patients with at least 200 (60) 101 (31) 78 (24) 38 (12) one AE that led to dose reduction Diarrhea 54 (16) 17 (5.1) 3 (0.9) 0 Palmar-plantar 38 (11) 14 (4.2) 2 (0.6) 1 (0.3) erythrodysesthesia syndrome Fatigue 33 (10) 12 (3.6) 11 (3.4) 5 (1.6) Hypertension 25 (7.6) 20 (6.0) 0 0 Source: ADAE, CSR Table 60

A dose interruption is defined as a withholding of study treatment that is intended to be temporary. Study treatment was interrupted due to an AE of 70% of patients in the cabozantinib group and 59% of patients in the everolimus group. The most frequently reported AEs that led to dose interruption are summarized in Table 44.

Table 44: XL184-308 $GYHUVH (YHQWV WKDW /HG WR 'RVH ,QWHUUXSWLRQV LQ RI 6XEMHFWV in Either Treatment Arm (All Treated Population) Cabozantinib Everolimus N=331 N=322 Preferred Term All grades Grade 3/4 All grades Grade 3/4 n (%) n (%) n (%) n (%) Patients with an AE 233 (70) 152 (46) 189 (59) 102 (32) Diarrhea 72 (22) 27 (8.2) 8 (2.5) 5 (1.6) Palmar-plantar erythrodysesthesia 46 (14) 24 (7.3) 5 (1.6) 2 (0.6) syndrome Fatigue 41 (12) 16 (4.8) 19 (5.9) 9 (2.8) Nausea 28 (8.5) 4 (1.2) 6 (1.9) 1 (0.3) Decreased appetite 25 (7.6) 4 (1.2) 4 (1.2) 0 Vomiting 25 (7.6) 4 (1.2) 4 (1.2) 1 (0.3) Hypertension 17 (5.1) 15 (4.5) 0 0 Source: ADAE, CSR Table 61

The dose of study drug was modified (reduced or interrupted) for 77% of subjects in the cabozantinib group and 61% of patients in the everolimus group. Adverse events that led to dose modifications of either cabozantinib or everolimus groups are summarized in Table 45.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 45: XL184-308 Adverse Events That /HG WR 'RVH 0RGLILFDWLRQ LQ % of Subjects in Cabozantinib Arm (All Treated Population) Cabozantinib Everolimus N=331 N=322 Preferred Term All grades Grade 3/4 All grades Grade 3/4 n (%) n (%) n (%) n (%) Number of patients with at least 255 (77) 170 (51) 197 (61) 109 (34) one AE that led to dose modifications Diarrhea 87 (26) 32 (9.7) 10 (3.1) 5 (1.6) Palmar-plantar erythrodysesthesia 53 (16) 24 (7.3) 5 (1.6) 2 (0.6) syndrome Fatigue 50 (15) 19 (5.7) 25 (7.8) 11 (3.4) Nausea 35 (11) 5 (1.5) 6 (1.9) 1 (0.3) Decreased appetite 29 (8.8) 4 (1.2) 6 (1.9) 1 (0.3) Vomiting 27 (8.2) 4 (1.2) 5 (1.6) 1 (0.3) Stomatitis 20 (6.0) 7 (2.1) 18 (5.6) 7 (2.2) Source: ADAE, CSR Table 62

Reviewer Comment:

There were greater amounts of dose interruptions, reductions, and modifications in the cabozantinib treated group than those treated with everolimus. However as of the data cutoff, median study drug exposure was approximately twice as long in the cabozantinib arm (32 weeks in the cabozantinib arm and 19 weeks in the everolimus arm). The median daily doses (relative dose intensities) of cabozantinib and everolimus in Study XL184-308 were 45 mg (76%) and 9.2 mg (92%), respectively. Dose reductions and interruptions were used in both arms to titrate the dose to individual patient tolerability.

The review team considered the utility of a dose finding study as a post marketing requirement or commitment due to the high rate of dose reductions in the cabozantinib group compared to the everolimus group. At the review team’s request, the applicant provided dose modeling metrics to determine the potential risk of decreasing efficacy with a decreased dose of cabozantinib. Exposure-response analyses support the dose selection of 60 mg daily. Doses higher than 60 mg daily were not considered, given that dose reductions were required in 60% of patients at 60 mg daily. Simulation suggested that lower starting doses such as 40 mg or 20 mg daily would likely compromise efficacy though the incidence of dose reductions may be lower than with the 60 mg dose. Furthermore, the Clinical Pharmacology reviewers determined that the exposure to cabozantinib was very similar across doses and indications.

In addition, despite the rate of dose reductions in the cabozantinib group, the rate of treatment discontinuations due to adverse events was only 10%, suggesting that most AE’s were managed with dose interruptions and supportive measures. Given that most

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

adverse events could be safely managed by clinicians, the review team did not feel that a post marketing requirement was indicated.

Based on the totality of evidence which includes the exposure data, the potential loss of efficacy a lower dose would impart, and the relative safety given clinicians ability to manage AEs, the review team determined that a dose finding study would not be warranted.

Please see Dr. Song’s review in Clinical Pharmacology for further discussion of this topic.

7.3.4 Significant Adverse Events

Grade 3-4 Adverse Events

Grade 3 and Grade 4 AEs occurring up to 30-days after last dose of study therapy occurred in 68.3% of patients (n=226) in the cabozantinib group and 57.8% of patients (n=186) in the everolimus group. The higher incidence of AEs in the cabozantinib arm was primarily due to a higher rate of hypertension, diarrhea, and palmar-plantar erythrodysesthesia syndrome. Frequently reported Grade 3 and 4 AEs (those reported IRU 1.5% of patients in cabozantinib arm) are summarized in Table 46.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 46: XL184-308 *UDGH DQG $GYHUVH (YHQWV .5% in Cabozantinib Group (All Treated Population) System Organ Class Cabozantinib Everolimus Adverse Event N=331 N=322 n(%) n(%) Total Patients with Event 226 (68.3) 186 (57.8) General Disorders And Administration Site Conditions Fatigue 30 (9.1) 22 (6.8) Asthenia 14 (4.2) 7 (2.2) General physical health deterioration 7 (2.1) 8 (2.5) Gastrointestinal Disorders Diarrhea 38 (11.5) 7 (2.2) Nausea 13 (3.9) 1 (0.3) Abdominal pain 12 (3.6) 4 (1.2) Stomatitis 8 (2.4) 7 (2.2) Vomiting 7 (2.1) 3 (0.9) Skin and Subcutaneous Tissue Disorders Palmar-plantar erythrodysesthesia syndrome 27 (8.2) 3 (0.9) Vascular Disorders Hypertension 49 (14.8) 10 (3.1) Blood and Lymphatic System Disorders Anemia 18 (5.4) 50 (15.5) Respiratory, Thoracic, and Mediastinal Disorders Dyspnea 10 (3.0) 14 (4.3) Pleural effusion 9 (2.7) 6 (1.9) Pulmonary embolism 8 (2.4) 1 (0.3) Pneumonia 5 (1.5) 13 (4.0) Metabolism and Nutrition Disorders Hypomagnesemia 16 (4.8) 0 (0) Hypokalemia 15 (4.5) 6 (1.9) Hyponatremia 13 (3.9) 8 (2.5) Hypophosphatemia 12 (3.6) 8 (2.5) Decreased appetite 9 (2.7) 3 (0.9) Hypocalcemia 6 (1.8) 0 (0) Musculoskeletal and Connective Tissue Disorders Back pain 7 (2.1) 7 (2.2) Nervous System Disorders Syncope 5 (1.5) 2 (0.6) Renal and Urinary Disorders Proteinuria 8 (2.4) 1 (0.3) Investigations Lipase increased 9 (2.7) 3 (0.9)

Reference ID: 3917720

Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

This section includes a discussion of specific adverse events that have been associated with TKIs, VEGF pathway inhibition, and other events with potentially serious consequences. The events of diarrhea, PPES, and hypertension were reported with high frequency in previous cabozantinib studies and have been commonly observed with VEGFR-TKIs. Less frequent but potentially life-threatening AEs that have been reported for patients receiving cabozantinib and other VEGF pathway inhibitors are as follows: GI perforations, fistulas, abscesses, hemorrhages, arterial thrombotic events, venous and mixed/unspecified thrombotic events, wound complications, osteonecrosis, proteinuria, and RPLS.

The adverse events specific to cabozantinib for studies XL184-308 (RCC), XL184-307 (CRPC), and XL184-306 (CRPC), and XL184-301 (MTC) are presented in Table 47. This summarizes the safety results of cabozantinib in randomized, controlled Phase 3 clinical trials across tumor types. Two different cabozantinib starting dose levels are reflected in these data: 140 mg capsule daily for MTC and 60 mg tablet daily for RCC and CRPC.

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 47 Safety for Medically Important Cabozantinib-Related Adverse Events across Tumor Types

Tumor Type RCC CRPC MTC (XL184-308) (XL 184-307 & XL184-306) (XL184-301) 60 mg 60 mg 140 mg N = 331 N = 741 N = 219 Category Any AE *UDGH Any AE *UDGH Any AE *UDGH n (%) n (%) n (%) n (%) n (%) n (%) GI perforation 3 (0.9) 2 (0.6) 9 (1.2) 8 (1.1) 7 (3.3) 7 (3.3) 1 Grade 5 AE Fistulas 4 (1.2) 1 (0.3) 9 (1.2) 3 (0.4) 10 (4.6) 5 (2.3) 3 Grade 5 AEs Intra-abdominal 4 (1.2) 4 (1.2) 11 (1.5) 8 (1.1) 5 (2.3) 1 (0.5) and pelvic abscesses Hemorrhage NA 7 (2.1) NA 35 (4.7) NA 7 (3.3) 2 Grade 5 AEs 6 Grade 5 AEs 2 Grade 5 AEs ATE 3 (0.9) 2 (0.6) 23 (3.1) 12 (1.6) 5 (2.3) 2 (0.9) VTE 24 (7.3) 12 (3.6) 95 (13.0) 66 (8.9) 12 (5.6) 10 (4.7) 3 Grade 5 AEs Wound 8 (2.4) 1 (0.3) 15 (2.0) 5 (0.7) 4 (1.9) 2 (0.9) complications Hypertension 128 (39) 52 (16) 213 (29) 149 (20) 70 (33) 18 (8.4) ONJ 2 (0.6) 1 (0.3) 22 (3.0) 9 (1.2) 3 (1.4) 1 (0.5) PPES 139 (42) 27 (8.2) 209 (28) 42 (5.7) 107 (50) 27 (13) Proteinuria 41 (12) 8 (2.4) 18 (2.4) 2 (0.3) 4 (1.9) 2 (0.9) RPLS 0 0 0 0 1 (0.5) 1 (0.5) QTc 1 (0.3) 0 11 (1.5) 3 (0.4) 5 (2.3) 1 (0.5) prolongation Diarrhea 245 (74) 38 (11) 381 (51) 55 (7.4) 135 (65) 34 (16) Source: ISS Table 65 ATE, arterial thrombotic events; CRPC, castration-resistant prostate cancer; GI, gastrointestinal; MTC, medullary thyroid cancer; ONJ, osteonecrosis of jaw; PPES, palmar-plantar erythrodysesthesia syndrome; RCC, renal cell carcinoma; RPLS, revers ble posterior leukoencephalopathy syndrome; VTE, venous thrombotic events

The populations for both Phase 3 CRPC studies were more frail than in the RCC Phase 3 study: elderly men (median age 65-69 years) who had advanced disease including bone metastases (eligibility requirement), some with visceral metastases, and were heavily pretreated (including a requirement for prior docetaxel and either abiraterone or enzalutamide). Patients in Study XL184-306 were further required to be experiencing pain at baseline despite optimization and of narcotic use. Bone metastases are the primary source of morbidity and major contributor to mortality for this population. In both Phase 3 CRPC studies, cabozantinib treated patients received the same dose and formulation as in the XL184-308 RCC study (ie, starting doses of 60 mg tablet qd).

Reviewer comments:

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

1. A review of medically important cabozantinib-related adverse events across tumor types suggested that some differences in rates of AEs may be due to tumor type. For example, the higher rate of GI perforations and fistulas in the MTC population when compared to the CRPC and RCC populations. MTC patients are more vulnerable to upper airway and GI tract complications due to the location of tumor burden in the chest/neck regions and applied local therapies, such as surgery and radiation.

2. The pattern of Grade 5 events of hemorrhage differed between the RCC and MTC studies and may reflect the study population differences with respect to sites of disease. The MTC events occurred in the upper respiratory/esophageal tract, a metastatic site that is not frequently involved in RCC patients. Known predisposing risk factors for severe hemorrhage include tumor invasion of major blood vessels.

3. Hypertension, diarrhea, and PPES were frequent events for cabozantinib-treated patients in all the Phase 3 studies. However there was a low incidence of hypertension, diarrhea, or PPES leading to treatment discontinuation in Study XL184-308 (0%, 0.9%, and 0.3%, respectively). This suggests that these events can be managed effectively with dose modification and supportive care in most cases.

A summary of the key toxicities associated VEGFR-TKI/VEGF inhibition experienced in Trial XL184-308 is summarized in Table 48.

Reference ID: 3917720

Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

x A persistent or significant incapacity or substantial disruption of the ability to conduct normal life functions x A congenital anomaly or birth defect

Based on these criteria, the review team determined that (b) (4) , (b) (4) could be omitted from the Warnings and Precautions section of the label. These events are included in Section 6 of the label in listings of adverse events.

2. The two Grade 5 incidences of hemorrhage in the cabozantinib group are discussed later in this section and in section 7.3.1.

Hypertension

The Applicant included the following terms: hypertension, blood pressure increased, and hypertensive crisis. The reviewer included blood pressure fluctuation. Adverse events of hypertension are summarized in Table 49.

The incidence of hypertension was higher in the cabozantinib group (130 patients [39.2%] compared with the everolimus group (24 patients [7.4%]. The severity of hypertension was also greater in the cabozantinib group, (53 patients [16%]) compared with everolimus (10 patients, [3.1%]). There were no Grade 4 or 5 hypertension events reported.

Table 49: XL184-308 Incidence of Hypertension Adverse Events (All Treated Population) Cabozantinib Everolimus N=331 N=322 n (%) n (%) Preferred Term Any AE Grade Grade Any AE Grade Grade 3/4 AE 5 AE 3/4 AE 5 AE Hypertension (all) 130 53 (16) 0 24 (7.4) 10 (3.1) 0 (39.2) Blood pressure 5 (1.5) 2 (0.6) 0 0 0 0 increased Hypertension 122 (37) 49 (15) 0 23 (7.1) 10 (3.1) 0

Hypertensive crisis 2 (0.6) 2 (0.6) 0 1 (0.3) 0 0 Blood pressure 1 (0.3)00000 fluctuation Source: ADAE.xpt

Two patients treated with cabozantinib had AE’s reported by the investigator as hypertensive crisis.

100

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

x Patient 3607-3841 experienced an AE reported as Grade 3 hypertensive crisis per the investigator on Study Day 32. The patient has no reported history of hypertension and had baseline blood pressure (BP) readings of systolic BP (SBP) 126 mm Hg and diastolic BP (DBP) 89 mm Hg. The last reported BP readings for the patient prior to the event on Study Day 29 was 138/80. There were no BP readings documented for the onset date of the AE. The last dose of cabozantinib prior to the event was taken on Study Day 31, and study treatment was interrupted in response to the event. The patient was administered captopril followed by perindopril and subsequently withdrew form study treatment on Day 71 without receiving further treatment with cabozantinib.

x Patient 4920-3695 experienced an AE reported as Grade 3 hypertensive crisis per the investigator on Study Day 197. The patient had a history of hypertension prior to randomization with a BP of 146/87. The patient experienced an AE of Grade 3 hypertension on Study Day 4. The patient received concomitant nitrendipine, ramapril, and amlodipine at various times while on study treatment, not concurrently, prior to the event of hypertensive crisis. Reported BP readings for the patient on the day of the event were 163/91. Subsequent to the hypertensive crisis event, the patient was administered nitrendipine. No action with study treatment was taken due to the AE, but the patient discontinued treatment due to study progression on the same day as the event.

No cabozantinib-treated patients had documented evidence of malignant hypertension, hypertensive urgency, or hypertensive emergency.

Reviewer comments:

1. Although two patients treated with cabozantinib had AEs reported by the investigator as hypertensive crisis, the narratives provided do meet the definition for hypertensive crisis in either patient.

2. The low incidence of hypertension leading to treatment discontinuation (See Table 39) suggests that these events were managed effectively with dose modification and supportive care in most cases.

3. Hypertension and hypertensive urgency is included in Section 5 Warnings and Precautions of the label given its high incidence in Trial XL184-308 and across the clinical program. Providers are advised to monitor blood pressure regularly And to discontinue cabozantinib for hypertensive crisis or severe hypertension that cannot be controlled with anti-hypertensive therapy.

Diarrhea

101

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

The Applicant included the following terms: diarrhea. The reviewer did not include any additional terms.

The median time to the first occurrence of diarrhea was 4.9 weeks in Study XL184-308.

The incidence of diarrhea events was 74% (245 patients) in the cabozantinib arm and 28% (89 patients) in the everolimus arm. Diarrhea was the most frequently reported AE for patients in the cabozantinib group. No diarrhea events of severity > Grade 3 were reported on either treatment arm of Study XL184-308. Diarrhea was reported as an SAE in 2.1% of cabozantinib treated patients. Dose modification due to diarrhea occurred in 26% of patients. Adverse events of diarrhea are summarized in Table 50.

Table 50: XL184-308 Incidence of Diarrhea (All Treated Population) Cabozantinib Everolimus N=331 N=322 n (%) n (%) Preferred Term Any AE Grade Grade SAE Any Grade Grade SAE 3/4 AE 5 AE Any AE 3/4 AE 5 AE Any Grade Grade Diarrhea 245 (74) 38 (11) 0 7 (2.1) 89 (28) 7 (2.2) 0 2 (0.6) Source: ADAE.xpt

Table 51 summarizes patients with diarrhea with serious adverse events.

102

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 51: XL184-308 Patients with Diarrhea Serious Adverse Events (All Treated Population)

Patient ID Sex Grade Onset Duration Action Dose Related Outcome (day) (day) Reduced 184308- M 3 249 6 Dose interrupted NYR 1417-3569 4 Medications given 184308- M 3 131 4 Medications given N Y R 1417-3912 184308- M 3 314 6 Dose interrupted NNR 2041-3089 Medications given 184308- M 2 175 24 Medications given N Y R 3233-3275 184308- F 3 66 11 Dose interrupted N Y R 3801-3713 184308- M 3 74 4 Medications given Y Y R 6112-3037 184308- M 3 31 8 Medications given Y Y R 6124-3525 Source: ADAE.xpt, CRFs Abbreviations: F, female; M, male; N, no; O, ongoing; R, resolved/recovered; Y, yes

Reviewer comments

1. Diarrhea was the most frequently observed AE among patients treated with cabozantinib and is typical of those observed with other TKI’s. Although the rate of Grade 3/4 events of diarrhea was low (11%) compared to the overall incidence of diarrhea (74%), the impact of lower grades of diarrhea on quality of life should not undervalued. In a sponsor administered health-related quality of life measure (FKSI- Total score), the biggest treatment difference was a worse score for cabozantinib for diarrhea, as related to treatment side effects.

2. The reviewer queried the applicant on the potential for studies evaluating the use of prophylactic loperamide or any other antidiarrheal agent as prophylaxis for symptoms of diarrhea while taking cabozantinib. The sponsor has not investigated this possibility, and noted concerns with the added risk of constipation (an adverse effect of both loperamide and cabozantinib), and the subsequent risk of ileus, or gastrointestinal perforations, which have been observed with cabozantinib.

3. For the label, the diarrhea is listed in Section 5 Warnings and Precautions section. Providers are advised to withhold cabozantinib in patients who develop intolerable Grade 2 diarrhea or Grade 3-4 diarrhea that cannot be managed with standard antidiarrheal treatments until improvement to Grade 1. Cabozantinib can then be resumed at reduced dose.

103

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Palmar-Plantar Erythrodysesthesia Syndrome (PPES)

The Applicant included the following terms: PPES. The reviewer included intermittent PPE.

Incidence of PPES was high in the cabozantinib group (139 patients [42%]); 19 patients (5.9%) had PPES in the everolimus group. No events of PPES > Grade 3 were reported for either treatment arm. Adverse events of PPES are summarized in Table 52.

Table 52 XL184-308 Incidence of Palmar-Plantar Erythrodysaesthesia Syndrome (All Treated Population) Cabozantinib Everolimus N=331 N=322 n (%) n (%) Preferred Term Any AE Grade Grade SAE Any Grade Grade SAE 3/4 AE 5 AE Any AE 3/4 AE 5 AE Any Grade Grade PPES 139 (42) 27 00193 (0.9) 0 0 (8.2) (5.9) Intermittent 1 (0.3)0000000 PPE Source: ADAE.xpt

Cutaneous toxicities other than PPES were reported with cabozantinib treatment. While rash is common both for VEGFR-TKIS and mTOR inhibitors such as everolimus, other skin toxicities such as ulcers and changes in pigmentation were more common in the cabozantinib treated patients.

104

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 53 summarizes dermatologic toxicities in both groups of patients.

Table 53 XL184-308 Incidence of Dermatologic Toxicities (All Treated Population) Cabozantinib Everolimus N=331 N=322 n (%) n (%) Preferred Term Any AE Grade 3/4 Grade Any AE Grade Grade AE 5 AE 3/4 AE 5 AE Rasha 76 (23) 2 (0.6) 0 144 (46.2) 3 (0.9) 0

Cutaneous toxicitiesb 133 (39.8) 2 (0.6) 0 89 (26.3) 1 (0.3) 0

Hair changesc 33 (9.9) 0 0 4 (1.2) 0 0

Nail changesd 9 (2.7) 0 0 19 (5.9) 0 0 Source: ADAE.xpt a Includes PT terms: genital rash, dermatitis contact, dermatitis, intermittent leg rash, rash erythematous, rash macular, rash papular, dermatitis acneiform, rash pruritic, rash maculopapular b Includes PT terms: skin ulcer, blister, wound, dry skin, skin lesion, skin exfoliation, dermatitis bullous, skin hypertrophy, cheilitis, actinic keratosis, skin disorder, erythema, skin irritation, furuncle, skin discoloration, chapped lips, dermatosis, skin hypopigmentation, erysipelas, skin depigmentation, skin discoloration, skin fissures, excoriation, skin infection, urticaria, pain of skin, eczema, folliculitis, skin toxicity, xerosis, acne c Includes PT terms: alopecia, hair color changes d Includes PT terms: nail dystrophy, nail infection, nail ridging, nail disorder, paronychia

Reviewer Comments:

1. For the purposes of the label, the applicant proposed including PPES in Section 6 Clinical Trials, however given the frequency of this adverse event in Trial XL184-308 and across the clinical program, the Clinical Review team recommended that PPES be included in Section 5 Warnings and Precautions. The applicant agreed to this modification and instructions on management of PPES are provided in the label.

2. Skin and cutaneous toxicities, including rash, occurred with high frequency (27%) in patients treated with cabozantinib, however rates of rash were higher in patients treated with everolimus (46%). Despite the high incidence of rash in both groups, almost all were Grade 1-2 events. This suggests that rash was managed by dose modification or interruption. Rash is included in the label in Section 6 Adverse Reactions, in a table OLVWLQJ DGYHUVH UHDFWLRQV ZKLFK RFFXUUHG LQ RI SDWLHQWV ZKR UHFHLYHG cabozantinib. It is worth noting that rash occurred with greater frequency in patients treated with everolimus (43%) than in patients treated with cabozantinib (23%).

3. Cutaneous toxicities such as skin disorder, skin discoloration, and skin hypopigmentation were reported with greater frequency in the cabozantinib treated group than the everolimus group. While none were greater than grade 1-2, the potential effects of skin changes on patients quality of life as it may relate to self-image should be

105

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

considered. Hair changes were reported in up to 10% of patients treated with cabozantinib compared with everolimus (1.2%), however all were Grade 1-2.

Renal Failure and Proteinuria

The Applicant included the following terms: acute prerenal failure, azotemia, glycosuria, nephropathy toxic, proteinuria, renal failure, renal failure acute, renal failure chronic, renal impairment. The reviewer included blood creatinine increased.

Proteinuria was reported for 12% of patients in the cabozantinib group and 9.3% of patients in the everolimus group. There were more Grade 3 events of proteinuria in the cabozantinib group (2.4%) compared to the everolimus group (0.3%). There were no Grade 4 or 5 proteinuria events on either arm. Proteinuria led to study treatment discontinuation in 3 patients (0.9%). The median time to first occurrence of proteinuria was approximately 4.1 weeks. Adverse events of renal failure and proteinuria are summarized in Table 54. Table 54 : XL184-308 Incidence of Selected Renal Failure and Proteinuria Adverse Events (All Treated Population) Cabozantinib Everolimus N=331 N=322 n (%) n (%) Preferred Term Any AE Grade Any AE Grade 3/4 AE 3/4 AE Blood creatinine 15 (4.5) 1 (0.3) 35 (11) 0 increased Glycosuria 0 0 1 (0.3) 0

Nephropathy toxic 1 (0.3) 0 46 (14) 0 Proteinuria 41 (12) 8 (2.4) 30 (9.3) 1 (0.3) Renal failurea 4 (1.2) 2 (0.6) 14 (4.3) 5 (1.5) Source: ADAE.xpt, ADSL.xpt a.includes PT terms acute renal failure, azotemia, renal failure, renal failure acute, renal failure chronic, renal impairment

Reviewer Comments:

1. Proteinuria is included in the label in Section 6 Adverse Reactions, in a table listing adverse reactions which ocFXUUHG LQ RI SDWLHQWV ZKR UHFHLYHG FDER]DQWLQLE Section 2.2 Dosage Adjustments instructs providers to permanently discontinue cabozantinib for nephrotic syndrome.

2. Proteinuria is an expected event in this patient population due to the high rate of prior nephrectomy and the requirement for VEGR-TKI therapy. While the rates of proteinuria

106

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

are greater in Trial XL184-308 than in other tumor types (Table 47), there were no events of nephrotic syndrome in this study. Previous events of nephrotic syndrome have occurred in the clinical program for cabozantinib.

Hepatobiliary Disorders

The Applicant included the following terms: bile duct obstruction, cholangitis, cholangitis acute, cholecystitis, cholecytitis acute, cholecystitis chronic, cholelithiasis, gallbladder pain, hepatic steatosis, hepatic vein thrombosis, hepatitis cholestatic, hepatocellular injury, hepatosplenomegaly, hepatotoxicity, hyperbilirubinemia, jaundice, liver tenderness, and portal vein thrombosis. The reviewer included bile acid malabsorption, steatorrhea, alanine transferase, alanine aminotransferase increased, aspartate aminotransferase, aspartate aminotransferase increased, blood bilirubin increased, gamma glutamyltransferase, gamma glutamyltranferase increased, hepatic enzyme increased, liver function test abnormal, transaminases increased, blood alkaline phosphatase, and blood alkaline phosphatase increased.

Table 55 summarizes the adverse events from the hepatobiliary disorders SOC.

107

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 55: XL184-308 Incidence of Adverse Events in the Hepatobiliary Disorders SOC (All Treated Population) Cabozantinib Everolimus N=331 N=322 n (%) n (%) Preferred Term Any AE Grade Grade Any AE Grade Grade 3/4 AE 5 AE 3/4 AE 5 AE Hepatobiliary disorders 14 (4.2) 7 (2.1) 0 10 (3.1) 4 (1.2) 0 Bile duct obstruction 2 (0.6) 2 (0.6) 0 0 0 0 Cholangitisa 2 (0.6) 2 (0.6) 0 1 (0.3) 1 (0.3) 0 Cholecystitisb 2 (0.6) 1 (0.3) 0 4 (1.2) 2 (0.6) 0 Hepatic steatosis 0 0 0 1 (0.3) 0 0 Thrombosisc 2 (0.6) 1 (0.3) 0 2 (0.6) 1 (0.3) 0 Hepatocellular injuryd 2 (0.6) 1 (0.3) 0 1 (0.3) 0 0 Hepatosplenomegaly 1 (0.3) 0 0 0 0 0 Hyperbilirubinemia 4 (1.2) 0 0 0 0 0 Jaundice 1 (0.3) 0 0 1 (0.3) 0 0 Liver tenderness 0 0 0 1 (0.3) 0 0 Bile acid 2 (0.6)00000 malabsoprtione ALT increased 53 (16) 8 (2.4) 0 19 (5.9) 1 (0.3) 0 AST increased 58 (18) 6 (1.8) 0 18 (5.6) 1 (0.3) 0 Blood bilirubin 11 (3.3) 5 (1.5) 0 1 (0.3) 0 0 increased Gamma 21 (6.3) 4 (1.2) 0 28 (8.7) 12 (3.7) 0 glutamyltransferase increased Liver function test 11 (3.3) 3 (0.9) 0 3 (0.9) 0 0 abnormalf Blood alkaline 16 (4.8) 1 (0.3) 0 18 (5.6) 1 (0.3) 0 phosphatase increased Source: ADAE.xpt, ADSL.xpt a includes PT terms cholangitis and cholangitis acute b includes PT terms cholecystitis, cholecystitis acute, cholecystitis chronic, cholelithiasis and gallbladder pain c includes PT terms hepatic vein thrombosis and portal vein thrombosis d includes PT terms hepatitis cholestatic, hepatocellular injury, hepatotoxicity e includes PT terms bile acid malabsorption and steatorrhea f includes PT terms liver function test abnormal, hepatic enzyme increased, and transaminases increased Reviewer comment:

/DE DEQRUPDOLWLHV RFFXUULQJ LQ RI SDWLHQWV ZKR UHFHLYHG FDER]DQWLQLE are listed in Section 6 Adverse Reactions of the label. AST increased, ALT increased, ALP increased, and GGT increased are all included. There were no confirmed cases of Hy’s Law in Trial XL184-308, which is further discussed in Section 7.4.2.

108

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Fistulas and GI perforations

The Applicant included the following terms: anal fistula, fistula, gastrointestinal fistula in the AE term category. The reviewer did not include any additional terms.

Fistulas were reported for four patients (1.2%) in the cabozantinib arm and no patients in the everolimus arm. One patient in the cabozantinib arm had a Grade 3 anal fistula; all other fistulas were < Grade 3.

There was one uncoded < Grade 3 AE of ‘fistula anus’ reported in the cabozantinib arm. One patient was noted to have an event of fistula which was coded ‘chin fistula’; this patient also had a chin fistula at baseline.

The Applicant included the following terms: appendicitis perforated, gastrointestinal perforation, intestinal perforation, and small intestinal perforation. The reviewer did not include any additional terms.

Gastrointestinal perforations were reported for three patients (0.9%) in the cabozantinib arm and two patients (0.6%) in the everolimus arm. Two patients in the cabozantinib arm had Grade 3 events (gastrointestinal perforation and intestinal perforation). In the everolimus arm, one patient had a Grade 4 intestinal perforation and the other had a Grade 5 gastrointestinal perforation.

Reviewer Comment:

1. As discussed in the beginning of this section, the rates of GI perforation and fistula were higher in the MTC population than in the RCC group. However given the severity of these events, both the applicant and the medical reviewers agreed that GI perforations and fistulas should be included in Section 5 Warnings and Precautions, however would not be included as a black box warning. Providers are instructed to monitor for symptoms, and discontinue cabozantinib for fistulas that cannot be adequately managed or perforations.

Hemorrhage

The Applicant included the following terms: anal hemorrhage, cerebral hematoma, conjunctival hemorrhage, epistaxis, extradural hematoma, eye hemorrhage, gastric hemorrhage, gingival bleeding, hemarthrosis, hematemesis, hematochezia, hematoma, hemorrhagic anemia, hemorrhoidal hemorrhage, injection site hematoma, lip hemorrhage, melena, penile hemorrhage, post procedural hemorrhage, pulmonary hemorrhage, rectal hemorrhage, renal hemorrhage, tongue hemorrhage, ulcer

109

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

hemorrhage, upper gastrointestinal hemorrhage, and vaginal hemorrhage. The reviewer did not include any additional terms.

7KH LQFLGHQFH RI *UDGH KHPRUUKDJLF HYHQWV ZDV LQ SDWLHQWV WUHDWHG ZLWK cabozantinib, and in 1.6% of patients treated with everolimus. The incidence of events of hemorrhage is summarized in Table 56. Table 56: XL184-308 Incidence of Hemorrhage (All Treated Population) Cabozantinib Everolimus N=331 N=322 n (%) n (%) Preferred Term Any AE Grade Grade Any AE Grade Grade 3/4 AE 5 AE 3/4 AE 5AE Cerebral hematoma 1 (0.3) 1 (0.3) 0 0 0 0 Epistaxis 12 (3.6) 0 0 46 (14) 0 0 Extradural 1 (0.3) 0 1 (0.3) 0 0 0 hematoma Gingival bleeding 4 (1.2) 0 0 0 0 0 Hemarthrosis 1 (0.3) 1(0.3) 0 0 0 0 Hematemesis 1(0.3) 0 0 0 0 0 Hematochezia 2 (0.6) 0 0 0 0 0 Hematoma 2 (0.6) 0 0 0 0 0 Hemorrhagic anemia 1 (0.3) 1 (0.3) 0 0 0 0 Hemorrhagea 18 (5.4) 2 (0.6) 0 6 (1.9) 3 (0.9) 0 Injection site 1 (0.3)00000 hematoma Melena 1 (0.3) 0 0 2 (0.6) 0 0 Post procedural 1 (0.3) 0 1 (0.3) 0 0 0 hemorrhage Source: ADAE.xpt, ADSL.xpt aIncludes PT terms anal, conjunctival, eye, gastric, hemorrhoidal, lip, penile, rectal, renal, tongue, ulcer, upper gastrointestinal, and vaginal hemorrhage.

Reviewer Comments:

1. Narratives of the two grade 5 AEs in the cabozantinib group, extradural hematoma, and post-procedural hemorrhage, are included in Section 7.3.1. The applicant assessed both events as not-treatment related. However the medical reviewer felt that both events were possibly related to cabozantinib treatment. Despite this, the rate of fatal hemorrhage was low or comparable to other approved VEGF-TKIs. Thus, the medical reviewer agreed with the applicant that hemorrhage could be included in Warnings and Precautions in lieu of a Boxed Warning.

110

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

2. For labeling purposes, hemorrhage is included in Section 5 Warnings and Precautions, and is listed first due to its severity. Providers are instructed to not administer cabozantinib to patient that have or are at risk for severe hemorrhage.

Thrombotic Events

The Applicant included the following terms in Arterial Thrombotic Events: carotid artery occlusion, carotid artery thrombosis, infarction, myocardial infarction, and transient ischemic attack. The reviewer included acute coronary syndrome.

There was a higher incidence of arterial thrombotic events in the cabozantinib group (5 patients [1.5%] in cabozantinib group, 2 patients [0.6%] everolimus group). All events in both treatPHQW JURXSV ZHUH *UDGH The incidence of events of arterial thrombotic events is summarized in Table 57.

Table 57: XL184-308 Incidence of Arterial Thrombotic Events (All Treated Population) Cabozantinib Everolimus N=331 N=322 n (%) n (%) Preferred Term Any AE Grade Grade SAE Any AE Grade Grade SAE 3/4 AE 5 AE Any 3/4 AE 5 AE Any Grade Grade Carotid artery 3 (0.9) 2 (0.6) 0 2 (0.6) 0 0 0 0 occlusiona Myocardial infarctionb 1 (0.3) 0 0 0 1 (0.3) 1 (0.3) 0 1 (0.3) Transient ischemic 1 (0.3) 0 0 1 (0.3) 0 0 0 0 attack (TIA) Acute coronary 0 0 0 0 1 (0.3) 1 (0.3) 0 1 (0.3) syndrome aincludes carotid artery occlusion and carotid artery thrombosis binlcudes myocardial infarction and infarction

Table 58 summarizes patients with arterial thrombotic events with serious adverse events.

111

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 58: XL184-308 Patients with Arterial Thrombotic Events SAE’s (Cabozantinib) Patient ID Sex Grade Onset Duration Action Dose Related Outcome (day) (day) Reduced 184308- M 3 64 105 Medication YNR 4805-3817 Procedure 184308- M 2 64 2 Dose Interrupted NNR 4805-3817 Medication 184308- M 3 225 5 Dose Interrupted NNR 1603-3779 Medication Source: ADAE.xpt, CRFs Abbreviations: F, female; M, male; N, no; O, ongoing; R, resolved/recovered; Y, yes

Patient 4805-3817 experienced a grade 2 serious transient ischemic attack, and a grade 3 serious carotid artery occlusion. The patient was hospitalized on Day 64 of treatment with cabozantinib for a multilocular angiogenic cerebral lesion (Grade 3 cerebrovascular disorder), as well as the carotid artery occlusion and TIA. The patient had a history of underlying atherosclerosis, and the investigator deemed the events as serious but unrelated to cabozantinib. It is possible that these events are related to drug, given the known safety profile of cabozantinib and with other drugs in its class. The patient remained on cabozantinib as of the data cut-off, and the cerebrovascular disorder had not resolved.

The Applicant included the following terms in Venous and Mixed Thrombotic Events: deep vein thrombosis, embolism, hemorrhoids thrombosed, hepatic vein thrombosis, intracranial venous sinus thrombosis, pelvic venous thrombosis, portal vein thrombosis, pulmonary embolism, splenic vein thrombosis, subclavian vein thrombosis, thrombophlebitis superficial, thrombosis, tumor thrombosis, venous thrombosis, and venous thrombosis limb.

There were more venous and mixed thrombotic events in the patients treated with cabozantinib. 7KH PRVW IUHTXHQW YHQRXV DQG PL[HG thrombotic events in the cabozantinib group were pulmonary embolism (3.9% cabozantinib, 0.3% everolimus) and deep vein thrombosis (2.4% cabozantinib, 0.9% everolimus). Grade 3 or 4 pulmonary embolism was reported for 2.7% of patients in the cabozantinib group and 0.3%, or 1 patient in the everolimus group. Pulmonary embolism and other venous and mixed thrombotic events are summarized in Table 59.

112

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 59: XL184-308 Incidence of Venous and Mixed Thrombotic Events (All Treated Population) Cabozantinib Everolimus N=331 N=322 n (%) n (%) Preferred Term Any AE Grade Grade SAE Any AE Grade Grade SAE 3/4 5 AE Any 3/4 AE 5 AE Any AE Grade Grade Deep vein thrombosisa 8 (2.4) 1 (0.3) 0 4 (1.2) 3 (0.9) 0 0 0 Embolism 1 (0.3) 0 0 0 1 (0.3) 0 0 0 Hemorrhoids 1 (0.3)0000000 thrombosed Thrombosisb 6 (1.8) 3 (0.9) 0 2 (0.6) 3 (0.9) 2 (0.6) 0 1 (0.3) Pulmonary embolismc 13 (3.9) 9 (2.7) 0 6 (1.8) 1 (0.3) 1 (0.3) 0 1 (0.3) Thrombophlebitis 1 (0.3) 0 0 0 1 (0.3) 0 0 0 superficial a includes deep vein thrombosis, thrombosis, venous thrombosis, and venous thrombosis limb b includes hepatic vein thrombosis, intracranial venous sinus thrombosis, pelvic venous thrombosis, portal vein thrombosis, splenic vein thrombosis, subclavian vein thrombosis, and tumor thrombosis c includes pulmonary embolism and pulmonary thrombosis

Table 60 summarizes patients with venous thrombotic events with serious adverse events.

113

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 60: XL184-308 Patients with Venous and Mixed Thrombotic Events SAE’s (Cabozantinib) Patient ID Sex Grade Onset Duration Action Dose Related Outcome (day) (day) Reduced 184308- M 3 57 57 Dose interrupted YYR 1357-3441 Medication 184308- M 3 279 65 Dose interrupted YYR 1461-3120 Medication 184308- M 3 366 __ Dose interrupted NYO 3371-3253 Medication 184308- M 3 150 __ Medication N Y O 4458-3593 184308- M 3 88 12 Dose interrupted NYR 4610-3772 Treatment discontinued Medication 184308- M 4 193 __ Medication N N O 6112-3177 died 184308- M 2 281 9 Dose interrupted YYR 1461-3120 Medication Procedure 184308- M 2 227 12 Medication N N R 3607-3549 184308- M 3 250 __ Medication N N O 4502-3670 died 184308- M 2 226 __ Medication N N O 5107-3183 184308- M 2 8 __ Medication N N O 4612-3134 184308- M 3 225 5 Dose interrupted NNR 1603-3779 Medication Source: ADAE.xpt, CRFs Abbreviations: F, female; M, male; N, no; O, ongoing; R, resolved/recovered; Y, yes

Among the cabozantinib-treated patients who experienced pulmonary embolisms AEs, three reported a second AE defined as a venous and mixed/unspecified thrombotic event; one patient experienced an AE of subclavian vein thrombosis, and two patients experienced DVT’s.

Reviewer comments:

1. Across the clinical program, there was increased rate of thrombotic events in each tumor type, including renal cell carcinoma, castrate resistant prostate cancer, and medullary thyroid cancer. The CRPC population experienced higher incidence of venous thrombotic events such as pulmonary embolism and DVT likely due to greater age, possible impacts of gender, greater immobility due to the presence of bone metastases, and the greater extent of pre-treatment.

114

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

2. Given the increased rate of venous thrombotic events in the cabozantinib group compared to the everolimus group in Trial XL184-308, the medical reviewer queried the applicant to determine if there were mitigating risk factors which may have contributed to this increased risk. There were 29 venous and mixed/unspecified events. The applicant was queried for baseline platelet count, and platelet count at time of thrombotic event. The applicant provided information for the 24 patients for which these events arose. Screening platelet values were available for all patients. Platelet values within 7 days of start date of the event was available for 21 patients.

Increased platelet values at screening were reported for 6 patients out of the 24 patients with venous thrombotic events. Two patients with normal screening platelet values had increased platelets within 7 days of the start date of each adverse event. One patient had sepsis which may have precipitated the reported increase of platelet values and is suggestive of reactive thrombocytosis. One patient experienced constipation and hemorrhoids which is a risk factor for the reported event of hemorrhoidal thrombosis.

Increased platelet levels are commonly observed in patients with cancer. The available information does not indicate that there is a clear correlation between increased platelet values and thromboembolic events.

3. For labeling purposes, thrombotic events are included in Section 5 Warnings and Precautions. Providers are instructed to discontinue cabozantinib in patients who develop an acute myocardial infarction or any other arterial thromboembolic complication. They are also informed of the higher rate of venous thromboembolism in cabozantinib (7.3%) treated patients compared to everolimus (2.5%) treated patients. Similarly, providers are informed of higher rates of pulmonary embolism (3.9%) in patients treated with cabozantinib versus everolimus (0.3%).

Osteonecrosis of the jaw (ONJ)

Two patients (0.6%) were reported by the applicant to have ONJ in the cabozantinib group; one patient experienced a Grade 3 SAE and other experienced a nonserious Grade 2 AE. The patient who experienced a Grade 3 SAE of ONJ has an ongoing AE of Grade 2 ONJ which was assessed as related to prior therapy. This patient who experienced the Grade 2 AE also had a history of ONJ prior to randomization.

ONJ was reported for two patients in the everolimus group. Both of these patients experienced Grade 3 SAEs. No AEs of ONJ were reported as Grade 2 or higher in either treatment arm.

Reviewer comments:

115

Reference ID: 3917720

Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Investigator-Sponsored Trials (ISTs) and National Cancer Institute Cancer Therapy Evaluation Program (NCI-CTEP) studies and in post marketing experience.

Reviewer comment:

RPLS is included in Section 5 Warnings and Precautions given its potential morbidity. Cases of RPLS have occurred in the clinical program. Providers are advised to perform an evaluation for RPLS in any patient presenting with seizures, headache, visual disturbances, confusion or altered mental function. Cabozantinib will be discontinued in patients who develop RPLS.

7.4 Supportive Safety Results

7.4.1 Common Adverse Events

Overall, AEs were experienced by 100% of patients in the cabozantinib group and all but one of the patients in the everolimus group. AEs occurring by MedDRA System Organ Classes (SOCs) up to 30-days after treatment discontinuation are shown in Table 61.

117

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 61: XL184-308 Summary of System Organ Class (All Treated Population)

System Organ Class Cabozantinib Everolimus N=331 N=322 n(%) n(%) Blood and lymphatic system disorders 77 (23.3) 137 (42.5) Cardiac disorders 26 (7.8) 22 (6.8) Ear and labyrinth disorders 20 (6.0) 11 (3.4) Endocrine disorders 76 (23.0) 4 (1.2) Eye disorders 29 (8.8) 31 (9.6) Gastrointestinal disorders 312 (94.3) 242 (75.2) General disorders and administrative site conditions 273 (82.5) 264 (82.0) Hepatobiliary disorders 14 (4.2) 10 (3.1) Immune system disorders 6 (1.8) 7 (2.2) Infections and infestations 138 (41.2) 149 (46.3) Injury, poisoning and procedural complications 55 (16.6) 33 (10.3) Investigations 198 (59.8) 140 (43.5) Metabolism and nutrition disorders 219 (66.2) 206 (64.0) Musculoskeletal and connective tissue disorders 180 (54.4) 166 (51.6) Neoplasms benign, malignant and unspecified (incl cysts 32 (9.7) 33 (10.2) and polyps) Nervous system disorders 174 (52.6) 114 (35.4) Psychiatric disorders 74 (22.4) 74 (23.0) Renal and urinary disorders 67 (20.2) 80 (24.8) Reproductive system and breast disorders 24 (7.3) 23 (7.1) Respiratory, thoracic, and mediastinal disorders 183 (55.3) 213 (66.2) Skin and subcutaneous tissue disorders 238 (71.9) 206 (64.0) Surgical and medical procedures 2 (0.6) 2(0.6) Vascular disorders 147 (44.4) 47 (14.6) Source: ADAE.xpt, ADSL.xpt, analysis generated by MAED

Treatment emergent adverse events (TEAEs) occurred in most patients in both the groups. Table 62 OLVWV 7($(V RFFXUULQJ LQ RI SDWLHQWV in the cabozantinib group occurring up to 30-days post treatment.

118

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 62: XL184-308 Treatment Emergent Adverse Events Reactions occurring in &DER]DQWLQLE *URXS $OO 7UHDWHG 3RSXODWLRQ

Cabozantinib Everolimus N=331 N=322 System Organ Class All Grades Grade 3-4 Grade 5 All Grades Grade 3-4 Grade 5 Adverse Event n (%) n (%) n (%) n (%) n (%) n (%) General Disorders And Administration Site Conditions Fatigue 186 (56.2) 30 (9.1) 0 150 (46.6) 22 (6.8) 0 Mucosal inflammation 63 (19.0) 3 (0.9) 0 73 (22.7) 11 (3.4) 0 Asthenia 62 (18.7) 14 (4.2) 0 50 (15.5) 7 (2.2) 0 Gastrointestinal Disorders Diarrhea 245 (74.0) 38 (11.5) 0 89 (27.6) 7 (2.2) 0 Nausea 166 (50.2) 13 (3.9) 0 90 (28.0) 1 (0.3) 0 Vomiting 106 (32.0) 7 (2.1) 0 45 (14.0) 3 (0.9) 0 Stomatitis 74 (22.4) 8 (2.4) 0 77 (23.9) 7 (2.2) 0 Constipation 83 (25.1) 1 (0.3) 0 62 (19.3) 1 (0.3) 0 Abdominal Paina 81 (24) 16 (5) 0 43 (13) 4 (1.2) 0 Dyspepsia 40 (12.1) 1 (0.3) 0 15 (4.7) 0 0 Skin And Subcutaneous Tissue Disorders PPE 139 (42.0) 27 (9.2) 0 19 (5.9) 3 (0.9) 0 Rash 89 (27) 2 (0.6) 0 148 (46) 2 (0.6) 0 Dry skin 37 (11.2) 0 0 32 (9.9) 0 0 Vascular Disorders Hypertensionb 130 (39.2) 53 (16) 0 24 (7.4) 10 (3.1) 0 Blood And Lymphatic System Disorders Anemia 56 (16.9) 18 (5.4) 0 123 (38.2) 50 (15.5) 0 Respiratory, Thoracic And Mediastinal Disorders Dysphonia 66 (19.9) 2 (0.6) 0 12 (3.7) 0 0 Dyspnea 63 (19.0) 10 (3.0) 0 92 (28.6) 14 (4.3) 0 Cough 60 (18.1) 1 (0.3) 0 107 (33.2) 3 (0.9) 0 Metabolism And Nutrition Disorders Decreased Appetite 152 (45.9) 9 (2.7) 0 109 (33.9) 3 (0.9) 0 Musculoskeletal And Connective Tissue Disorders Back pain 56 (16.9) 7 (2.1) 0 47 (14.6) 7 (2.2) 0 Pain in extremity 47 (14.2) 4 (1.2) 0 25 (7.8) 1 (0.3) 0 Muscle spasms 42 (12.7) 0 0 15 (4.7) 0 0 Arthralgia 38 (11.5) 1 (0.3) 0 46 (14.3) 4 (1.2) 0 Nervous System Disorders Dysgeusia 78 (23.6) 0 0 30 (9.3) 0 0 Headache 37 (11.2) 1 (0.3) 0 39 (12.1) 1 (0.3) 0 Dizziness 36 (10.9) 0 0 21 (6.5) 0 0 Endocrine Disorders Hypothyroidism 68 (20.5) 0 0 2 (0.6) 1 (0.3) 0 Renal and Urinary Disorders Proteinuria 41 (12.4) 8 (2.4) 0 30 (9.3) 1 (0.3) 0 Investigations Weight decreased 103 (31.1) 6 (1.8) 0 40 (12.4) 0 0 PPE, palmar-plantar erythrodysesthesia syndrome

119

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

a includes PT terms abdominal pain, abdominal pain upper, abdominal pain lower and abdominal discomfort bincludes hypertension, blood pressure increased, hypertensive crisis, blood pressure fluctuation

Source: Created by reviewer using MAED and ADAM datasets on 2/10/16: ADAE (AEBODSYS, AEDECOD) and ADSL (TRTA)

The review of safety evaluated additional potential toxicities of study drug therapy through analyses of the incidence of AEs based on hierarchical composites of MedDRA preferred terms (i.e. high level terms) and hierarchical composites of MedDRA high level terms (i.e. high level group terms) in each treatment group. The common high level terms in the cabozantinib group were diarrhea (74%), asthenic conditions (71%), nausea and vomiting symptoms (56.5%), appetite disorders (45.9%), and skin and subcutaneous conditions (42%). Table 63 and Table 64 summarize the incidence of high level terms and high-OHYHO JURXS WHUPV UHVSHFWLYHO\ RFFXUULQJ LQ RI WKH cabozantinib group compared to the everolimus group occurring up to 30-days post treatment.

120

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 63: XL184-308 7UHDWPHQW (PHUJHQW $GYHUVH (YHQWV 10%) by High Level Term (All Treated Population)

High Level Term Cabozantinib Everolimus N=331 N=322 n(%) n(%) Diarrhea (excluding infective) 245 (74.0) 89 (27.6) Asthenic conditions 235 (71.0) 188 (58.4) Nausea and vomiting symptoms 187 (56.5) 108 (33.5) Appetite disorders 152 (45.9) 110 (34.2) Skin and subcutaneous conditions NEC 139 (42.0) 19 (5.9) Vascular hypertensive disorders NEC 122 (36.9) 23 (7.1) Musculoskeletal and connective tissue pain and discomfort 118 (35.7) 100 (31.1) Physical examination procedures and organ system status 107 (32.3) 47 (14.6) Gastrointestinal atonic and hypomotility disorders NEC 96 (29) 65 (20.2) Upper respiratory tract signs and symptoms 93 (28.1) 32 (9.9) Liver function analyses 87 (26.3) 43 (13.4) Sensory abnormalities NEC 82 (24.8) 33 (10.3) Stomatitis and ulceration 79 (23.9) 97 (30.1) Gastrointestinal and abdominal pains (excl oral and throat) 77 (23.3) 41 (12.7) Coughing and associated symptoms 74 (22.4) 116 (36.0) Thyroid hypofunction disorders 68 (20.5) 2 (0.6) Breathing abnormalities 66 (19.9) 96 (29.8) Mucosal findings abnormal 65 (19.6) 75 (23.3) Rashes, eruptions, and exanthems NEC 58 (17.5) 112 (34.8) Anemias NEC 56 (16.9) 125 (38.8) Magnesium metabolism disorders 54 (16.3) 5 (1.6) Potassium imbalance 52 (15.7) 36 (11.2) Dermal and epidermal conditions NEC 51 (15.4) 38 (11.8) Urinary abnormalities 47 (14.2) 42 (13.0) Pain and discomfort NEC 46 (13.9) 44 (13.7) Upper respiratory tract infections 44 (13.3) 43 (13.4) Dyspeptic signs and symptoms 42 (12.7) 15 (4.7) Muscle related signs and symptoms NEC 42 (12.7) 17 (5.3) General signs and symptoms NEC 41 (12.4) 33 (10.3) Joint related signs and symptoms 40 (12.1) 52 (16.2) Neurological signs and symptoms NEC 39 (11.8) 21 (6.5) Headaches NEC 38 (11.5) 40 (12.4) Tissue enzyme analyses NEC 37 (11.2) 25 (7.8) Lower respiratory tract and lung infections 35 (10.6) 42 (13.0) Source: ADAE.xpt, ADSL.xpt, analysis generated by MAED

121

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 64: XL184-308 7UHDWPHQW (PHUJHQW $GYHUVH (YHQWV E\ +LJK /HYHO *URXS Term (All Treated Population)

High Level Group Term Cabozantinib Everolimus N=331 N=322 n(%) n(%) Gastrointestinal motility and defecation conditions 277 (83.7) 138 (42.9) General system disorders NEC 269 (81.3) 257 (79.8) Gastrointestinal signs and symptoms 228 (68.9) 140 (43.5) Respiratory disorders NEC 167 (50.5) 178 (55.3) Appetite and general nutritional disorders 152 (45.9) 112 (34.8) Skin and subcutaneous tissue disorders NEC 143 (43.2) 23 (7.1) Neurologic disorders NEC 140 (42.3) 80 (24.8) Epidermal and dermal conditions 133 (40.2) 180 (55.9) Infections – pathogen unspecified 124 (37.5) 131 (40.7) Musculoskeletal and connective tissue disorders NEC 123 (37.2) 105 (32.6) Vascular hypertensive disorders 123 (37.2) 24 (7.5) Physical examination and organ system status topics 107 (32.3) 47 (14.6) Oral soft tissue conditions 96 (29) 103 (32.0) Hepatobiliary investigations 87 (26.3) 43 (13.4) Bone, calcium, magnesium, and phosphorous metabolism 86 (26.0) 38 (11.8) disorders Electrolyte and fluid balance conditions 80 (24.2) 53 (16.5) Thyroid gland disorders 76 (23.0) 2 (0.6) Muscle disorders 69 (20.9) 43 (13.4) Skin appendage conditions 66 (19.9) 58 (18.0) Urinary tract signs and symptoms 62 (18.7) 67 (20.8) Anemias nonhemolytic and marrow depression 56 (16.9) 127 (39.4) Joint disorders 46 (13.9) 53 (16.5) Enzyme investigations NEC 38 (11.5) 25 (7.8) Headaches 38 (11.5) 41 (12.7) Injuries NEC 36 (10.9) 20 (6.2) Source: ADAE.xpt, ADSL.xpt, analysis generated by MAED

7.4.2 Laboratory Findings

Abnormalities in hematology tests were primary Grade 1 to 2 in severity in the cabozantinib group. Decreased white blood cells and decreased absolute neutrophil count occurred more frequently in the cabozantinib group compared with the everolimus group. However decreases in hemoglobin, lymphocytes, and platelets were more frequent in the everolimus treated group. Abnormalities in liver function tests were also primarily Grade 1 to 2 severity. Grade 3-4 increases in liver function tests (AST, ALT, and ALP) occurred more frequently in the cabozantinib group compared to the

122

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

everolimus group. Abnormalities in renal function were more frequent in the everolimus group (71%) compared to the cabozantinib group (58%).

Table 65 summarizes ODERUDWRU\ DEQRUPDOLWLHV RFFXUULQJ LQ RI SDWLHQWV ZKR received cabozantinib.

Table 65 XL184-308 Treatment Emergent Laboratory AbnormalitiHV 2FFXUULQJ LQ of Patients (Cabozantinib) Test Cabozantinib Everolimus N=331 N=322 n (%) n (%) Any AE Grade 3/4 Any AE Grade AE 3/4 AE Chemistry AST increased 74 3 40 <1 ALT increased 68 3 32 <1 Creatinine increased 58 <1 71 0 Triglycerides increased 53 4 73 13 Hypophosphatemia 48 8 36 5 Hyperglycemia 37 2 59 8 Hypoalbuminemia 36 2 28 <1 ALP increased 35 2 29 1 Hypomagnesemia 31 7 4 <1 Hyponatremia 30 8 26 6 GGT increased 27 5 43 9 Hematology White blood cells decreased 35 <1 31 <1 Absolute neutrophil count 31 2 17 <1 decreased Hemoglobin decreased 31 4 71 17 Lymphocytes decreased 25 7 39 12 Platelets decreased 25 <1 27 <1 Source: Applicant Response to FDA Information Request ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma glutamyl transferase

The potential for cabozantinib to induce liver injury was assessed by the incidence of potential Hy’s Law cases, and by the incidence of AEs related to hepatotoxicity.

A summary of patients who met laboratory screening for potential drug-induced liver injury (DILI; concurrent ALT or AST > 3x the ULN, total bilirubin > 2 x ULN, and ALP < 2

123

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

X ULN [potential Hy’s Law cases] or concurrent ALT or AST > 3 x ULN, total bilirubin > [ 8/1 DQG $/3 [ 8/1 LQ SUHVHQWHG LQ Table 66.

Table 66: XL184-308 Potential DILI Cases by Laboratory Screening Criteria (All Treated Population) Cabozantinib Everolimus N=331 N=322 n (%) n (%)

Potential DILI case: >3 x ULN ALT or AST, >2 x ULN Total bilirubin, and: <2 x ULN ALP (total patients; potential Hy’s 00 Law cases) [ 8/1 $/3 WRWDO SDWLHQWV 2 (0.6) 1 (0.3) [ ULN ALP 0 0 Source: ISS Table 50, CRF

There were no cases that met the screening criteria for Hy’s Law in either treatment arm. There were two patients in the cabozantinib group who had ALT or AST > 3 x ULN, total bilirubin > 2 x ULN, DQG FRQFXUUHQW $/3 [ 8/1 %RWK SDWLHQWV KDG FRQIRXQGLQJ factors:

Patient 3910-3100 experienced an elevation of transaminases 30 days after initiation of treatment with cabozantinib. On Study Day 32, treatment with cabozantinib was interrupted due to fever, asthenia, PPES, and mucositis. On Study Day 36, after taking ciprofloxacin for 3 days and acetaminophen for 2 days, the patient met the screening FULWHULD $/7 DQG $67 ! [ 8/1 WRWDO ELOLUXELQ ! [ 8/1 DQG FRQFXUUHQW $/3 [ ULN). The clinical characteristics of the abnormalities were consistent with cholestatic hepatitis which was confirmed by histology from a liver biopsy. Following discontinuation of cabozantinib, ciprofloxacin, and acetaminophen and initiation of immunosuppressive treatment with methylprednisone, the subject’s ALT, AST, and bilirubin laboratory values decreased to near normal levels. The case was confounded by the use of concomitant medications of ciprofloxacin and acetaminophen immediately after the initial elevation of liver enzymes which may have contributed to the cholestatic hepatitis event. An additional potential cause could have been delayed autoimmune-mediated hepatitis associated with prior nivolumab treatment; the patient discontinued nivolumab 2 to 3 months prior to enrolling in the current study.

Patient 1361-3312 had laboratory assessments that met the screening criteria of ALT DQG $67 ! [ 8/1 WRWDO ELOLUXELQ ! [ 8/1 DQG FRQFXUUHQW $/3 [ 8/1 RQ 6WXG\ Day 351. The patient had tumor metastases in the pancreas at baseline and elevated amylase and lipase levels. Treatment with cabozantinib had been interrupted on Study

124

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Day 338 due to recurrent and worsening of ALT and AST levels without an increase in bilirubin or significant changes in ALP. On Study Day 358, the patient was hospitalized for uncontrolled abdominal pain. An abdominal CT scan showed bile duct obstruction by pancreatic head mass. The patient underwent endoscopic retrograde cholangiopancreatography with biliary stent placement. Following this procedure, ALT, AST, bilirubin, ALP, serum amylase and lipase levels were resolving, and treatment with cabozantinib was discontinued due to PD. A temporal relationship exists between the initial elevations of ALT and AST and cabozantinib treatment; however the malignant bile duct obstruction due to disease progression further increased transaminase levels and resulted in elevated cholestatic markers (ALP and bilirubin).

Reviewer comment:

1. There were no Hy’s Law cases identified. The incidence of screening events for potential cases which met Hy’s Law criteria was comparable on both study arms suggesting no marked increase in the risk of hepatotoxicity associated with the use of cabozantinib.

7.4.3 Vital Signs

Based on analyses of mean value and mean change at baseline at each cycle, no clinically meaningful differences in heart rate or temperature were observed during the course of treatment with either study group.

See Section 7.3.5 for a discussion regarding hypertension.

The effect of cabozantinib on body mass composition has not been formally evaluated. Table 67 VXPPDUL]HV WKH LQFLGHQFH RI ORVV RI RI EDVHOLQH ERGy weight in Study XL184-308. The median time to this degree of weight loss was similar in either treatment arm.

125

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 67 XL184- ,QFLGHQFH RI :HLJKW /RVV RI %DVHOLQH %RG\ :HLJKW All Treated Population) Cabozantinib Everolimus N=331 N=322 n (%) n (%) 3DWLHQWV H[SHULHQFLQJ ZHLJKW ORVV YV EDVHOLQe body 152 (46) 69 (19) weight, n (%) 0HGLDQ WLPH WR ZHLJKW ORVV YV EDVHOLQH ERG\ ZHLJKW 141 133 (days) Source: ISS Table 58

7.4.4 Electrocardiograms (ECGs)

The incidence of events of QTc prolongation was low for both treatment groups (one patient each).

Patient 4929-3536 in the cabozantinib group had a Grade 2 nonserious AE of ECG QT prolonged on Study Day 59. The patient was receiving cabozantinib at a dose of 60 mg daily at the time of onset of the AE, and no modifications in study treatment were implemented as a result of the event. No QTC correction by the Fridericia’s formula (QTcF) > 500 ms was reported for this patient, and the absolute increase in QTcF from baseline for this patient was unknown.

7.4.5 Special Safety Studies/Clinical Trials

There were no special safety studies or clinical trials in this application.

7.4.6 Immunogenicity

Review of clinical safety does not raise concerns of immunogenicity with this small molecule. No formal immunogenicity studies were required or were done.

7.5 Other Safety Explorations

7.5.1 Dose Dependency for Adverse Events

Refer to Clinical Pharmacology review.

126

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

7.5.2 Time Dependency for Adverse Events

Median times to submission specific adverse events are summarized in Table 68.

Table 68 XL184-308 Time to First Occurrence of Submission Specific Adverse Events Adverse Event Cabozantinib (60 mg) N=331 Time to First Occurrence (weeks) Median (25th, 75th percentiles) GI perforation 100.0 (7.0, 40.0) Fistulas 30.3 (26.2, 35.4) +HPRUUKDJH *UDGH 20.9 (4.6, 35.3) Arterial thrombotic events 11.4 (9.1, 32.1) Venous and mixed/unspecified thrombotic events 16.1 (8.1, 25.8) Wound complications 11.0 (5.6, 25.1) Hypertension 3.0 (2.0, 6.1) Osteonecrosis 10.0 (7.0, 13.0) PPES 3.4 (2.3, 6.1) Proteinuria 4.1 (2.4, 8.1) Diarrhea 4.9 (2.7, 8.1) QT prolongation 8.4 (8.4, 8.4) GI, gastrointestinal; PPES, palmar-plantar erythrodysesthesia Source: ISS Table 28

Reviewer comment:

The median time to occurrence of common adverse events such as hypertension, diarrhea, and PPES was 3.0, 4.9, and 3.4 weeks respectively. However, the median time to first dose reduction was 55 days (range 10,355), or approximately 7.8 weeks. This reflects clinical practice, in which events such as hypertension, diarrhea, and PPES were initially managed by either supportive medication such as antihypertensive or antidiarrheal therapy.

7.5.3 Drug-Demographic Interactions

Subgroup analyses based on age, gender, and race are shown below. All AEs appear WR EH VLPLODU DFURVV DJHV \HDUV DQG DJH \HDUV

Treatment emergent adverse events by subgroup, including age, gender, and race and ethnicity, are summarized in Table 69 and Figure 10.

127

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 69 XL184-308 Treatment Emergent Adverse Events by Subgroup (All Treated Population)

(Safety) Subgroup Cabozantinib Everolimus N=331 N=322 n(%) n(%) Sex Male 254 (100) 236 (99.6) Female 77 (100) 84 (100) Age Group <65 197 (100) 193 (100) 134 (100) 128 (99.2) Race White 270 (100) 257 (99.6) Black or African American 6 (100) 3 (100) Asian 21 (100) 25 (100) Other 34 (100) 35 (100) Missing 0 (100) 1 (100) Ethnicity Hispanic or Latino 19 (100) 18 (100) Not Hispanic or Latino 278 (100) 267 (99.6) Missing 34 (100) 36 (100) Region North American 119 (100) 118 (100) Europe 167 (100) 150 (99.3) Rest of the World 45 (100) 53 (100) Source: ADSL, ADAE

128

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Figure 10 XL184-308 Treatment Emergent Adverse Events by Subgroup (All Treated Population)

Source: ADSL, ADAE The x axis is on the log scale.

Treatment emergent adverse events leading to dose reduction by subgroup, including age, gender, and race and ethnicity, are summarized in Table 70 and Figure 11.

129

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Table 70 XL184-308 Treatment Emergent Events Leading to Dose Reduction by Subgroup (All Treated Population) (Safety) Subgroup Cabozantinib Everolimus N=331 N=322 n(%) n(%) Sex Male 153 (60.2) 51 (21.5) Female 47 (61.0) 27 (32.1) Age Group <65 112 (56.9) 41 (21.2) 88 (65.7) 37 (28.7) Race White 159 (58.9) 65 (25.2) Black or African American 4 (66.7) 2 (66.7) Asian 17 (81.0) 3 (12.0) Other 20 (58.8) 8 (22.9) Ethnicity Hispanic or Latino 11 (57.9) 5 (27.8) Not Hispanic or Latino 170 (61.2) 65 (24.3) Missing 19 (55.9) 8 (22.2) Region North American 66 (55.5) 31 (26.3) Europe 102 (61.1) 39 (25.8) Rest of the World 32 (71.1) 8 (15.1) Source: ADSL, ADAE

Figure 11 XL184-308 Treatment Emergent Events Leading to Dose Reduction by Subgroup (All Treated Population)

Source: ADSL, ADAE The x axis is on the log scale.

Reviewer comment:

130

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Treatment emergent events leading to dose reduction were more common in the cabozantinib group compared to the everolimus group. This exploratory analysis shows that Asian patients treated with cabozantinib experienced a higher rate of dose reductions (81%) than those treated with everolimus (12%). Furthermore, their rate of dose reductions was higher than in other ethnic groups. This analysis is exploratory in nature, and is limited by the small number of Asian patients treated with cabozantinib (n=21), as well as with everolimus (n=25).

Treatment emergent serious adverse events by subgroup, including age, gender, and race and ethnicity, are summarized in Table 71 and Figure 12.

Table 71 XL184-308 Treatment Emergent Serious Adverse Events by Subgroup (Safety) Subgroup Cabozantinib Everolimus N=331 N=322 n(%) n(%) Sex Male 101 (39.8) 97 (40.9) Female 30 (39.0) 41 (48.8) Age Group <65 72 (36.5) 80 (41.5) 59 (44.0) 59 (45.7) Race White 107 (39.6) 115 (44.6) Black or African American 2 (33.3) 2 (66.7) Asian 7 (33.3) 12 (48.0) Other 15 (44.1) 9 (25.7) Ethnicity Hispanic or Latino 3 (15.8) 8 (44.4) Not Hispanic or Latino 108 (38.8) 121 (45.1) Missing 20 (58.8) 10 (27.8) Region North American 38 (31.9) 43 (36.4) Europe 75 (44.9) 68 (45.0) Rest of the World 18 (40.0) 28 (52.8) Source: ADSL, ADAE

131

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Figure 12 XL184-308 Treatment Emergent Serious Adverse Events by Subgroup

Reviewer comment:

Serious adverse events of any grade up to 30-days after last dose of study therapy, regardless of causality, were similar in both treatment groups (39.6% cabozantinib, 43.2% everolimus). This exploratory analysis shows that African American or Black patients treated with everolimus experienced a higher rate of serious adverse events (66.7%) than dose treated with cabozantinib (33.3%). However this analysis is exploratory in nature, and is limited by the small number of African American or Black patients treated with cabozantinib (n=6) and everolimus (n=3) who experienced SAE’s.

7.5.4 Drug-Disease Interactions

Refer to FDA Clinical Pharmacology Review.

7.5.5 Drug-Drug Interactions

Refer to FDA Clinical Pharmacology Review.

The administration of cabozantinib with agents that are strong CYP3A4 inhibitors should be used with caution. Chronic administration of agents that are strong CYP34A inducers with cabozantinib should be avoided.

An internally initiated FDA review for cases of drug interaction with cabozantinib identified a literature case report of a drug interaction between cabozantinib and warfarin in a renal cell carcinoma treated patient. Upon further review it was determined that the patient was treated with COMETRIQ off-label, which is a different formulation and dose, approved for the treatment of medullary thyroid cancer.

Currently, there is no direct clinical evidence of increased of drug interactions between CABOMEYTX and warfarin.

7.6 Additional Safety Evaluations

7.6.1 Human Carcinogenicity

Refer to FDA Nonclinical Review.

132

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

To fulfill a PMR for NDA 203756, the Applicant evaluated the carcinogenicity potential of XL184 in a 26-week Tg.rasH2 mouse study. No test article-related increases in the incidence of neoplastic lesions were reported in this model at doses up to 15 mg/kg/day, the highest doses tested (1.2-fold, the nominal MRHD of 60 mg/day of cabozantinib). XL184, was not found to be mutagenic or clastogenic (per Dr. Margaret Brower’s review NDA# 203756).

7.6.2 Human Reproduction and Pregnancy Data

No clinical studies with cabozantinib have been conducted in pregnant or lactating women. No clinical data on exposure during pregnancy or lactation are available. It is not known whether cabozantinib is excreted in human milk. Therefore lactating females were excluded from clinical trials of cabozantinib.

7.6.3 Pediatrics and Assessment of Effects on Growth

Not applicable for this application.

7.6.4 Overdose, Drug Abuse Potential, Withdrawal and Rebound

Limited data regarding overdose have been obtained in patients receiving cabozantinib, and overdoses for cabozantinib-treated patients on Phase 3 clinical trials employing the 60-mg tablet starting dose have been infrequent. Through the data cutoff of May 22, 2015, one patient took 180 mg cabozantinib for 1 day; on Studies XL184-307 and XL184-306, one patient each in the respective cabozantinib arms took 120 mg cabozantinib for day 1. A drug-related SAE of Grade 3 diarrhea was reported shortly after the overdose of the XL184-307 patient, and that patient subsequently underwent a dose reduction from 60 mg to 40 mg qd. The diarrhea reduced in severity after 7 days but persisted as a Grade 1 or 2 AE for several weeks. Concurrently with the diarrhea, a Grade 3 AE of treatment-related general physical health deterioration was also reported for this patient. For the overdose events in Studies XL184-308 and XL184-306, no treatment-related AEs or SAEs were reported within 30 days after the overdose.

There is no evidence that this class of drug is habit forming or could lead to dependence. There is no anticipated risk of abuse.

Clinical data are not available to evaluate the effect on disease or clinical biomarkers after withdrawal of cabozantinib.

7.7 Additional Submissions / Safety Issues

Not applicable for this application.

133

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

8 Postmarket Experience

Cabozantinib capsules (Cometriq) were first approved by the FDA on November 29, 2012 for the treatment of patients with progressive, metastatic MTC at a dose of 140 mg qd. Cometriq was made commercially available in the United States on January 24, 2013. On March 21, 2014, cabozantinib capsules (Cometriq) at the 140 mg dose received approval through the centralized procedure by the European Commission for the treatment of adults with progressive, unresectable locally advanced or metastatic MTC.

The post-marking patient population through May 22, 2015 comprised 1149 total patients exposed including approximately 1083 in the US, 42 in the EU, and 24 from other countries.

Through May 22, 2015, patients in the US marketed setting have received cabozantinib for the treatment of thyroid cancer (n=453) as well as malignancies other than the approved indication, including prostate cancer (n=184), renal cancer (n=183), hepatocellular cancer (n=19), and lung cancer (n=61). In the EU, patients have thus far received marketed drug for MTC (n=11), pheochromocytoma (n=1), and HCC (n=1). Cumulatively, 587 serious adverse reactions (SARs) have been reported in the post- marketing setting through May 22, 2015. No new safety findings bearing on the known overall safety profile were identified.

Through May 22, 2015, 75 post-marketing SARs for 49 cases were received in patients who received Cometriq off-label for the indication of renal cancer (including RCC and malignant neoplasm of the renal pelvis). With the exception of unknown cause of death (death [n=11]), pneumonia (n=4), dehydration (n=3), rectal hemorrhage (n=3), hypertension (n=2), hypotension (n=2), vomiting (n=2), and pain in extremity (n=2), the occurrence of any individual SAR was limited to one event. After the May 22, 2015 cutoff, one unconfirmed case of posterior reversible encephalopathy syndrome (PRES; also called RPLS) was reported by a non-study physician via the post-marketing process for a patient who was enrolled in Study XL184-308. The report was not contemporaneous with the event (made > 1 year afterward) and there was inconsistent information in the report regarding the date of the event relative to study treatment. The patient also had confounding factors including receipt of a prior VEGR-TKI and radiation for brain metastases. There is no evidence of imaging supporting the diagnosis of RPLS, and the event was not confirmed by the study investigator. Additional follow-up is ongoing, at the time of this review.

134

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

9 Appendices

9.1 Literature Review/References

1. SEER Stat Fact Sheets: Kidney and Renal Pelvis Cancer. http://seer.cancer.gov/statfacts/html/kidrp.html 2015 (2015). 2. Ryan Q. FDA clinical review of everolimus for the treatment of advanced renal cell carcinoma. http://www.accessdata.fda.gov/drugsatfda docus/nda/2009/022334s000 MedR P1.pdf (2009). 3. Mckee A. FDA clinical review of axitinib for the treatment of metastatic renal cell cancer.http://www.accessdata.fda.gov/drugsatfda docs/nda/2012/202324Orig1s000Me dR.pdf (2012). 4. Kane, R.C. et al. Sorafenib for the treatment of advanced renal cell carcinoma. Clin Cancer Res 12, 7271-8 (2006). 5. FDA approves Opdivo to treat advanced form of kidney cancer http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm473971.htm 6. Guidance for Industry Warnings and Precautions, Contraindications, and Boxed Warning Sections of Labeling for Human Prescription Drug and Biological Products – Content and Format http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guida nces/UCM075096.pdf

9.2 Labeling Recommendations

Although labelling negotiations are ongoing, the following is a summary of key labeling decisions, which at the time of this review were agreed upon by the applicant and the medical review team:

Section 1 Indications and Usage

Cabozantinib is indicated for the treatment of patients with advanced renal cell carcinoma who have received prior anti-angiogenic therapy. The rationale for this recommendation is discussed is Section 6.1 of this review.

Boxed Warning

Hemorrhage is will described in the label in Section 5 Warnings and Precautions and not as a Boxed Warning. The rationale for this recommendation is discussed in Section 7.3.5 of this review.

Section 5 Warnings and Precautions

135

Reference ID: 3917720 Clinical Review Michael Brave, Harpreet Singh NDA 208692 Cabometyx (cabozantinib)

Proteinuria and wound complications are not included in Section 5. These adverse events are listed in Section 6 Adverse Reactions. Diarrhea is included in the Warnings and Precautions section of the label. The rationale for these recommendations is discussed in Section 7.3.5 of this review.

9.3 Advisory Committee Meeting

The Division did not obtain the advice of the Oncologic Drug Advisory Committee (ODAC) for this application.

136

Reference ID: 3917720 ------This is a representation of an electronic record that was signed electronically and this page is the manifestation of the electronic signature. ------/s/ ------B HARPREET SINGH 04/15/2016

MICHAEL H BRAVE 04/15/2016

JULIA A BEAVER 04/15/2016

Reference ID: 3917720 CLINICAL FILING CHECKLIST FOR NDA/BLA or Supplement

NDA/BLA Number: 208692 Applicant: Exelixis, Inc. Stamp Date: December 22, 2015 Drug Name: Cabozantinib NDA/BLA Type: NDA Supplement N/A

On initial overview of the NDA/BLA application for filing:

Content Parameter Yes No NA Comment FORMAT/ORGANIZATION/LEGIBILITY 1. Identify the general format that has been used for this Electronic CTD application, e.g. electronic common technical document (eCTD). 2. Is the clinical section legible and organized in a manner to X allow substantive review to begin? 3. Is the clinical section indexed (using a table of contents) X and paginated in a manner to allow substantive review to begin? 4. For an electronic submission, is it possible to navigate the X application in order to allow a substantive review to begin (e.g., are the bookmarks adequate)? 5. Are all documents submitted in English or are English X translations provided when necessary? LABELING 6. Has the applicant submitted a draft prescribing information X that appears to be consistent with the Physician Labeling Rule (PLR) regulations and guidances (see http://www.fda.gov/Drugs/GuidanceComplianceRegulatory Information/LawsActsandRules/ucm084159 htm SUMMARIES 7. Has the applicant submitted all the required discipline X summaries (i.e., Module 2 summaries)? 8. Has the applicant submitted the integrated summary of X safety (ISS)? 9. Has the applicant submitted the integrated summary of X efficacy (ISE)? 10. Has the applicant submitted a benefit-risk analysis for the X product? 11. Indicate if the Application is a 505(b)(1) or a 505(b)(2). X 505(b)(2) Applications 12. If appropriate, what is the relied upon listed drug(s)? X 13. Did the applicant provide a scientific bridge demonstrating X the relationship between the proposed product and the listed drug(s)/published literature? 14. Describe the scientific bridge (e.g., BA/BE studies) X DOSAGE 15. If needed, has the applicant made an appropriate attempt to X The applicant made an determine the correct dosage regimen for this product (e.g., appropriate attempt to appropriately designed dose-ranging studies)? determine the correct Study Number: XL184-020 dosage as evidenced Study Title: A Phase I, Open Label, Randomized, Single- by the fact that Trial Dose, Three-Treatment, Parallel Pharmacokinetic Study of 308 used a different Cabozantinib (XL184) Tablet Formulation Administered as dose than was a Single Oral Dose in Healthy Adult Subjects approved for Sample Size: 63 Medullary Thyroid Treatment Arms: 20 mg, 40 mg, or 60 mg tablets (XL184) Cancer. Whether they File name: Clinical Filing Checklist for NDA 208692 1

Reference ID: 3875369 CLINICAL FILING CHECKLIST FOR NDA/BLA or Supplement

Content Parameter Yes No NA Comment Location in submission: Module 5.3.3.1 have determined the optimal dose will be a review issue. EFFICACY 16. Do there appear to be the requisite number of adequate and X well-controlled studies in the application?

Pivotal Study #1 XL184-308: A Phase 3, Randomized, Controlled Study of Cabozantinib (XL184) vs Everolimus in Subjects with Metastatic Renal Cell Carcinoma that has Progressed after Prior VEGFR Tyrosine Kinase Inhibitor Therapy

Indication: As monotherapy for the treatment of patients with unresectable advanced or metastatic renal cell carcinoma (RCC) following one prior vascular endothelial growth factor (VEGF) targeted therapy

17. Do all pivotal efficacy studies appear to be adequate and X well-controlled within current divisional policies (or to the extent agreed to previously with the applicant by the Division) for approvability of this product based on proposed draft labeling? 18. Do the endpoints in the pivotal studies conform to previous X Primary Endpoint: Agency commitments/agreements? Indicate if there were PFS not previous Agency agreements regarding Secondary Endpoint: primary/secondary endpoints. ORR and OS 19. Has the application submitted a rationale for assuming the X Sites outside of North applicability of foreign data to U.S. population/practice of America enrolled 63% medicine in the submission? of patients trial XL184-308. Foreign sites in the trial have relatively similar standard in practice of medicine to the U.S. In addition, this trial was conducted according to GCP. SAFETY 20. Has the applicant presented the safety data in a manner X consistent with Center guidelines and/or in a manner previously requested by the Division? 21. Has the applicant submitted adequate information to assess X the arrhythmogenic potential of the product (e.g., QT interval studies, if needed)? 22. Has the applicant presented a safety assessment based on all X current worldwide knowledge regarding this product? 23. For chronically administered drugs, have an adequate X Exposures were not number of patients (based on ICH guidelines for exposure1) chronic as defined by

File name: Clinical Filing Checklist for NDA 208692 2

Reference ID: 3875369 CLINICAL FILING CHECKLIST FOR NDA/BLA or Supplement

Content Parameter Yes No NA Comment been exposed at the dosage (or dosage range) believed to be ICH guidelines. efficacious? 24. For drugs not chronically administered (intermittent or X short course), have the requisite number of patients been exposed as requested by the Division? 25. Has the applicant submitted the coding dictionary2 used for X mapping investigator verbatim terms to preferred terms? 26. Has the applicant adequately evaluated the safety issues that X are known to occur with the drugs in the class to which the new drug belongs? 27. Have narrative summaries been submitted for all deaths and X adverse dropouts (and serious adverse events if requested by the Division)? OTHER STUDIES 28. Has the applicant submitted all special studies/data X requested by the Division during pre-submission discussions? 29. For Rx-to-OTC switch and direct-to-OTC applications, are X the necessary consumer behavioral studies included (e.g., label comprehension, self selection and/or actual use)? PEDIATRIC USE 30. Has the applicant submitted the pediatric assessment, or X Applicant requested a provided documentation for a waiver and/or deferral? waiver for pediatric age subsets. PREGNANCY, LACTATION, AND FEMALES AND MALES OF REPRODUCTIVE POTENTIAL USE 31. For applications with labeling required to be in Pregnancy X Section 5.4 in and Lactation Labeling Rule (PLLR) format, has the Summary of Clinical applicant submitted a review of the available information Safety in Module 2 regarding use in pregnant, lactating women, and females addresses this issue. and males of reproductive potential (e.g., published No clinical studies literature, pharmacovigilance database, pregnancy registry) with cabozantinib in Module 1 (see have been conducted http://www.fda.gov/Drugs/DevelopmentApprovalProcess/D in pregnant or evelopmentResources/Labeling/ucm093307 htm)? lactating women. ABUSE LIABILITY 32. If relevant, has the applicant submitted information to X assess the abuse liability of the product? FOREIGN STUDIES 33. Has the applicant submitted a rationale for assuming the X See #19 above applicability of foreign data in the submission to the U.S. population? DATASETS 34. Has the applicant submitted datasets in a format to allow X

1 For chronically administered drugs, the ICH guidelines recommend 1500 patients overall, 300-600 patients for six months, and 100 patients for one year. These exposures MUST occur at the dose or dose range believed to be efficacious. 2 The “coding dictionary” consists of a list of all investigator verbatim terms and the preferred terms to which they were mapped. It is most helpful if this comes in as a SAS transport file so that it can be sorted as needed; however, if it is submitted as a PDF document, it should be submitted in both directions (verbatim -> preferred and preferred -> verbatim). File name: Clinical Filing Checklist for NDA 208692 3

Reference ID: 3875369 CLINICAL FILING CHECKLIST FOR NDA/BLA or Supplement

Content Parameter Yes No NA Comment reasonable review of the patient data? 35. Has the applicant submitted datasets in the format agreed to X previously by the Division? 36. Are all datasets for pivotal efficacy studies available and X complete for all indications requested? 37. Are all datasets to support the critical safety analyses X available and complete? 38. For the major derived or composite endpoints, are all of the X raw data needed to derive these endpoints included? CASE REPORT FORMS 39. Has the applicant submitted all required Case Report Forms X in a legible format (deaths, serious adverse events, and adverse dropouts)? 40. Has the applicant submitted all additional Case Report X Forms (beyond deaths, serious adverse events, and adverse drop-outs) as previously requested by the Division? FINANCIAL DISCLOSURE 41. Has the applicant submitted the required Financial X Disclosure information? GOOD CLINICAL PRACTICE 42. Is there a statement of Good Clinical Practice; that all X clinical studies were conducted under the supervision of an IRB and with adequate informed consent procedures?

IS THE CLINICAL SECTION OF THE APPLICATION FILEABLE? __Yes______

If the Application is not fileable from the clinical perspective, state the reasons and provide comments to be sent to the Applicant.

Please identify and list any potential review issues to be forwarded to the Applicant for the 74- day letter.

Reviewing Medical Officer Date

Clinical Team Leader Date

File name: Clinical Filing Checklist for NDA 208692 4

Reference ID: 3875369 ------This is a representation of an electronic record that was signed electronically and this page is the manifestation of the electronic signature. ------/s/ ------B HARPREET SINGH 01/20/2016 Filing Checklist - clinical

JULIA A BEAVER 01/20/2016

Reference ID: 3875369