`
`CENTER FOR DRUG EVALUATION AND
`RESEARCH
`
`
`APPLICATION NUMBER:
`203085Orig1s000
`
`STATISTICAL REVIEW(S)
`
`
`
`
`
`
`
`

`

`
`
`
`
`U.S. Department of Health and Human Services
`Food and Drug Administration
`Center for Drug Evaluation and Research
`Office of Translational Sciences
`Office of Biostatistics
`
`
`S T A T I S T I C A L R E V I E W A N D E VA L U A T I O N
`CLINICAL STUDIES
`
`NDA #:
`Drug Name:
`Indication(s):
`Applicant:
`Date(s):
`
`203085
`Stivarga® (Regorafenib)
`Metastatic colorectal cancer
`Bayer HealthCare Pharmaceuticals
`Submission: 4/27/2012
`PDUFA: 9/27/2012
`Review Priority:
`Priority
`
`
`Biometrics Division:
`Division of Biometrics V
`Statistical Reviewer:
`Huanyu (Jade) Chen
`Concurring Reviewers: Kun He, Team Leader
`Rajeshwari Sridhara, Division Director
`
`Division of Oncology Products 2
`Shan Pradhan, M.D., Clinical Reviewer for Efficacy
`Kaushik Shastri, M.D., Clinical Reviewer for Safety
`Steve Lemery, M.D., Team Leader
`Patricia Keegen, M.D., Division Director
`Project Manager:
`Monica Hughes
`
`
`Keywords: stratified log-rank test, interim analysis, K-M curve, Cox regression
`
`
`
`Medical Division:
`Clinical Team:
`
`Reference ID: 3182820
`
`

`

`3
`
`Table of Contents
`1 EXECUTIVE SUMMARY .................................................................................................................................... 5
`2
`INTRODUCTION .................................................................................................................................................. 6
`2.1
`OVERVIEW......................................................................................................................................................... 6
`2.2
`DATA SOURCES ................................................................................................................................................. 6
`STATISTICAL EVALUATION OF STUDY 14387............................................................................................ 7
`3.1
`DATA AND ANALYSIS QUALITY ........................................................................................................................ 7
`3.2
`EVALUATION OF EFFICACY ............................................................................................................................... 7
`3.2.1
`OBJECTIVE..................................................................................................................................................... 7
`3.2.2
`STUDY DESIGN .............................................................................................................................................. 7
`3.2.3
`EFFICACY MEASURES.................................................................................................................................... 9
`3.2.4
`SAMPLE SIZE CONSIDERATIONS .................................................................................................................. 10
`3.2.5
`INTERIM ANALYSIS...................................................................................................................................... 10
`3.2.6
`STATISTICAL METHODOLOGIES................................................................................................................... 11
`3.2.7
`APPLICANT’S RESULTS AND FDA STATISTICAL REVIEWER’S FINDINGS/ COMMENTS................................. 11
`3.2.7.1 PATIENT POPULATION AND DISPOSITION..................................................................................................... 11
`3.2.7.2 BASELINE CHARACTERISTICS...................................................................................................................... 12
`3.2.7.3 PRIMARY EFFICACY ENDPOINT – OS........................................................................................................... 15
`3.2.7.4 KEY SECONDARY ENDPOINT – PFS............................................................................................................. 16
`3.2.7.5 KEY SECONDARY ENDPOINT – ORR ........................................................................................................... 18
`3.3
`EVALUATION OF SAFETY................................................................................................................................. 18
`3.4
`BENEFIT/RISK RATIO....................................................................................................................................... 18
`4 FINDINGS IN SPECIAL/SUBGROUP POPULATIONS ................................................................................ 19
`4.1
`OS SUBGROUP ANALYSIS................................................................................................................................ 19
`SUMMARY AND CONCLUSIONS ................................................................................................................... 20
`5.1
`STATISTICAL ISSUES........................................................................................................................................ 20
`5.2
`COLLECTIVE EVIDENCE................................................................................................................................... 20
`5.3
`CONCLUSIONS AND RECOMMENDATIONS ........................................................................................................ 20
`5.4
`LABELING RECOMMENDATION ........................................................................................................................ 21
`
`5
`
`
`
`
`
`Reference ID: 3182820
`
`2
`
`

`

`
`LIST OF TABLES
`
`
`Table 1 Planned Stopping Criteria and Alpha Spending at the Interim and Final Analyses of OS ................................ 10
`Table 2. Actual Stopping Criteria and Alpha Spending at the Interim and Final Analyses of OS ................................. 10
`Table 3. Patient Population and Disposition (ITT)......................................................................................................... 12
`Table 4. Baseline Demographics Characteristics (ITT).................................................................................................. 12
`Table 5. CRF Stratification Factors and Misclassifications at IVRS.............................................................................. 13
`Table 6 Baseline Disease Characteristics (ITT).............................................................................................................. 14
`Table 7. Prior Anti-Cancer Drugs................................................................................................................................... 15
`Table 8. OS Analyses (ITT)............................................................................................................................................ 15
`Table 9. Sensitivity Analyses for OS.............................................................................................................................. 15
`Table 10. Number of Death by Planned Region............................................................................................................. 16
`Table 11. FDA and Applicant’s PFS Analyses (ITT)..................................................................................................... 17
`Table 12. Sensitivity Analysis for PFS........................................................................................................................... 18
`Table 13. ORR Analyses (ITT) ...................................................................................................................................... 18
`Table 14. OS (Months) Subgroup Analysis.................................................................................................................... 19
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`3
`
`Reference ID: 3182820
`
`

`

`LIST OF FIGURES
`
`Figure 1: Trial Schema ..................................................................................................................................................... 8
`Figure 2. K-M Curves for OS......................................................................................................................................... 16
`Figure 3. K-M Curves for PFS ....................................................................................................................................... 17
`
`
`
`
`Reference ID: 3182820
`
`4
`
`

`

` EXECUTIVE SUMMARY
`
` 1
`
`
`In this New Drug Application (NDA), the applicant is seeking a regular approval of Stivarga®
`(Regorafenib), a novel oral multi kinase inhibitor targeting cancer cells and the tumor micro-
`environment, for the treatment of metastatic colorectal cancer (mCRC) in patients who have been
`previously
`treated with,
` for fluoropyrimidine-based
`chemotherapy, anti-VEGF therapy, and if KRAS wild type, an anti-EGFR therapy.
`
`The pivotal study 14387 (CORRECT) was a randomized, double blinded, placebo-controlled
`multinational phase III trial evaluating the efficacy and safety of regorafenib in combination with
`best supportive care (BSC) relative to placebo in combination with BSC. The primary endpoint
`was overall survival (OS). The key secondary endpoints were progression free survival (PFS),
`objective response rate (ORR), and disease control rate (DCR). A total of 760 patients were
`randomized in a 2:1 allocation (Regorafenib: 505; placebo: 255) in 16 countries and 114 active
`centers (18 US centers).
`
`The data and analyses from the study 14387 demonstrated that the regorafenib and BSC
`combination (REG/BSC) had statistically significant improvements in the OS when compared with
`placebo and BSC combination (PBO/BSC). The stratified log-rank test p-value for OS comparison
`was 0.0102 compared with the allocated alpha of 0.018 at the second interim analysis. The median
`OS was 6.4 (95% CI: 5.8, 7.3) months for the REG/BSC arm and 5.0 (95% CI: 4.4, 5.8) months for
`the PBO/BSC arm. The stratified Cox proportional HR was 0.77 with 95% CI (0.64, 0.93).
`
`Based on the data and analyses from the study 14837, the REG/BSC arm demonstrated a
`statistically significant improvement in the primary endpoint OS, compared with the PBO/BSC
`arm. Whether the data and analyses from the current submission demonstrate an overall favorable
`risk-benefit profile is deferred to the clinical team reviewing this application.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Reference ID: 3182820
`
`5
`
`(b) (4)
`
`

`

`2 INTRODUCTION
`
`Stivarga® (Regorafenib) is a novel oral multi kinase inhibitor which targets cancer cells and the
`tumor micro-environment. In this New Drug Application (NDA), the applicant is seeking to a
`regular approval for the treatment of metastatic colorectal cancer (mCRC) in patients who have
`been previously treated with,
` for fluoropyrimidine-based
`chemotherapy, anti-VEGF therapy, and if KRAS wild type, an anti-EGFR therapy. The pivotal
`study 14387 (CORRECT) was a randomized, double blinded, placebo-controlled multinational
`phase III trial.
`
`2.1 Overview
`
`Colorectal cancer (CRC) is one of the most common cancers worldwide. Standard treatments exist
`for first and second line CRC therapy. However, additional treatments which show a clinical
`benefit for metastatic CRC patients whose disease has progressed need to be developed in order to
`fulfill the unmet medical need in this seriously ill patient population.
`
`According to the applicant’s report, regorafenib (REG) is a novel oral multi kinase inhibitor which
`targets cancer cells and the tumor micro-environment. It inhibits tumor growth, progression and
`metastases by inhibiting the proliferation of tumor cells, the formation of new tumor vasculature
`and stromal signaling. The substance was selected based on its kinase inhibition profile which
`includes angiogenic (vascular endothelial growth factor receptor [VEGFR] 2/3, TIE-2
`[angiopoietin receptor]), stromal (platelet derived growth factor receptor [PDGFR]-ß, fibroblast
`growth factor receptor [FGFR]) and oncogenic (KIT, RET and BRAF) (receptor tyrosine) kina.
`
`Regorafenib in combination with best supportive care (BSC) (REG/BSC) compared with placebo
`in combination with BSC (PBO/BSC) was evaluated in study 14387 for patients with mCRC who
`have been previously treated with, or are not considered candidates, for fluoropyrimidine-based
`chemotherapy, anti-VEGF therapy, and if KRAS wild type, an anti-EGFR therapy. This study was
`a randomized, double blinded, placebo-controlled multinational phase III trial comparing the
`efficacy and safety of REG/BSC therapy.
`
`Study 14387 was conducted at 114 centers within 16 countries. A total of 760 patients were
`randomized in a 2:1 allocation (REG/BSC: 505; PBO/BSC: 255). The randomization was
`centralized and stratified by prior treatment with vascular endothelial growth factor (VEGF)
`targeting drugs, time from diagnosis of metastatic disease, and geographical region. The cut-off
`date for the efficacy analysis was July 21, 2011.
`
`2.2 Data Sources
`
`The electronic submission including protocols, statistical analysis plan, study reports, and analysis
`datasets for the original NDA submission are located on the network with network path:
`\\CDSESUB1\EVSPROD\NDA203085\.
`
`
`
`
`Reference ID: 3182820
`
`6
`
`(b) (4)
`
`

`

`3 STATISTICAL EVALUATION OF STUDY 14387
`
`Part of the text, tables and figures presented in this section are adapted from the Applicant’s
`Clinical Study Report (CSR).
`
`3.1 Data and Analysis Quality
`
`At the original submission, the applicant did not submit SAS programs and adequate
`documentations for data definition. The primary efficacy dataset was in the long format, which
`needed extra data manipulation to conduct efficacy analysis. In addition, some important disease
`characteristics were not included in the derived demographic dataset. Upon this reviewer’s request,
`the applicant resubmitted the adequate documentations, derived datasets, and analysis programs.
`This reviewer was able to duplicate the analysis variable derivation and reproduce applicant’s
`summary statistics. No further data resubmission was requested.
`
`3.2 Evaluation of Efficacy
`
`3.2.1 Objective
`
`The primary objective of 14387 was to evaluate whether patients receiving REG/BSC would have
`clinical benefit of overall survival (OS) more than those receiving the PBO/BSC. The secondary
`objectives were to compare progression free survival (PFS), objective response rate (ORR), and
`disease control rate (DCR) between the two treatment groups.
`
`Reviewer’s Comments:
`As stated by the applicant in the IND communication dated on April 24, 2012, the DCR would not
`be used for the labeling claim. This review focuses on the evaluation of efficacy results on the
`primary endpoint OS and the key secondary endpoints PFS and ORR.
`
`3.2.2 Study Design
`
`Study 14387 was a randomized, double blinded, placebo-controlled multinational phase III trial
`evaluating the efficacy and safety of regorafenib (160 mg QD with 3 weeks on and 1 week off) in
`combination with best supportive care (BSC) relative to placebo in combination with BSC
`(PBO/BSC) in patients with mCRC who have been previously treated with, or are not considered
`candidates, for fluoropyrimidine-based chemotherapy, anti-VEGF therapy, and if KRAS wild type,
`an anti-EGFR therapy. Figure 1 presents the trial schema.
`
`
`
`
`Reference ID: 3182820
`
`7
`
`

`

`Figure 1: Trial Schema
`
`
`
`Note: Adapted from Figure 7-1 in CSR
`
`Approximately 690 patients in approximately 120 centers were planned to be randomized in a 2:1
`ratio (REG/BSC: 460; PBO/BSC: 230) in order to observe 582 OS events. The randomization was
`centralized and stratified by prior treatment with vascular endothelial growth factor (VEGF)
`targeting drugs (yes/no), time from diagnosis of metastatic disease (≥18 months vs. <18 months),
`or geographical region (region 1: North America, Western Europe, Israel and Australia, versus
`region 2: Asia, versus region 3: South America, Turkey and Eastern Europe). Asian was planned to
`randomize no more than 250 patients.
`
`The main inclusion criteria were:
`- Age ≥18 years, ECOG performance status 0–1, life expectancy ≥ 3 months
`- Patients with metastatic CRC (Stage IV)
`- Histological/cytological documentation of adenocarcinoma of colon or rectum
`- Disease progression during or within 3 months after the last administration of approved
`standard therapies or intolerance
`
`
`Patients would be treated until one of the following occurs:
`- Progressive disease (PD), per RECIST criteria V1.1, or clinical progression
`- Death
`- Unacceptable toxicity
`- Disposition including withdraw consent form or physician decision
`- Substantial non-compliance with the protocol
`
`
`Reviewer’s Comments:
`1. There was discordance on the region’s definition by different versions of SAP. Under IND
`75642 SN 309 submission for the final SAP, the applicant stated that the official
`randomization code for region (or any stratification factor) was never modified during the
`course of this study.
`2. The applicant did not open clinical sites in either New Zealand or South Africa, and were
`subsequently deleted from SAP versions 2.0 and 2.1.
`3. None was assigned to South America or Turkey.
`4. All patients received prior VEGF therapy.
`
`
`
`
`Reference ID: 3182820
`
`8
`
`

`

`3.2.3 Efficacy Measures
`
`The primary endpoint was OS, defined as the time from randomization to death. Subjects alive at
`the time of analysis would be censored at the last date known to be alive. After discontinuation, all
`the patients would be followed monthly for survival until death. If a subject was lost to follow up
`and there was no contact after randomization, this subject would be censored at day 1.
`
`One of the key secondary endpoints was PFS per investigator (INV) assessment. The PFS was
`defined as the time from date of randomization to first observed disease progression (radiological
`or clinical) or death due to any cause, if death occurred before progression was documented.
`Patients were counted as death if patient died within 16+1 weeks. The actual tumor scan date was
`used for the calculation of PFS. If a tumor assessment was performed over more than one day, the
`earliest date would be used for the calculation of PFS. Every effort was planned to be made to
`obtain radiologic imaging. In those cases where patients were unable to obtain radiologic
`examinations due to deterioration of medical condition, the clinical PD was reported by the
`investigator. The date of clinical progression was used for the determination of the date of
`progression.
`
`Tumors were planned to be measured at baseline and at 8 week intervals according to RECIST,
`version 1.1 during the active treatment period. Treatment with regorafenib after PD could be
`continued per investigator’s decision. For patients who discontinued study treatment without PD,
`available tumor assessments would be recorded in the CRF until documented PD. Additionally, the
`administration of any anti-cancer drugs in follow-up must be recorded in the CRF.
`
`For patients without progressive disease (PD), PFS was censored:
`- at the date of last actual tumor evaluation
`- at day 1, if
`patients who were alive without any post-baseline tumor assessments or
`
`patients who were alive without neither post-baseline radiological tumor evaluation nor
`
`no clinical progression
`- on the date of the last evaluable scan before 2 consecutive missed or non-evaluable
`assessments. This rule was also applied to patients who died later than 16+1 weeks post
`randomization
`- on the date of the last evaluable tumor assessment, if patient died without PD and occurred
`within the 16+1 weeks of the last evaluable tumor assessment
`- on the date of the last evaluable tumor assessment for patients who discontinue or withdraw
`early without documented PD or death event
`- on the date of the last scan performed or tumor assessment prior to the change of anti-cancer
`therapy In this case, death was considered a PFS event.
`
`
`Reviewer’s Comments:
`Due to potential subjectivity in clinical assessment, the investigator defined clinical PD event is
`not included as PD event for purposes of determining PFS. Progression is defined by the objective
`pathologic or radiological findings. This reviewer considered PFS results excluding all of the
`clinical PD events as FDA’s primary analysis on PFS.
`
`
`
`9
`
`Reference ID: 3182820
`
`

`

`The other key secondary endpoint was ORR per investigator (INV) assessment, defined as the
`percentage of subjects with complete response (CR) or partial response (PR).
`
`3.2.4 Sample Size Considerations
`
`The trial was designed to have 90% power to detect a hazard ratio (HR) of 0.75 with a two-sided
`alpha of 0.05 and 2:1 randomization ratio, assuming a median OS of 4.5 months for the PBO/BSC
`arm and 6 months for the REG/BSC arm. Assuming an accrual rate of 30 patients per month after
`an initial 4 months ramp up period with 3% projected drop off rate, it was estimated that 582 OS
`events were needed for the final OS analysis, which could be expected from a total accrual of 690
`patients. This trial was planned to reach its primary endpoint (PFS) in approximately 32 months.
`
`3.2.5 Interim Analysis
`
`According to SAP, two interim analyses were planned. The first interim OS analysis for futility
`was planned at approximately 174 deaths (30%) at 15.5 months. The second OS interim analysis
`for efficacy and futility was planned at approximately 408 deaths (70%) at 23.5 months. The Lan-
`DeMets alpha spending function with an O’Brien-Fleming type of boundary was used to adjust
`alpha for the 2nd efficacy interim and final analyses. The futility boundaries were calculated
`independently for the interim analyses. Table 2 summarizes the stopping criteria and alpha
`spending for planed interim and final analyses.
`
`Table 1 Planned Stopping Criteria and Alpha Spending at the Interim and Final Analyses of OS
`Stopping Boundaries of HR
`Time
`# Event Efficacy (≥ Lower) Futility (≤ Upper) Nominal Alpha (two-sided)
`1st
`175 (30%)
`
`1.33
`-
`2nd
`408 (70%)
`0.77
`0.91
`0.015
`Final
`582
`0.84
`
`0.045
`
`
`Reviewer’s Comments:
`1. The applicant did 2 interim analyses before the final analysis was conducted. Table 2
`summarizes alpha allocation based on actual conducted analyses. Based on the actual
`number of events, the reviewer used alpha spending value 0.018 for the 2nd interim
`analysis of OS.
`
`
`Table 2. Actual Stopping Criteria and Alpha Spending at the Interim and Final Analyses of OS
`Stopping Boundaries of HR
`Time
`# Event Efficacy (≥ Lower) Futility (≤ Upper) Nominal Alpha (two-sided)
`1st
`301 (52%)
`
`1.01
`-
`2nd
`432 (74%)
`0.79
`0.90
`0.018
`Final
`582
`0.84
`
`0.044
`
`2. One December 23, 2011, the DMC recommended the applicant to stop this study based on
`the 2nd interim analysis results. As of 2nd interim analysis cut off dated on July 21, 2011, the
`efficacy boundary had been crossed.
`
`
`
`
`
`Reference ID: 3182820
`
`10
`
`

`

`3.2.6 Statistical Methodologies
`
`Intent to Treat (ITT) population was defined as all randomized patients. The ITT population was
`the primary analysis population for the efficacy analyses.
`
`Efficacy Analysis Method for OS
`
`The analysis for OS was performed using a stratified log-rank test, stratified by the same
`stratification factors as used for randomization: prior treatment with VEGF targeting drugs
`(yes/no), time from diagnosis of metastatic disease (TFDMD) (≥ 18 months vs. <18 months), and
`geographical region 1 (North America, Western EU, Israel and Australia) versus region 2 (Asia)
`versus region 3 (South America, Turkey and Eastern EU). The median OS and survival curves
`were estimated using the Kaplan-Meier (KM) method. The KM estimates at different time points
`with corresponding 95% confidence intervals as well as the differences of these estimates were
`calculated. The hazard ratio (HR) and 95% confidence interval (CI) of REG/BSC over the
`PBO/BSC were estimated by a stratified Cox regression procedure.
`
`Efficacy Analysis Method for PFS
`
`The PFS analysis method was identical to OS analysis.
`
`Efficacy Analysis Method for ORR
`
`The analysis for ORR was performed using a Cochran-Mantel-Haenszel (CMH) test adjusting for
`the same stratification factors at randomization. ORR estimates and 95% confidence intervals
`would be estimated for each treatment group. The difference of ORR between the REG/BSC and
`PBO/BSC arm and the corresponding 95% confidence intervals would also be calculated.
`
` hierarchical procedure of testing secondary endpoints in the order of PFS and ORR was
`proposed.
`
`Reviewer’s Comments:
`Because all patients received prior VEGF therapy, this reviewer excluded prior VEGF therapy
`from the stratification log-rank test and stratified CMH test.
`
`3.2.7 Applicant’s Results and FDA Statistical Reviewer’s Findings/ Comments
`
`3.2.7.1 Patient Population and Disposition
`
`Study 14387 was conducted at 105 centers in 15 countries worldwide. A total of 760 patients were
`randomized in a 2:1 allocation (REG/BSC: 505; PBO/BSC: 255). Table 3 presents the study
`population and patient disposition.
`
`
` A
`
`
`
`Reference ID: 3182820
`
`11
`
`

`

`Table 3. Patient Population and Disposition (ITT)
`
`REG/BSC
`N
`505
`Never treated
`5 (1%)
`Ongoing
`52 (10%)
`Disposition
`448 (89%)
`Adverse event (non-treatment related)
`42 (8%)
`Adverse event (treatment related)
`43 (9%)
`Progressive disease
`1 (<1%)
`Radiological Progressive disease
`315 (62%)
`Clinical progressive disease
`20 (4%)
`Death
`7 (1%)
`Withdrawal by subject
`16 (3%)
`Physician Decision
`2(<1%)
`Protocol Violation
`2(<1%)
`
`PBO/BSC
`255
`2 (<1%)
`9 (4%)
`244 (96%)
`7(3%)
`23 (9%)
`0
`192 (75%)
`13 (5%)
`4 (2%)
`5 (2%)
`0
`0
`
`
`Reviewer’s Comments:
`1. Five patients (140010006, 200070005, 200090002, 220070006, 280050010) randomized to
`the REG/BSC arm and 2 patients (200080002, 240040006) randomized to the PBO/BSC
`arm did not receive their allocated treatment.
`2. By the time of cut-off date for the 2nd interim analysis, there were approximately 10% and
`4% patients still on study treatment in the REG/BSC arm and the PBO/BSC arm.
`3. The majority of the discontinuations were associated with progressive disease (PD).
`Among them, 507 patients (67%) had PD identified by radiology.
`4. Discontinuations were imbalanced between the REG/BSC and PBO/BSC arms. The
`placebo arm had more PD, and REG/BSC arm had more AE.
`
`12
`
`
`3.2.7.2 Baseline Characteristics
`
`Table 4 presents the patient baseline demographic characteristics.
`
`Table 4. Baseline Demographics Characteristics (ITT)
`REG/BSC PBO/BSC
`
`505
` 255
`N
`60.7 (10.1)
`60.1(10.0)
`Age (yr) Mean (SD)
` Median (min - max) 61 (22-82) 61 (25-85)
` ≥ 65
`196 (39%)
`89 (35%)
`Female
`194 (38%) 102 (40%)
`Race White
`392 (78%) 201 (79%)
` Asian
`76 (15%)
`35 (14%)
` Other
`37 (7%)
`19 (7%)
`US
`47 (9%)
`36 (14%)
`
`Reviewer’s Comments:
`
`
`Reference ID: 3182820
`
`

`

`1. Patients were balanced by race and gender.
`2. There were more patients with age of 65 years or older in the REG/BSC arm.
`3. There were more U.S. patients enrolled in the PBO/BSC arm.
`
`
`Table 5 summarizes the case report form (CRF) stratification factors and misclassifications at
`interactive voice response system (IVRS).
`
`Table 5. CRF Stratification Factors and Misclassifications at IVRS
`REG/BSC
`PBO/BSC
`
`N
`505
`255
`Region 1
`420(83%)
`212 (83%)
`2
`69 (14%)
`35 (14%)
`3
`16 (3%)
`8 (3%)
`2+3
`85 (17%)
`43 (17%)
`Time from first diagnosis of metastatic disease to randomization
`<18 months
`91 (18%)
`49 (19%)
`≥18 months
`414 (82%)
`206 (81%)
`Prior anti-VEGF therapy
`505 (100%)
`255 (100%)
`IVRS Misclassification
`18 (4%)
`15 (6%)
`Region 1: North America, Western Europe, Israel and Australia, versus; Region 2: Asia;
`Region 3: South America, Turkey and Eastern Europe
`
`Reviewer’s Comments:
`1. Overall, the discordance rate between IVRS and CRFs was 4% (33). As part of the
`applicant’s data collection and monitoring, patients were reclassified based on the source
`documents. These reclassified strata were used for statistical analyses in the CSR. In the
`section 3.2.3.6.3, the impact of stratification misclassification is discussed.
`2. Only 3% of patients enrolled in region 3, this reviewer also conducted sensitivity analyses
`which used region 1 vs. 2+3 instead of planned region (1 vs. 2 vs. 3) in the stratification
`log-rank test and stratified CMH test.
`3. Per SAP, this study plan was designed to enroll no more than 250 (37%)) patients in Asia
`(region 2). Due to quick enrollment in region 1, 14% patients were indeed enrolled in Asia.
`
`
`Table 6 summarizes the important baseline disease characteristics in the ITT population.
`
`
`
`
`Reference ID: 3182820
`
`13
`
`

`

`Table 6 Baseline Disease Characteristics (ITT)
`
`
`REG/BSC
`505
`505 (100%)
`265 (52%)
`240 (48%)
`273 (54%)
`205 (41%)
`4 (1%)
`41 (8%)
`22 (4%)
`493 (98%)
`2 (<1%)
`5 (1%)
`323 (64%)
`151 (30%)
`30 (6%)
`
`PBO/BSC
`255
`255 (100%)
`146 (57%)
`109 (43%)
`157 (62%)
`94 (39%)
`2 (1%)
`25 (10%)
`10 (4%)
`245 (96%)
`3 (1%)
`4 (2%)
`172 (68%)
`69 (27%)
`14 (5%)
`
`N
`Stage IV
`ECOG PS 0
` 1
`KRAS mutation Yes
` No
`BRAF Mutation Yes
` No
`Impaired Renal
`Histology Adenocarcinoma
` Adenocarcinoma in situ
` Mucinonous carcinoma
`Primary Site of Disease Colon
` Rectum
` Colon and Rectum
`Time from most recent PD/relapse to randomization
`6.2 (6.5)
`6.5 (5.7 )
` Mean (STD)
`5.0 (0.1-50.0 ) 4.6 (0.3-52.1 )
` Median (min-max)
`505 (100%)
`255 (100%)
`Prior Surgical Therapeutic Procedure
`135 (27%)
`78 (31%)
`Prior Radiotherapy
`302 (60%)
`157 (62%)
`Prior treatment lines >3
`0
`0
`Prior Systemic anti-cancer Therapy 0-1
`82(16%)
`39(15%)
`2
`121(24%)
`59(23%)
`3
`127(25%)
`64(25%)
`4
`76(15%)
`40(16%)
`5
`99(20%)
`53(21%)
` ≥6
`Prior Systemic anti-cancer Therapy on or after diagnosis of metastatic disease
`1
`16 (3%)
`5 (2%)
`2
`119 (24%)
`58 (23%)
`3
`125 (25%)
`72 (28%)
`4
`113 (22%)
`49 (19%)
`5
`60 (12%)
`32 (13%)
` ≥6
`72 (14%)
`39 (15%)
`
`
`Reviewer’s Comments:
`1. In terms of KRAS mutation, there were 8% more patients in the PBO/BSC arm than those
`in the REG/BSC arm.
`2. There were more patients with an ECOG PS of 0 in the PBO/BSC arm than those in the
`REG/BSC arm.
`
`
`Table 7 summarizes the distribution of prior anti-cancer drug.
`
`
`
`
`Reference ID: 3182820
`
`14
`
`

`

`Table 7. Prior Anti-Cancer Drugs
`
`N
`Fluoropyrimidine
`Bevacizumab
`Irinotecan
`Oxaliplatin
`Panitumumab and/or Cetuximab
`KRAS Wide Type
`KRAS Unknown
`KRAS Mutation
`
`REG/BSC
`505
`505 (100%)
`505 (100%)
`505 (100%)
`505 (100%)
`264/505 (52.3%)
`204/205 (99.5%)
`27/27 (100.0%)
`33/273 (12.1%)
`
`PBO/BSC
`255
`255 (100%)
`255 (100%)
`255 (100%)
`255 (100%)
`121/255 (47.5%)
`94/94 (100.0%)
`4/4 (100.0%)
`23/157 (14.6%)
`
`
`Reviewer’s Comments:
`1. All patients received prior treatment with fluoropyrimidine-, oxaliplatin-, and irinotecan-
`based chemotherapy, and with bevacizumab.
`2. Per protocol, patients with KRAS wide type tumor should have got anti-EGFR antibody
`therapy (cetuximab and/or panitumumab). All but one patient with K-Ras mutation-
`negative tumors received panitumumab or cetuximab.
`
`REG/BSC
`505
` 275 (55%)
`6.4
`(5.8, 7.3)
`
`Un-stratified
`HR (95% CI)
`P-value
`0.77 (0.63, 0.93)
`0.0077
`
`0.77 (0.63, 0.93)
`
`0.0077
`
`
`3.2.7.3 Primary Efficacy Endpoint – OS
`
`Table 8 presents the efficacy analysis for OS with a total of 432 (57%) death events. The
`REG/BSC treated patients demonstrated a statistically significant difference in OS compared with
`the PBO/BSC treated patients based on a stratified log-rank test with a p-value 0.0102. The
`median OS was 6.4 months (95% CI: 5.8, 7.3) for the REG/BSC arm and 5.0 months (95% CI: 4.4,
`5.8) for the PBO/BSC arm. The stratified hazard ratio was 0.77 with 95% CI (0.64, 0.94).
`
`Table 8. OS Analyses (ITT)
`PBO/BSC
`
`255
`N
`157 (62%)
`Number of deaths, n (%)
`5.0
`Median Overall Survival (months)
`(4.4, 5.8)
`95% CI
`HR (95% CI) b
`0.77 (0.64, 0.94)
`Stratified Log-Rank Test P-value a,b
`0.0102
`a Stratified by planned stratification factors: geographic