CENTER FOR DRUG EVALUATION AND
`RESEARCH
`
`
`APPLICATION NUMBER:
`
`203085Orig1s000
`
`SUMMARY REVIEW
`
`
`
`
`

`

`Division Director Summary Review
`
`Se tember 20, 2012
`Patlicia Kee- an
`From
`
`Subject
`Division Director Summary Review
`NDA #
`NDA 203085
`
`A licant Name
`
`Date of Submission
`
`PDUFA Goal Date
`
`Proprietary Name /
`Established (USAN) Name
`Dosa _e Forms / Stren_ h
`
`Proposed Indication(s)
`
`Recommended Action for NME:
`
`Ba er Healthcare Pharmaceuticals, Inc.
`
`Auril 27, 2012
`
`October 27, 2012
`
`Stivarga Tablets /
`regorafenib
`Tablets for oral administration/ 40 111
`
`For the treatment of patients with metastatic colorectal
`cancer (mCRC) who have been previously treated with,
`M“) fluoropyrimidine—
`based chemotherapy, an anti-VEGFR therapy, and, if
`KRAS wild ‘0 e, an anti-EGFR thera o
`
`
`
`——0ND Action Packa ' e, includin- :
`
`Names of disci I line reviewers
`
`OND=0flice ofNew Drugs
`0NDQA=Oflice ofNew Drugs Quality Assessment
`OPDP=0flice of Prescription Drug Promotion
`0815: Office of Surveillance and Epidemiology
`DMH’A=Division ofMedication EII‘OI Prevention and Analysis
`OSI=0flice of Scientific Investigations
`DRISK=Division ofRisk Managennmt
`CDTIFCtoss-Discipline Team Leader
`
`NDA 203085
`
`Division Directory Summary Review
`
`Page 1 of 23
`
`Reference ID: 3192218
`
`

`

`Division Director Summary Review
`Introduction
`
`1.
`
`Regorafenib (Stivarga Tablets, Bayer) is a small molecule inhibitor of multiple membrane-
`bound and intracellular kinases (multi-kinase inhibitor) involved in a wide range of normal
`cellular functions and in pathologic processes, such as oncogenesis, tumor angiogenesis, and
`maintenance of the tumor microenvironment. The kinase inhibition profile of regorafenib
`affect the angiogenic (VEGFR 2/3, TIE2), stromal (PDGFR-ß, FGFR) and oncogenic (KIT,
`RET and B-RAF) cellular processes and pathways.
`
`The clinical efficacy and safety of regorafenib were primarily supported by a single clinical
`trial (Protocol 14387; “CORRECT”) enrolled 670 patients with metastatic colorectal cancer
`with disease progression following all FDA-approved therapy. The results of this single trial
`were considered sufficient to serve as the sole trial in support of this NDA since it was a large
`multicenter study with consistency of the treatment effects across study subsets; met both the
`primary endpoint of overall survival as well as one of the key secondary efficacy endpoints,
`progression-free survival, which involves different events; and the effects on survival and
`progression-free survival were statistically very persuasive.
`
`CORRECT was an international, multicenter, randomized (2:1), double-blind, placebo-
`controlled, trial comparing the effect of regorafenib at a dose of 160 mg once daily for 3 weeks
`(days 1-21) of a 28-day cycle plus best supportive care (BSC) (n=505) to matching placebo
`plus BSC (n=255) on overall survival (primary endpoint). Key secondary endpoints were
`progression-free survival, objective response rate, and response duration.
`
`The CORRECT trial demonstrated statistically significant improvements in both overall
`survival and in progression-free survival for regorafenib treatment patients over those
`receiving best supportive care alone, however there was inadequate tumor shrinkage among
`regorafenib-treated patients, as determined by RECIST criteria, to consider this a part of the
`clinical benefit of this drug.
`
`
`NDA 203085
`
`Division Directory Summary Review
`
`Page 2 of 23
`
`Reference ID: 3192218
`
`

`

`Efficacy Outcomes
`Overall Survival
`Number of deaths, n (%)
`Median Overall Survival (months)
`95% CI
`HR (95% CI)
`Stratified Log-Rank Test P-value a,b
`Progression-free Survival
`Number of Death or Progression, n (%)
`Median Progression-free Survival (months)
`95% CI
`HR (95% CI)
`Stratified Log-Rank Test P-value a
`Overall Response Rate
`Overall response, n (%)
`95% CI
`
`Stivarga + BSC
`(N=505)
`
`Placebo + BSC
`(N=255)
`
` 275 (55%)
`6.4
`(5.8, 7.3)
`
`157 (62%)
`5.0
`(4.4, 5.8)
`0.77 (0.64, 0.94)
`0.01
`
`417 (83%)
`2.0
`(1.9, 2.3)
`
`231 (91%)
`1.7
`(1.7, 1.8)
`0.49 (0.42, 0.58)
`<0.0001
`
`5 (1%)
`0.3%, 2.3%
`
`1 (0.4%)
`0%, 2.2%
`
`
`
`The most frequently observed adverse drug reactions (≥30%) in regorafenib-treated patients
`are asthenia/fatigue, decreased appetite and food intake, palmar-plantar erythrodysesthesia
`(hand-foot syndrome), diarrhea, mucositis, weight loss, infection, hypertension and dysphonia.
`The most frequent laboratory abnormalities are cytopenias (anemia, thrombocytopenia, and
`lymphopenia), liver dysfunction (hyperbilirubinemia, transaminitis), and metabolic
`derangements (hypocalcemia, hypophosphatemia, and hypokalemia). The most serious
`adverse drug reactions of regorafenib in the CORRECT trial, occurring at an increased
`incidence in regorafenib-treated patients and placebo-treated patients, respectively, were Grade
`3 palmar-plantar erythrodysesthesia (17% vs. 0), fatal hepatotoxicity (1.6% vs. 0.4%),
`myocardial ischemia and infarction (1.2% vs. 0.4%), and fatal hemorrhage (0.8% vs. 0).
`Across the clinical trials safety database of 1100 patients, serious adverse drug reactions with
`regorafenib were identified at the following rates: gastrointestinal perforation (0.6%), fatal
`drug-induced liver injury (0.3%), hypertensive crisis (0.18%), and reversible posterior
`leukoencephalopathy (0.09%). These adverse drug reaction profile for regorafenib appear to
`be arise primarily from its inhibition of the VEGF pathway (i.e., hypertension, RPLS, cardiac
`ischemia/infarction, hemorrhage, viscus perforation, fistula formation, dysphonia) and of the
`EGFR pathway (rash), although some of the common and serious adverse drug reactions of
`regorafenib are seen in drugs both with and without known kinase inhibition (e.g.,
`hepatotoxicity, asthenia/fatigue, decreased appetite and food intake, palmar-plantar
`erythrodysesthesia, diarrhea, and mucositis) and cannot be attributed to a specific mechanism.
`
`All review disciplines recommended approval. The approval was based on a single, adequate
`and well-controlled trial that showed a highly robust effect on 23% relative reduction in the
`immediate risk of death and 51% relative reduction in the immediate risk of disease
`progression or death. While the absolute magnitude of the treatment effects on survival
`(difference of 1.4 months in median survival times) and progression-free survival (difference
`of 1.2 weeks in median progression-free survival times) are small, the ability of any single
`
`NDA 203085
`
`Division Directory Summary Review
`
`Page 3 of 23
`
`Reference ID: 3192218
`
`

`

`agent to demonstrate efficacy in this heavily pre-treated population represents clinical benefit,
`when considered in the context of serious adverse drug reactions occurring in fewer than 1%
`of patients and common toxicities already considered acceptable with other approved agents
`for the treatment of metastatic colorectal cancer (e.g., palmar-plantar erythrodysesthesia,
`nausea/vomiting, mucositis, diarrhea, and hypertension) and which are generally manageable
`with dose modification and symptomatic treatment.
`
`2. Background
`
`Proposed indication
`
`In 2012, there will be an estimated 103, 170 new cases of colon cancer, 40, 290 new cases of
`rectal cancer,, and an estimated 51, 690 deaths from colon or rectal cancers1 While the
`mortality from colorectal cancer has decreased in the past 50 years, approximately half the
`decline in mortality rates (from 28 deaths per 100,000 to 17 deaths per 100,000) is attributed to
`screening and early diagnosis2. The identification of new systemic treatments for patients with
`metastatic disease has improved short-term outcomes but not long-term cure rates. The
`standard of care in the United States for the treatment of metastatic colorectal cancer includes
`first-line and second-line treatment with fluoropyrimidine-based combination chemotherapy
`(FOLFOX or FOLFIRI) administered with bevacizumab for the majority of patients.
`Cetuximab and panitumumab are indicated for the treatment of patients with metastatic
`colorectal cancer in which the tumor does not contain mutations in the c oncogene (K-Ras
`wild-type), either as an addition to combination chemotherapy for initial treatment (cetuximab)
`or as monotherapy in patients with recurrent, chemotherapy-refractory disease (cetuximab,
`panitumumab). . The very elderly or those with co-morbid conditions which may render
`intensive treatment intolerable, are generally treated either with combinations of approved
`drugs (5-flurouracil and leucovorin, capecitabine, oxaliplatin, irinotecan, with or without anti-
`EGFR directed antibodies) or with single agent therap.
`
`Bayer has requested approval for the proposed indication:
`“For the treatment of patients with metastatic colorectal cancer (mCRC) who have
`been previously treated with,
`, fluoropyrimidine-
`based chemotherapy, an anti-VEGFR therapy, and, if KRAS wild type, an anti-EGFR
`therapy.”
`
`There are no FDA-approved therapies for the proposed indication, which was adequately
`reflected by the patient population enrolled in the primary efficacy trial. Thus this patient
`population represents a disease with a clear, unmet medical need.
`
`Regulatory History of NDA
`
`July 19, 2006: Submission for IND 75642
`
`
`1 http://www.cancer.gov/cancertopics/types/colon-and-rectal
`2 http://www.cancer.gov/cancertopics/factsheet/cancer-advances-in-focus/colorectal
`
`NDA 203085
`
`Division Directory Summary Review
`
`Page 4 of 23
`
`Reference ID: 3192218
`
`(b) (4)
`
`

`

`
`September 3, 2009: End-of-Phase 2 meeting
`Key agreements regarding the proposed registration trial
`• The primary objective should be overall survival; an earlier assessment of the treatment
`effect may be obtained in Phase 2 trial or through an interim analysis for futility in the
`planned Phase 3 trial
`• A single trial could support an NDA if well-conducted & designed, with statistically
`persuasive results so that a second trial is unethical or infeasible
`• Control arm of best supportive care plus placebo was acceptable in patients no longer
`responding to approved therapies or standard combination regimens
`• Sample size is adequate based on assumptions regarding treatment effects
`• The proposed Phase 3 trial, supported by Phase 2 studies in mCRC, would not support
`claims for
`, given the
`number of approved drugs for mCRC and trial design
`• Pharmacokinetic studies, including food effect studies, supporting the NDA should
`evaluate regorafenib and the metabolites M2 and M5
`• Hepatic impairment studies and studies to assess possible drug interactions should also be
`conducted
`
`
`January 22, 2010: Special Protocol Assessment (SPA) No Agreement letter issued for
`Protocol 14387; “CORRECT” trial. Areas of outstanding disagreement or requiring further
`clarification:
`(cid:131)
` of the planned final analysis and
`Inclusion of a futility interim analysis earlier than
`inclusion in the protocol of a single interim analysis for overall survival for efficacy.
`(cid:131) Whether a 1.5 month difference in median overall survival times would be considered
`“clinically significant” is a review issue.
`(cid:131) Clarify the proportion of patients to be enrolled in the United States
`
`April 13, 2010: Based on FDA’s draft responses to the Type A meeting to discuss FDA’s
`January 22, 2010 SPA No Agreement letter, Bayer cancelled the Type A meeting and will
`submit the revised protocol
`
`
`April 13, 2010: Submission of the revised Protocol 14387 under a request for Special Protocol
`Assessment. The request for review under SPA was withdrawn on May 3, 2010.
`
`
`June 10, 2011: Fast Track designation granted for regorafenib for the treatment of patients
`with metastatic colorectal cancer (mCRC) after failure of standard therapies
`
`August 23, 2011: Pre-NDA meeting
`Key agreements regarding the proposed NDA content and format:
`(cid:131) Proposed nonclinical program acceptable to support NDA
`(cid:131) Proposed content/format for ISE and ISS acceptable
`(cid:131) Pooled analysis of efficacy not required (based on one major efficacy trial)
`(cid:131)
`ISS should include serious adverse event information from all regorafenib trials
`(monotherapy and combination therapy in patients with cancer, healthy volunteer trials).
`ISS would contain all data from regorafenib monotherapy studies and used pooled data to
`
`NDA 203085
`
`Division Directory Summary Review
`
`Page 5 of 23
`
`Reference ID: 3192218
`
`(b) (4)
`
`(b) (4)
`
`

`

`assess for risks in subgroups (e.g., by age, gender, organ dysflmction). Case report forms
`required for all serious adverse events, miless event is a manifestation of disease
`progression.
`I Datasets to be provided in CDISC, SD'IM
`I Data cut-off between clinical data an in 120—day safety update should not exceed 6 months
`from date of 120-day safety update submission.
`I Bayer will make “best efforts” to include QTc study report in the initial NDA submission
`I Bayer encouraged to include “biomarker report” in the initial NDA submission.
`
`(m4)
`
`I Proposed rolling submission scheduled for NDA is acceptable.
`
`April 27, 2012: NDA 203085 submitted.
`I Approval requested under 505 (b)(1) of the Federal Food. Drug and Cosmetic Act for “the
`treatment of patients with metastatic colorectal cancer (CRC) who have been previously
`treated with.
`(ma), fluoropyrimidine-based chemotherapy.
`an anti-VEGF therapy. and, if KRAS wild type. an anti-EGFR therapy.
`I Priority review requested based on results of study 14387 which demonstrate that
`Regorafenib is the first agent to demonstrate a statistically significant improvement in
`overall survival (OS) for patients that have previously been treated with all approved
`therapies for metastatic colorectal cancer (including fluoropyrimidine-based chemotherapy,
`an anti-VEGF therapy. and. if KRAS wild type, an anti—EGFR therapy) while presenting a
`manageable adverse event profile. The survival advantage is clinically meaningful and
`establishes Regorafenib as the only treatment for these patients. If approved. Regorafenib
`has the potential to provide a safe and effective treatment in a disease where no satisfactory
`alternate therapy exits.
`I Request for Waiver of Pediatric Studies submitted to waive the requirement to assess the
`safety and effectiveness of the drug product in children 16 years of age and below in
`accordance with 21 CFR 314.55. We request the waiver on the basis that studies are
`impossible or highly impractical, as the number of pediatric patients is extremely small and
`the proposed indication, colorectal cancer, qualifies for a disease-specific waiver.
`
`June 25, 2012: Filing letter issued, containing notification of priority review designation and
`deficiencies identified.
`
`3.
`
`CMC
`
`I concur with the conclusions reached by the chemistry reviewers regarding the acceptability
`of the manufacturing of the drug product and drug substance. Manufacturing site inspections
`were acceptable. Stability testing supports an expiry of 36 months. There are no outstanding
`issues that preclude approval.
`
`M")
`Regorafenib monohydrate is the drug substance.
`thus regorafenib is the active drug substance
`
`NDA 203085
`
`Division Directory Summary Review
`
`Page 6 of 23
`
`Reference ID: 3192218
`
`

`

`present in the regorafenib film-coated tablets. The final drug product, Stivarga tablets for oral
`administration is formulated as light pink oval shaped tablets debossed with "BAYER" on one
`side and "40" on the other. Each tablet contains 40 mg of regorafenib. Stivarga tablets are
`supplied in packages containing three bottles, with each bottle containing 28 tablets, for a total
`of 84 tablets per package. The drug product should be at room temperature in the original
`bottle containing a desiccant.
`4. Nonclinical Pharmacology/Toxicology
`
` I
`
` concur with the conclusions reached by the pharmacology/toxicology reviewers that there are
`no outstanding pharmacology/toxicology issues that preclude approval.
`
`The pharmacology/toxicology reviewers stated that the NDA adequate non-clinical
`information to support the NDA for the proposed intended use. While the application did not
`contain carcinogenicity studies or a complete battery of reproductive toxicology studies, these
`are not required for products to be indicated for the treatment of advanced, incurable cancers.
`Similarly, the finding of potential mutagenic effects for a major metabolite (M2) did not
`require a specific Warning based on the indicated population.
`
`The NDA contained the reports for nonclinical primary pharmacology studies confirming the
`claimed effects of regorafenib and its two major metabolites (M2 and M5) on kinase
`inhibition, via biochemical assays or cellular assays, examining the phosphorylation of
`downstream targets, to establish kinase inhibition at clinically achievable exposures in humans
`at the recommended dose for multiple kinase targets (see product labeling). Both the M-2 and
`M-5 metabolites showed inhibitory activity equal to or greater than the activity of the
`regorafenib. In addition, in vivo evaluation of anti-angiogenic effects were evaluated in rats
`and mice,
`
`The application also contained reports of repeat dose toxicology studies in rodents and dogs.
`Toxicologic findings demonstrated both rats and dogs which were also observed in patients
`with cancer involved the gastrointestinal tract (vomiting, diarrhea, decreased motility),
`hematopoietic/lymphoid system (marrow hypocellularity, neutropenia, thrombocytopenia, and
`lymphopenia), atrophy of lymphoid organs), the reproductive system (atrophy), hepatic
`enzyme elevation with histopathologic changes in the liver, cutaneous toxicity (dyskeratosis,
`hyperkeratosis, acanthosis, dermatitis, and alopecia), and skeletal system.
`
`Findings identified in animals that have not been confirmed in clinical trials of adults with
`cancer include renal toxicity (glomerulopathy, tubular degeneration/regeneration, tubular
`dilation, and interstitial fibrosis), skeletal changes (changes in dentin and epiphyseal growth
`plates), reproductive toxicity (increased necrotic corpus lutea and atrophy in the ovaries in
`females and decreased weight of the testes, prostate, and seminal vesicles and retarded
`maturation of the testes along with aspermia/oligospermia in the epididymides in males),
`histopathologic changes in the adrenal glands, and hypothyroidism.
`
` A
`
` report of a safety pharmacology study was also submitted to the NDA. This study did not
`identify significant cardiotoxicity.
`
`NDA 203085
`
`Division Directory Summary Review
`
`Page 7 of 23
`
`Reference ID: 3192218
`
`

`

`
`Embryofetal studies were conducted in Wistar rats and Himalayan rabbits demonstrated
`increased post-implantation loss and teratogenic effects including skeletal and cardiovascular
`malformations and renal findings of dilation of the renal pelvis or hydronephrosis at exposures
`significantly lower than the human exposure at the recommended daily dose. Based on these
`findings, and consistent with current practices in the Division of Hematology-Oncology
`Toxicology, Pregnancy category D was recommended.
`
` A
`
` distribution study in pregnant rats documented regorafenib exposure in the fetus, with
`greater regorafenib concentrations in fetal adrenal glands and brain as compared to the
`maternal blood and increased concentrations of regorafenib or its active metabolites in
`maternal mammary fluid as compared to the blood. Based on these studies, labeling directs
`lactating mothers to discontinue nursing while taking regorafenib.
`
`The pharmacology/toxicology and maternal health team agreed that, based on embryofetal and
`teratogenic effects observed in general toxicology studies, in which female rats administered
`regorafenib at dose levels resulting in exposures similar to those observed in humans at the
`clinically recommended dose, product labeling should indicate the potential risks of impaired
`fertility in both men and women. Dr. Helms noted that these animals were not followed for a
`sufficient period to determine reversibility (persistent findings noted at the 4-week recovery
`period without additional follow-up). Again, given the indicated population, the findings and
`limitations of the findings (i.e., based on animal data) will be conveyed in product labeling.
`
`5.
`
` Clinical Pharmacology
`
` I
`
` concur with the conclusions reached by the clinical pharmacology/biopharmaceutics reviewer
`that there are no outstanding clinical pharmacology issues that preclude approval.
`
`The NDA contained clinical pharmacology data from two dose-finding trials, evaluating
`continuous dosing and a three-week on/one-week off dosing schedule, three drug interaction
`studies, one food effect study and one bioequivalence trial comparing the pharmacokinetic of
`the tablet form used in the major efficacy trial with that of the “to-be-marketed tablet. FDA
`accepted the NDA for filing prior to the completion of additional clinical pharmacology trials
`and expected analyses of exposure-response and population pharmacokinetics because of the
`findings of improved survival in the efficacy trial for a population with an unmet medical
`need.
`
`Following oral administration, regorafenib undergoes enterohepatic circulation. It is highly
`protein bound (99.5%), as are the two major metabolites (M2 and M5) of regorafenib, both of
`which are clinically active. Regorafenib is primarily metabolized by CYP3A4 and UGT1A9
`and about 71% of a single radiolabeled dose (24% as metabolites) was excreted in feces. The
`mean elimination half-lives of regorafenib, M2, and M5 are 28 hours, 25 hrs and 51 hrs,
`respectively. Hepatic elimination appears to be the major route of elimination for regorafenib.
`
`
`NDA 203085
`
`Division Directory Summary Review
`
`Page 8 of 23
`
`Reference ID: 3192218
`
`

`

`The bioavailability of regorafenib and its active metabolites are affected by the presence of
`food (fasted vs. fed) and the fat content (low vs. high-fat meal) when regorafenib is taken with
`food. Since the major efficacy trial which provides substantial evidence of effectiveness of
`regorafenib was performed with the direction to take regorafenib following a low-fat meal, and
`in light of the food-effects, product labeling recommends that regorafenib be administered
`following a low-fat meal.
`
`Pharmacokinetic data obtained in patients with mild renal impairment (n=10) or mild, Child-
`Pugh A (n=4) or moderate/ Child-Pugh B (n=10) hepatic impairment do not suggest altered
`clearance requiring dose adjustments. However, Bayer will be required to conduct trials
`assessing pharmacokinetics in patients with severe renal impairment and severe hepatic
`impairment.
`
`Pharmacokinetic studies were conducted to evaluate for interactions between regorafenib and
`irinotecan, between regorafenib and 5-fluorouracil, and between regorafenib and oxaliplatin.
`There was no evidence of a pharmacokinetic interaction with fluoropyrimidines. Regorafenib
`and its metabolites inhibited UGT1A9 and inhibited UGT1A1 in vitro; exposure to irinotecan
`and its major active metabolite, SN-38, were increased by 28% and 44%, respectively when
`irinotecan was administered following regorafenib. Exposure to oxaliplatin was increased by
`39% when oxaliplatin was administered following regorafenib. The mechanism for this
`apparent interaction is unknown. Because regorafenib is indicated for use as a single agent,
`these interactions are not included in product labeling.
`
`Additional pharmacokinetic trials demonstrated interactions between regorafenib and
`ketoconazole and between regorafenib and rifampin. Administration of ketoconazole increased
`the exposure of regorafenib by 33% and decreased the mean AUC of M2 and M5 each by
`93%. Administration of rifampin decreased exposure of regorafenib by 50%, increased
`exposure of M5 by 264%, and had no apparent effect on exposure of M2. This data is
`described in product labeling based on the potential for co-administration with regorafenib of
`drugs that are strong inhibitors or strong inducers the CYP3A4 enzyme.
`
`Regorafenib or its active metabolites M2 or M5 inhibited CY2B6, CYP2C8, CYP2C9,
`CYP2C19, CYP2D6, or CYP3A4 in vitro. The effect of regorafenib on the PK of CYP2C8,
`CYP2C9, and CYP3A4 substrates are under evaluation in an ongoing study. Regorafenib did
`not induce cytochrome P450 activity in vitro.
`
`PMRs and PMCs
`All post-marketing requirements and commitments were focused on ensuring adequate
`characterization in the pharmacokinetics of regorafenib to ensure safe dosing
`recommendations based on food effects, drug interactions, organ dysfunction, and
`demographics (e.g., age, gender, race).
`
`The required post-marketing trials under 505(o) and the agreed-upon post-marketing
`commitments requested by the Clinical Pharmacology review staff are summarized in section
`13, of this review.
`
`
`NDA 203085
`
`Division Directory Summary Review
`
`Page 9 of 23
`
`Reference ID: 3192218
`
`

`

`6. Clinical Microbiology
`
`Not applicable for dosage form (oral tablet)
`7. Clinical/Statistical-Efficacy
`
`Protocol 14387, titled “A randomized, double-blind, placebo-controlled phase III study of
`regorafenib plus best supportive care (BSC) versus placebo plus BSC in patients with
`metastatic colorectal cancer (CRC) who have progressed after standard therapy” provides the
`data supporting this NDA. Additional clinical trials data included in the application serve to
`further characterize the adverse drug reaction profile and, for the dose-finding trials, to
`establish the tolerability of the proposed dose and schedule, 160 mg orally, once daily for 21
`days of each 28 day cycle (3-weeks on/1-week off). This schedule was selected over a
`continuous daily dosing schedule based on higher regorafenib exposure and a perception of
`higher activity (higher disease control rate).
`
`Protocol 14387 was a randomized (2:1), double-blind, placebo-controlled trial. 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).
`
`Key inclusion criteria were age 18 years or older, ECOG performance status of 0 or 1,
`metastatic adenocarcinoma of colon or rectum with disease progression during or within 3
`months after the last administration of an FDA-approved drug(s) for colorectal cancer or
`intolerance to such drugs.
`
`Patients were randomized to regorafenib 160 mg or matching placebo, administered orally,
`once daily on days 1-21 of each 28-day treatment cycle. Study drug administration continued
`until objective disease progression (per RECIST), clinical progression, unacceptable toxicity,
`or death. Treatment could also be terminated for withdrawal of patient consent, physician
`decision or non-compliance with the protocol.
`
`The planned sample size of 690 patients was designed to detect a hazard ratio (HR) of 0.75 for
`overall survival after 582 deaths, with a two-sided alpha of 0.05 and 90% power, given the 2:1
`randomization ratio. This was based on the assumed median overall survival times of 6
`months and 4.5 months for the regorafenib- and placebo-treated arms, respectively. Two
`formal interim analyses were planned for overall survival; the first interim analysis would be
`conducted for “futility” after approximately 174 deaths (30% of the planned 582 deaths for the
`final analysis), while the second interim analysis would be conducted for both futility and to
`terminate the trial early for efficacy, at approximately 408 deaths (70% of the planned 582
`deaths for the final analysis). The type 1 error rate was preserved through adjustment for
`multiplicity based on the O'Brien-Fleming-type error spending function.
`
`
`NDA 203085
`
`Division Directory Summary Review
`
`Page 10 of 23
`
`Reference ID: 3192218
`
`

`

`Results
`
`Protocol 14387 enrolled 780 patients at 105 clinical sites across 15 countries; there were 505
`patients randomized to regorafenib and 255 patients randomized to placebo, which constitutes
`the intent-to-treat population for the primary and key secondary efficacy analysis. The first
`patient was enrolled on April 30, 2010. The data cut-off date for efficacy analyses was July
`21, 2011. Baseline demographic and prognostic information (abstracted from the statistical
`review) are presented in the following table:
`
`
`NDA 203085
`
`Division Directory Summary Review
`
`Page 11 of 23
`
`Reference ID: 3192218
`
`

`

`Baseline Demographics and Disease Characteristics
`by Treatment Arm CORRECT Trial
`
`.
`.
`.
`.
`Demographic or Disease Characteristic
`Age (years)
`Median
`
`> 65 years
`Gender
`Female
`
`Regorafenib
`(11:505)
`
`61
`
`196 39%
`
`194 38%
`
`392 (78%)
`76 (15%)
`
`420 (83%)
`47 (9%)
`69 (14%)
`
`Placebo
`(n=255)
`
`60
`
`89 35%
`
`102 40%
`
`201 (79%)
`35 (14%)
`
`212 (83%)
`36 (14%)
`35 (14%)
`
`Time from 1 diagnosis of metastatic disease to
`randomization
`
`< 18 months
`> 18 months
`
`91 (18%)
`
`49 (19%)
`
`—__505 (100%)
`
`255 (100%)
`
`ECOG PS
`
`K—Ras tumor status
`
`wild-type
`mutant
`
`265 (52%)
`240 48%
`
`273 (54%)
`205 41%
`
`146 (57%)
`109 43%
`
`157 (62%)
`94 39%
`
`
`
`Primary site
`172 (68%)
`323 (64%)
`Colon
`694(27%)
`151 (30%)
`Rectum
`Colon and Rectum
`——_505 (100%)
`——_135 (27%)
`PrioYrLines ofSystemic Anti-Cancer Therapy for
`
`255 (100%)
`
`78 (31%)
`
`Metastatic Disease
`
`0-1
`
`Prior Treatment with
`
`Fluoropyrimidine
`Oxaliplatin
`Irinotecan
`Panitumumab or cetuximab
`
`K—Ras wild-type
`K-Ras status - unknown
`K-Ras mutant
`
`16 (3%)
`119 (24%)
`125 (25%)
`113 (22%)
`60 (12%)
`
`505 (100%)
`505 (100%)
`505 (100%)
`
`204/205 (99.5%)
`27/27 (100%)
`33/273 (12%)
`
`5 (2%)
`58 (23%)
`72 (28%)
`49 (19%)
`32 (13%)
`
`255 (100%)
`255 (100%)
`255 (100%)
`
`94/94 (100%)
`4/4 (100%)
`23/157 (15%)
`
`NDA 203085
`
`Division Directory Summary Review
`
`Page 12 of 23
`
`Reference ID: 3192218
`
`

`

`The trial was terminated at the frrst interim analysis of overall survival for efficacy, after 432
`deaths (74% of the planned 582 deaths for the final analysis). Based on the O'Brien-Fleming—
`type error spending function, the boundary was considered to have been crossed (< p=0.018) at
`the interim analysis which yielded a hazard ratio of 0.77, p=0.0102 stratified log-rank test.
`Efficacy was tested for the secondary endpoints of progression-free survival, which was also
`statistically significant, and for overall response rate, which was not significantly different
`between arms. The results of the key efficacy analyses are summarized in the following table,
`abstracted from the statistical review.
`
`Efficacy Results from the CORRECT Trial
`
`_
`
`Regorafenib
`
`Placebo
`
`Overall Survival
`
`
`
`Stratified Lo ' -Rank Test -va1ue
`
`Pro 1 ression-free Survival
`
`Overall Res t onse Rate
`
`Overall res 0 use, 11 %
`
`l 0.4%
`
`95% CI
`
`0.3%, 2.3%
`
`0%, 2.2%
`
`In exploratory subset analyses based on demographic parameters (age, gender, race) and on
`prognostic factors (ECOG performance status, K-Ras mutation status, time from metastatic
`cancer diagnosis of less than 18 months or 18 months or longer, nmnber of prior lines of
`chemotherapy), there was consistent evidence of a treatment effect favoring the regorafenib
`arm for overall survival.
`
`Kaplan—Meier curves for the two treatment arms for overall survival and for progression-free
`survival, respectively, abstracted from the statistical review, are presented below.
`
`NDA 203085
`
`Division Directory Summary Review
`
`Page 13 of 23
`
`Reference ID: 3192218
`
`

`

`Kaplan-Meier Curves for Overall Survival, by Treatment Arm, for the CORRECT trial
`
`Kaplan-Meier Curves for Progression-Free Survival, by Treatment Arm, for the
`CORRECT trial
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`NDA 203085
`
`Division Directory Summary Review
`
`Page 14 of 23
`
`Reference ID: 3192218
`
`

`

`The results described above, demonstrate a statistically persuasive and clinically meaningful
`increase in overall survival in patients for whom there is no FDA-approved treatment. The
`effects were supported by consistent trends in improved overall survival in relevant patient
`subgroups and evidence of a significant improvement in progression-free survival.
`
`8. Safety
`
`Safety evaluation across multiple trials
`
`There was adequate data in the application to assess the risks of regorafenib treatment. The
`evaluation of safety in this application was supported primarily by data from the CORRECT
`trial in which 500 patients with mCRC received regorafenib and safety was compared with the
`253 patients with mCRC who received placebo