`
`CENTER FOR DRUG EVALUATION AND
`
`RESEARCH
`
`APPLICATION NUMBER:
`
`21-660
`
`STATISTICAL REVIEWgS!
`
`

`

`
`
`
`US. Department of Health and Human Services
`Food and Drug Adminisuation
`Center for Drug Evaluation and Research
`Office of Pharmacoepidemiology and Statistical Science
`Office of Biostaxistics
`
`STATISTICAL REVIEW AND EVALUATION
`
`CLINICAL STUDIES
`
`NBA/Serial Number:
`
`21-660 I N000
`
`Drug Name:
`
`Indication(s):
`
`Abraxanem (ABI-OOT)
`
`[unit strength = 100 mg/50 mL vial]
`
`Metastatic Breast Cancer
`
`Applicant:
`
`American BioScience, Inc.
`
`Date(s):
`
`Submission Date: March 8, 2004
`
`PDUFA Date: January 8, 2005
`
`Review Priority:
`
`Standard.
`
`Biometrics Division:
`
`Division of Biometrics I (HFD-710)
`
`Statistical Reviewer:
`
`Peiling Yang, PhD.
`
`Concurring Reviewers:
`
`Team Leader: Rajeshwari Sridhara, PhD.
`
`Acting Division Director: Kooros Mahjoob, PhD.
`
`Medical Division:
`
`Clinical Team:
`
`Division of Oncology Drug Product (HFD- 150)
`
`Reviewer: Nancy Scher, MD.
`
`Team Leader: Ramzi Dagher, MD.
`
`Project Manager:
`
`Ms. Sheila Ryan
`
`Keywords: Superiority.
`
`

`

`TABLE of CONTENTS
`
`
`1.1
`1 2
`1 3
`
`CONCLUSIONS AND RECOMMENDATIONS ...................................................................................... 1
`BRIEF OVERVIEW OF CLINICAL STUDIES ....................................................................................... I
`STATISTICAL ISSUES ANDFINDINGS
`l
`
`2
`
`INTRODUCTION ............................................................................................................................... 4
`
`2.1
`2.2
`
`OVERVIEW ..................................................................................................................................... 4
`DATA SOURCES ............................................................................................................................. 6
`
`3
`
`STATISTICAL EVALUATION ...................................................................................................... 7
`
`3 .1
`
`EVALUATION OF EFFICACY ............................................................................................................ 7
`3.1.1
`Study Design ............................................................................................................................. 7
`
`3.1.2
`Efl'icacy Endpoints.................................................................... 3
`
`3.1.3 Hypothesis Test and Sample Size Considerations ........................................................... 9
`3.1.4
`Statistical Anaiysis Plan ........................................................................................................... 9
`
`Sponsor '5 Results and Reviewer's Comments
`.................. I2
`3.1.5
`Patient Disposition ..................................................
`...... 12
`3.1.5.1
`
`
`...... 13
`Analysis Population ..
`3.1.5.2
`
`.......... 14
`3.1.5.3
`Demographics ..................................
`
`..
`...... 15
`3.1.5.4
`Other Baseline Characteristics.
`Primary Endpoint: Reconciled TargetLesionResponseRain (reeTLRR) ................. 17
`3.1.5.5
`
`Secondary Endpoint: Investigator-Assessed Overall Response Rate (invDRR)....
`315.6
`21
`
`.....
`...
`3.1.5.7
`Secondary Endpoint. Duration of Response................
`23
`Secondary Endpoint: Time to Disease Progression.
`...... 28
`3.1.5.8
`
`.......... 32
`Secondary Endpoint: Overall Survival .......................
`3.1.5.9
`EVALUATION OF SAFETY ............................................................................................................. 33
`
`3.2
`
`4
`
`FINDINGS IN SPECIALJSUBGROUP POPULATIONS ............................................................ 34
`
`4.1
`
`GENDER, RACE AND AGE,............................................................................................................. 34
`
`4.1.! Gender .................................................................................................................................... 34
`4.1.2 Race .................................................................................................................................. 34
`4. 1. 3 Age.......................................................................................................................................... 35
`OTHER SPECIAUSUBGROUP POPULATIONS .................................................................................. 36
`4.2
`
`Line ofTherapy............................................................................................................. 36
`4.2. 1
`4. 2. 2
`Prior Anthracyline Therapy.................................................................................................... 37
`4.23
`Patient Populationin Taro! Indication .................................................................................. 38
`4. 2. 4 Country................................................................................................................................... 39
`
`5
`
`SUMMARY AND CONCLUSIONS
`
`....................................................................... 40
`
`5.1
`5.2
`
`STATIST1CAL ISSUES AND COLLECTIVE EVIDENCE ...................................................................... 40
`CONCLUSIONS AND RECOMMENDATIONS .................................................................................... 41
`
`
`NDA21660/N000Abraxanc{A31007)
`” 7
`'
`7
`__
`"
`i
`
`
`
`

`

`
`
`LIST OF TABLES
`
`TABLE 1 : REVIEWER'S RESULTS OF FDA-CONFIRMED RECTLRR (ALL RANDOMIzED PATIENTS) .................. 2
`TABLE 2: FACTORS 0F POTENTIAL INFLUENCE ON RESPONSE ....................................................................... 1 1
`TABLE 3: SPONSOR'S SUMMARY OF PATIENT DISPOSITION ........................................................................... 12
`TABLE 4: SPONSOR'S SUMMARY OF PATIENT POPULATIONS FOR ANALYSIS .........................................
`13
`
`TABLE 5: SPONSOR'S RESULTS OF PATIENT DEMOGRAPHICS (SPONSOR-DEFINED ITT POPULATION)........... 14
`TABLE 6: SPONSOR‘S RESULTS OF INITIAL DIAGNOSIS OF BREAST CANCER (SPONSOR-DEFINED ITI‘
`POPULATION) ....................................................................................................................................... 15
`TABLE 7: SPONSOR'S RESULTS OF PRIOR THERAPIES AT BASELINE (ON SPONSOR’S ITT POPULATION) ........ 16
`TABLE 8: SPONSOR‘S RESULTS OF HISTORY OF PRIOR MEIASTATIC TREATMENTS AT BASELINE (ON
`SPONSOR’S I'IT POPULATION) .............................................................................................................. 16
`TABLE 9: SPONSOR'S RESULTS OF RECONCILED TARGET LESION RESPONSE BY RESPONSE TYPE ( SPONSOR—
`DEPINED ITT POPULATION) .................................................................................................................. 17
`
`TABLE 10: SPONSOR'S RESULTS OF RECTLRR (SPONSOR—DEFINED ITT POPULATION)...
`.................. 18
`TABLE 1 1: SPONSOR'S RESULTS OF RECTLRR IN PATIENTS RECEIVING Ifi-LINE THERAPY (SPONSOR-
`DEFINED ITT POPULATION) ................................................................................................................. 18
`TABLE 12: SPONSOR'S RESULTS OF RECTLRR IN PATIENTS WHO MET TAXOL INDICATION (SPONSOR-
`DEPINED ITT POPULATION) .................................................................................................................. 19
`TABLE 13: DISCREPANCY IN RESPONSE STATUS BETWEEN FDA AND SPONSOR ........................................... 19
`TABLE 14: REVIEWER'S RESULTS OF FDA-CONFIRMED RECTLRR (ALL RANDOMIZED PATIENrs) .............. 20
`TABLE 15: REVIEWER'S RESULTS OF FDA-CONFIRMED RECTLRR IN PATIENTS RECEIVING lg-LINE
`THERAPY (ALL RANDOMIEED PATIENTS) ............................................................................................. 20
`TABLE 16: REVIEWER'S RESULTS OF FDA-CONFIRMED RECTLRR IN PATIENTS WHO MET TAXOL
`INDICATION (ALL RANDOMIZBD PATIENTS) ......................................................................................... 21
`TABLE 17: SPONSOR'S COMPARISON OF ”WORK AND REcTLRR.................................................................. 21
`TABLE 18: SPONSOR'S RESULTS OF INVESTIGATOR-ASSESSED OVERALL RESPONSE BY RESPONSE TYPE (ON
`SPONSOR-DEFINED ITT POPULATION) ................................................................................................. 22
`TABLE 19: SPONSOR'S RESULTS OF INVORR (SPONSOR-DEFINED ITT POPULATION) ................................... 22
`TABLE 20: SPONSOR'S RESULTS OF INVORR IN PATIENTS RECEIVING 1“—LINE THERAPY (S PONSOR—DEPINED
`ITT POPULATION) ............................................................................................................................... 23
`TABLE 21: REVIEWER'S RESULTS OF [NI/ORR (ALL RANDOMIZED PATIENTS)...
`..23
`TABLE 22: SPONSOR'S RESULTS OF DURATION OF RESPONSE ON RESPONDERS WHOHAD RECONCILED
`TARGET LESION RESPONSE .................................................................................................................. 24
`TABLE 23: SPONSOR'S RESULTS OF DURATION OF RESPONSE ON RESPONDERS WHO HAD OVERALL
`RESPONSE BASED ON INVESTIGATOR RESPONSE ASSESSMENT DATASET INCLUDING CYCLES > 6....... 25
`TABLE 24: REVIEWER'S RESULTS DURATION OF RESPONSE ON FDA-CONFIRMED RESPONDERS (BASED ON
`RECONCILED ASSESSMENT THROUGH CYCLE 6) ................................................................................... 25
`TABLE 25: REVIEWER'S RESULTS OP DURATION OF RESPONSE ON FDA-CONFIRMED RESPONDERS (BASED ON
`INVESTIGATOR‘S ASSESSMENT INCLUDING CYCLES BEYOND 6 OR WORLDCARE ASSESSMENT THROUGH
`CYCLE 6) .............................................................................................................................................. 27
`TABLE 26: SPONSOR'S RESULTS OF TIME TO DISEASE PROGRESSION ........................................................... 29
`TABLE 27: REVIEWER‘S RESULTS OF TIME TO DISEASE PROGRESSION (ALL RANDOMIZED PATIENTS)......... 30
`TABLE 28: SPONSOR'S RESULTS OF OVERALL SURVIVAL ON SPONSOR-DEFINED ITT POPULATION ............. 32
`TABLE 29: REVIEWER'S RESULTS OP OVERALL SURVIVAL (ALL RANDOMIZED PATIENTS) ........................... 33
`TABLE 30: SUBGROUP ANALYSIS BY RACE: REVIEWBR'S EXPLORATORY RESULTS OF FDA-CONFIRMED
`RECTLRR (ALL RANDOMIZED PATIENTS) ............................................................................................ 34
`TABLE 31: SUBGROUP ANALYSIS BY AGE: REVIEWER'S EXPLORATORY RESULTS OF FDA-CONFIRMED
`REcTLRR (ALL RANDOMIZEDPATIENTS) 35
`TABLE 32: SUBGROUP ANALYSIS BY LINE OF THERAPY: REVIEWER'S EXPLORATORY RESULTS OF FDA-
`CONPIRMED RECTLRR (ALL RANDOMIEED PATIENTS) ........................................................................ 36
`TABLE 33: SUBGROUP ANALYSIS BY PRIOR ANTHRACYCLINE THERAPY: REVIEWER'S EXPLORATORY
`RESULTS OF FDA-CONFIRMED RECTLRR (ALL RANDOMIEEB PATIENrs) ........................................... 37
`
`
`
`W N
`
`DA 21-660 IN 000 Abraxane (ABI—OO?)
`
`II
`
`

`

`
`
`TABLE 34; SUBGROUP ANALYSIS BY TAXOL INDICATION: REVIEWER'S EXPLORATORY RESULTS OF FDA-
`CONFIRMED REc’I‘LRR (ALL RANDOMIZED PATIENTS) ........................................................................ 38
`TABLE 35: SUEGROUP ANALYSIS BY COUNTRY: REVIEWER’S EXPLORATORY RESULTS 0? FDA-CONFIRMED
`RECTLRR (ALL RANDOMIZED PATIENTS) ............................................................................................ 39
`
`APPEARS THIS WAY
`0N om e 1 NA].
`
`
`
`"
`
`'
`
`" '
`
`ni
`
`
`NDA21—660 AbrWABI)
`
`"
`
`"
`
`
`
`

`

`
`
`LIST of FIGURES
`
`FIGURE 1: REVIEWER'S KAPLAN-MEIER CURVES OF DURATION OF RESPONSE FOR FDA-CONFIRMED
`RESPONDERS (BASED ON RECONCILED ASSESSMENT THROUGH CYCLE 6) ........................................... 26
`FIGURE 2: REVIEWER'S KAPLAN-MEIER CURVES OF DURATION OF RESPONSE FOR FDA-CONFIRMED
`RESPONDERS (BASED ON INVESTIGATOR‘S ASSESSMENT INCLUDING CYCLES BEYOND 6 OR
`WORLDCARE ASSESSMENT THROUGH CYCLE 6) .................................................................................. 23
`FIGURE 3: REVIEWER'S KAPLAN-MEIER CURVES OF TIME To DISEASE PROGRESSION (INVESTIGATOR‘S
`ASSESSMENT INCLUDING CYCLE > 6; ALL RANDOMIZED PATIENTS) ................................................... 3 1
`FIGURE 4: REVIEWER‘S KAPLAN-MEIER CURVES OF TIME TO DISEASE PROGRESSION (RECONCILED
`ASSESSMENT THROUGH CYCLE 6; ALL RANDOMIZED PATIENTS)......................................................... 31
`FIGURE 5: REVIEWER'S KAPLAN-MEIER CURVES 0F OVERALL SURVIVAL (ALL RANDOMIZED PATIENTS) .. 33
`
`APPEARS THIS WAY
`ON ORIGINAL
`
`
`NA 21—60 AbraxancAU31)-007
`'
`'
`" '
`”
`iv
`
`

`

`1 EXECUTIVE SUMNIARY
`
`1.1 CONCLUSIONS AND RECOMMENDATIONS
`
`
`
`In this reviewer’s opinion, the study results from this randomized, multi-center, open-label, and
`phase III trial support the efficacy claim based on the primary endpoint, reconciled target lesion
`response rate.
`
`It is to be noted that this NDA was filed under Section 505 (b)(2) of the Federal Food, Drug and
`Cosmetic Act, referencing the label, efficacy and safety ofTaxol (paclitaxel) Injection. Taxol
`was approved in April 1994 “for the treatment ofbreast cancer afterfailure ofcombination
`chemotherapyfor metastatic disease or relapse within 6 months ofadjuvant chemotherapy. Prior
`therapy should have included an anthracycline unless clinically contraindicated”. The sponsor’s
`proposed indication is:
`'
`F" ‘_
`_
`,
`_
`an-
`The appropriateness of the patient
`population proposed by the sponsor is deferred to the clinical reviewer.
`
`1.2 BRIEF OVERVIEW OF CLINICAL STUDIES
`
`lltbrzntancTM (ABI-007), cytotoxic antineoplastic, was developed as a cremophor-free formulation
`of paclitaxel. It is to be administered intravenously over 30 minutes every 3 weeks. Patients may
`be treated for up to 6 cycles and patients without progression could be treated for a longer period.
`
`In this NDA submission, efficacy data were collected by the sponsor in two trials: CAOIZ-O and
`CAOOZ-O. Trial CA012-0 was a randomized, multi-center, open—label, Phase III trial, conducted
`at 70 sites, located in Russia/Ukraine, Canada, the US. and the United Kingdom. A total of 460
`patients with metastatic breast cancer were randomized to receive either ABI-OO? or Taxol. This
`trial included patients who had received first—line or subsequent-line of treatment, and some
`patients had not received prior anthracycline therapy. Trial CAOOZ—O was a phase 2 trial,
`conducted in patients with metastatic breast cancer and was to provide support to the efficacy and
`safety of ABI—OO7.
`
`1.3
`
`STATISTICAL ISSUES AND FINDINGS
`
`In this NDA, Study CAOlZ-O was the only randomized pivotal study conducted to establish
`efficacy and safety. A total of 460 patients were randomized to this study. Of those, 233 patients
`were randomized to the ABI-OO? arm and 227 patients to the Taxol arm. The primary efficacy
`endpoint was response rate based on reconciled (investigators and independent radiology experts)
`assessment of target lesions through cycle 6. The primary analysis was a sequential test with the
`following pre-specified testing order: non-inferiority test in the whole study population,
`superiority test in the whole study population, and superiority test in the subgroup of patients who
`received study treatment as the 1“ line therapy. Based on the FDA clinical reviewer’s
`adjudication of response status, there were 50 and 25 responders in the ABI-O? arm and the Taxol
`arm, respectively. The observed response rates Were 21.5% and l 1.1%, respectively and the
`estimated ratio of response rates (ABI-OO'I/Taxol) was 1.899 with a 95% confidence interval of
`“mm-W
`NDA 2l—660 [N 000 Abraxane (ABI-OO?)
`l
`
`
`
`
`
`

`

`1.228 — 2.93 7. This suggests the superiority of ABI-007 with respect to the primary endpoint in
`the whole study population (Table 1 below or Table 14 for details)
`
`This NDA was filed under Section 505 (b)(2) of the Federal Food, Drug and Cosmetic Act,
`referencing the label, efficacy and safety of Taxol (paclitaxel) Injection. A total of 272 patients
`(58%) in Study CAO 12-0 met the Taxol indication. Of those, 129 patients were randomized to
`the ARI—007 arm and 143 patients to the Taxol arm. Based on the FDA clinical reviewer’s
`adjudication of response status, there were 20 and 12 responders in the ARI-07 arm and the Taxol
`arm, respectively. The observed response rates were 15.5% and 8.4%, respectively and the
`estimated ratio of response rates (ABI-007I'Taxol) was 1.848 with a 95% confidence interval of
`0.941 ~ 3.628. Although not statistically significant, the result from this subgroup of patients
`appeared trending towards the same direction as from the whole study population (Table 16).
`
`Table 1: Reviewer's Results of FDA-Confirmed recTLRR (All Randomized Patients)
`
`Reconciled Target Lesion Response Assessment
`Dataset
`
`ARI-007
`[N= 2331
`
`Taxol
`[N= 227]
`
`
`
`
`No of FDA—Confirmed Responders ——
`Response Rate
`215%
`11.1%
`(95% Binomial Confidence Interval)
`(l6. 19%- 26.73%)
`(6.94% -- 15.09%)
`
`Ratio of Response Rates (ABI—007/Taxol)
`(95% Confidence Interval)l
`
`1.899
`(i .-228 2.937)
`
`0003
`
`95% confidence interval of the ratio in superiority analysis based on the stratified Cochran-Mantel-
`Haenszel (CMH) test, stratified by 1‘I line vs. > 1St line therapy.
`b P-value from the stratified CMH test.
`
`
`
`
`Time to disease progression was a secondary endpoint. This endpoint was measured by this
`reviewer from the date of randomization, as opposed to the date of the first study dose that was
`used by the Sponsor. When documentation of disease progression was based on investigator’s
`assessment including cycles beyond 6, 45.9% of the patients randomized to the ABI-OO'i and
`54.6% of the patients randomized to the Taxol arm had progressive disease. The observed
`median time to disease progression was 21.9 weeks and 16.4 weeks (5.02 months and 3.77
`months) respectively for the ABI-OO'I and Taxol arms respectively. When documentation of
`disease progression was based on reconciled assessment through cycle 6, 39.4% of the patients
`randomized to the ABl—007 and 52.0% of the patients randomized to the Taxol arm had
`progressive disease. The observed median time to disease progression was 17.0 weeks and 15.6
`weeks (3.90 months and 3.57 months) respectively for the ABI-007 and Taxol arms respectively.
`In both approaches of documenting progressive disease, there was statistically significant
`difference in distributions of time to disease progression between treatment arms (Table 27;
`logrank p-values < 0.04; point estimates of hazard ratio (ABI-007/Taxol) S 0.76).
`
`Duration of response and overall survival were secondary endpoints. Data are not mature for a
`meaningful comparison between treatment arms as summarized below.
`
`
`NBA 21 6-60/N 000 Abraxane(ABi007) "
`"" ’
`"
`’ "
`2"
`
`

`

`Statistical Issues:
`
`
`
`[11
`
`Duration of Response. Since the primary endpoint was reconciled target lesion response
`rate, a consistent analysis of duration of response would be based on reconciled assessment
`However, the reconciled assessment was available only for the first 6 cycles. By the end of
`cycle 6, only 11 (14.7%) of the 75 FDA-confirmed responders had progressive disease
`(Table 24). Since most of the responders had not developed progressive disease, the median
`duration of response was unobservable in both treatment arms. When documentation of
`progressive disease was based on the either WorldCare assessment through cycle 6 or
`investigator’s assessment including cycles beyond 6, there were only 34.7% of the
`responders who subsequently had documented progressive disease with observed median
`durations of response 23.4 weeks (5.38 months) and 27.0 weeks (6.20 months) for the ABI-
`007 and the Taxoi anus, respectively (Table 25). Data are not mature for a meaningful
`comparison between treatment arms with respect to duration of response.
`
`[2]
`
`Overall survival. Only 34% (= 157/460) of the patients died by the end of the trial (Table
`28). Data are not mature for a meaningful comparison between treatment arms with respect
`to overall survival.
`
`APPEARS THIS WAY
`
`(IN ORIGINAL
`
`
`
`
`
`NDA 21-660 IN 000 Abraxane (ARI-007)
`
`
`
`

`

` 2
`
`INTRODUCTION
`
`2.1 OVERVIEW
`
`This NDA was filed under Section 505 (b)(2) of the Federal Food , Drug and Cosmetic Act,
`referencing the label, efficacy and safety of Taxol (paclitaxel) Injection. Taxol was first approved
`in December 1992 for ovarian cancer and was subsequently approved for additional indications,
`including the breast cancer indication approved in April 1994: “for the treatment ofbreast
`cancer afterfailure ofcombination chemotherapyfor metastatic disease or relapse within 6
`months ofaajiuvant chemotherapy. Prior therapy should have included an anthracycline unless
`clinically contraindicated’.
`
`AbraxaneTM (ABI-OO'I'), cytotoxic antincoplastic, was developed as a cremophor-free formulation
`of paclitaxel. It is to be administered intravenously over 30 minutes every 3 weeks. Patients may
`be treated for up to 6 cycles and patients without progression could be treated for a longer period.
`The sponsor’s proposed indication is:
`._....
`
`In this NDA submission, efficacy data were collected by the sponsor in two trials: CAOIZ-O and
`CAOOZ-O. Trial CA012—0 was a randomized, multi-center, open-label, Phase III trial, conducted
`at 70 sites, located in Russia/Ukraine, Canada, the US. and the United Kingdom. A total of 460
`patients with metastatic breast cancer were randomized to receive either ABI-OO'l or Taxol. This
`trial included patients who had received first-line or subsequent-line of treatment, and some
`patients had not received prior anthracycline therapy. Trial CAOOZ-O was a phase 2 trial,
`conducted in patients with metastatic breast cancer and was to provide additional support to the
`efficacy and safety of ABI-OO'l.
`
`Since Study CA012-0 was the only comparative study of efficacy, it is the only study reviewed
`by this statistical reviewer. The following bulletsl summarize important milestone in product
`development with emphasis on Trial CAOIZ-O that may be relevant to statistical review. A more
`detailed summary of milestones is deferred to the clinical review.
`
`. May 12, 1998: The original Investigational New Drug Application was submitted under
`IND No. 55974.
`
`. May 3, 2001: An Initial draft of the phase 3 protocol (CA012-0) was discussed. FDA
`responses to the sponsor’s submitted questions indicate acceptance of the proposed phase
`3 protocol, with changes to the statistical analysis plan. Since ARI-007 and Taxol “have
`the some active ingredient._.the Division would accept a single randomized, non-
`inferiority phase 3 study, maintaining at least 75% ofthe Taxol eflect, with response rate
`as a primary endpoint. This phase 3 stuafil, along with phase 2 data (study CA002)
`showing similar activity, would support approval ofABI—007 in a second line metastatic
`breast cancer setting.” FDA stated that ABI should “study a sufficient subset ofpatients
`(at least 100 treated with ARI—007) in Taxol '5 approved indication.”
`
`'
`
`June 11, 2001 (SN. 070): Phase 3 protocol (CA012-0) submitted, entitled “A Controlled
`Randomized, Phase III, Multicenter, Open Label Study ofARI—007 (A Cremophor-Free,
`
`I Contents in some of the bullets were extracted from the Clinical Review by Dr. Scher.
`
`
`NDA 21-660 [N 000 Abraxane (ABl-OO'I')
`
`4
`
`

`

`
`
`
`
`Protein Stabilized, Nanoparticle Paclitaxel) and Tara! in Patients with Metastatic Breast
`Cancer."
`
`June 19, 2001 (SN. 071): Protocol amended to change treatment duration from a
`maximum of 3 cycles to 6 cycles and to change end of study from 9 weeks to 15 weeks.
`The design is for “210 evaluable patients (230 enrolled) per treatment arm with at least
`100 patients per arm that have been previously treated with anthracyclines.”
`
`October 29, 2001: First patient randomized to Study CA012-O (per the sponsor’s data
`set “survival.xpt”).
`
`March 6, 2002 (SN. 115): Protocol Amendment I for Study CA012-0 was submitted,
`where changes included
`
`—
`
`_
`—
`
`—
`
`Changing definition of HT population from exposure to 2 2 cycles to Z 1 cycle of
`therapy;
`only target lesions (TLs) to be evaluated for the primary endpoint;
`imaging studies to assess response after cycles 2, 3 and 5, with confirmation of
`response at weeks 9 and 15;
`overall response to include TLs and non-TLS.
`
`July 30, 2002 (SN. 161): Protocol Amendment 2 for Study CA012-0 was submitted.
`WorldCare designated central image reader.
`
`October 8, 2002: Interim analysis for Study CA012-0 by Data Monitoring Committee
`after response assessed for 105 patients treated for 2 2 cycles in each arm. No change in
`sample size.
`
`December 9, 2002 (SN. 198): Statistical Analysis Plan, version 2.1, was submitted.
`
`March 18, 2003: Last patient randomized to Study CAOIZ-O (per the sponsor’s data set
`“survival.xpt").
`
`March 19, 2003 (SN. 217): Pre—NDA meeting. Methods for response analysis were
`attached in the submission. FDA emphasized the importance, for a 505(b)(2) submission,
`of documenting that study patients had failed combination chemotherapy for metastatic
`breast cancer or relapsed within 6 months of adjuvant chemotherapy. “Previous
`chemotherapy should have included an anthracycline unless contraindicated,” and the
`reason should be indicated, Randomization within each country was to be divided into
`two strata, anthracycline naive and anthracycline treated.
`
`April '7, 2003: Data cut-off date for Study CAOlZ-O (per the sponsor’s Study Report);
`patients with ongoing benefit continued treatment beyond 6 cycles.
`
`September 3, 2003: Data lock date for Study CA012—0.
`
`October 20, 2003 (SN. 27S): Statistical Analysis Plan, version 3.0, was submitted,
`where 3 sets of response data (investigator, independent radiology group, and reconciled)
`were defined. The applicant specified that “reconciled response” was the “primary
`response dataset.”
`
`W N
`
`DA 21—660 [N 000 Abraxane (ABI-OO?)
`
`5
`
`
`
`

`

`
`
`2.2 DATA SOURCES
`
`Data used for review are from the electronic submission received between November 2003 and
`
`October 2004. The network path is \\Cdsesubl\1121660\N 000\” in the EDR.
`
`APPEARS TH!S WAY
`0H ORlGINAL
`
`APPEARS THIS WAY
`0N ORIGINAL
`
`
`
`NDA 21—660 IN 000 Abraxane (ABI-OO?)
`
`6
`
`

`

`3 STATISTICAL EVALUATION
`
`This Section only considers the comparative study CAOIZ-O.
`
`3.1
`
`EVALUATION OF EFFICACY
`
`3.1.1
`
`STUDY DESIGN
`
`
`
`This study was designed as a controlled, randomized, multi-center, open-Iabel, Phase III,
`outpatient, non-inferiority study to evaluate the safety/tolerability and anti-tumor effect of
`intravenously (IV) administered ABI-OO? compared to that of Taxol in patients with metastatic
`breast cancer.
`
`Inclusion criteria of study population included the following:
`
`I
`I
`
`I
`
`I
`
`I
`
`patient is female, non-pregnant and not lactating, and 2 18 years of age;
`patient has a histologically or cytologically confirmed measurable metastatic breast
`cancer and is a candidate for paclitaxel therapy in accordance with standard of care;
`if patient has received Taxol or docetaxel as adjuvant therapy, she has not relapsed with
`breast cancer within l year of completing adjuvant Taxol or docetaxel;
`patient has no other malignancy within the past 5 years, except non-melanoma skin
`cancer, cervical intraepithelial neoplasia (GIN), or in-situ cervical cancer (CIS);
`patient is a suitable candidate for single—agent paclitaxel treatment.
`
`Complete inclusion criteria, as well as exclusion criteria, are referred to the Sponsor’s Study
`Report or FDA Clinical Review by Dr. Scher.
`
`Patients who received anthracycline prior to study enrollment were required to have a 4-week
`interval between the last dose of anthracycline and the first dose of study drug. Eligible patients
`were randomized in a 1:1 ratio to receive either of the following treatments:
`
`Investigational arm:
`I
`I Control arm:
`
`ABI—OO'I', 260 mg/m2 IV over 30 minutes,
`Taxol, 175 mg/m2 IV over 3 hours.
`
`Randomization was stratified by
`
`'
`I
`
`country;
`history of anthracycline use (anthracycline-exposed vs. Ina‘I‘ve).
`
`Patients were to be treated for up to 6 cycles (3 weeks per cycle) and patients without progression
`could be treated for a longer period, at the discretion of the investigator. Patients were to be
`assessed with imaging studies for response after cycles 2, 3, and 5, with confirmation of response
`at weeks 9 and [5.
`
`WM
`NDA 21-660 IN 000 Abraxane (ABl-OO'I')
`7
`
`
`
`

`

`
`
`3.1.2
`
`EFFICACY ENDPOINTS
`
`The primary efficacy endpoint and many secondary efficacy endpoints pertained to anti-tumor
`response. In this section, definitions of anti-tumor response are briefly described, followed by the
`definitions of efficacy endpoints.
`
`Methods of tumor assessment included radiologic imaging or clinical evaluation. Lesions at the
`Baseline Visit were to be classified as either target or non-target lesions. At baseline, 21 maximum
`of 5 target lesions per organ and 10 target lesions in total were to be identified. The following
`defines several types of response per target or non-target lesions:
`
`I Target Response. This was defined as achieving CR (complete response) or PR (partial
`response) of target lesions based on Response Evaluation Criteria in Solid Tumors
`(RECIST) guideline.
`I Non—target Response. Complete non-target response was defined as the disappearance
`of all non-target lesions based on RECIST guideline for non-target lesions response
`criteria.
`
`I Overall Response. This was determined based on target response, non-target response,
`and absence or- appearance of new lesions.
`
`Two versions (v. 2 and v. 3) of Statistical Analysis Plan were submitted to FDA during the drug
`development.
`In the latest version (v. 3), the sponsor defined the following 3 sets of response
`data:
`
`. Response data as assessed by the investigator;
`I Response data as assessed by an independent radiology laboratory (WorldCare);
`I Reconciled response assessment data based upon independent assessment of images
`and investigator responses.
`
`The reconciled response assessment data was to be the primary response dataset for efficacy
`evaluation and the other two sets were considered as the secondary response datasets.
`
`The primary efficacy endpoint was
`
`I
`
`recTLRR: the percentage of patients who achieved confirmed complete or partial
`target lesion response using the reconciled response assessment data set.
`
`Secondary efficacy endpoints were
`
`I
`
`invORR: percentage of patients who achieved complete or partial overall response
`using the investigator’s assessment data set;
`time to disease progression;
`I
`patient survival;
`I
`percentage of patients who achieved each target lesion response of CR, PR, SD, or PD;
`I
`percentage of patients who achieved each overall response of CR, PR, SD, or PD;
`I
`time to first complete or partial target lesion response;
`I
`time to first complete or partial overall response;
`I
`I
`duration of complete or partial target lesion response;
`
`601-007) 7
`_.__
`"
`8"
`
`

`

`
`
`duration of first complete or partial overall response;
`number of cycles of therapy to maximum target lesion response;
`number of cycles of therapy to maximum overall response;
`duration of CR, PR, or SD for target lesion response;
`duration of CR, PR, or SD for overall response; and
`QOL evaluated by changes from baseline in scores on the ECOG (Zubrod) performance
`status scale, EORTC-QLQ-C30, and weight.
`
`3.1.3 HYPOTIIESiS TEST AND SAMPLE SIZE CONSIDERATIONS
`
`The study was designed for testing non-inferiority hypothesis. The null hypothesis was that
`recTLRR in the ABI-007 arm was no larger than 75% of that in the Taxol arm. The burden of the
`proof to establish non-inferiority was to provide statistical evidence against the null hypothesis.
`If non-inferiority was established, then superiority test was to proceed. If superiority was
`established, then superiority test was to further proceed only in the subgroup of patients who
`received study drug as the first—line therapy for metastatic cancer.
`
`The initial sample size was 210 (approximately 230 enrolled) evaluable patients per treatment
`arm with at least 100 patients per arm that had been previously treated with anthracycline. This
`was based on the assumptions that the Taxoi response rate was assumed to be 30% and the ABI-
`007 response rate was assumed to 36% (a relative improvement of a fifih).
`
`An interim analysis to re—estimate sample size was performed after approximately 105 patients
`had been treated for a minimum of two cycles in each arm, and had undergone assessment of
`response. Sample size was not changed from the initial estimate.
`
`3.1.4
`
`STATISTICAL ANALYSIS PLAN
`
`The primary analysis for the primary endpoint recTLRR was based on the stratified Cochran-
`Mantel-Haenszel (CMH) test, stratified by whether study treatment was iSt line therapy versus >
`1“ line therapy. The primary analysis was the multiple nested testing as summarized below:
`
`I
`
`I
`
`.
`
`Step 1: Test for non—inferiority of ABI-OO? compared to TAXOL. Non-inferiority was
`to be established if the lower bound of the confidence limit for the ratio of response rates
`(ABI-007 I TAXOL) was greater than 0.75. if non-inferiority was established, then
`proceed to Step 2