Double-Blind, Randomized Trial Comparing the Efficacy
`and Tolerability of Fulvestrant Versus Anastrozole in
`Postmenopausal Women With Advanced Breast Cancer
`Progressing on Prior Endocrine Therapy: Results of a
`North American Trial
`
`By C.K. Osborne, J. Pippen, $.E. Jones, L.M. Parker, M. Ellis, 5. Come, $.Z. Gertler, J.T. May, G. Burton, I. Dimery,
`A. Webster, C. Morris, R. Elleolge, and A. Buzdar
`
`Purpose: To compare the efficacy and tolerability of
`fulvestrant (formerly ICI 182,780) with anastrozole in the
`treatment of advanced breast cancer in patients whose
`disease progresses on prior endocrine treatment.
`Patients and Methods:
`In this double-blind, double-
`dummy, parallel-group study, postmenopausal pa-
`tients were randomized to receive either an intramus-
`cular injection of fulvestrant 250 mg once monthly or a
`daily oral dose of anastrozole 1 mg. The primary end
`point was time to progression (TTP). Secondary end
`points included obiective response (OR) rate, duration
`of response (DOR), and tolerability.
`Results: Patients (n = 400) were followed for a me-
`dian period of 16.8 months. Fulvestrant was as effec-
`tive as anastrozole in terms of TTP (hazard ratio, 0.92;
`95.14% confidence interval [CI], 0.74 to 1.14; P = .43);
`median TTP was 5.4 months with fulvestrant and 3.4
`months with anastrozole. OR rates were 17.5% with
`both treatments. Clinical benefit rates (complete re-
`
`HE SELECTIVE estrogen receptor modulator (SERM)
`tamoxifen (Nolvadex; AstraZeneca, Wilmington, DE)
`is well established as a highly effective treatment for pre-
`and postmenopausal patients with either advanced or early
`breast cancer.
`1 Tamoxifen has also been shown to be
`
`effective in reducing the incidence of breast cancer in
`patients at risk of developing the disease2 and in women
`
`From the Breast Center at Baylor College of Medicine, Methodist
`Hospital, US Oncology, andM.D. Anderson Cancer Center, Houston, and
`Baylor—Sammons Cancer Center, Dallas, TX; Dana—Farber Cancer Insti-
`tute and Beth Israel Deaconess Medical Center, Boston, MA; Lombardi
`Cancer Center, Washington, DC; Ottawa Regional Cancer Center, Ot-
`tawa, Ontario, Canada; Hematology Oncology Association VA Ltd,
`Richmond, VA; Louisiana State University Health Science Center, Shreve-
`port, LA; and AstraZeneca, Macclegield, United Kingdom.
`Submitted October 10, 2001; accepted March 22, 2002.
`Supported by a grant from AstraZeneca Pharmaceuticals, Wilming-
`ton, DE.
`This article was published ahead ofprint at www.fco.org.
`Address reprint requests to CK. Osborne, MD, Breast Center,
`Baylor College of Medicine, 1 Baylor Plaza, MS 600, Houston, TX
`77030; email: kosborne@breastcenter.tmc.edu.
`© 2002 by American Society of Clinical Oncology.
`0732-183X/02/201 6-3386/$20. 00
`
`sponse + partial response + stable disease 2 24 weeks)
`were 42.2% for fulvestrant and 36.1% for anastrozole
`(95% Cl, -4.00% to 16.41%; P = .26). In responding
`patients, median DOR (from randomization to progres-
`sion) was 19.0 months for fulvestrant and 10.8 months
`for anastrozole. Using all patients, DOR was signifi-
`cantly greater for fulvestrant compared with anastro-
`zole; the ratio of average response durations was 1.35
`(95% Cl, 1.10 to 1.67; P < 0.01). Both treatments were
`well tolerated.
`Conclusion: Fulvestrant was at least as effective as
`anastrozole, with efficacy end points slightly favoring
`fulvestrant. Fulvestrant represents an additional treat-
`ment option for postmenopausal women with ad-
`vanced breast cancer whose disease progresses on
`tamoxifen therapy.
`J Clin Oncol 20:3386-3395. © 2002 by American
`Society of Clinical Oncology.
`
`with ductal carcinoma-in-situ.3 Patients who have tumor
`
`progression or develop resistance to tamoxifen are often
`treated with second-line hormonal therapy. The treatment
`options currently available comprise the third generation of
`oral, selective nonsteroidal aromatase inhibitors including
`anastrozole, letrozole, and the steroidal agent exemestane.
`Fulvestrant (Faslodex; AstraZeneca, Macclesfield, United
`Kingdom) is a “pure” estrogen antagonist with a novel
`mode of action, distinct from that of tamoxifen or any other
`antiestrogen currently available. Fulvestrant,
`like tamox-
`ifen, binds to estrogen receptors (ERs) competitively. How-
`ever, in contrast to tamoxifen, fulvestrant’s binding leads to
`rapid degradation and loss of ER protein.4 Furthermore,
`fulvestrant antagonizes all of the transactivating functions
`of the receptor, whereas tamoxifen blocks only one, a
`feature that contributes to its estrogen agonist activity in
`some tissues.4 Accordingly, fulvestrant is the first in a new
`class of antiestrogens—an ER downregulator—and is de-
`void of agonist activity.5 Fulvestrant has greater potency
`than tamoxifen at inhibiting the growth of breast tumors and
`doubles the time to the development of resistance in a
`xenograft murine model of human breast cancer.6 It also
`inhibits growth of tamoxifen-resistant tumors in this mod-
`
`3386
`
`Journal of Clinical Oncology, Vol 20, No l6 (August l5), 2002: pp 3386-3395
`DOI: lO.l200/JCO.2002.l0.058
`
`AstraZeneca Ex. 2029 p. 1
`Mylan Pharms. Inc. V. AstraZeneca AB IPR2016-01324
`
`

`
`FULVESTRANT v ANASTROZOLE IN ADVANCED BREAST CANCER
`
`3387
`
`el.6 In primary breast cancer patients who received a single
`injection of fulvestrant (at doses of 50, 125, or 250 mg)
`14 to 21 days before the initial tumor resection, fulves-
`trant produced a dose-dependent reduction in both ER
`and progesterone receptor (PgR) expression.7 In contrast,
`a separate group of patients in the same study who
`received tamoxifen 20 mg orally before tumor resection
`showed an increase in PgR expression, thereby confirm-
`ing the partial estrogen agonist activity of tamoxifen. A
`phase II study in postmenopausal women with advanced
`breast cancer whose disease progresses after tamoxifen
`therapy given as adjuvant or for advance disease showed
`that subsequent treatment with fulvestrant was associated
`with durable responses.8‘10
`This study provides the first opportunity to compare the
`relative efficacy of ER suppression with the ER downregu-
`lator fulvestrant with that of anastrozole, as second-line
`therapy in patients with potentially horrnone-dependent
`advanced breast cancer.
`
`PATIENTS AND METHODS
`
`Sudy Design
`
`The study (trial 0021) was a randomized, double-blind, double-
`dummy, phase III trial conducted in North America. The trial was
`originally designed to compare two doses of fulvestrant (125 mg and
`250 mg per month) as an intramuscular injection with anastrozole as a
`1 mg/d oral dose. A nonblinded, open-label trial using the same drug
`doses and a similar protocol (trial 0020) was conducted concurrently in
`Europe, South Africa, and Australia (see accompanying article in this
`issue of the Journal of Clinical Onoo|ogy).“
`A preliminary data summary and an interim analysis were planned
`and conducted to determine the clinical activity of fulvestrant 125 mg,
`which had not been previously tested. Therefore both trials included a
`preliminary data summary stage after the first 30 subjects in the
`fulvestrant 125-mg group (combined from both trials) had been treated
`and followed up for 3 months. This interim assessment showed
`insufficient evidence of clinical activity for fulvestrant 125 mg with no
`objective tumor responses at 3 months. The independent data monitor-
`ing committee therefore recommended that recruitment to the fulves-
`trant 125-mg treatment arm be stopped. Patients already recruited into
`the 125-mg arm in trial 0021 were permitted to remain on fulvestrant
`125 mg or withdraw from the trial and be placed on other treatments at
`the discretion of their clinician. These patients were not monitored
`further for efficacy. The lack of an objective response in the low-dose
`fulvestrant arm also suggests that response due to tamoxifen with-
`drawal in this study must be uncommon. As a consequence of dropping
`this treatment arm, the protocol for the study was amended to compare
`fulvestrant 250 mg with anastrozole 1 mg.
`An interim analysis was conducted when 170 progressions or deaths
`had occurred across the remaining arms and time to progression (TTP)
`was formally analyzed. The rate of objective response (OR; defined as
`complete response [CR] + partial response [PR] using Union Intema-
`tionale Contre le Cancer criteria) and adverse event (AE) data were
`summarized. As a result of the interim analysis, the independent data
`monitoring committee recommended that the trial should continue.
`
`The primary end point of the comparison between the two drugs was
`TTP. Secondary end points included OR, duration of response (DOR),
`time to treatment failure (TTF), time to death (TTD), and tolerability.
`Other secondary end points were quality of life, symptomatic response,
`and pharmacokinetics. Other end points included clinical benefit (CR -l-
`PR + stable disease [SD] 2 24 weeks) and duration ofclinical benefit.
`All data are reported here except pharmacokjnetics, which will be
`reported elsewhere.
`
`Patient Population
`All patients were postmenopausal women with locally advanced or
`metastatic breast cancer whose disease had progressed on adjuvant
`endocrine therapy with an antiestrogen or whose disease had pro-
`gressed after first-line endocrine therapy for advanced disease. All
`women had a life expectancy of longer than 3 months and tumors with
`evidence of hormone sensitivity (ie, prior sensitivity to hormonal
`therapy or known ER or PgR positivity).
`For inclusion in the trial, patients had to have a World Health
`Organization performance status of S 2, histologic or cytologic
`confirmation of breast cancer, objective evidence of recurrence or
`progression of disease that was not amenable to curative treatment, and
`the presence of at least one measurable or assessable (nonmeasurable)
`lesion. All patients had to be postmenopausal (ie, 2 60 years old or
`aged 2 45 years with arnenorrhea for > 12 months or follicle-
`stimulating hormone levels within postmenopausal range, or having
`undergone a bilateral oophorectomy).
`the presence of life-
`Exclusion criteria included the following:
`threatening metastatic visceral disease (defined as extensive hepatic
`involvement) or any degree of brain or leptomeningeal involvement;
`symptomatic pulmonary lymphangitic spread; prior treatment for breast
`cancer with fulvestrant or any aromatase inhibitor; more than one prior
`endocrine medical treatment for advanced breast cancer; extensive
`radiation therapy or cytotoxic treatment within the past 4 weeks;
`estrogen replacement therapy within 4 weeks of randomization; treat-
`ment with luteinizing hormone—releasing hormone analogs within 3
`months before randomization; and any concurrent medical illness or
`laboratory abnormalities that would compromise safety or prevent
`interpretation of results.
`Subjects taking bisphosphonates for bone disease were permitted
`to enter the trial, but their bone lesions were not considered to be
`assessable for response, although they were assessable for progres-
`sion. Initiation of bisphosphonate treatment during the trial was
`discouraged but allowed in the absence of objective evidence of
`progression. If bisphosphonates were commenced, bone lesions
`were assessed only for progression.
`All patients provided written informed consent, and the relevant
`ethical committees approved the studies.
`
`Trial Treatments
`
`Fulvestrant was supplied in vials as a sirrgle-dose, castor oil—based,
`5% solution. Each vial contained 250 mg of fulvestrant at a concen-
`tration of 50 mg/mL in a volume of 5 mL. The matched placebo was
`5 mL of the oily excipient. Fulvestrant 250 mg or matching placebo
`was administered slowly as a 2.5-mL injection into each buttock.
`Injections were given once a month, which was defined as every 28
`days (i 3 days).
`Anastrozole (Arimidex) 1 mg and matching placebo were supplied
`as round, white, film—coated tablets and administered orally once daily.
`Medical personnel saw all patients on a monthly basis because all
`patients required fulvestrant or placebo injections.
`
`AstraZeneca Ex. 2029 p. 2
`
`

`
`3388
`
`OSBORNE ET AL
`
`Patients continued treatment until objective disease progression or
`other events required withdrawal; at such time, trial treatment was
`stopped and standard therapy was initiated at the discretion of the
`treating physician. Thereafter, patients were followed up until death.
`Patients who withdrew from trial treatment before progression were
`followed up until objective disease progression and death.
`All patients were seen by a physician to make objective tumor
`assessments every 3 months until evidence of either objective disease
`progression or death. Patients with skin and soft tissue lesions were also
`assessed every month during the first 3 months of treatment.
`
`Stati sti cal Methodology
`The trial was designed to detect the superiority of fulvestrant 250 mg
`in terms of efficacy and tolerability compared with anastrozole 1 mg in
`postmenopausal women with advanced breast cancer.
`The final analysis was scheduled to occur when 340 events (ie,
`objective disease progression or death) had occurred across the two
`groups. This provided 90% power to detect a hazard ratio (HR) 2 1.43
`or S 0.70 for fulvestrant treatment compared with anastrozole treat-
`ment, at a significance level of 5%. It was therefore planned to recruit
`392 patients (196 in each treatment group) to achieve the required
`number of events.
`The efficacy analyses Were performed according to randomized
`treatment (ie, “intention to trea ”) using anominal significance level of
`5%. However, for the TTP and OR analyses, the significance level was
`adjusted to 4.86% because of the preliminary data summary of OR and
`the interim analysis of TTP. As a result, the 95% confidence intervals
`(CIs) were adjusted accordingly to 95.14%. All significance levels are
`two-sided.
`Although not described in the protocol, fulvestrant was retrospec-
`tively compared with anastrozole for noninferiority for OR, TTP, and
`TTF. Because of the interim analysis, a one-sided CI of 97.57% was
`used for the evaluation of TTP and OR. For the analysis of TTF, a
`one-sided CI of 97.5% was used. These limits are identical to using the
`upper limit ofthe 95.14% two-sided CI from the analysis of TTP, the
`lower limit of the 95.14% two-sided CI for the difference in response
`rates for OR, and the upper limit of the 95% two-sided CI for TTF.
`For previous United States regulatory submissions of hormonal
`treatments for advanced breast cancer, the requirements for showing
`noninferiority for TTP were based on the upper one-sided confidence
`limit for the TTP HR not being greater than 1.25 (ie, a potential
`deficiency of > 25% for the experimental treatment had to be ruled
`out).
`In the same submissions,
`the requirement for demonstrating
`noninferiority in terms of response rate was based on ruling out a
`deficiency in the difference in response rates of more than 10%.
`Consequently, these criteria have been used to assess noninferiority of
`fulvestrant relative to anastrozole in this trial.
`'|'|'P. TTP was defined as the time from randomization until
`objective disease progression or death from any cause before progres-
`sion. Subjects who had not progressed at the time of analysis were
`right-censored using the last assessment date. Treatments were com-
`pared using Cox’s proportional hazards regression model (including the
`covariates age, performance status, measurable compared with non-
`measurable disease, receptor status, previous response to hormone
`therapy, previous use of cytotoxic chemotherapy, and use of bisphos-
`phonate therapy for bone disease). A global test was performed to
`determine whether there were significant treatment-by-baseline covari-
`ate interactions. The estimate of the treatment effect was expressed as
`an HR (fulvestrant/anastrozole), together with the corresponding CI
`and P value. TTP was also summarized using Kaplan-Meier curves for
`each treatment group, and the median TTP was calculated.
`
`TTF. TTF was defined as the number of days from randomization
`until the earliest occurrence of disease progression, death from any
`cause, or withdrawal from trial treatment for any reason. Patients
`whose treatment had not failed at the time of analysis were right-
`censored in the analysis at the time oftheir last assessment. Any patient
`who did not receive any trial therapy was assigned an uncensored TTF
`of zero days. Statistically, TTF was analyzed in the same way as TTP.
`OR rate. Responders were defined as those patients with a CR or
`PR. To qualify as a responder, the patient had to satisfy the criteria for
`CR or PR on one visit with no evidence of disease recurrence or death
`within 4 weeks after assessment. Treatment differences in OR were
`assessed by comparing the proportion of responders using a logistic
`regression model (with the same covariates as for TTP). The estimate
`of the treatment effect
`is expressed as an odds ratio (fulvestrant/
`anastrozole),
`together with the corresponding Cl and P value.
`In
`addition, an estimate of the difference in response rates (fulvestrant/
`anastrozole) and corresponding CI was also produced.
`DOR. The DOR was defined for responding patients only as the
`period of time from randomization to the first observation of disease
`progression. Patients who died before reaching progression were
`classified as completing their response at time of death. The DOR was
`summarized using Kaplan-Meier curves for each treatment group, and
`the median DOR was also calculated for each group.
`No statistical comparison was performed for DOR in only those
`patients responding to treatment, because this is not a randomized
`comparison. Rather, all patients were included in a statistical analysis
`of DOR, defined for responders as the time from onset of response to
`disease progression and for nonresponders as zero. These data were
`also summarized using Kaplan-Meier curves.
`Clinical benefit. Clinical benefit was defined as the sum of CR -1-
`PR + SD 2 24 weeks. Although a formal analysis of clinical benefit
`was not protocoled, treatment differences in the rate of clinical benefit
`were retrospectively assessed in the same way as that of OR rate. The
`duration of clinical benefit was presented as for DOR.
`TTD. As specified in the protocol, TTD (overall survival) will be
`analyzed when more than 50% of the patients have died. At the time of
`this data analysis, only 34.5% of patients had died; therefore, no formal
`statistical analyses were conducted.
`
`Tolerability
`Any detrimental change in a patient’s condition subsequent to them
`entering the trial and during the follow-up period afier the final
`treatment (8 weeks after last injection or 30 days after the last tablet,
`whichever was the greater), which was not unequivocally due to
`progression of disease, was considered to be an AE. No formal
`statistical analyses were performed on the safety data from this
`individual trial. However, a plarmed statistical analysis of predefined
`AEs was performed on the combined data from this trial and the
`multinational trial; this will be reported elsewhere. The most common
`AEs
`(occurring at an incidence of Z 10%) and most common
`drug-related AEs are reported here by treatment received.
`
`Quality of Life
`Quality of life (QOL) was assessed using the Functional Assessment
`of Cancer Therapy (FACT)—Breast questionnaire, which is composed
`of the FACT-General QOL tool for cancer patients plus the breast
`cancer subscale. This questionnaire has been extensively validated in
`respect to psychometric properties and sensitivity to clinical chang-
`es12’13 and is in use in a number of large breast cancer treatment trials
`in the United States and Europe.
`
`Astrazeneca Ex. 2029 p. 3
`
`

`
`FULVESTRANT v ANASTROZOLE IN ADVANCED BREAST CANCER
`
`3389
`
`The analysis was undertaken on data collected up to the date of
`progression using the trial outcome index (TOI) within the FACT-
`Breast. This measure is the sum of the functional well-being, physical
`well-being and breast cancer subscale dimensions of the questionnaire.
`The difference in TOI over time between the fulvestrant 250-mg
`group and the anastrozole 1-mg group was compared using a general-
`ized linear mixed model (ie, a random coetficients model) with the
`same covaiiates as for TTP. A graph of the mean TOI (: standard
`deviation) over time was also produced.
`
`RESULTS
`
`Pati ents
`
`A total of 400 patients randomized to either fulvestrant
`250 mg (n = 206) or to anastrozole (n = 194) were
`followed for a median period of 16.8 months. The majority
`(95% of the fulvestrant group and 96% of the anastrozole
`group) had been treated previously with tamoxifen either as
`adjuvant therapy or as initial therapy for advanced disease.
`Ninety-four patients in each group had received endocrine
`therapy as adjuvant treatment. Of these, 67 patients in the
`fulvestrant group and 75 patients in the anastrozole group
`stopped treatment less than 365 days before randomization.
`The characteristics of the patients are presented in Table
`1. Patients in the fulvestrant and the anastrozole groups
`were similar for age, weight, breast cancer history, and ER
`and PgR status (Table 1).
`
`Eificacy
`
`TTP. At the time of analysis, 83.5% of the fulvestrant
`group and 86.1% of the anastrozole group had experienced
`disease progression. There was no significant difference for
`TTP between the two treatment groups (HR, 0.92‘, 95.14%
`Cl, 0.74 to 1.14‘, P = .43). The HR (0.92) indicates that the
`risk of progression (over a given period of time) for patients
`randomized to fulvestrant was 8% lower than it was for
`
`patients randomized to anastrozole. The 95.14% Cl indi-
`cates that the risk of progression for patients randomized to
`fulvestrant 250 mg could be between 26% lower and 14%
`higher than it is for patients randomized to anastrozole.
`These data demonstrate noninferiority of fulvestrant relative
`to anastrozole. Median TTP was 5.4 months for fulvestrant
`
`and 3.4 months for anastrozole (Fig 1).
`TTF. The majority of treatment failures were due to
`objective disease progression (94%), and accordingly, the
`Kaplan-Meier curves for TTP and TTF are very similar. For
`fulvestrant, there were 164 treatment failures (79.6%) be-
`cause of disease progression‘, for anastrozole, there were
`163 (84.0%). Other reasons for treatment failures in-
`cluded AEs, protocol noncompliance, and withdrawal of
`informed consent. TTF was similar for the two groups,
`with there being no significant difference between them
`
`Table 1. Demographic and Pretreatment Characteristics
`Fulvestrant
`Anastrozole
`250 mg
`I mg/cl
`[n : 206)
`(n : 194)
`No.
`%
`No.
`%
`
`Characteristic
`
`Age, years
`Mean
`Range
`Weight, kg
`Mean
`Range
`Prior treatment
`
`63
`33-89
`
`72
`37-1 27
`
`62
`36-94
`
`73
`43-134
`
`Cytotoxic chemotherapy
`Endocrine therapy for advanced
`disease
`Acljuvant endocrine therapy
`Hormone receptor status
`ER and/or PgR+ve
`ER/PgR unknown
`ER/PgR-ve
`Metastatic or recurrent disease at baseline
`Breast
`Skin
`Bone
`Liver
`Lung
`Lymph nodes
`Other
`Extent ol metastatic or recurrent disease at
`baseline
`Soft tissue only
`Bone only
`Visceral only
`Lymph node only
`Not recorded
`Mixed‘
`Measurable disease‘l'
`No measurable disease
`
`129
`1 10
`
`122
`
`179
`13
`14
`
`8
`43
`90
`47
`63
`58
`22
`
`12
`47
`39
`15
`1
`92
`1 14
`92
`
`62.6
`53.4
`
`59.2
`
`86.9
`6.3
`as
`
`3.9
`20.9
`43.7
`22.8
`30.6
`28.2
`10.7
`
`5.8
`22.8
`18.9
`7.3
`0.5
`44.7
`55.3
`44.7
`
`122
`97
`
`116
`
`169
`15
`10
`
`8
`41
`85
`45
`60
`56
`8
`
`13
`43
`45
`17
`2
`74
`107
`87
`
`62.9
`50.0
`
`59.8
`
`87.1
`7.7
`5.2
`
`4.1
`21.1
`43.8
`23.2
`30.9
`28.9
`4.1
`
`6.7
`22.2
`23.2
`8.8
`1 .0
`38.1
`55.2
`44.8
`
`NOTE. Patients may be in more than one category.
`Abbreviations: ER, estrogen receptor; PgR, progesterone receptor.
`*Mixed is delined as breast and/or a combination of skin, bone, liver, lung,
`or lymph nodes.
`lMeasurable lesions were lesions that were clinically measurable in two
`perpendicular axes with at least one dimension being 2 2.5 cm or measurable
`using imaging in two perpendicular axes with at least one dimension being 2
`1 .0 cm.
`
`(HR, 0.96; 95% CI, 0.77 to 1.19; P = .69) (Fig 2). The
`data also satisfy the criteria for noninferiority. Median
`TTF was 4.6 months for fulvestrant (n = 206) and 3.3
`months for anastrozole (n = 194).
`OR rate and clinical benefit. Fulvestrant resulted in an
`OR in 36 patients (17.5%), while anastrozole produced an
`OR in 34 patients (17.5%)
`(Table 2). There was no
`statistically significant difference in OR between fulvestrant
`and anastrozole (difference in response rates, 0.17%‘,
`95.14% CI,
`-6.31% to 9.30%). The lower CI shows
`
`AstraZeneca Ex. 2029 p. 4
`
`

`
`
`
`Median TTP: Fulvestrant 5.4 months
`Anastrozoie 3.4 months
`
`OSBORNE ET AL
`
`Fig I. Kaplan-Meier estimates
`for time to progression.
`
`3390
`
`O 01
`
`
`
`Proportionnotprogressed
`
` Fulvestrant 250 mg
`
`1
`----- Anastrozoie1 mg
`
`24
`30
`36
`
`Time to progression (months)
`
`noninferiority of fulvestrant relative to anastrozole. The
`odds ratio for achieving a11 OR i11 the fulvestrant group
`versus the anastrozole group was 1.01 (95.14% CI, 0.59 to
`1.73; P = .96).
`Clinical benefit rates of 42.2% and 36.1% were observed
`
`for fulvestrant and anastrozole, respectively (Table 2), with
`the analysis showing no statistically significant difference
`(difference in clinical benefit
`rates,
`5.83%‘, 95% Cl,
`-4.42% to 9.36%; P = .26).
`Extended follow-up was performed in order to obtain
`more complete information for DOR (median follow-up,
`21.3 months). The median DOR, as measured from random-
`ization to progression, in those patients who responded to
`treatment was 19.0 months for fulvestrant (11 = 36) and 10.8
`months for anastrozole (n = 34). The Kaplan-Meier curves
`for the DOR are shown in Fig 3. In addition, DOR using all
`
`patients—where DOR was defined as from the onset of
`response to disease progression for responders and zero for
`nonresponders—was significantly greater for fulvestrant
`compared with anastrozole (ratio of average response dura-
`tions, 1.35', 95% CI, 1.10 to 1.67; P < .01). The Kaplan-
`Meier curves for DOR in all patients are shown in Fig 4.
`The median duration of clinical benefit was 12.9 months for
`
`fulvestrant (n = 87) and 10.9 months for anastrozole (n =
`70) (Fig 5).
`'|'|'D. At the time of this data analysis, a similar
`number of deaths had occurred in each treatment group
`(fulvestrant, n = 73 [35.4%]', anastrozole, n = 65
`[33.5%]). However, as specified in the protocol, TTD
`(overall survival) will be analyzed when more than 50%
`of the patients have died. Consequently, no formal
`statistical analyses have been conducted on these data. In
`
`
`
`Median TTP: Fulvestrant 4.6 months
`Anastrozole 3.3 months
`
`
`
`Proportionnotfailed
`
`Fig 2. Kaplan-Meier estimates
`for time to treatment failure.
`
`
`
`Fulvestrant 250 mg
`----- Anastrozo|e1 mg
`
`
`24
`30
`36
`
`Time to treatment failure (months)
`
`AstraZeneca Ex. 2029 p. 5
`
`

`
`FULVESTRANT v ANASTROZOLE IN ADVANCED BREAST CANCER
`
`3391
`
`Table 2. Best Objective Responses for Fulvestrant 250 mg IM or
`Anastrozole 1 mg Orally ocl
`Fulvestrant
`(n : 206)
`
`Anastrozole
`(n : 194)
`
`CR
`PR
`Total
`SD 2 24 Weeks
`SD < 24 weel<s
`
`Not progressed
`Progressed
`Total
`Clinical beneiil (CR + PR
`+ SD 2 24 weeks)
`
`No.
`
`10
`26
`36
`51
`3
`
`11
`105
`170
`87
`
`°/o
`
`4.9
`1 2.6
`17.5
`24.8
`1.5
`
`5.3
`51 .0
`82.5
`42.2
`
`No.
`
`7
`27
`34
`36
`1
`
`19
`104
`160
`70
`
`%
`
`3.6
`13.9
`17.5
`18.6
`0.5
`
`9.8
`53.6
`82.5
`36.1
`
`Abbreviations: CR, complete response; PR, partial response; SD, stable
`disease.
`
`line with the study protocol, TTD analyses will be
`presented in a future publication.
`
`Tolerability
`
`Both fulvestrant and anastrozole were well tolerated,
`with only five fulvestrant patients (2.5%) and five anas-
`trozole patients (2.6%) withdrawing because of AEs. The
`profile of AEs (thought to be drug-related or not) was
`similar for the two drugs, and patients in both treatment
`groups reported a wide range of ABS. The most fre-
`quently reported AEs, graded according to Coding Sym-
`bols
`for Thesaurus of Adverse Reaction Terms
`
`(COSTART) classification, are shown in Table 3 and
`included asthenia, nausea, pain, headache, vasodilatation,
`pharyngitis, dy spnea, back pain, bone pain, and vomiting.
`The incidence and severity (most were mild) of events
`
`were generally similar between treatment groups. The
`most common drug-related AEs are shown in Table 4.
`The incidence of hot flashes was similar for both drugs
`(23.5% for
`fulvestrant and 24.9% for anastrozole).
`The incidence of weight gain (fulvestrant V anastrozole:
`1.5% V 1.6%), vaginitis (3.4% V2.6%), and thromboem-
`bolic disease (3.4% V 6.7%) was low for both fulvestrant
`and anastrozole.
`Joint disorders
`(including arthritis,
`arthralgia, and arthrosis) were reported by 9.3% and
`13.5% of patients in the fulvestrant and anastrozole
`groups, respectively.
`Injection site effects. Both patient groups received in
`total 6,908 monthly injections: 3,752 injections in the
`fulvestrant group and 3,156 placebo injections in the anas-
`trozole group. Fifty-five patients (27.0%) receiving fulves-
`trant and 45 patients (23.3%) receiving anastrozole (plus
`placebo injection) reported injection site reactions. Overall,
`86 fulvestrant courses (4.6%) of the total of 1,879 and 71
`placebo courses (4.4%) of the total of 1,624 resulted in an
`injection site event. Most of these events were charac-
`terized as injection site pain, reaction, and/or inflamma-
`tion. None of these events was serious, and only one
`patient complained of severe injection site pain and
`requested to be withdrawn from fulvestrant treatment (the
`investigator did not consider
`the event
`to be drug
`related). Two anastrozole-treated patients given placebo
`injections were also withdrawn because of AEs related to
`their injections.
`
`QOL
`
`A graph of the mean TO1 (i standard deviation) over
`time is shown in Fig 6. The analysis of QOL data up to the
`date of progression showed that QOL was maintained over
`
`esti-
`3. Kaplan-Meier
`Fig
`mates for duration of response.
`
`AstraZeneca Ex. 2029 p. 6
`
`Median DOR: Fulvestrant 19.0 months
`Anastrozole 10.8 months
`
`1.0
`
`0.9
`
`0.8
`
`0.7
`
`0.6
`
`0.5
`
`0.4
`03
`0.2
`
`0.1
`
`
`
`Proportionresponding
`
`_
`
`‘
`
`. _ . _ . _ .
`
`I
`
` — Fulvestrant 250 mg
`—-—-- Anastrozole 1 mg
`
`
`
`18
`
`24
`
`30
`
`36
`
`40
`
`Duration of response (months)
`
`

`
`3392
`
`0'35
`:3co
`
`OSBORNE ET AL
`
`Ratio of average response durations 1.35
`(95% CI, 1.10 to 1.67: P<.01)
`
`53NU‘!
`
`.0N
`
`0.05 Fig
`
`
`
`Proportionresponding
`
`0.15
`
`:3
`
`esti-
`4. Kaplan-Meier
`mates for duration of response
`(all patients).
`
`Duration of response (months)
`
`the two treatments were not statistically
`time and that
`significantly different.
`
`DISCUSSION
`
`This North American, randomized, double-blind, double-
`dummy, phase III study in postmenopausal women with
`advanced breast cancer was designed to compare the effi-
`cacy and tolerability of fulvestrant with that of the well-
`established third-generation aromatase inhibitor anastro-
`zole. Although the results did not show superiority of
`fulvestrant over anastrozole, the primary analysis of TTP
`showed fulvestrant to be at least as effective as anastrozole.
`
`The CIs observed for TTP, TTF, and OR for the two
`hormone agents allow us to rule out inferiority of fulvestrant
`to anastrozole. Furthermore, DOR using all patients was
`
`significantly greater in the fulvestrant group than in the
`anastrozole group. The clinically durable response observed
`with fulvestrant is in accordance with preclinical observa-
`tions in tumor models, where fulvestrant was shown to
`double the DOR compared with tamoxifen or withdrawal of
`estrogen,6 as well as with phase II clinical data,” where
`long responses were observed in tamoxifen-resistant, ad-
`vanced breast cancer patients. Furthermore, in the clinical
`benefit group, where patient numbers were higher,
`the
`duration of clinical benefit also favored fulvestrant. With
`
`respect to other end points of the study, fulvestrant showed
`a numerical improvement in TTP (median, 5.4 months) and
`TTF (median, 4.6 months) compared with anastrozole
`(median, 3.4 and 3.3 months, respectively), although the
`analyses of TTP and TTF showed no statistically significant
`
`esti-
`5. Kqplan-Meier
`Fig
`mates tor duration of clinical
`benefit (DOCB).
`
`AstraZeneca Ex. 2029 p. 7
`
`Median DOCB: Fuivestrant 12.9 months
`Anastrozole 10.9 months
`
`
`
` Fulvestrant 250 mg
`—.—.. Anastrozo|e1 mg
`
`
`30
`36
`40
`0
`6
`12
`18
`24
`
`0 U1
`
`
`
`Proportionwithclinicalbenefit
`
`Duration of clinical benefit (months)
`
`

`
`FULVESTRANT v ANASTROZOLE IN ADVANCED BREAST CANCER
`
`3393
`
`differences. OR was identical in the fiilvestrant and anastrozole
`
`groups (17.5%). A greater number of patients in the fulvestrant
`group achieved long-term SD, giving a slightly higher clinical
`benefit rate (42.2% for fulvestrant V 36.1% for anastrozole),
`although the analysis of clinical benefit rate showed no
`statistically significant difference. Thus, trends for all major
`end points favored fulvestrant. Reassuringly, the summary data
`of QOL showed that the use of fulvestrant in this patient
`population rriaintained QOL as well as anastrozole did. It is
`pre