`DOI 10.1007/s10549-01 1- 1947-7
`
`CLINICAL TRIAL
`
`Dose-dependent change in biomarkers during neoadjuvant
`endocrine therapy with fulvestrant: results from NEWEST,
`a randomized Phase II study
`
`Irene Kuter - Julia M. W. Gee - Roberto Hegg - Christian F. Singer - Rajendra A. Badwe -
`Elizabeth S. Lowe - Ugochi A. Emeribe - Elizabeth Anderson - Francisco Sapunar -
`Pauline Finlay - Robert I. Nicholson - José Bines - Nadia Harbeck
`
`Received: 23 December 2011/ Accepted: 26 December 2011/ Published online: 28 January 2012
`© Springer Science+Business Media, LLC. 2012
`
`Abstract NEWEST (Neoadjuvant Endocrine Therapy for
`Women with Estrogen-Sensitive Tumors) is the first study to
`compare biological and clinical activity of fulvestrant 500
`versus 250 mg in the neoadjuvant breast cancer setting. We
`hypothesized that fulvestrant 500 mg may be superior to
`250 mg in blocking estrogen receptor (ER) signaling and
`growth. A multicenter, randomized, open-label, Phase II
`study was performed to compare fulvestrant 500 mg
`(500 mg/month plus 500 mg on day 14 of month 1) versus
`fulvestrant 250 mg/month for 16 weeks prior to surgery in
`postmenopausal women with ER+ locally advanced breast
`cancer. Core biopsies at baseline, week 4, and surgery were
`
`On behalf of the NEWEST Investigators.
`The details of the investigators participating in the study are given in
`Appendix.
`
`Electronic supplementary material The online version of this
`article (doi:10.1007/s10549-O11-1947-7) contains supplementary
`material, which is available to authorized users.
`
`)
`I. Kuter (
`Massachusetts General Hospital, Professional Office Building
`228, 55 Fruit Street, Boston, MA 02114, USA
`e—mail: ikuter@partners.org
`
`J. M. W. Gee A P. Finlay A R. I. Nicholson
`Tenovus Centre for Cancer Research, Welsh School of
`Pharmacy, Cardiff University, Cardiff, Wales, UK
`
`R. Hegg
`School of Medicine, University of Sao Paulo and Hospital Pérola
`Byington, Sao Paulo, Brazil
`
`C. F. Singer
`Division of Special Gynaecology, Medical University of Vienna,
`Vienna, Austria
`
`R. A. Badwe
`Tata Memorial Hospital, Mumbai, India
`
`assessed for biomarker changes. Primary endpoint: change
`in Ki67 labeling index (LI) from baseline to week 4 deter-
`mined by automated computer imaging system (ACIS).
`Secondary endpoints: ER protein expression and function;
`progesterone receptor (PgR) expression; tumor response;
`tolerability. ER and PgR were examined retrospectively
`using the H score method. A total of 211 patients were
`randomized
`(fulvestrant
`500 mg:
`n : 109;
`250 mg:
`n = 102). At week 4, fulvestrant 500 mg resulted in greater
`reduction of Ki67 LI and ER expression versus 250 mg
`(-78.8 vs. —47.4% [p < 0.0001] and —25.0 vs. —13.5%
`[p = 0.0002], respectively [ACIS]); PgR suppression was
`not significantly different (-22.7 vs. -17.6; p = 0.5677).
`However, H score detected even greater suppression of ER
`(-50.3 vs. —13.7%; p < 0.0001) and greater PgR suppres-
`sion (-80.5 vs. -46.3%; p = 0.0018) for fulvestrant 500
`versus 250 mg. At week 16, tumor response rates were 22.9
`and 20.6% for fulvestrant 500 and 250 mg, respectively,
`with considerable decline in all markers by both ACIS and
`E. S. Lowe « U. A. Emeribe
`AstraZeneca, Wilmington, DE, USA
`
`E. Anderson A F. Sapunar
`Formerly AstraZeneca Pharmaceuticals, Macclesfield, UK
`
`I. Bines
`Instituto de Cancer, Rio de Janeiro, Brazil
`
`N. Harbeck
`Frauenklinik der Technischen Universitat Munchen, Munich,
`Germany
`
`Present Address:
`N. Harbeck
`Breast Centre, Department of Obstetrics and Gynaecology,
`University of Cologne, Cologne, Germany
`
`‘E Springer
`
`AstraZeneca Ex. 2077 p. 1
`Mylan Pharms. Inc. V. AstraZeneca AB IPR2016-01324
`
`
`
`238
`
`Breast Cancer Res Treat (2012) 133:237—246
`
`H score. No detrimental effects on endometrial thickness or
`
`bone markers and no new safety concerns were identified.
`This provides the first evidence of greater biological activity
`for fulvestrant 500 versus 250 mg in depleting ER expres-
`sion, function, and growth.
`
`Keywords Estrogen receptor-positive breast cancer -
`Fulvestrant 500 mg - Fasl0dex® - Neoadjuvant -
`Biomarkers
`
`Introduction
`
`Endocrine therapy is commonly used in the neoadjuvant
`setting to attempt to downstage large primary tumors and
`permit breast conserving surgery [2]. This setting also
`enables assessment of tumor response in situ and allows for
`further tailoring of subsequent adjuvant therapy based on
`the biological characteristics of the individual tumor.
`Fulvestrant is an estrogen receptor (ER) antagonist with
`no known agonist effects. Data from the recent COmpa1isoN
`of Fulvestrant In Recurrent or Metastatic breast cancer
`
`(CONFIRM) study showed that a high-dose regimen of
`fulvestrant 500 mg was associated with a significantly longer
`progression-free survival than the 250 mg regimen (hazard
`ratio [HR] = 0.80, 95% confidence interval [CI] 0.68-0.94;
`p = 0.006), corresponding to a 20% reduction in the risk of
`progression [5]. These data led to the approval of fulvestrant
`500 mg (500 mg on day 0, 14, 28, and every 28 days
`thereafter) for the treatment of postmenopausal women with
`locally advanced or metastatic breast cancer who have pro-
`gressed or recurred after prior anti-estrogen therapy. In the
`first-line setting, the randomized Phase II Fulvestrant flRst-
`line Study comparing endocrine Treatments (FIRST) study
`demonstrated that fulvestrant 500 mg is at least as effective
`as anastrozole in terms of clinical benefit
`(odds ratio
`[OR] = 1.30, 95% CI 0.72-2.38; p : 0.386), and has a
`similar objective response rate (36.0 vs. 35.5%, respectively)
`[11]. In a preplanned follow—up analysis reporting mature
`data, time to progression was 23.4 months for fulvestrant
`500 mg compared with
`13.1 months
`for anastrozole
`(HR = 0.66, 95% CI 0.47—O.92;p = 0.01) [13].
`Two presurgical studies have previously shown that
`treatment with fulvestrant
`leads
`to a dose-dependent
`downregulation of ER, depletion of the ER-regulated pro-
`tein progesterone receptor (PgR), and reduction in prolif-
`erative activity as indicated by the Ki67 labeling index 0.1)
`with doses up to 250 mg [4, 12]. It was expected, therefore,
`that neoadjuvant therapy with a high-dose regimen of ful-
`vestrant 500 mg would further increase biological activity
`on ER expression, function, and growth.
`Against this background, the current study was designed
`to evaluate the effects of neoadjuvant endocrine therapy
`
`@ Springer
`
`with fulvestrant 500 mg and fulvestrant 250 mg in terms of
`biological
`activity (Ki67 LI, ER,
`and PgR),
`tumor
`response, and tolerability in postmenopausal women with
`locally advanced breast cancer.
`Imaging
`The ChromaVisionTM Automated Cellular
`System (ACIS) used in this study is an image analysis
`system that can detect and count individual pixels of two
`chromogen colors used to stain histological sections. The
`use of automated image analysis systems has become more
`frequent over recent years, although it has never previously
`been used for biomarker measurement in a fulvestrant trial
`
`setting. Therefore, we also used an established manual
`scoring method (H score). Since the H score method has
`been used effectively in previous fulvestrant studies and
`those of other endocrine agents [12], its use in the present
`study enabled subsequent cross-study comparisons to be
`made when considering the effects of fulvestrant observed
`here in the neoadjuvant setting.
`
`Patients and methods
`
`Study design and patients
`
`NEWEST (Neoadjuvant Endocrine Therapy for Women
`with Estrogen-Sensitive Tumors; 9238lL/0065) was a
`randomized, open-label, multicenter, Phase
`II
`smdy
`involving postmenopausal women with newly diagnosed,
`ER-positive,
`locally advanced breast cancer who had
`received no prior breast cancer treatment (NCT0093002).
`Eligible patients (intent-to-treat
`[ITT] population) were
`randomly assigned to receive neoadjuvant treatment with
`either fulvestrant 500 mg/month (plus 500 mg on day 14 of
`month 1) or
`fulvestrant 250 mg/month for 16 weeks
`immediately before surgery.
`Women had to be postmenopausal (260 years old, or age
`345 years with amenorrhea for 212 months, or follicle-
`stimulating hormone and estradiol levels within postmeno-
`pausal range, or prior bilateral oophorectomy). Other key
`inclusion criteria were: histologically or cytologically con-
`firmed invasive breast cancer; ER-positive disease as
`determined locally; operable or potentially operable locally
`advanced tumor (T2, 3, 4b, N(L3, M0); tumor size 22 cm;
`willingness to undergo biopsy procedures and surgery; and
`World Health Organization performance status (L2. Key
`exclusion criteria were: any previous treatment for breast
`cancer; inoperability; multifocal disease (>2 major tumor
`nodules); presence of metastatic disease; other current
`malignancy or prior malignancy within the previous 3 years;
`abnormal laboratory values; any severe concurrent condi-
`tion; history of bleeding diathesis or need for long-term anti-
`coagulant therapy; or treatment with a non-approved or
`experimental dmg within 4 weeks of randomization.
`
`AstraZeneca Ex. 2077 p. 2
`
`
`
`Breast Cancer Res Treat (2012) 133:237—246
`
`239
`
`All patients provided written informed consent prior to
`registration. The study was conducted in accordance with
`the Declaration of Helsinki and with local ethics committee
`
`approval at each participating center (36 centers in Austria,
`Brazil, Germany,
`India,
`the United Kingdom, and the
`United States).
`
`Treatment
`
`Eligible patients were randomized 1:1 to receive either
`fulvestrant 500 mg or fulvestrant 250 mg for 16 weeks
`preceding the surgery. Fulvestrant 500 mg was given as
`two 5-mL intramuscular (IM) injections, one in each but-
`tock, on days 0, 14, 28, and every 28 days thereafter for
`16 weeks. Fulvestrant 250 mg was given as one 5-mL IM
`injection, in the buttock, on days 0, 28, and every 28 days
`thereafter for 16 weeks. Patients in the fulvestrant 250 mg
`arm did not receive additional (fulvestrant placebo) injec-
`tions. At the completion of 16 weeks of treatment, patients
`underwent definitive surgery (lumpectomy or mastectomy).
`
`Study objectives
`
`The primary objective of the study was to compare the
`effects of fulvestrant 500 and 250 mg on expression of the
`proliferation marker Ki67 after 4 weeks of treatment.
`Secondary objectives included: effects on ER and PgR
`expression,
`tumor response, and tolerability; effects on
`endometrial thickness and uterine dimensions; effects on
`serum markers of bone turnover (bone-specific alkaline
`phosphatase [ALP], C-terminal
`telopeptides of type-1
`collagen [CTX—l], and procollagen type 1 N propeptide
`[PINP]); and downstaging assessed by a comparison of the
`actual surgery performed at 16 weeks with the likely sur-
`gery predicted at study entry. Ki67 index and ER and PgR
`expression were also assessed at 16 weeks to monitor for
`sustained fulvestrant activity.
`
`MIB-1 anti-Ki67 antibody, the IDS anti-ER antibody or the
`PgR 636 anti-PgR antibody (all supplied by Dako, Ely, UK).
`Binding of the primary antibodies was visualized using an
`avidin—biotin complex and the chromogen 3,3’-diamino-
`benzide. The sections were lightly counterstained with
`hematoxylin before being dehydrated and mounted. Quality
`control slides were included in all assays to ensure consis-
`tency.
`In the first
`instance,
`the immunohistochemical
`staining of the tissues was assessed using the ChromaVi—
`sionTM ACIS. This system detects and determines the
`intensity and counts individual pixels of the two chromogen
`colors used in the immunohistochemical procedures (in this
`case, brown = positive; blue = negative). Wherever pos-
`sible, ten representative fields across each tumor specimen
`were scored; in cases where ten fields could not be obtained,
`every available tumor cell was included in the analysis. The
`Ki67 LI was defined as the percentage of tumor cell nuclei
`positively stained with intensity above a predetermined
`threshold. In the case of ER and PgR, the mean intensity as
`well as the percentage of positively stained nuclei was
`calculated and combined to produce a proprietary histo-
`score. ER and PgR expression were also assessed retro-
`spectively on the same stained tissue samples using the
`H score method which is derived by microscopic assessment
`of the percentage of tumor cells in each of five staining
`categories
`(negative, very weak, weak, moderate and
`strong) to give an H score ranging from 0 to 300 [8, 12]. This
`assessment was performed at the Tenovus Centre for Cancer
`Research by two experienced observers (JMWG and PF)
`who were blinded to the ACIS and clinical outcome data
`
`and reached a consensus for each slide. Sequential samples
`from each patient were evaluated at the same time to ensure
`comparative assessment of tumor histology wherever pos-
`sible. To ensure the analysis was robust, only paired Sarn-
`ples for both the ACIS and H score methods were included.
`Any samples with non-specific staining or unacceptably low
`cellularity were eliminated from analysis.
`
`Study assessments
`
`Assessment of clinical response
`
`Assessment of biomarkers
`
`Core biopsies, using an 11- to 14-gauge needle, were taken
`at baseline, at week 4, and at surgery (week 16). These
`tumor cores were routinely formalin—fixed and paraffin—
`embedded locally, with central
`immunohistochemical
`assessment of changes in Ki67, ER, and PgR expression at
`each time point using well-established methods [7]. Briefly,
`5-um sections of pre- and post-treatment tissue samples
`were dewaxed in xylene and rehydrated through graded
`alcohols after which endogenous peroxidase was blocked.
`Following heat-mediated antigen retrieval and blocking of
`non-specific binding, the sections were incubated with the
`
`During the 16-week treatment phase, patients underwent
`clinical breast examination every 4 weeks. Tumor volume
`was measured by 3D ultrasound at baseline, week 4, and
`after 16 weeks of treatment before definitive surgery.
`Optional
`tumor measurements by magnetic resonance
`imaging (MRI) were obtained at baseline and 16 weeks.
`Tumor response was defined as complete response (dis-
`appearance of all lesions), partial response (365% reduc-
`tion in tumor volume by
`3D ultrasound), disease
`progression (373% increase in tumor volume), or stable
`disease (neither partial response nor disease progression)
`[17]. Objective responders were those patients with a
`complete response or partial response.
`
`@ Springer
`
`AstraZeneca Ex. 2077 p. 3
`
`
`
`240
`
`Statistical analysis
`
`Sample size calculation was based on the primary endpoint.
`Based on a 5.36% (::0.616) reduction in Ki67 values fol-
`lowing treatment with fulvestrant 250 mg in Study 018
`(which compared the short—term biological effects of ful-
`vestrant vs. tamoxifen) [12], a sample size of 80 patients
`per group would provide 80% power to detect a difference
`of 0.274 in log-transformed Ki67 values at 4 weeks for
`fulvestrant 500 mg relative to 250 mg at the two-sided,
`p : 0.05 significance level. Data for the efficacy endpoints
`were analyzed and summarized on an ITT basis. Treatment
`differences in Ki67 LI between fulvestrant 500 and 250 mg
`were assessed using analysis of Variance (ANOVA),
`modeling natural log-transformed changes from baseline
`Ki67 L1 to Ki67 Ll at week 4. Also, a post hoc ANOVA
`was used to assess the effects of fulvestrant 500 mg and
`fulvestrant 250 mg on ER and PgR expression derived by
`the ACIS method. For easier interpretation of the data,
`treatment effects (least squares mean and Cls) were back-
`transformed and expressed as percentages. Mean percent-
`age changes in H scores were calculated from baseline to
`weeks 4 and 16 using the manually derived score data.
`Differences in tumor response were analyzed using logistic
`regression. The safety population consisted of all patients
`who received at least one dose of study drug. Only patients
`with a baseline endometrial thickness 55 mm were inclu-
`ded in the statistical analysis of this safety endpoint.
`
`Talerability
`
`The frequency and severity of adverse events (AEs) were
`recorded throughout the study and up to 8 weeks after the
`last
`injection. Changes
`from baseline in endometrial
`thickness
`and uterine dimensions were
`assessed at
`
`16 weeks using transvaginal ultrasound (in all patients with
`an intact uterus). Patients with apparent thickening of the
`endometrium (>5 mm) or with suspicious ovarian findings
`were referred to a gynecologist
`for advice, but were
`allowed to continue the study unless the investigator
`decided otherwise. Serum was collected for analysis of
`bone CTX—l (a marker of bone resorption) and of both
`ALP and PINP (markers of bone formation), which were
`assessed twice at baseline (before randomized treatment),
`then every 4 weeks until surgery.
`
`Results
`
`Patients
`
`Breast Cancer Res Treat (2012) 133:237-246
`
`receive fulvestrant 250 mg. The first subject was enrolled
`on 7 February 2005 and the last subject completed the
`study on 9 July 2007. Patient disposition throughout the
`study is shown in Supplemental Fig. 1. Overall, 990% of
`patients had ER-positive disease and only one patient in
`each group had unknown ER status. Patient demographics
`and characteristics
`at baseline were similar between
`
`groups, as outlined in Table 1. The mean age of patients
`enrolled was 67 years and 85.3% were Caucasian.
`
`Biological activity
`
`Fulvestrant 500 mg reduced mean Ki67 L1 to a signifi-
`cantly greater extent
`than fulvestrant 250 mg (mean
`
`Table 1 Patient demographics and characteristics at baseline
`
`Mean age, years (range)
`Age category, It (%)
`<65 years
`365 years
`Race (%)
`Caucasian
`Black
`Oriental
`Other
`WHO performance status (%)
`Unknown
`0
`1 or 2
`ER/PgR status (%)
`ER-1-/PgR+
`ER-1-/PgR—
`ER or PgR unknown
`Primary tumor stage (%)
`T2
`T3/T4b
`Unknown
`Tumor grade (%)
`1
`2
`3
`Unassessable, missing or not done
`Intact uterus, (%)
`Yes
`No
`Unknown
`
`Fulvestrant,
`500 mg
`(n : 109)
`
`Fulvestrant,
`250 mg
`(n : 102)
`
`66.9 (47-94)
`
`66.8 (47-87)
`
`46 (42.2)
`63 (57.8)
`
`92 (84.4)
`5 (4.6)
`1 (0.9)
`11 (10.1)
`
`2 (1.9)
`19 (17.4)
`88 (80.7)
`
`76 (69.7)
`23 (21.1)
`10 (9.2)
`
`53 (48.6)
`55 (50.5)
`1 (0.9)
`
`12 (11.0)
`56 (51.4)
`18 (16.5)
`23 (21.1)
`
`87 (79.8)
`16 (14.7)
`6 (5.5)
`
`44 (43.1)
`58 (56.9)
`
`88 (86.3)
`3 (2.9)
`3 (2.9)
`8 (7.8)
`
`2 (2.0)
`16 (15.7)
`84 (82.4)
`
`72 (70.6)
`20 (19.6)
`10 (9.8)
`
`51 (50.0)
`50 (49.0)
`1 (10)
`
`9 (8.8)
`52 (51.0)
`21 (20.6)
`20 (19.6)
`
`82 (80.4)
`14 (137)
`6 (5.9)
`
`A total of 211 women were included in the study; 109 were
`randomized to receive fulvestrant 500 mg and 102 to
`
`ER estrogen receptor, PgR progesterone receptor, WHO World Health
`Organization
`
`@ Springer
`
`AstraZeneca Ex. 2077 p. 4
`
`
`
`Breast Cancer {es Treat (2012) 133:237—246
`
`Week 4
`
`Week 16
`
`
`Mean%change
`
`
`
`frombaselineinKi67labelingindex
`
`(ACIS)
`
`.10 _
`.20 _
`.30 _
`.40 _
`
`-50 -
`-so —
`-7o —
`
`-so—
`
`'90 ‘
`
`n
`
`as
`
`,,=6O
`
`p < 0.0001
`
`Fulvestrant 500 mg
`Fulvestrant 250 mg
`
`(7:26
`
`Fig. 1 Effects of fulvestrant 500 mg and fulvestrant 250 mg on Ki67
`labeling index after 4 and 16 weeks of treatment (intent—to—treat).
`ACIS Automated Cellular Imaging System
`
`Table 2 Effects of fulvestrant 500 mg and fulvestrant 250 mg on
`Ki67 labeling index after 4 weeks of treatment (ITT)
`Fulvestrant
`Fulvestrant
`500 mg
`250 mg
`(n = 109)
`(n = 102)
`
`241
`
`longer treatment time point. Fulvestrant 500 mg reduced
`mean PgR expression to a greater extent than fulvestrant
`250 mg at week 4 (Fig. 2c, d; Table 3). These differences
`reached statistical significance using the H score method
`(-80.5 vs. -46.3%; p : 0.0018; Table 3; Fig. 2d) but
`were not statistically significant according to ACIS data
`(-22.7 vs.
`-17.6%; Table 3; Fig. 2c). At week 16,
`decreases in PgR were observed relative to baseline, but
`there was no significant difference in PgR expression for
`fulvestrant 500 mg compared with fulvestrant 250 mg
`using either ACIS (-29.2 vs. -30.5%; Fig. 2c) or H score
`methods (-88.0 vs. -84.5%; p = 0.6445; Fig. 2d).
`
`Clinical activity
`
`At weeks 4 and 16, tumor response rates in the ITT popu-
`lation were numerically higher with fulvestrant 500 mg than
`with fulvestrant 250 mg (17.4 vs. 11.8% at week 4 and 22.9
`vs. 20.6% at week 16, respectively) (Table 4). In a post hoc
`analysis of evaluable patients with abaseline and a 16-week
`assessment (n : 69 in both arms), tumor response rates were
`36.2 and 30.4% for fulvestrant 500 mg and fulvestrant 250
`mg, respectively (Table 4). Overall, only 13% of evaluable
`patients progressed during the 16 Weeks of therapy (fulve-
`strant 500 mg n = 8; fulvestrant 250 mg n = 10).
`
`60
`-78.8
`
`63
`-47.4
`
`-70.8 to -84.6
`-17.5
`
`-28.6 to -61.3
`-10.5
`
`Tolerability
`
`Evaluable patientsa, n
`Mean percent reduction from
`baseline
`95% CI
`Absolute reduction from
`baseline
`95% CI
`
`-15.7 to -18.8
`
`-6.3 to -13.6
`
`p valueb
`
`<0.0001
`
`Ki67 labeling index was determined by ChromaVisionTM Automated
`Cellular Imaging System (ACIS)
`CI confidence interval, ITT intent—to—treat
`3 Patients for whom data were available at both baseline and 4-week
`time points
`b From ANOVA, modeled on the natural log-transformed change
`from baseline with treatment as a model term
`
`-47.4%,
`-78.8 vs.
`from baseline:
`percent change
`p < 0.0001) after 4 weeks of treatment (Fig. 1; Table 2).
`This corresponded with a significantly greater reduction in
`mean ER expression at week 4 for fulvestrant 500 mg
`compared with fulvestrant 250 mg using both ACIS and
`H scoring methods G3ig. 2a, b; Table 3). However,
`the
`magnitude of reduction caused by fulvestrant 500 mg
`detected by H score (-50.3%) was greater
`than that
`detected by ACIS (-25.0%). At week 16, reductions in
`mean Ki67 LI (-77.4 vs. -628%; Fig. 1) as well as mean
`ER expression by ACIS (-36.5 vs. -31.3%; Fig. 2a) and
`H score (-45.2 vs. -56.1%; Fig. 2b) were observed for
`both fulvestrant 500 mg and fulvestrant 250 mg, but the
`differences between the doses were not significant at this
`
`In total, 208 patients were eligible for assessment of tol-
`erability. Both treatments were well
`tolerated over the
`16-week treatment period. Treatment-related AEs were
`experienced by 37.4 and 30.7% of patients and treatment-
`related serious AEs by 0.9 and 3.0% of patients in the
`fulvestrant
`500 mg
`and
`fulvestrant 250 mg groups,
`respectively. Only two AEs (one per group) led to with-
`drawal; neither Was thought to be treatment-related (one
`transient ischemic attack; one pulmonary embolism). One
`patient randomized to fulvestrant 250 mg experienced an
`AE leading to death during the posttreatment follow-up
`period that was also not considered to be treatment-related
`(cause of death unknown, possibly cardiac-related). The
`most common AEs are described in Table 5.
`
`Both doses of fulvestrant reduced endometrial thickness,
`with changes after 16 weeks of treatment similar between
`fulvestrant 500 mg and fulvestrant 250 mg groups (Sup-
`plemental Table 1). Serum bone marker levels were similar
`Within and between the two groups throughout the study,
`with neither dose producing substantial changes in any of the
`three bone markers assessed (ALP, CTX, and PINP) (Sup-
`plemental Fig. 2). Few patients reported receiving prior
`medications
`(bisphosphonates, corticosteroids, hormone
`replacement therapy) that might confound interpretation of
`bone or endometrial data (fulvestrant 500 mg: 8 patients;
`
`@ Springer
`
`Astrazeneca Ex. 2077 p. 5
`
`
`
`242
`
`Breast Cancer Res Treat (2012) 133237-246
`
`300
`250
`200
`
`(a)
`E on
`E E
`g L0
`g N
`‘<6’
`'8
`9 as
`-- 01$
`C” E ‘ 150
`3 0 V
`-E o 2
`g L0
`O E
`E g
`‘D $
`E >
`II 2LIJ
`
`100
`
`50
`O
`
`--
`
`-
`
`’
`
`.
`
`"
`
`-
`
`T
`
`Fulvestrarr 500 mg
`Fulvestrarr 250 mg
`
`_
`
`T
`_._
`L
`
`__
`
`--
`__
`"
`n=108 n=101 n=60 n=63 n=26 n=3i
`0 weeks
`4 weeks
`i6 weeks
`Time
`
`(b)
`
`
`
`
`
`ERindexfollowingtreatmentwith
`
`300
`N01O
`
`NoO
`
`01O
`
`0
`
`fulvestrant500mgand250mg
`
`(Hscore)
`
`I
`
`-I
`
`Fulvestrant 500 mg
`Fulvestrant 250 mg
`
`l
`
`n=108 n=101 n=59
`n=63 n=23 n=25
`0 weeks
`4 weeks
`16 weeks
`Time
`
`***p < 0.001 (fulvestrant 500 mg vs 250 mg)
`
`***p < 0.001 (fulvestrant 500 mg vs 250 mg)
`
`(C)
`
`(d)
`
`
`
`PgFlindexfollowingtreatmentwith
`
`
`
`
`
`300
`250
`
`200
`150
`
`100
`
`-0
`
`fulvestrant500mgand250mg (ACIS)
`
`Fulvestrant 500 mg
`Fulvestrant 250 mg
`
`_
`
`__
`
`'
`
`.-
`
`--
`
`*
`
`T—.—
`
`l
`-
`n=76 n=72 n=43 n=45 n=19 n=2i
`4 weeks
`16 weeks
`0 weeks
`Time
`
`
`
`PgFlindexfollowingtreatmentwith
`
`
`
`
`
`300
`250
`
`200
`
`1 so
`
`100
`
`50
`
`0
`
`fulvestrant500mgand250mg
`
`(Hscore)
`
`Fulvestrant 500 mg
`Fulvestrant 250 mg
`
`u
`
`T
`
`.-.
`
`"
`
`__
`__
`..
`n=45 n=17 n=16
`n=76 n=72 n=42
`0 weeks
`4 weeks
`16 weeks
`Time
`
`**p < 0.01 (fulvestrant 500 mg vs 250 mg)
`
`Fig. 2 Effects of fulvestrant 500 mg and fulvestrant 250 mg after 4
`and 16 weeks of treatment
`(intent—to—treat) on: estrogen—receptor
`expression by a ACIS and b H score, and progesterone receptor
`
`expression by c ACIS and d H score. ACIS Automated Cellular
`Imaging System; ER estrogen receptor; PgR progesterone receptor.
`Horizontal line in box plots represents the median Value
`
`fulvestrant 250 mg: 13 patients). Overall, AEs were con-
`sistent with the known toxicity profile of fulvestrant and no
`new safety concerns were identified.
`
`Discussion
`
`NEWEST is the first study to compare both the biological
`and clinical activity of fulvestrant 500 and 250 mg regi-
`mens in the neoadjuvant breast cancer setting. Fulvestrant
`500 mg reduced mean Ki67 L1 to a significantly greater
`extent than fulvestrant 250 mg at week 4 (p < 0.0001)—the
`primary endpoint. There was also a significantly greater
`reduction in ER (p < 0.0001) and PgR (p = 0.0018)
`expression at week 4 with fulvestrant 500 mg using the
`H score method. These data provide the first indication that
`fulvestrant 500 mg has
`significantly greater biological
`activity on proliferative activity and ER and PgR expression
`than fulvestrant 250 mg in this setting. The results are
`consistent with the dose-dependent clinical effect
`for
`
`fulvestrant, with increased clinical efficacy recently repor-
`ted for the higher fulvestrant 500 mg dose regimen [5, 14].
`At week 16, similar degrees of reduction in Ki67 LI, ER,
`and PgR expression were observed with both doses. How-
`ever, the lower numbers of paired samples available at this
`time point (16-31; Fig. 2) and the resulting loss of statis-
`tical power to detect a treatment difference may be con-
`founding the current analysis. This study required adequate
`tumor biopsies to be taken at the specified time points
`(baseline, week 4, and at surgery), and only paired samples
`were analyzed. Since many centers were reluctant to permit
`taking more than two or three cores at each time point, and
`some cores contained limited tumor material, biomarkers
`could unfortunately not be measured on every patient at
`every time point. In addition, 18 patients in the fulvestrant
`500 mg group and 23 in the fulvestrant 250 mg group did
`not undergo surgeiy at week 16 as per protocol, which also
`contributed to the low number of paired samples available.
`The greater effect of fulvestrant 500 mg on Ki67, ER,
`and PgR at 4 weeks may be at least partly attributable to
`
`@ Springer
`
`AstraZeneca Ex. 2077 p. 6
`
`
`
`Breast Cancer Res Treat (2012) 133:237—246
`
`Table 3 Effects of fulvestrant 500 mg and fulvestrant 250 mg on ER and PgR expression after 4 weeks of treatment (ITT)
`ACIS
`H score
`
`ER
`
`Evaluable patientsa, n
`Mean percent reduction from baseline
`95% CI
`
`p valueb
`PgR
`Evaluable patientsa, n
`Mean percent reduction from baseline
`95% CI
`
`p valueb
`
`Fulvestrant
`500 mg
`(n = 109)
`
`Fulvestrant
`250 mg
`(n = 102)
`
`Fulvestrant
`500 mg
`(n = 109)
`
`60
`-25.0
`-21.0 to -28.8
`
`0.0002
`
`43
`-22.7
`-9.5 to -33.9
`
`0.5677
`
`63
`-13.5
`-9.0 to -17.8
`
`45
`-17.6
`-4.0 to -29.3
`
`58
`-50.3
`-39.9 to -58.9
`
`<0.0001
`
`31
`-80.5
`-69.5 to -87.6
`
`0.0018
`
`243
`
`Fulvestrant
`250 mg
`(n = 102)
`
`60
`-13.7
`-4.0 to -28.4
`
`34
`-46.3
`-17.5 to -65.0
`
`ACIS Automated Cellular Imaging System, CI confidence interval, ER estrogen receptor, ITT intent—to—treat, PgR progesterone receptor
`3 Patients for whom data were available at both baseline and 4-week time points
`b From ANOVA, modeled on the natural log—transformed change from baseline with treatment as a model term
`
`Table 4 Tumor response by 3D ultrasound (ITT)
`
`Time of assessment
`
`Response n (%)
`
`Fulvestrant 500 mg
`
`ITT population
`(n = 109)
`
`Evaluable for
`response
`
`Fulvestrant 250 mg
`_
`ITT population
`(n = 102)
`
`Evaluable for
`response
`
`Odds ratio (CI)
`and p value
`
`Week 4
`
`Not determined‘
`
`12 (11.0)
`
`Objective responseb
`Stable disease
`Disease progression
`Not evaluable
`
`19 (17.4)
`65 (59.6)
`7 (6.4)
`6 (5.5)
`
`Week 16
`
`Not determined‘
`
`40 (36.7)
`
`Objective responseb
`Stable disease
`Disease progression
`Not evaluable
`
`25 (22.9)
`32 (29.4)
`8 (7.3)
`4 (3.7)
`
`CI confidence interval, ITT intent—to—treat
`
`97 (100)
`(19.6)
`(67.0)
`(7.2)
`(6.2)
`
`69 (100)
`(36.2)
`(46.4)
`(11.6)
`(5.8)
`
`7 (6.9)
`
`12 (11.8)
`77 (75.5)
`6 (5.9)
`0 (0.0)
`
`33 (32.4)
`
`21 (20.6)
`38 (37.3)
`10 (9.8)
`0 (0.0)
`
`95 (100)
`(12.6)
`(81.1)
`(6.3)
`(0.0)
`
`69 (100)
`(30.4)
`(55.1)
`(14.5)
`(0.0)
`
`1.68 (0.77—3.70)
`p = 0.1933°
`
`1.30 (0.64—2.64)
`p = 0.4705°
`
`3 For patients who did not have data at baseline or time of assessment; bpatients with complete or partial responses (all but one patient (250 mg
`group, week 16) had partial responses); Cfrom logistic regression, modeled on objective response rate, with treatment as a model term
`
`the loading element of the fulvestrant 500 mg regimen,
`since steady-state plasma levels should have been reached
`at 4 weeks, whereas steady-state levels would not have
`been reached with fulvestrant 250 mg at this time [10].
`Here, we used two methods (ACIS and H score) of
`assessing ER and PgR expression. While interpretation of
`H score results is dependent on the experience and ability of
`the assigned experts, automated systems such as ACIS have
`been developed with a view to reducing intra- and inter-
`observer variability. However,
`the use of the H score
`
`technique does enable comparison with previous fulvestrant
`studies that have also used this methodology successfully.
`Interestingly, our analyses showed significant differences
`in sensitivity between ACIS and H score. Automated scoring
`by ACIS reduced the spread of individual measurements,
`making it difficult to detect a significant difference in
`treatment effect or to discriminate between fulvestrant
`
`doses, particularly for PgR. In contrast, H score data had a
`wider spread measured on a continuous scale encompassing
`both percentage positive and intensity (without a threshold
`
`@ Springer
`
`AstraZeneca Ex. 2077 p. 7
`
`
`
`244
`
`Breast Cancer Res Treat (2012) 133:237—246
`
`Table 5 Most common adverse events
`patients)
`
`(occurring in >5% of
`
`In terms of clinical benefit, the overall tumor response
`rates in NEWEST were similar between treatment arms and
`
`Fulvestrant 500 mg
`(n = 107) %
`
`Fulvestrant 250 mg
`(n = 101) %
`
`Hot flash
`Injection—site pain
`Fatigue
`Nausea
`Headache
`Hypertension
`Procedural pain
`Pain in extremity
`Cough
`Diarrhea
`Vomiting
`Peripheral edema
`Postoperative wound
`infection
`
`Back pain
`Pyrexia
`
`15 (14.0)
`16 (15.0)
`15 (14.0)
`11 (10.3)
`10 (9.3)
`7 (6.5)
`8 (7.5)
`5 (4.7)
`8 (7.5)
`6 (5.6)
`3 (2.8)
`6 (5.6)
`5 (4.7)
`
`4 (3.7)
`7 (6.5)
`
`10 (9.9)
`4 (4.0)
`5 (5.0)
`7 (6.9)
`8 (7.9)
`8 (7.9)
`5 (5.0)
`7 (6.9)
`4 (4.0)
`6 (5.9)
`8 (7.9)
`5 (5.0)
`6 (5.9)
`
`6 (5.9)
`2 (2.0)
`
`for positivity), and so differences at each time point could
`be more easily detected, with increased discrimination
`between doses. This is consistent with findings reported
`from a previous study (018) involving fulvestrant and using
`H score, where dose-dependent reductions in ER (-39, -
`50, and -59%) and PgR (-12, -52, and -67%) expression
`were reported 2-3 weeks after single doses of fulvestrant
`(50, 125, and 250 mg, respectively) [12]. However,
`the
`magnitude of reduction in ER (-59%) and PgR (-67%)
`expression induced by fulvestrant 250 mg in Study 018 was
`greater than that reported with fulvestrant 250 mg at week 4
`in this study (-13.5 and -46.3% for ER and PgR, respec-
`tively). This is undoubtedly due to differences in baseline
`H score values, which were much higher in NEWEST
`compared with Study 018, and which can be attributed to
`the increased sensitivity of the more recent immunohisto-
`chemical methods for determining ER and PgR status.
`These differences have impact on the magnitude of change
`in expression levels detected and therefore on any sub-
`sequent cross-trial comparisons made. Nonetheless, the data
`reported here support the hypothesis that higher fulvestrant
`doses lead to greater ER downregulation and subsequently
`decreased ER function in postmenopausal women with
`hormone—receptor positive breast cancer. Despite the high
`sensitivity and reproducibility of results reported previously
`with ACIS in both breast [1, 18] and colorectal tissue [9],
`our results suggest that H scoring may be a more suitable
`approach for examining changes in ER and PgR expression
`associated with neoadjuvant endocrine treatment.
`
`@ Springer
`
`were comparable to those reported in other neoadjuvant
`studies (24% for anastrozole and 20% for tamoxifen at
`3 months; 35% for letrozole and 25% for tamoxifen at
`4 months) [6, 16]. Direct comparison between studies is
`challenging because of variability in assessment tech