Randomized Comparison of Tamoxifen and Two
`Separate Doses of Toremifene in Postmenopausal
`Patients With Metastatic Breast Cancer
`
`By Daniel F. Hayes, J.A. Van Zyl, Anne Hacking, Louis Goedhals, W.R. Bezwoda, James A. Mailliard,
`Stephen E. Jones, Charles L. Vogel, Robert F. Berris, Irving Shemano, and John Schoenfelder
`
`Purpose: To perform a randomized three-arm com-
`parison of tamoxifen (TAM; 20 mg/d) and two separate
`doses of toremifene (TOR; 60 mg/d [TOR60] and 200
`mg/d [TOR200]) in postmenopausal patients with hor-
`mone receptor-positive or -unknown metastatic breast
`cancer.
`
`Materials and Methods: Six hundred forty-eight pa-
`tients with hormone receptor-positive or -unknown met-
`astatic breast cancer were randomly assignedto receive
`TAM (n = 215), TOR60 (n = 221), or TOR200 (n = 212).
`Results: The combined response rates (by intent to
`treat) were as follows: TAM, 44%; TOR60, 50%; and
`TOR200, 48%. Complete and partial response rates were
`as follows: TAM, 19%; TOR60, 21%, and TOR200, 23%
`(not statistically different). Median times to progression
`and overall survival were notsignificantly different. Ad-
`verse events(lethal, serious but nonlethal, and important
`
`but non-life-threatening) were similar in all three arms,
`except that patients in the TOR200 arm hada statistically
`significantly increased rate of nausea (37% v 26% and
`26% for TOR200, TAM, and TOR60,respectively; P =
`.027). Quality-of-life assessments were not different
`amongthethree arms.
`Conclusion: The activity, toxicity, and side effects of
`TOR in postmenopausal womenwith hormonereceptor-
`positive or -unknown metastatic breast cancer are simi-
`lar if not equivalent to those of TAM. We detected no
`clear evidence of a dose-response effect for TOR. TOR60
`is an effective and safe agent for the treatment of post-
`menopausal women with hormone receptor-positive
`metastatic breast cancer and can be considered analter-
`native to TAM asfirst-line treatment for such patients.
`J Clin Oncol 13:2556-2566. © 1995 by American So-
`ciety of Clinical Oncology.
`
`REATMENTOF PATIENTSwith metastatic breast
`cancer is palliative and may consist of either local
`or systemic therapies.'? Approximately 50% to 60% ofall
`postmenopausalpatients will have hormonally responsive
`disease.? Of the available hormone therapies, the anties-
`trogen tamoxifen (TAM) is generally considered to be
`the first-line treatment of choice because of its excellent
`efficacy-to-toxicity ratio.'
`Toremifene (TOR) is a triphenylethylene derivative
`that was developed to improve the therapeutic-to-toxic
`ratio of antiestrogens.*” Like TAM, TOR has both anti-
`estrogenic and estrogenic activities in preclinical in vitro
`and in vivo studies.*!° Also, like TAM, TOR binds with
`high affinity to cytoplasmic estrogen receptors.”
`Phase I studies of TOR have demonstrated that it is
`
`From the Dana-Farber Cancer Institute, Boston, MA; Mamma
`Clinic, Cape Town; Groote Schuur Hospital, Cape Town; National
`Hospital, Bloemfontein; University of Witwatersrand, Johannesburg,
`South Africa; Creighton Cancer Center, Omaha, NE; Sammons Can-
`cer Center, Baylor University Medical Center, Dallas, TX; Compre-
`hensive Cancer Research Group,
`Inc, Mt Sinai Comprehensive
`Breast Center, Miami, FL; Colorado Cancer Research Program,
`Denver, CO; and Pharmacia, Inc, Columbus, OH.
`Submitted January 23, 1995; accepted June 5, 1995.
`Supported by Pharmacia, Inc (formerly Adria Laboratories), Co-
`lumbus, OH.
`Address reprint requests to Daniel F. Hayes, MD, Dana-Farber
`Cancer institute, 44 Binney St, Boston, MA 02115.
`© 1995 by American Society of Clinical Oncology.
`0732-183X/95/1310-0013$3.00/0
`
`generally well tolerated, with a clinical toxicity profile
`similar to that of TAM.'' Phase II trials of TOR in
`patients with estrogen receptor (ER)-positive or -un-
`known advanced breast cancer have demonstrated that
`
`doses of 60 and 240 mg/d produced response rates up to
`68%.'*"" A randomized phase II trial of three separate
`doses of TOR (20, 40, and 60 mg/d) suggestedless effi-
`cacy at 20 mg/d, but similar response rates with 40 and
`60 mg/d."*
`Because of the favorable and promising phase I and II
`data, a worldwide phase III trial to compare TOR at two
`doses, 60 mg/d (TOR60) and 200 mg/d (TOR200) with
`TAM at 20 mg/d wasinitiated in November 1988. This
`trial was open to postmenopausal women with ER-posi-
`tive or progesterone receptor (PgR)-positive or ER/PgR-
`unknown tumors with measurable or assessable meta-
`
`static breast cancer. In this trial, we observed that the
`efficacy and toxicity of TORare similar to those of TAM.
`
`MATERIALS AND METHODS
`
`Patient Selection
`
`Eligible patients included postmenopausal or perimenopausal
`women with a histologically documented prior history of breast
`cancer that was ER- and/or PgR-positive at either the primary tumor
`or metastatic site, or for which the ER and PgR status were unknown.
`Patients must have had either bidimensionaily measurable metastatic
`breast cancer or assessable lytic bone metastases. Patients may have
`had prior adjuvant chemotherapy, but could not have had prior hor-
`mone or cytotoxic chemotherapy for recurrent/metastatic disease,
`although TAM treatment for < 14 days before entry was allowed.
`Prior adjuvant TAM wasallowed,but the interval between discontin-
`uation of adjuvant TAM treatment and relapse/entry onto trial was
`
`2556
`
`Journalof Clinical Oncology, Vol 13, No 10 (October), 1995: pp 2556-2566
`
`Downloaded from ascopubs.org by 151.194.33.114 on February 17, 2017 from 151.194.033.114
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`AstraZeneca Exhibit 2039 p. 1
`InnoPharma Licensing LLC v. AstraZeneca AB IPR2017-00904
`Fresenius-Kabi USA LLC v. AstraZeneca AB IPR2017-01910
`
`

`

`TOREMIFENE v TAMOXIFEN IN METASTATIC BREAST CANCER
`
`2557
`
`required to be = 12 months. Patients must have had a performance
`status of 0 to 2 by the ECOGscale. Patients were not eligible for
`the trial if they were actively menstruating, had serum bilirubin
`levels more than 2 mg/dL or AST levels = 100 U/L,orif they had
`brain metastasis, inflammatory breast carcinoma, or lymphangitic
`pulmonary metastasis. Patients were also excluded if they had had
`asecond primary malignancy within the 5 years preceding entry onto
`this trial. Signed, informed consent was obtained before enrollment.
`Ofnote, accrual wasoriginally initiated for two separate but iden-
`tical multiinstitutional studies. However, since they were performed
`under the auspices of a single sponsor, and since entry criteria,
`protoco] conduct, data management, and auditing were identical, the
`resulis have been combined and analyzed as a single study.
`Patients were stratified by whether they had bone-only metastases
`(with or without other nonmeasurable disease) or nonbony assessable
`disease and were randomly assigned to treatment with TAM at 20
`mg orally per day (TAM), TOR60orally, TOR200orally. Following
`baseline evaluations, patients were reevaluated every 8 weeks, in-
`cluding history, physical examination, ocular examination, perfor-
`mance status, chest radiograph, diagnostic imaging of previously
`documented sites of disease (bone scintigraphy, bone radiographs,
`and liver imaging), complete blood cell counts, serum chemistries,
`antithrombin-III (ATIDlevels, and subjective patient ratings (Visual
`Analog Scale [VAS] and pain assessment and analgesic require-
`ments). Serial assessments of initially detected tumor sites were
`continued every 8 weeks for 48 weeks, and then every 12 weeks.
`Serial assessments of bone disease were performed every 16 weeks
`in patients with known bonydisease if no increase in bone symptoms
`or serum calcium was noted.
`Patients with measurable disease were assessed by their primary
`care physician at each evaluation to have a complete response or
`partial response,
`to be stable, or to have progressive disease ac-
`cording to World Health Organization (WHO)criteria.’® Quality-of-
`life assessments were evaluated using several parameters, as follows:
`serial changes in Eastern Cooperative Oncology Group (ECOG)
`performance status; mood, pain, and enjoyment of life subjective
`analysis by VAS; analgesic requirement as assessed by the ECOG
`Analgesics Requirement Scale; tumor-specific symptoms; solicited
`clinical toxicities; other treatment-emergent symptoms; and serious
`adverse events. These data were collected prospectively as part of
`the required clinical data reporting.
`
`Patients
`
`(TAM, n = 1; TOR6O, n = 2; TOR200, n = 4); seven refused to
`continue or were lost to follow-up evaluation (TAM, n = 4; TOR60,
`n = 3; and TOR200, n = 0); and four did not comply with the
`treatment regimen (TAM, n = 2; TOR60, n = 1; and TOR200, n
`= 1). (3) Progressive disease/changing clinical course or insufficient
`data: 48 patients were treated, but follow-up time or data collection
`wasinsufficient to evaluate response. In eight patients, disease pro-
`gressed within 4 weeks of entry (TAM, n = 4; TOR60, n = 3; and
`TOR200, n = 1). Ninepatients suffered early death on study (TAM,
`n = 2; TOR60, n = 3; and TOR200, n = 4). Two received radiation
`therapy to only assessable lesion within the first 8 weeks (TAM, n
`= 2; TOR60, n = 0; and TOR200, n = 0). Twenty-nine patients
`were not assessable for response because insufficient data were col-
`lected to assess this end point (TAM, n = 9; TOR60, n = 12;
`and TOR200, n = 8). (4) Physician-related protocol violations: two
`patients were not assessable due to major protocol violation. One
`patient on TAM wasincorrectly treated, and a second patient on
`TAM was removed from study at her physician’s discretion.
`Response rates are provided for all patients by intent to treat and
`only for assessable patients. All other data are presented for all
`patients on study by intent to treat.
`The three treatment arms were similar regarding race, ER and
`PgR content, site of dominant disease, disease-free interval between
`primary diagnosis and first recurrence, and performance status
`(Table 1).
`
`Statistical Analysis
`Data were double-key—entered and subjected to both manual and
`computerized checks for logic and consistency before being made
`available for statistical analysis. Since the primary objective of the
`study was to compare each of the TOR groups with the TAM group,
`the treatment group comparisons (TOR60 v TAM and TOR200 v
`
`Table 1. Patient Characteristics by Intent to Treat
`Treatment Arm
`TOR6O
`{n = 221)
`No.
`%
`
`TAM
`{n = 215)
`No.
`%
`
`Characteristic
`
`TOR200
`(n = 212}
`No.
`%
`
`Age, years
`Mean
`Range
`Mean no.of metastatic organ
`sites
`
`él
`35-85
`1.72
`
`6.1
`
`63
`37-88
`1.85
`
`5.6
`
`62
`40-85
`1.77
`
`6.4
`
`Six hundred forty-eight patients with metastatic breast cancer were
`enrolled onto this trial at 129 sites in six countries and randomized
`Disease-free interval (mean years}
`to oneof three arms (TAM, n = 215; TOR60, n = 221; and TOR200,
`Dominantsite of disease
`n = 212). Accrual began November 11, 1988 and was completed
`Visceral*
`August 31, 1991. Of these patients, 546 (84%) were deemed assess-
`Bone
`able for efficacy analysis, as follows: TAM, 172 patients (80%);
`Soft tissue
`TOR60, 187 patients (85%); and TOR200, 187 patients (88%). One
`ERlevel (fmol/mg)
`hundred two patients were considered nonassessable for response
`8
`16
`6
`14
`10
`21
`< 10
`evaluation for the following reasons: (1) Administrative: 34 patients
`
`
`10-100 79=37 83 38 75 35
`
`
`
`were determined retrospectively to be ineligible according to proto-
`> 100
`51
`24
`64
`29
`48
`23
`col miles. The most commonreasons included negative ER and PgR
`Unknown
`64
`30
`60
`27
`73
`34
`status, metastatic skin lesions less than 1 cm, patient received prior
`PgRlevel (fmol/mg)
`therapy for metastases, or liver function tests above stated limit
`<10
`(TAM, n = 8; TOR60, n = 10; and TOR200, n = 3). Seven patients
`10-100
`were registered, but never received therapy (TAM, n = 5; TOR60,
`> 100
`n = 0; and TOR200, n = 2). Six patients did not have assessable
`Unknown
`or measurable metastatic disease (TAM, n = 4; TOR60, n = 0; and
`TOR200, n = 2). (2) Received insufficient therapy: 18 patients were
`taken off study very early. Seven suffered an early adverse event
`
`81
`96
`35
`
`38
`45
`16
`
`86
`100
`35
`
`39
`45
`16
`
`82
`90
`38
`
`39
`42
`18
`
`35
`58
`53
`69
`
`17
`37
`«16
`24
`53
`27
`28
`62
`25
`32 & 31
`
`34
`58
`40
`80
`
`16
`27
`19
`38
`
`“Data not available for all patients: TAM, n = 212; TORSO, n = 221;
`TOR200, n = 210.
`
`Downloaded from ascopubs.org by 151.194.33.114 on February 17, 2017 from 151.194.033.114
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`AstraZeneca Exhibit 2039 p. 2
`
`

`

`2558
`
`HAYES ET AL
`
`TAM) were treated as separate and independent. No adjustments
`were made for multiple comparisons. As stated in the protocol,
`efficacy analyses were conducted when (approximately) 70% of en-
`rolled patients had experienced progressive disease. The correspond-
`ing data cutoff date was August 31, 1992. Safety analyses were
`based on all data that were available for analysis as of December7,
`1993.
`
`The pairwise treatment group comparisonswith respect to qualita-
`tive variables were made using either the x’ test or Fisher’s exact
`test. The two-sample ¢ test was used to make comparisons with
`respect to quantitative variables. Lifetime analyses (ie, time to pro-
`gression, response duration, and survival) were made using standard
`Kaplan-Meier methods. All P comparisons were two-sided and were
`conducted at the .05 level of significance.
`
`RESULTS
`
`Therapeutic Outcome
`
`The response rates for the three treatment arms are
`listed in Table 2. Response rates were notstatistically
`different between either of the TOR arms and the TAM
`
`arm, whether they were evaluated by intent to treat or
`by assessable patients only. The overall response rate
`(complete response, partial response, and stable disease)
`was 50%. The combined response rates when evaluated
`by intent to treat for the three arms were as follows:
`TAM, 44%; TOR60, 50%; and TOR200, 48%. If only
`complete responses and partial responses are considered,
`the respective response rates were 19%, 21%, and 23%.
`These differences were notstatistically significant.
`When only assessable patients were analyzed, com-
`bined response rates were 53%, 56%, and 54%, respec-
`tively, for TAM, TOR60, and TOR200. Complete and
`partial response rates were 24%, 24%, and 26%, respec-
`tively, for the three arms.
`
`The median times to progression byintent to treat were
`175 days (5.8 months), 168 days (5.6 months), and 167
`days (5.6 months) for TAM, TOR60, and TOR200,re-
`spectively (Table 3). The differences in the durations of
`the times to progression were notstatistically significant
`(Fig 1).
`Although overall survival was not a primary end point
`of this study, no statistically significant difference was
`observed for patients treated on the three arms (Table 3
`and Fig 2). Median overall survival times were 950 days
`(31.7 months), 1,145 days (8.3 months), and 904 days
`(30.1 months) for patients treated with TAM, TOR60,
`and TOR200, respectively. Patients on the TOR200 arm
`fared slightly less well than those on TOR60 or TAM.
`The hazards ratios for mortality were as follows: TAM/
`TOR60,1.04 (95% confidence interval, 0.76 to 1.42); and
`TAM/TOR200, 0.81 (95% confidence interval, 0.60 to
`1.10). However, these small differences were not statisti-
`cally significant (Wilcoxon P values: TAM/TOR60 = .8
`and TAM/TOR200 = .2).
`The median response duration between treatment arms
`for those patients who had a partial or complete response
`or stable disease (TAM, n = 41; TOR60, n = 47; TOR200,
`n = 48) was determined by intentto treat (Table 3). At the
`time of this analysis, 16, 19, and 19 patients were continuing
`on TAM, TOR60, or TOR200, respectively, without evi-
`dence of disease progression. The median response duration
`from time of randomization was 577 days (19.1 months),
`509 days (16.9 months), and 554 days (18.4 months) for
`the TAM, TOR60, and TOR200 patients, respectively (Fig
`3). The transient difference between the response duration
`curves approached but did not reachstatistical significance,
`
`Table 2. Response Rates for TAM, TOR60, and TOR200
`by Intent to Treat
`Treatment Arm
`TOR6O
`No.
`
`TAM
`No.
`
`%
`
`TOR200
`No.
`%
`
`%
`
`Variable
`
`Table 3. Time to Progression and Overall Survival for Patients on TAM,
`TOR60, and TOR200byIntent to Treat
`Treatment Arm
`TOR60
`No.
`
`TAM
`No.
`
`%
`
`TOR200
`No.
`%
`
`%
`
`Variable
`
`All patients
`Total no.
`Complete response
`Partial response
`Stable disease
`Primary progressive disease
`Notavailable
`Complete + partial response
`Assessable patients
`Total no.
`Complete response
`Partial response
`Stable disease
`Primary progressive disease
`Complete + partial response
`
`215
`11
`30.
`53
`89
`32
`Al
`
`172
`v
`30
`50
`81
`4l
`
`222
`14
`33.
`63
`90
`21
`47
`
`6
`(15
`29
`Ad
`10
`21
`
`212
`1
`37
`53.
`89
`22
`48
`
`5
`«18
`25
`842
`~«#+10
`23
`
`187
`7
`13
`(17
`32.
`59-32
`83
`44
`45
`24
`
`187
`é
`a}
`37,20
`53,
`28
`86
`946
`48
`26
`
`5
`14
`25
`=)
`15
`19
`
`6
`«18
`29
`47
`24
`
`All patients
`Total no.
`
`215
`
`221
`
`212
`
`Median time fo progression
`(days)*
`Progressedt
`Medianoverall survival
`(days)
`Dead
`Responders}
`Total no.
`Median response duration
`{days)*
`Progressedt
`*From randomization.
`TAttime of analysis.
`#Complete and partial responders.
`
`175
`150
`
`950
`81
`
`4l
`
`577
`25
`
`168
`160
`
`72
`
`167
`155
`
`73
`
`1,145
`76
`
`34
`
`904
`95
`
`A5
`
`47
`
`48
`
`509
`28
`
`60
`
`554
`29
`
`=—60
`
`70
`
`38
`
`«61
`
`Downloaded from ascopubs.org by 151.194.33.114 on February 17, 2017 from 151.194.033.114
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`AstraZeneca Exhibit 2039 p. 3
`
`

`

`TOREMIFENE v TAMOXIFEN IN METASTATIC BREAST CANCER
`
`2559
`
`}-—~-~Ahe- -----------------+--------+--+
`
`
`
`(—-O—)
`OTT games meee* (x)
`a
`
`0
`200
`400
`600
`800
`1000
`
`1200
`
`1400
`
`B 1.0
`wn
`
`n ®S
`
`0.8
`
`ou
`2 06
`L=
`
`&a
`
`0.4
`
`oL
`Oo
`§ 0.2
`c
`8.
`© 0.0
`& 0:
`
`Time in Days
`
`Fig 1. Time to progression for patients treated with TAM (+++, n = 215), TOR60 (OOO, n = 221), and TOR200 (xxx, n = 212).
`Postmenopausal hormone receptor-positive or -unknown patients with metastatic breast cancer were randomly assigned to 1 of 3 arms as
`designated. Progression determined from time of study entry (P values: log-rank = .95; Wilcoxon = .96}.
`
`favoring the TAM group over TOR60 (Wilcoxon P value
`for TAM v TOR60 = .08, for TAM v TOR200 = .2; log-
`rank P values = .3 and .3, respectively).
`Outcomes for all patients were analyzed by ER and
`PgR content. As expected, response rates, times to pro-
`gression, and overall survival for patients on each arm
`were superior for ER-positive patients when compared
`with those whose tumors were ER-negative. However, no
`statistically significant differences were observed when
`these outcomes were compared amongthe three treatment
`groups (TAM, TOR60, and TOR200) for patients in the
`following subgroups: ER-, PgR-positive; ER-positive,
`PgR-negative; ER-, PgR-positive; ER-positive, PgR-un-
`known; and ER-, PgR-unknown.
`
`Tumor Fiare
`
`Clinical tumorflare, defined as a transient increase in
`bone and/or musculoskeletal pain requiring an elevated
`analgesic requirement, cutaneous erythema,
`increased
`skin lesion site, and/or hypercalcemia within 2 weeks of
`starting the study drug, has been previously described for
`TAM.” Clinical tumor flare was assessed in 597 of 648
`patients (TAM, n = 192; TOR60, n = 206; and TOR200,
`n= 199), Of these, 105 (17.5%) experiencedclinical flare
`(TAM, 36 of 192 patients [19%], TOR60, 32 of 206
`[16%]; and TOR200 37 of 199 [19%]). None of these
`differences wasstatistically significant.
`
`Adverse Events
`
`Thirty-six patients died on study or within 30 days of
`the last drug dose (TAM, n = 8 [4%]; TOR60, n = 19
`[9%]; and TOR200, n = 10 [5%]). Of these deaths, 19
`werefelt to be secondary to progressive metastatic breast
`cancer (TAM, n = 6; TOR60, n = 9; and TOR200, n =
`4). Other causes of death included hypercalcemia (TAM,
`n = 1; thromboembolism (TAM, n = 1; TORSO, n = 3;
`and TOR200, n = 1); sepsis (TOR60, n = 1); gastrointes-
`tinal bleeding (TOR60, n = 1; and TOR200, n = 1);
`cardiovascular events (TOR60, n = 1; and TOR200, n =
`1); cerebrovascular events (TOR60, n = 1; and TOR200,
`= 1); acute pericarditis (TOR200, n = 1); and causes
`not determined (TOR60, n = 3; and TOR200, n = 1). In
`final analysis, deaths not due to breast cancer (treatment-
`related, possibly treatment-related, or not determined)
`were similarly distributed, with no significant differences
`among the three arms (TAM, n = 2 [1%]; TOR60, n =
`9 [4%]; and TOR200, n = 6 [3%)).
`Serious but nonlethal adverse events are listed in Table
`4. The incidence of thromboembolic events was similar.
`
`Likewise, cardiac events occurred at a similar rate in the
`three treatment arms. In this regard, baseline circulating
`ATIII levels for a selected group of patients (N = 532;
`TAM, n = 172; TOR60, n = 182; and TOR200, n =
`178) were similar (means: 110.2 U/mL, 110.8 U/mL, and
`
`Downloaded from ascopubs.org by 151.194.33.114 on February 17, 2017 from 151.194.033.114
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`AstraZeneca Exhibit 2039 p. 4
`
`

`

`2560
`
`1.0
`
`
`
`SurvivalProbabilityoOoOoO&o©
`
`o ho
`
`od oO
`
`HAYES ET AL
`
`
`
`0
`
`200
`
`400
`
`600
`
`800
`
`1000
`
`1200
`
`1400
`
`Time in Days
`
`Fig 2. Overall survival for patients treated with TAM (+ ++, n = 215), TOR60 (COO, n = 221), and TOR200 [x x x, n = 212). Postmenopausal
`hormonereceptor-positive or -unknownpatients with metastatic breast cancer were randomly assigned to 1 of 3 arms as designated. Survival
`determined from time of study entry (P values: log-rank = .81; Wilcoxon = .74).
`
`112.2 U/mL, respectively). Serial ATIII data were col-
`lected from 347 patients at week 8 and from 240 patients
`at the time they were taken off treatment. Mean ATIII
`levels decreased at week 8 to 98.4 U/mL, 99.2 U/mL,
`and 102.4 U/mL andat time off treatment to 97.5 U/mL,
`99.6 U/mL, and 97 U/mL for TAM, TOR60, and
`TOR200, respectively. The trend of decreasing ATIII lev-
`els after treatment wasnotsignificant for any of the three
`groups(analysis of variancetest forall three groups), and
`there were no significant differences between each of the
`TOR groups and the TAM group(f test).
`Most hepatic abnormalities observed in the TAM or
`TOR6O arms could be related to progressive metastatic
`breast cancer. However, a slightly increased incidence
`of AST abnormalities (= 100 IU/L) not associated with
`progressive disease was noted in the TOR200 arm (10%)
`when compared with the TAM arm (2%). Two patients
`were removed from the TOR200 arm because of marked
`multiple liver function abnormalities. These abnormalities
`were temporally related to initiation of TOR and resolved
`on discontinuation of the drug. Elevations in calcium lev-
`els occurred in 3%, 3%, and 5% of the TAM, TOR60,
`and TOR200 patients, respectively. These differences
`were notstatistically significant.
`important but non—life-threatening side effects that oc-
`curred at any time on study were prospectively assessed
`as part of the protocol (Table 5). Data are available for
`
`most, but notall, of the patients on each arm of the study
`(TAM, 203 of 215 patients, TOR60, 215 of 221; and
`TOR200, 207 of 212). The incidence of hot flashes, vagi-
`nal bleeding, vaginal discharge, peripheral edema, vom-
`iting, and dizziness was similar in all three arms, regard-
`less of whether all reports of these side effects or only
`those of a moderate to severe nature were considered.
`
`Nausea occurred in 20% of women on the TOR200 arm,
`compared with 14% for patients on the TAM and TOR60
`arms (P = .125, Fisher’s exact test).
`The most common unsolicited and subjectively re-
`ported side effects that were assessed by the on-site inves-
`tigators to be possibly related to the drug or of indetermi-
`nate cause were pain and asthenia (TAM, n = 38 [18%];
`TOR6O, n = 52 [24%]; and TOR200, n = 50 [24%)}).
`Others included anorexia, headache, diarrhea, vaginitis,
`rash, pruritis, depression, and insomnia. The incidence of
`all of these was = 5%, and each occurred with similar
`frequencyin al] three arms. Because these were collected
`as Spontaneous nonsolicited comments from the patients,
`and collection depended on the physicians’
`recording
`them, no statistical analysis is provided. However, none
`of these complaints was substantially more common in
`one arm than in the other two.
`
`Ocular abnormalities have been previously reported to
`be associated with the use of TAM, and were observed
`in phase I and phase II trials of TOR.'!?! Therefore, pro-
`
`Downloaded from ascopubs.org by 151.194.33.114 on February 17, 2017 from 151.194.033.114
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`AstraZeneca Exhibit 2039 p. 5
`
`

`

`TOREMIFENE v TAMOXIFEN IN METASTATIC BREAST CANCER
`
`2561
`
`
`
`(—O—)
`
`——a
`
`go 1.0
`
`nC
`
`c S2
`
`al
`
`0.8
`
`0.6
`
`0.4
`
`£02
`o
`
`£ 2
`
`Cc
`©@
`
`io
`
`Oo
`Cc
`oO
`
`Oo
`
`o
`
`o oO
`
`b-~-~~—~-—-—~~Dep~)------------+-------
`
`
`ee eemee eeeea (--x--)
`L—. (+
`
`
`1000
`1200
`1400
`0
`200
`400
`600
`800
`
`‘|
`
`Fig 3. Response duration for patients who responded to TAM (+++, n = 41), TOR60 (OOO, n = 47), and TOR200 (xxx, n = 48). For
`those patients who had a complete or partial response, response duration was determined from the time of onset of the response (P values:
`log-rank = .15; Wilcoxon = .08).
`
`Time in Days
`
`spective ophthalmologic evaluation of all patients was
`required for participation in this study (Table 6). The
`incidence of treatment-emergentcataracts (either present
`at baseline but worsened on study, or appeared on study)
`
`was 7%, 10%, and 7% for patients on TAM, TORGO,
`and TOR200, respectively. These differences were not
`Statistically significant. Visual-field abnormalities and dry
`eyes occurred infrequently in all three arms, and again
`the differences were notstatistically different. Ofinterest,
`two of 215 patients (< 1%) on TAM,four of 221 (2%)
`Table 4. Serious Events on Study (with or without mortality}
`on TOR60,and eight of 212 (4%) on TOR200 developed
`Treatment Arm
`corneal keratopathies. These keratopathies were charac-
`.
`:
`ar
`:
`TOR200
`TORSO
`TAM
`terized by whorl-like patterns on the subepithelial portion
`(n= 212)
`in = 221)
`tn = 215}
`Noe ONO ONO of the cornea and pigment deposition. Thesedid not result
`in visual abnormalities and did not worsen even when
`5
`2
`7
`3
`5
`2
`:
`:
`TOR was continued after the diagnoses were made. In
`2
`0
`0
`.
`.
`‘
`all patients, corneal keratopathies resolved after discon-
`1
`1
`0
`2
`5
`2
`tinuation of treatment.
`0
`1
`0
`In summary, 21 patients (3%) were withdrawn from
`°
`;
`;
`the study because of toxicity (TAM, n = 3; TOR60, n =
`-
`6; TOR200, n = 12). The precise reasons for removal of
`1
`2
`3
`:
`.
`.
`0
`0
`1
`patients from each arm are listed in Table 7.
`1
`2
`J
`.
`.
`Quality-of-Life Assessment
`Data from phase I and II trials that were available
`during the design of this study suggested that the response
`we
`:
`we
`data and toxicity profiles for TOR might be similar to
`those of TAM. Data from this trial confirmed this hypoth-
`esis. Therefore, prospective evaluations of quality-of-life
`measures were performed to documentfurther the pallia-
`tive efficacy of TOR (Table 8). The effect of treatment
`
`Event
`Thromboembolic events
`a.
`a:
`Thrombophlebitis/phlebitis
`Thrombosis
`Pulmonary embolism
`Transientischemic attack
`“ Cerebrovascular accident
`it
`aravac evens
`.
`Myocardialinfarction
`Arrhythmias
`Congestive heart failure
`Elevated liver function tests*
`Alkaline phosphatase
`AST
`Totalbilirubin
`Elevated calciumt
`
`;
`
`;
`
`,
`
`19F
`410
`110
`24
`4 zu 5
`4
`2
`3
`1
`6
`3
`6
`3
`
`35017
`22
`10#
`7
`3
`11
`5
`
`*Abnormality criteria: alkaline phosphatase = 200 IU/L; AST = 100
`IU/L; total bilirubin = 2 mg/dl.
`tAbnormality criteria: calcium > 11.5 mg/dL.
`#P < .05 compared with TAM arm.
`
`Downloaded from ascopubs.org by 151.194.33.114 on February 17, 2017 from 151.194.033.114
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`AstraZeneca Exhibit 2039 p. 6
`
`

`

`2562
`
`Table 5. Prospectively Assessed Side Effects/Toxicities of TAM, TOR60,
`and TOR200 {drug-related or indeterminate cause}
`Treatment Arm
`TOR6O
`{n = 221}
`
`TAM
`(n = 215)
`
`TOR200
`(n = 212)
`
`No.
`
`60
`27
`8
`
`32
`6
`
`11
`3
`
`30
`10
`
`%
`
`28
`13
`4A
`
`15
`3
`
`5
`]
`
`14
`5
`
`2
`1
`
`No.
`
`75
`39
`5
`
`28
`8
`
`11
`7
`
`31
`10
`
`9
`5
`
`%
`
`34
`17
`2
`
`13
`4
`
`5
`3
`
`14
`5
`
`4
`2
`
`No.
`
`63
`28
`5
`
`34
`6
`
`16
`5
`
`43
`12
`
`6
`3
`
`%
`
`30
`13
`2
`
`16
`3
`
`8
`2
`
`20*
`6
`
`3
`]
`
`HAYES ET AL
`
`Table 7. Patients Withdrawn From Protocolfor Toxicities
`TAM
`TOR60
`
`Toxicity
`(n = 215}
`{n = 221)
`3
`6
`
`TOR200
`[n = 212}
`12
`
`All
`
`Hypercalcemia
`Abnormalleft ventricular function
`CNS symptoms*
`Depression
`Gastrointestinal symptomst
`Thrombophlebitis/pulmonary
`embolus
`Myocardial infarction
`Miscellaneous
`
`1
`
`]
`
`1
`
`1
`1
`1
`
`1
`1
`1
`
`4
`2
`4
`1
`
`]
`
`“includes headache, incoordination, ataxia, dizziness, blurred vision,
`and transient ischemic attack.
`tincludes nausea and/or vomiting.
`¥Includes rash and arthritic symptoms.
`
`Toxicity
`Hot flashes
`All
`Moderate/severe
`Vaginal bleeding
`Vaginal discharge
`All
`Moderate/severe
`Edema
`All
`Moderate/severe
`Nausea
`All
`Moderate/ severe
`Vomiting
`All
`Moderate/severe
`Dizziness
`All
`Moderate/severe
`
`4
`3
`
`15
`4
`
`7
`2
`
`20
`7
`
`9
`3
`
`20
`6
`
`9
`3
`
`*P < .05 compared with TAM arm.
`
`regimen on performance status was assessed by con-
`structing Kaplan-Meier survival curves from time of ran-
`domization to nonambulantstatus on study (Fig 4). These
`curves werestatistically identical. At the median follow-
`up time, the percentages of patients who had progressed
`to ECOG performance status 3 or 4 at the time of this
`analysis were 7%, 7%, and 11% for TAM, TOR6Q, and
`TOR200, respectively. These differences were not statis-
`tically significant.
`Subjective assessmentof quality of life was determined
`by means of the VAS, grading from 0 to 10, with 10
`representing the best possible condition. No statistically
`significant differences were observed among the three
`arms for measures of enjoymentoflife, pain, and mood
`
`Measure
`
`Table 6. Ocular Effects/Toxicities of TAM, TOR60, and TOR200
`Treatment Arm
`TOR200
`TOR60
`TAM
`(n = 215)
`{n = 221}
`{n = 212)
`No.
`%
`No.
`%
`No.
`%
`
`16
`7
`22
`10
`15
`7
`Cataract*
`10
`5
`8
`4
`4
`2
`Visual-field abnormality
`2
`1
`A
`2
`8
`4
`Corneal keratopathy
`2
`1
`3
`1
`1
`<1
`Glaucoma
`0
`0
`0
`Retinopathy
`
`Dry eyes 8 16 7 20 9 16
`
`
`
`*Treatment-emergent.
`
`
`
`
`
`(Table 8). Likewise, analgesic requirements for the three
`arms were also similar (Table 8).
`Patients were also assessed for changes in protocol-
`solicited tumor-specific symptoms, including weight loss
`and gain and bonepain,as well as for bone-related perfor-
`mance status and analgesic requirement scores. No statis-
`tically significant differences were noted amongthe three
`arms (data not shown).
`
`DISCUSSION
`
`In this study, we have demonstrated that TOR has simi-
`lar efficacy and toxicities as those of its parent compound
`
`Table 8. Comparison of Quality-of-life Measures for Patients on TAM,
`TOR60, and TOR200
`Treatment Arm
`TOR6O
`No.
`
`TOR200
`No.
`
`%
`
`%
`
`Measure
`
`TAM
`
`No.
`
`%
`
`ECOG performance
`status
`Total no.
`
`Progressed to 3 or 4
`on study
`
`VAS
`Total no.
`Improvementin
`Enjoymentoflife
`Pain
`Mood
`Analgesics required
`Total no. of patients
`Mean score + SDT
`
`215
`
`221
`
`212
`
`14
`
`175
`
`68
`62
`78
`
`7
`
`39
`35
`45
`
`7
`
`16
`
`189
`
`8646"
`76
`40
`89
`47
`
`24
`
`190
`
`80
`72
`96
`
`VW
`
`42
`38
`51
`
`195
`0.71 + 0.8
`
`205
`0.76 + 0.8
`
`200
`0.79 + 0.9
`
`*n = 188 for enjoymentoflife, TOR6O arm.
`tMeanscore while on study: 0, no analgesics; 1, nonnarcotic analgesic;
`2, oral narcotic analgesic; 3, parenteral narcotic analgesic; 4, uncontrolled
`pain.
`
`Downloaded from ascopubs.org by 151.194.33.114 on February 17, 2017 from 151.194.033.114
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`AstraZeneca Exhibit 2039 p. 7
`
`

`

`TOREMIFENE v TAMOXIFEN IN METASTATIC BREAST CANCER
`
`
`
`ProportionofPatientsEventFree
`
`©2°©©=nN&fon)or)°
`
`© Oo
`
`2563
`
`0
`
`200
`
`400
`
`600
`
`800
`
`1000
`
`1200
`
`1400
`
`Time in Days
`
`Fig 4. Decline of performance status for patients treated with TAM (+++, n = 214], TOR60 (OOO, n = 221}, and TOR200 (xxx, n =
`212). Serial changes in performancestatus from baseline to ECOG 3 or 4 were determined prospectively for patients on each arm of the study
`from time of study entry (P values: log-rank = .84; Wilcoxon = .82).
`
`in postmenopausal patients with metastatic, hormone re-
`ceptor-positive (or -unknown) breast cancer. Further-
`more, our data suggest that the quality of life of patients
`on TOR wassimilar to that of patients assigned to the
`TAM arm. Although preclinical and phase II data had
`suggested the possibility of a dose-response relationship
`for TOR,the results of this clinical trial failed to detect
`such a relationship.
`Patients with metastatic breast cancer are frequently
`treated with a series of systemic therapies, which are
`usually chosen based on potential efficacy and toxicity
`predictions.” In general, hormonetherapies are associated
`with fewer side effects and toxicities than is chemother-
`apy. Therefore, patients are more likely to achieve pallia-
`tive benefit from hormonetherapies if anticipated efficacy
`is satisfactory. In this regard, certain biochemical and
`clinical features have been identified that allow selection
`of patien