High- Versus Standard-Dose Megestrol Acetate in Women With
`Advanced Breast Cancer: A Phase III Trial of the Piedmont
`
`Oncology Association
`
`By Hyman B. Muss, L. Douglas Case, Robert L. Capizzi, M. Robert Cooper, Julia Cruz, Don Jackson,
`Fred Richards ||, Bayard L. Powell, Charles L. Spurr, Douglas White, Patricia Zekan, Sally Read,
`Shelia Cates—Wilkie, James Bearden, John McCullough, Richard Callahan, Kenneth Karb, Jim Atkins,
`Bart Paschal, Bill Ramseur, John lusk, and Valerie Stanley
`
`One hundred seventy-two patients with advanced
`breast cancer were randomized to receive oral stan-
`
`dard-dose megestrol acetate (MA), 160 mg/d or
`high-dose MA, 800 mg/d. All but two patients had
`one prior trial of tamoxifen therapy for either meta-
`static disease (74%) or as adiuvant treatment (26%).
`Pretreatment characteristics were similar for both
`arms. High-dose MA resulted in a superior complete
`plus partial response rate (27% v 10%, P = .005),
`time to treatment failure (median, 8.0 v 3.2 months,
`P = .019), and survival (median, 22.4 v 16.5 months,
`P = .04) when compared with standard-dose ther-
`apy. These differences remained significant after ad-
`iustment
`for other covariates. Thirty-four patients
`were given high-dose MA after failure of standard-
`
`dose MA treatment, and none responded. Weight
`gain was the most distressing side effect, with 13% of
`standard-dose and 43% of high-dose patients gaining
`more than 20 lbs. Four maior cardiovascular events
`occurred in patients receiving high-dose treatment
`and one in patients given standard doses. Other
`toxicity was modest. High-dose MA may represent a
`significant improvement in secondary endocrine ther-
`apy for advanced breast cancer patients refractory to
`initial endocrine treatment, but its use on a regular
`basis should be reserved until these results are con-
`firmed by other clinical trials.
`J Clin Oncol 8:1797-1805. © 1990 by American Soci-
`ety of Clinical Oncology.
`
`NDOCRINE therapy represents a major
`modality for the treatment of metastatic
`breast cancer. Historically, ablative procedures,
`including oophorectomy, adrenalectomy, and hy-
`pophysectomy, provided objective remissions in
`20% to 50% of patients. Later, estrogens, andro-
`gens, progestins, and corticosteroids were added
`with similar results."2 Currently, tamoxifen, an
`antiestrogen represents the most commonly used
`first-line endocrine therapy because of its ease of
`administration, minimal
`toxicity, and similar
`efficacy to older agents.3
`Unfortunately, almost all patients treated with
`tamoxifen will ultimately progress and require
`salvage treatment with either further endocrine
`therapy or chemotherapy. Both medroxyproges—
`terone acetate (MPA) and megestrol acetate
`(MA) have been extensively used as salvage
`treatments.“’5 In randomized trials, both MPA
`and MA have had similar response rates to
`tamoxifen when used as first-line therapy. MA
`has demonstrated response rates ranging from
`14% to 56% when used as primary treatment and
`from 6% to 23% when used as secondary therapy.5
`These salvage response rates with MA are simi-
`
`lar to those achieved with other agents including
`aromatase inhibitors, androgens, and estrogens.
`Two phase III trials of high-dose MPA have
`demonstrated higher response rates than those
`using standard dosesfi‘7 Cavalli et al6 compared
`1,000 mg/ d intramuscularly (IM) with 500 mg
`IM twice weekly.
`In 184 assessable patients,
`there was a 33% response rate to high-dose
`therapy and a 15% response to the low-dose
`therapy (P < .01). However, time to progression
`and survival did not differ significantly between
`the two arms. Pannuti et 317 compared 1,500 mg
`IM daily with 500 mg IM daily. The proportion
`of patients with complete or partial response or
`
`From the Piedmont Oncology Association and Cancer
`Center of Wake Forest University. Bowman Gray School of
`Medicine, Winston—Salem, NC.
`Submitted March I, 1990; accepted May 4, 1990.
`Supported in part by National Institutes ofHealth Grants
`CA-12197 and CA-37378, National Cancer Institute, Be-
`thesda, MD.
`Address reprint requests to Hyman B. Muss, MD, Cancer
`Center of Wake Forest University, Bowman Gray School of
`Medicine, 300 S Hawthorne Rd, Winston—Salem, NC 27103.
`© 1990 by American Society of Clinical Oncology.
`0732-183X/90/081 1 ~0009$3.00/0
`
`Journal of Clinical Oncology, Vol 8, No 11 (November), 1990: pp 1797-1805
`
`1797
`
`Downloaded from ascopubscrg by 86.188.169.218 on February 2, 2017 from 086.188.169.218
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`AstraZeneca Exhibit 2148 p. 1
`InnoPharma Licensing LLC V. AstraZeneca AB IPR2017-00904
`Fresenius-Kabi USA LLC V. AstraZeneca AB IPR2017-01910
`
`

`

`l 798
`
`MUSS ET AL
`
`stable disease was significantly greater for the
`high-dose regimen. Other investigators,"’9 how-
`ever, have not shown superiority of high-dose
`MPA compared with lower dosages. The develop-
`ment of MA, a potent and well-absorbed oral
`progestin, has led to renewed interest in clinical
`trials using oral progestins.lo Serum levels of
`progestin activity are substantially higher for
`MA than for equivalent doses of MPA.“'12 MA
`has also been extremely well tolerated, with the
`major toxicity being weight gain. The suggestion
`that high-dose MPA might be associated with
`higher response rates coupled with the superior
`oral availability of MA compared with MPA led
`to the development of a phase III trial by the
`Piedmont Oncology Association (POA) compar-
`ing high-dose with standard-dose MA therapy. A
`preliminary report of this trial has been pre-
`sented.13
`
`PATIENTS AND METHODS
`
`randomized
`The study was designed as a phase III,
`five-stage group sequential
`trial and was based on the
`assumption that the response rate to standard-dose MA as
`salvage therapy would be 20%. A maximum of 150 evaluable
`patients (75 per arm) was necessary to allow detection of an
`increase in the response rate to 40% in the high-dose arm with
`80% power at the 5% one-sided level of significance. Patients
`were stratified into three groups prior to randomization:
`group 1 had received prior hormonal therapy and responded,
`group 2 had received prior hormonal therapy without re-
`sponse, and group 3 had received prior adjuvant hormonal
`therapy and relapsed. Patients relapsing after adjuvant
`therapy were eligible for the trial at time of relapse.
`Eligibility required that patients be at least 18 years of age,
`have histologically confirmed breast cancer with progressive
`metastatic disease, have one previous trial of hormonal
`therapy other than a progestin, have measurable or evaluable
`disease, and have less than three prior chemotherapy regi-
`mens. Either the estrogen or progesterone receptor from the
`primary tumor or a metastatic site had to be positive or both
`had to be unknown with the exception that patients who had
`receptor—negative tumors but had responded to prior hor-
`monal
`therapy were eligible. Those receiving concurrent
`progestin therapy for nonmalignant disease or those with
`brain metastasis as the only evidence of tumor recurrence
`were excluded. Patients were allowed on trial if they had any
`prior malignancy regardless of type without evidence of
`recurrence for longer than 5 years, skin malignancy exclud-
`ing melanoma, or cancer of the cervix treated greater than 3
`years prior to protocol entry without evidence of recurrence.
`Prior concomitant radiation therapy was permissible pro-
`vided there was evaluable or measurable disease outside the
`treatment field. Informed consent meeting Cancer Center,
`institutional, and federal guidelines was required.
`Patients were randomized to either MA, a 160 mg tablet
`(Megace; Bristol-Myers, Evansville, IN), once daily or high-
`
`dose MA, five 160 mg tablets (800 mg) daily. For patients
`receiving high-dose therapy, two tablets of the drug were
`administered in the morning, and one tablet at lunch, dinner,
`and bedtime. The bioequivalcnce of the 160 mg investiga-
`tional tablet has been shown to be 97% of the 40 mg four
`times a day dosage.” Patients were maintained on therapy
`unless removed for toxicity or disease progression. Dose
`modifications were made only for excessive weight gain
`(generally an increase in 10% of prestudy body weight) on the
`high-dose arm; such patients were continued on therapy at
`160 mg daily. Patients randomized to the standard treatment
`arm were allowed to cross over to the high-dose arm following
`progression. Patients were removed from study for grade 3 or
`4 granulocyte or platelet suppression, nausea and vomiting
`not easily controlled with antiemetics, persistent vaginal
`bleeding, uncontrolled hypertension, or hypercalcemia for
`longer than 3 weeks.
`Patients were seen every 4 weeks for follow-up. Palpahle
`lesions and chest x-rays showing metastatic disease were
`reevaluated every 4 weeks to assess response. For patients
`with disease on bone scan, bone survey, liver scan, or other
`imaging modality, studies were required to be repeated every
`12 weeks for two separate time periods and then every 6
`months. Patients who had stable disease in one disease site
`were not required to have repeat follow-up studies unless
`disease progression was clinically suggested. Strict Interna-
`tional Union Against Cancer criteria were used to document
`response.ls In order to qualify for complete or partial re-
`sponse, patients must have had repeat evaluation of all
`previously documented metastatic sites, with changes meet-
`ing complete and partial regression criteria.
`Difierences in pretreatment characteristics and response
`outcomes between treatment groups were assessed using
`Wilcoxon rank-sum tests for continuous variables and Fish-
`er’s exact tests for categorical variables. Logistic regression
`was used to determine which variables were significantly
`associated with response and to assess the effect of treatment
`regimen adjusted for other covariates. Log-rank tests were
`used to compare unadjusted survival and time to treatment
`failure distributions between regimens. Cox’s proportional
`hazards regression model was used to determine which
`variables were significantly associated with survival or time to
`treatment
`failure and to assess the eflect of treatment
`regimen adjusted for other covariables. Estimates of median
`follow-up time were calculated using the Korn method.”
`Analysis of weight gain changes over time was done using
`methods described by Espelandl7 and Espeland et a1.”
`
`RESULTS
`
`Between September 1985 and October 1988,
`172 patients were accrued to this study. Of these
`patients, two were ineligible: one had estrogen
`and progesterone receptor (ER/PR)-negative dis-
`ease without prior response to endocrine therapy
`and a second did not have histologically con-
`firmed breast cancer, leaving 170 qualified pa—
`tients available for analysis. The final update of
`these data was completed in November 1989.
`Follow-up ranged from 0.1 to 47+ months with
`
`Downloaded from ascopubs.org by 86.188.169.218 on February 2, 2017 from 086.188.169.218
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`AstraZeneca Exhibit 2148 p. 2
`
`

`

`MEGESTROL ACETATE AND BREAST CANCER
`
`1 799
`
`an estimated median of 21.3 months.16 Except
`for two patients who had received prior diethyl-
`stilbestrol, all patients had received tamoxifen.
`Descriptive statistics for the pretreatment char-
`acteristics are shown in Table 1. Most of these
`
`characteristics are fairly evenly balanced be-
`tween the two arms. However, more patients who
`were black were randomized to the standard-
`
`dose regimen (P = .072), and all six patients
`with poor performance status (three) received
`standard-dose MA initially (P = .029). Disease
`site data are shown in Table 2. For purposes of
`coding dominant site, breast, soft
`tissue, and
`lymph node metastasis were considered soft tis-
`sue; bone and bone marrow metastasis (with or
`without soft tissue involvement) were considered
`bone; and other sites (with or without soft tissue
`or bone involvement) were considered viscera.
`There were no significant differences in disease
`sites, number of sites, or dominant site between
`arms.
`
`Table 2. Disease Sites
`
`Standardoose
`No.
`%
`
`High-Dose
`No,
`%
`
`Total
`Site
`Breast (contralateral)
`Soft tissue*
`Lymph nodes
`Bone
`Bone marrow
`Liver
`
`lung parenchyma
`Pleural effusion
`Brain
`Other
`No. sites/patient
`1
`2
`3
`4
`Dominant site
`Soft tissue
`Bone
`Viscera
`
`86
`
`1 0
`32
`9
`65
`5
`1 3
`
`2O
`10
`1
`3
`
`35
`26
`1 9
`6
`
`7
`40
`39
`
`100
`
`12
`37
`10
`76
`6
`1 5
`
`23
`12
`1
`3
`
`41
`3O
`22
`7
`
`8
`47
`45
`
`84
`
`9
`20
`15
`60
`1
`1 0
`
`12
`7
`1
`3
`
`43
`30
`9
`2
`
`1 1
`47
`26
`
`100
`
`1 1
`24
`18
`71
`1
`1 2
`
`14
`8
`1
`4
`
`51
`36
`1 1
`2
`
`13
`56
`31
`
`Table I. Pretreatment Characteristics
`
`*lncludes skin, subcutaneous, and muscle involvement.
`
`Characteristic
`
`Total patients
`Median age, years
`Range
`Age _>_ 50
`Race
`White
`Black
`Performance status
`0
`1
`2
`3
`DFI
`0
`0-2 years
`> 2 years
`ER/ PR status
`+ / +
`+ / _
`+ /?
`— / +
`— / —
`Unknown
`Hormonal strata
`
`Standard-Dose
`No.
`%
`
`High-Dose
`No.
`%
`
`86
`
`100
`
`84
`
`100
`
`63
`37-84
`
`87
`
`81
`19
`
`28
`55
`10
`7
`
`19
`26
`56
`
`36
`29
`1 3
`7
`o
`1 6
`
`75
`
`70
`16
`
`24
`47
`9
`6
`
`16
`22
`48
`
`31
`24
`1 1
`6
`o
`14
`
`62
`38-82
`68
`
`77
`7
`
`26
`46
`12
`O
`
`10
`27
`47
`
`39
`14
`9
`5
`l
`1 6
`
`Response data are presented in Table 3.
`Twenty-two of 80 evaluable patients on high-
`dose MA (28%) responded compared with eight
`of 81 evaluable standard-dose patients (10%)
`(P = .005). There were two complete responders
`to high-dose MA and one to the standard-dose
`regimen. Logistic regression was used to assess
`treatment effect after adjustment for other cova-
`riables. Age, race, performance status, disease-
`free interval (DFI), ER/PR status, physician
`group (POA v medical school), prior therapy,
`dominant site, and number of sites were included
`in the analysis. ER and PR were initially ex-
`cluded due to the substantial number of missing
`values. Treatment regimen, adjusted for these
`variables, was significantly associated with re-
`
`Table 3. Response Versus Initial Treatment
`Standarerose
`High-Dose
`No.
`%
`No.
`‘36
`
`No. evaluable
`Response
`Complete response
`Partial response
`Stable disease
`Progression
`95% CI for response
`
`81
`
`1
`7
`31
`42
`
`100
`
`1
`9
`38
`52
`4-19
`
`80
`
`2
`20
`37
`21
`
`100
`
`2
`25
`46
`26
`18-39
`
`NOTE. 95% CI for response overall, 13%-26%.
`Abbreviation: CI, confidence interval.
`
`81
`
`92
`8
`
`31
`55
`14
`O
`
`12
`32
`56
`
`46
`17
`1 1
`6
`1
`1 9
`
`37
`36
`27
`
`50
`57
`
`34
`30
`22
`
`Prior Rx — response
`Prior Rx — no response
`Prior hormone — adiuvant
`Prior treatment
`42
`50
`43
`Radiation
`48
`51
`44
`Chemotherapy
`Abbreviations: Rx, treatment; DFI, disease-free interval.
`
`40
`35
`26
`
`31
`30
`23
`
`Downloaded from ascopubs.org by 86.188.169.218 on February 2, 2017 from 086.188.169.218
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`AstraZeneca Exhibit 2148 p. 3
`
`

`

`1 800
`
`MUSS ET AL
`
`sponse (P = .0102). A backward-stepping algo-
`rithm was used to remove nonsignificant covari-
`ates from the model. DFI (P = .0008) was the
`only covariate significantly associated with re-
`sponse. Performance status was of borderline
`significance (P = .0943). Treatment regimen,
`adjusted for DFI, was significantly associated
`with response (P = .0011). The odds of respond-
`ing to treatment were approximately 4.2 times
`greater for high-dose patients. ER and PR were
`then included,
`thus reducing the number of
`observations available in the analysis. DFI
`(P = .004) and ER were significant (P = .0180);
`treatment regimen, adjusted for these variables,
`was still significant (P = .0155).
`Only 16 of the 170 qualified patients remain on
`their initial treatment regimen, six patients on
`standard dose and 10 on high dose. Of those
`removed from the standard-dose regimen, 74%
`had disease progression, 15% died while on
`treatment, 2% were removed for toxicity, and 1%
`refused further treatment. Of those removed
`
`from the high-dose regimen, 71% had disease
`progression, 5% were removed for toxicity, 4%
`were removed per decision of the treating physi-
`cian, 4% died while on study, 1% refused further
`treatment, and 2% were removed for other rea-
`sons. Three of the deaths on the low-dose regi-
`men and one of the three deaths on the high-dose
`regimen were due to causes other than cancer.
`Time to treatment failure was calculated as
`
`the time from initial therapy to failure on the
`initial arm or to the last date of contact. Patients
`were considered treatment failures if removed
`
`from the study for any reason above. Time to
`failure estimates are shown in Fig 1. The overall
`estimated median time to failure was 5.1 months;
`the estimated median was 3.2 months for the
`
`standard-dose regimen and 8.0 months for the
`high-dose regimen (P = .0185). Cox’s propor-
`tional hazards regression model was used to
`determine which covariates were significantly
`associated with time to treatment failure and to
`
`assess treatment effect after adjustment for other
`variables. Treatment regimen, adjusted for all
`the other covariates except receptor status, was
`significantly associated with time to failure
`(P = .0088). The same backward-stepping strat-
`egy described previously was used for this analy-
`sis. Dominant site of disease (P = .0001), prior
`chemotherapy (P = .0005), and strata (P =
`.0434) were all significantly associated with time
`to failure. Patients with bone-dominant disease
`
`failure.
`times to treatment
`have the longest
`Treatment regimen, adjusted for these variables,
`was also significantly associated with time to
`failure (P = .0078). Standard-dose patients are
`at approximately 1.56 times the risk of failure
`per unit
`time as compared to the high-dose
`patients. ER and PR were then included in the
`analysis, but neither was significantly associated
`with time to failure. Time to failure estimates are
`
`Fig 1. Time to treatment
`failure v time.
`(-—) high dose;
`(—) standard dose.
`
`\
`
`n Failed
`
`Median
`
`High Dose
`Standard Dnse
`
`84
`BE
`
`74
`SO
`
`8.0
`3.2
`p < 0.02
`
`SD
`
`1.4
`0.3
`
`Range
`
`0.5-36.8
`O.1-36.9+
`
`
`
`Failure Time
`
`(Months)
`
`
`
`ProportionFailureFree0OOOOC
`
`C 0
`
`Downloaded from ascopubsorg by 86.188.169.218 on February 2, 2017 from 086.188.169.218
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`AstraZeneca Exhibit 2148 p. 4
`
`

`

`MEGESTROL ACETATE AND BREAST CANCER
`
`1801
`
`Total
`Site
`Skin
`Bone
`Visceral
`Strata
`Prior Rx, response
`Prior Rx, no response
`Prior Rx, adiuvant
`Chemotherapy
`No
`Yes
`
`n
`
`86
`
`7
`40
`39
`
`34
`30
`22
`
`42
`44
`
`Table 4. Time To Treatment Failure in Months
`Standard-Dose
`Median
`
`SD
`
`n
`
`Fail
`
`80
`
`7
`36
`37
`
`31
`28
`21
`
`37
`43
`
`3.2
`
`3.6
`4.0
`2.7
`
`4.1
`3.2
`2.7
`
`4.4
`2.8
`
`0.3
`
`1.0
`1 .4
`0.4
`
`1 .4
`0.4
`0.4
`
`1 .7
`0.2
`
`84
`
`1 1
`47
`26
`
`31
`30
`23
`
`36
`48
`
`High-Dose
`Median
`
`8.0
`
`5.1
`9.5
`3.0
`
`1 1 .0
`6.9
`7.5
`
`8.9
`7.3
`
`Fail
`
`74
`
`1 1
`39
`24
`
`25
`28
`21
`
`31
`43
`
`SD
`
`1 .4
`
`1.6
`l .1
`1 .9
`
`2.5
`2.3
`3.2
`
`2.1
`1 .9
`
`shown in Table 4 for the significant covariates
`noted above.
`
`Thus far, 89 of the 170 qualified patients have
`died. Survival data are plotted in Fig 2. The
`overall estimated median survival time was 18.5
`
`months; the estimated median was 16.5 months
`
`for the standard-dose patients and 22.4 months
`for the high-dose patients (P = .0388). Cox’s
`proportional hazards regression model was used
`to determine which covariates were significantly
`associated with survival and to assess the effect of
`
`treatment regimen after adjustment for covari-
`ates. The same variables used in response assess—
`ment were included in this analysis. Treatment
`regimen, adjusted for all these covariates except
`receptor status, was of borderline significance
`
`(P = .0903). A backward-stepping algorithm
`was used to remove nonsignificant variables from
`the model. Prior chemotherapy (P = .0061),
`dominant site of disease (P = .0381), DFI
`(P = .0174),and physician location (P = .0290)
`were the significant covariates. No prior chemo-
`therapy, soft tissue disease, a DFI of greater than
`2 years, and being treated in the community were
`the favorable characteristics. Treatment regi-
`men, adjusted for these covariates, was signifi-
`cantly associated with survival (P = .0190). Stan-
`dard-dose patients were at approximately 1.68
`times the risk of death per unit time compared
`with high-dose patients. ER and PR were then
`included in the analysis, but neither was signifi-
`cantly associated with survival. Survival esti-
`
`n Expired Median
`
`SD
`
`Flange
`
`High Dose
`Standard Dose
`
`E4
`BE
`
`39
`50
`
`22.4
`15.5
`p < 0.04
`
`1.0-42 . 9+
`3.4
`3.3 O.1-47.3+
`
`
`
`L1101\lA\__.\.1
`
`1
`
`Is
`
`H he
`
`0o0
`
`o
`
`
`
`ProportionSurviving
`
`0
`
`Lu _i_.._
`i
`o m
`
`c:
`
`Survival vtime. (—-) high
`Fig 2.
`dose; (—) standard dose.
`
`o O ,4.i
`
`12
`
`r—'
`18
`
`‘r—v
`24
`
`_—T
`30
`
`—r-
`
`Survival Time
`
`(Months)
`
`35
`
`42
`
`45
`
`Downloaded from ascopubsorg by 86.188.169.218 on February 2, 2017 from 086.188.169.218
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`AstraZeneca Exhibit 2148 p. 5
`
`

`

`1802
`
`MUSS ET AL
`
`Total
`Dominant site
`Skin
`Bone
`Visceral
`DFl*
`0
`0—2
`> 2
`Prior chemotherapy
`No
`Yes
`Physician group‘f
`BGSM
`POA
`
`n
`
`86
`
`7
`40
`39
`
`16
`22
`48
`
`42
`44
`
`35
`51
`
`Table 5. Survival Time in Months
`Standard-Dose
`Median
`
`so
`
`Died
`
`50
`
`2
`24
`24
`
`9
`16
`25
`
`20
`30
`
`24
`26
`
`16.5
`
`——
`17.3
`10.3
`
`17.9
`7.5
`19.1
`
`19.4
`14.4
`
`15.3
`17.9
`
`3.3
`
`—
`2.7
`1.9
`
`4.7
`1.9
`1.8
`
`—
`4.5
`
`3.8
`5.8
`
`11
`
`84
`
`11
`47
`26
`
`10
`27
`47
`
`35
`48
`
`30
`54
`
`High-Dose
`Median
`
`Died
`
`39
`
`3
`20
`16
`
`4
`16
`19
`
`14
`25
`
`20
`19
`
`22.4
`
`-
`25.2
`16.2
`
`22.4
`17.2
`25.2
`
`27.6
`17.2
`
`13.9
`25.8
`
`so
`
`3.4
`
`—
`6.5
`7.1
`
`—
`3.1
`2.4
`
`4.0
`4.1
`
`3.6
`2.3
`
`Abbreviation: BGSM, Bowman Gray School of Medicine.
`*Time in years.
`TPatients treated at BGSM v in community (90A).
`
`mates are shown in Table 5 for the significant
`covariates listed above.
`
`Thirty-four patients have crossed over from
`standard—dose to high-dose therapy following
`disease progression. Two are not assessable for
`response. Of the 32 assessable patients, 10 re-
`mained stable, and 22 had tumor progression.
`Only two patients remain on treatment. The
`median time to treatment failure for this group is
`2.0 months (range, 0.5 to 13.8 months).
`
`Toxicity
`
`toxicities
`Hematologic and gastrointestinal
`were minimal. Hemoglobin levels were some-
`what lower for the standard—dose patients, though
`the difference is probably not clinically meaning-
`ful (mean, 11.9 v 12.6, respectively; P = .009).
`There was an increased incidence of edema
`
`among high-dose patients (P = .001).
`Weight gain was the most common toxicity
`associated with MA administration and is pre-
`sented in Table 6. Weight gain was more pro-
`nounced among high-dose patients (P< .001)
`and increased with time on study. Mean weight
`changes from baseline are shown in Fig 3. The
`weight change increased over
`time and was
`significantly greater
`for high-dose patients
`(P < .05). Weight loss was not significantly dif-
`ferent between the two treatments (P = .4690).
`Twelve (14%) of the high—dose patients (six
`responders and six stable) required close modifi—
`
`cations due to weight gain from 2 to 10 months
`after entry. Weight stabilized or decreased in
`nine of these 12 and response was maintained.
`Blood pressure increased on both regimens, the
`increase being somewhat greater for the high-
`dose patients (though the differences between
`
`Table 6. Weight Changes on Initial Treatment
`Standard-Dose
`High-Dose
`No.
`(%)
`No.
`(96)
`
`No. patients
`Weight gain (lbs)
`Mean (SD)
`Median
`Range
`0
`0-10
`1020
`20-30
`30—40
`40—50
`50-60
`60-70
`Weight gain leading to
`dose modification
`Weight loss (lbs)
`Mean (SD)
`Median
`Range
`0
`1-10
`1 0-20
`20—30
`30 +
`
`77
`
`(1 00)
`
`79
`
`(1 00)
`
`8.2
`5.0
`0-52
`28
`23
`16
`7
`2
`0
`1
`0
`
`0
`
`3.8
`1 .0
`0—22
`31
`35
`9
`2
`0
`
`(10.4)
`
`(36)
`(30)
`(21)
`(9)
`(3)
`(0)
`(1)
`(0)
`
`19.5
`18.0
`0—63
`9
`15
`21
`15
`1 0
`6
`1
`2
`
`(0)
`
`12
`
`(5.6)
`
`(40)
`(45)
`(1 2)
`(3)
`(0)
`
`5.4
`0.0
`0—52
`43
`22
`8
`3
`3
`
`(15.1)
`
`(11)
`(19)
`(27)
`(19)
`(1 3)
`(8)
`(1)
`(3)
`
`(15)
`
`(10.1)
`
`(54)
`(28)
`(1 0)
`(4)
`(4)
`
`Downloaded from ascopubscrg by 86.188.169.218 on February 2, 2017 from 086.188.169.218
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`AstraZeneca Exhibit 2148 p. 6
`
`

`

`MEGESTROL ACETATE AND BREAST CANCER
`
`i 803
`
`30
`
`
`
`
`
`Bottom Line — standard Dose)
`(Top Line — High Dose.
`in Sample
`Number
`43
`54
`54
`11
`21
`20
`15
`15
`21
`40
`33
`27
`
`
`53
`45
`57
`4B
`35
`42
`34
`EB
`31
`25
`25
`20
`
`
`o
`1
`2
`
`Time
`
`(months)
`
`4.)
`.C
`U)
`-r—| 20
`a)
`
`3 C
`
`-r—I
`
`10
`
`0
`
`CI.)
`CD
`u
`.C
`L)
`
`Cr
`
`C
`r0
`CD
`2
`
`~10
`
`Fig 3. Mean change in
`weight v time. (---) high dose;
`(——) standard dose; P < .05.
`
`regimens were not statistically significant). The
`mean maximum systolic and diastolic pressure
`increases for the standard- and high-dose arms
`were 17 and 9 mm Hg for the standard-dose arm
`and 23 and 11 mm Hg for the high-dose arm.
`Nine percent of patients on each arm had eleva-
`tions in diastolic pressure of greater than 20 mm
`Hg. Although causation cannot be implied, larger
`weight gains were associated with larger blood
`pressure increases.
`Five major cardiovascular events occurred dur-
`ing therapy. Deep vein thrombosis occurred in
`two patients on the high-dose arm and one on the
`standard-dose arm after 11 weeks to 9 months of
`
`treatment. One patient receiving high-dose treat-
`ment had a fatal myocardial
`infarction and
`another a thrombotic stroke. Edema was re-
`
`ported in seven patients (9%) on standard-dose
`and 25 (31%) on high-dose treatment. Vaginal
`bleeding while on treatment was seen in five
`women (5%) on standard-dose and seven (8%) on
`high-dose therapy.
`
`DISCUSSION
`
`this study showed that high-
`In conclusion,
`dose MA was superior to standard doses in terms
`of response, time to treatment failure, and sur-
`vival when used as a secondary treatment for
`advanced breast cancer. The 10% response to
`standard-dose MA in this trial is lower than the
`
`response rates of 15% to 30% generally reported
`
`by others in this clinical setting”‘25 but is similar
`to previously published response rates of 6%
`reported by our group26 and 12% reported by
`Ingle et al.27 However, the 95% CI for response in
`our trial of 4% to 19% overlaps the response rates
`reported by most authors, suggesting that these
`differences in response among different series
`may not be that meaningful. The significant
`improvement in response rate with our high-dose
`regimen was underscored by the marked prolon-
`gation in time to treatment failure for this group.
`The median time to treatment failure of 3.2
`
`months for standard-dose therapy in this trial is
`similar to that of 3 months,25 3.5 months,20 3.7
`months,26 and 4.5 months21 reported by others.
`Indeed, the time to treatment failure of 8 months
`for the high-dose regimen is similar to that for
`MA when used as first-line therapy21’22’24'26 and
`almost double that reported by others using MA
`as second-line therapy. The prolongation of sur-
`vival for high-dose treated patients was unantici-
`pated. The reason for the significant improve-
`ment in median survival of almost 6 months for
`
`this group is unclear. Also patients treated in the
`community survived significantly longer
`than
`those treated at the university center. This may
`be related to patient selection but again the
`reasons for this difference are uncertain.
`
`Hematologic and gastrointestinal toxicities sec-
`ondary to both standard- and high-dose treat-
`ment were modest. The most distressing side
`
`Downloaded from ascopubs.0rg by 86.188.169.218 on February 2, 2017 from 086.188.169.218
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`AstraZeneca Exhibit 2148 p. 7
`
`

`

`l 804
`
`MUSS ET AL
`
`efl'ect of MA remains weight gain. Almost 45%
`of patients on the high-dose regimen gained 20 or
`more pounds in weight as opposed to 13% taking
`the standard-dose. Excessive weight gain led to
`dose modifications (generally to 160 mg/d) in
`14% of high-dose patients without clearly alter-
`ing response. This suggests that induction with
`high-dose therapy followed by standard-dose
`treatment might ameliorate this side effect with-
`out diminishing eflicacy. Blood pressure in-
`creased for most patients on both regimens, but
`these changes were generally modest and did not
`appear to result in any clinically manifest toxic-
`ity. Cardiovascular events, mainly deep venous
`thrombosis, were reported in patients on both
`regimens. An increased tendency for thrombosis
`may be related to progestin use,28 but the low
`incidence of this side effect and lack of a clearly
`defined pathophysiologic mechanism associating
`progestin use with thrombosis does not support a
`compelling cause and effect relationship. More-
`over, patients with metastatic cancer have a
`higher risk of thromboembolic complications.”
`Until more data are available, patients taking
`MA should be closely monitored for the develop-
`ment of such toxicity.
`Patients receiving standard-dose therapy who
`were subsequently crossed over
`to high-dose
`therapy did not achieve complete or partial
`
`responses in this study. Although two patients
`still remain on the high-dose arm after crossover,
`and might achieve response, the response rate
`would still be extremely low. This is contrary to
`the experience of Tchekmedyian et al30 who re-
`ported one complete and two partial responses to
`high-dose MA in 12 patients who had progressed
`on standard-dose treatment; however, these au—
`thors treated the majority of their patients with
`MA dosage of 1,600 mg daily, possibly account-
`ing for these differences. Our data suggest, how-
`ever,
`that to achieve maximum benefit after
`failure of primary endocrine therapy, high-dose
`MA should be used as subsequent treatment.
`Although high-dose therapy was significantly
`more efficacious than standard-dose treatment in
`
`is premature to
`this trial, we believe that it
`recommend it as standard treatment. The sub-
`
`stantial weight gain associated with this regimen
`is likely to be psychologically deleterious to many
`women. In addition, corroboration of this trial is
`necessary, and a similar trial is now in progress
`by the Cancer and Leukemia Group B (CALGB
`protocol 8741). Also, the POA is ready to em-
`bark on a phase III trial comparing high-dose
`MA and tamoxifen as first-line treatment. Fur-
`
`ther data from that trial should help define the
`role of high-dose MA in the hormonal manage-
`ment of advanced breast cancer.
`
`REFERENCES
`
`1. Ingle JN: Additive hormonal therapy in women with
`advanced breast cancer. Cancer 53:766-777, 1984
`2. Henderson IC: Endocrine therapy of metastatic breast
`cancer, in Harris JR, Hellman S, Henderson IC, et al (eds)
`Breast Diseases. Philadelphia, PA, Lippincott, 1987 pp
`398—428
`3. Manni A: Tamoxifen therapy of metastatic breast
`cancer. J Clin Lab Med 109:290—299, 1987
`4. Haller DG, Glick JH: Progestational agents in ad-
`vanced breast cancer: An overview. Semin Oncol 13:2-8,
`1986 (suppl)
`5. Sedlacek, SM: An overview of megestrol acetate for the
`treatment of advanced breast cancer. Semin Oncol 15:33-13,
`1988 (suppl 1)
`6. Cavalli F, Goldhirsch F, Jungi F, et al: Randomized
`trial of low- versus high-dose medroxyprogesterone acetate in
`the induction treatment of postmenopausal patients with
`advanced breast cancer. J Clin Oncol 2:414—419, 1984
`7. Pannuti F, Martoni A, Di Marco AR, et al: Prospective,
`randomized clinical trial of two different high dosages of
`medroxyprogesterone acetate (MAP) in the treatment of
`metastatic breast cancer. Eur J Cancer 152593-601, 1979
`8. Gallagher J, Cairnduff F, Smith IE: High dose versus
`
`low dose medroxyprogesterone acctate: A randomized trial in
`advanced breast cancer. Eur J Cancer Clin Oncol 2321895-
`1900, 1987
`9. Della Cuna GR, Calciati A, Strada MRB, et al: High
`dose medroxyprogesterone acetate (MPA) treatment in met—
`astatic carcinoma of the breast: A dose-response evaluation.
`Tumori 64:143-149, 1978
`10. Schacter L, Rozencweig M, Canetta R, et al: Meges-
`trol acetate: Clinical experience. Cancer Treat Rev 16:49-63,
`1989
`11. Lundgren S, Kvinnsland S, Utaaker E, et al: EtTect of
`oral high-dose progestins on the disposition of antipyrine,
`digitoxin, and warfarin in patients with advanced breast
`cancer. Cancer Chemother Pharmacol 18:270~275, 1986
`12. Miller AA, Bechter R, Schmidt CG: Plasma concentra—
`tions of medroxyprogesterone acetate and megesterol acetate
`during long-term follow-up in patients treated for metastatic
`breast cancer. J Cancer Res Clin Oncol 114:186-190, 1988
`13. Muss H, Case D, Cates-Wilkie S, et al: High (HiMEG)
`vs standard Dose (S-MEG) oral progestin therapy (megestrol
`acetate, MegaceT) for metastatic breast cancer (MBC): A
`phase III trial of the Piedmont Oncology Association (POA).
`Proc Am Soc Clin Oncol 8:22, 1989 (abstr)
`
`Downloaded from ascopubsorg by 86.188.169.218 on February 2, 2017 from 086.188.169.218
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`AstraZeneca Exhibit 2148 p. 8
`
`

`

`MEGESTROL ACETATE AND BREAST CANCER
`
`l 805
`
`l4. Gaver RC, Pittman KA, Reilly CM, et al: Bioequiva-
`lence evaluation of new megestrol acetate formulations in
`humans. Semin Oncol 12:17—19, 1985 (suppl 1)
`15. Hayward JL, Carbone PP, Heuson JG, et al: Assess-
`ment of response to therapy in advanced breast cancer.
`Cancer 39:1289-1293, 1977
`16. Korn EL: Censoring distributions as a measure of
`