`
`Receptor Positive Patients
`
`A Randomized Trial Comparing Tamoxifen Alone Versus Tamoxifen Plus CMF
`
`W. R. BEZWODA. FCP (S.A.)," D. DERMAN, FCP (S.A.)." N. G. DE MOOR. DMR,T
`M. LANGE, FRCSJ AND J. LEVIN,DMR§
`
`In a controlled clinical trial, 52 patients with ER positive metastatic breast cancer were randomly
`assigned to receive initial treatment with either tamoxifen alone followed by sequential cyclophospha-
`mide, methotrexate, and S-fluoronracil (CMF) after tumor progression or concurrent tamoxifen plus
`CMF. All 52 patients entered are eligible and 50 patients were currently assessable. Response rates
`(CR + PR) were similar for the two treatment groups (15/24 tamoxifen alone, 17/26 tamoxifen
`+ CMF). The median durations of survival for the two treatment groups 17.7 months and 17.1 months
`were also not significantly different. Significant correlations were found between level of ER and re-
`sponse rate and also between level of ER and response duration. From the result of this investigation
`it would appear that there is no advantage to the addition of chemotherapy prior to an adequate trial
`of hormone therapy in patients with metastatic ER positive breast cancer.
`Cancer 50:2747-2750, 1982.
`
`CONSIDERABLECONTROVERSY exists regarding choice
`
`of initial treatment for the patient with recurrent
`or metastatic breast cancer. Early chemotherapy as first
`line treatment has its advocates, particularly since the
`report by Cooper1 of a high rate of complete tumor
`regression following aggressive combination chemo-
`therapy. It should, however, be noted that this early re-
`port dealt with the treatment of hormone resistant pa-
`tients and it must be admitted by even the most ardent
`proponents of chemotherapy that few, if any, patients
`are cured after metastases have been found.2 Since che-
`
`motherapy is certainly associated with more patient dis-
`comfort and a higher morbidity, other investigators have
`pointed to the potential advantages of a period of control
`of the disease by means of hormonal manipulation,3
`reserving chemotherapy for the moment when hormone
`treatment becomes ineffective. The quality of survival
`has always been the major justification for this view.
`One problem in regard to this approach however has
`been the relatively low response rate to hormone ma-
`nipulation in unselected patients with metastatic breast
`
`cancer. Response rates have usually been of the order
`of 30%,4 although patient selection by the use of certain
`empiric clinical criteria does allow higher response rates
`to be achieved.5 More recently the introduction of hor-
`mone receptor assays has allowed for the exclusion from
`hormone treatment patients with a poor chance of re-
`sponse. When estrogen receptor (ER) negative patients
`were excluded from hormone treatment, a response rate
`of about 60% can be expected in those who have a pos-
`itive test for cytoplasmic estrogen receptor.6 The intro-
`duction of anti-estrogen compounds has also simplified
`hormone treatment since these drugs are as effective as
`major endocrine ablation7 or additive hormone treat-
`ment,8 yet without many of the side effects of these other
`hormonal therapies. It seemed, therefore. that there was
`a need to study the efficacy of sequential treatment with
`hormones and chemotherapy and compare this to con-
`current hormone and chemotherapy in ER positive pa-
`tients.
`
`Patients and Methods
`
`From the Joint Breast Clinic, Johannesburg Hospital, University of
`the Witwatersrand, Johannesburg. South Africa.
`" Department of Medicine.
`T Department of Radiation Oncology.
`1 Department of Surgery.
`§ Department of Nuclear Medicine.
`Drs. Bezwoda, de Moor and Devin are in receipt of grants from the
`National Cancer Association of South Africa.
`Accepted for publication October 19, I981.
`
`Fifty-two female patients attending the Breast Clinic
`of the Johannesburg Hospital for treatment of recurrent
`or metastatic breast cancer were studied. The age of the
`patients ranged from 27—72 years (mean, 50.5 years) all
`but seven were postmenopausal and none had had prior
`oopherectomy. The study population included 31 black
`and 21 white patients. All the patients included in this
`
`0008—543X/82/1215/2747 $1.00 © American Cancer Society
`
`2747
`
`Genentech 2073
`
`Hospira v. Genentech
`|PR2017-00737
`
`Genentech 2073
`Hospira v. Genentech
`IPR2017-00737
`
`
`
`2748
`
`CANCER December 15
`
`1982
`
`Vol. 50
`
`TABLE 1. Characteristics of the Patients Entered into the Trial of
`Tamoxifen versus Tamoxifen + CMF
`
`Schedule A
`Tamoxifen alone
`
`Schedule B
`Tamoxifen
`+ concurrent CMF
`
`53.7 i 7.9
`
`50.1 t 9.8
`
`3
`
`0
`21
`15
`9
`6.27 1 3.14
`
`7
`2
`2
`2
`16
`21
`
`3
`
`l
`24
`12
`12
`5.9 i 2.6
`
`1
`
`1
`2
`2
`2
`15
`17
`
`Mean age
`Menopausal status
`Premenopausal
`Perimenopausal (within
`1 year of LMP)
`Postmenopausal
`Black
`White
`Tumor-free interval
`Metastatic sites
`Bone
`Lung
`Pleura
`Liver
`Lymph nodes
`Soft tissue
`
`trial had a positive assay for cytoplasmic estrogen re-
`ceptor (ER) on either metastatic or recurrent tumor. The
`assay used is based on the method described by McGuire
`et al.10 and has been previously reported.9 An ER con-
`centration of >3 fmol per mg of protein plus a binding
`index of >12% is regarded as a positive result in our
`laboratory. Other criteria for entry into the trial included
`histologic proof of recurrent or metastatic breast cancer,
`adequate documentation of the extent of the disease
`prior to starting therapy, and no previous treatment with
`either hormones or chemotherapy.
`After entry into the trial patients were randomly a1-
`located to receive tamoxifen alone as initial treatment
`
`(Group A) or tamoxifen plus CMF (cyclophosphamide,
`methotrexate, 5-fluorouracil) given concurrently (Group
`B). In both treatment schedules the dose of tamoxifen
`was 20 mg twice daily while the CMF schedule consisted
`of cyclophosphamide 100 mg/m2 orally X 14 days,
`methotrexate 35 mg/m2 IV day 1 and day 8, and 5-
`fluorouracil 350 mg/m2 IV day l and day 8. On disease
`
`TABLE 2. Response to Treatment in Trial of Tamoxifen
`versus Tamoxifen Plus CMF
`
`Schedule A
`Sequential Tamoxifen alone
`
`Response to Tamoxifen
`prior to chemotherapy
`
`No.
`
`24
`5
`10
`9
`
`%
`
`21
`42
`37
`
`Total
`CR
`PR
`NR
`
`Response
`after cross
`to CMF
`
`No.
`
`9
`0
`3
`6
`
`%
`
`0
`33
`67
`
`Percentage CR + PR 70
`
`Schedule B
`Concurrent
`Tamoxifen
`+ CMF
`
`No.
`
`26
`3
`14
`9
`
`%
`
`1
`1
`54
`35
`
`progression, those patients who received tamoxifen only
`were crossed over to receive CMF while those who re-
`
`ceived concurrent chemotherapy plus hormone therapy
`and those who had failed after the sequential treatment
`were taken off the trial but continued to be followed up.
`Further chemotherapy using different drugs or hor-
`monal manipulation was allowed at this point according
`to the investigators discretion. The criteria used for defi-
`nition of initial extent of the disease and response to
`treatment were those recommended by the Task Force
`on Breast Cancer of the UICC.”
`
`Twenty-four patients (15 black and nine white pa-
`tients) were treated with schedule A (sequential tamox-
`ifen alone) and 28 patients (16 black and 12 white) were
`treated with schedule B (concurrent tamoxifen plus
`CMF). The mean number of metastatic sites involved
`per patient was 1.9. Further clinical details are shown
`in Table 1. There were no significant differences between
`the two treatment groups in regard to age, menopausal
`status, tumor-free interval, site of metastases, or number
`of metastatic sites involved. The morphologic diagnoses
`included 35 poorly differentiated ductal carcinomas, ten
`anaplastic carcinomas, four infiltrating lobular carci-
`nomas, two medullary carcinomas, and one giant cell
`carcinoma.
`The trial was commenced in October 1977 and the
`
`period of follow-up is to January 1981. All patients gave
`informed consent prior to starting treatment. The trial
`protocol was approved by the Committee for Ethics in
`Human Research of the Faculty of Medicine, University
`of the Witwatersrand and conformed to the principles
`of the Declaration of Helsinki.
`
`Results
`
`All 52 patients randomised were eligible for entry into
`the trial. Tumor responses (either complete or partial
`remission) were seen in 34/50 patients (68%). Two pa-
`tients had entered the trial too recently for an assessment
`of response and survival to be made. Details of the re—
`sponse rates for the two regimens are shown in Table
`2. There were no significant differences between the two
`treatment schedules.
`
`Schedule A
`
`Fifteen of the 24 patients treated with the sequential
`regimen responded to the initial part of therapy with
`tamoxifen alone. Two of the remaining nine who were
`unresponsive to hormonal
`treatment subsequently
`achieved partial response after the addition of CMF.
`Seven of the 15 patients who were responsive to ta-
`moxifen, subsequently showed disease progression and
`were then treated with CMF. Only one of the seven
`patients responded to chemotherapy. Median duration
`
`
`
`No. 12
`
`TREATMENT OF METASTATIC BREAST CA
`
`Bezwoda et al.
`
`2749
`
`of survival in this group was 17.1 months and the me-
`dian duration of response (CR + PR) was 12.7 months.
`
`Schedule B
`
`Seventeen of 26 patients treated with concurrent ta-
`moxifen and CMF responded to therapy. The median
`durations of survival (17.7 months) and response in re-
`sponders (7.4 months) were not significantly different
`from those achieved by patients treated with sched-
`ule A.
`
`Median survival of all nonresponding patients was
`only 4.2 months. Analysis of the data to compare the
`response rate and survival of the patients in the two
`treatment groups according to race (black versus white)
`also failed to show any significant differences. For those
`patients receiving the sequential regimen of treatment
`(tamoxifen followed by CMF), the major portion of the
`survival experience was related to the period of hormone
`treatment. The three partial remissions seen after the
`addition of CMF (two following NR on tamoxifen alone,
`one following 15 months CR on tamoxifen alone) lasted
`for 14, 12, and 11 months, respectively.
`The distribution of ER concentrations was found to
`
`be markedly skewed. Logarithmic transformation of the
`values normalised this distribution and statistical anal-
`
`ysis was accordingly performed using logarithmically
`transformed data. ER concentrations in the different
`
`groups are expressed as geometric means i 1 SD. The
`ER concentrations ranged from 4—298 fmol/mg protein.
`The geometric mean of the ER concentrations in pa-
`tients treated with schedule A was 27 fmol/mg protein
`(:1 SD, 21—35) and this did not differ significantly from
`the geometric mean of the ER concentrations in patients
`treated with schedule B, 28 fmol/mg protein (i1 SD,
`22-35). Since there appeared to be no differences in
`composition or in outcome between the two treatment
`groups further analyses were performed using the pooled
`data. Significantly higher levels of estrogen binding were
`found among patients who responded to treatment (F
`= 4.08, dfl = 2; df2 = 47; P < 0.025) with the highest
`level 41 fmol/mg protein (: 14—124) being found amongst
`the 11 complete responders. The 24 patients who had
`a partial response to treatment showed a mean level of
`32 fmol/mg protein (i1 1—91) while the value for non-
`responders was only four fmol/mg protein (:5—39, geo-
`metric mean il SD). There were also significant cor-
`relations between the duration and initial response and
`the ER concentration (r = 0.322; P < 0.0125) and the
`length of survival and the ER concentration (r = 0.32;
`P < 0.0125).
`Analyses of the survival data according to different
`levels of ER (Fig. 1) showed a progressive and significant
`prolongation of survival with successively higher ER in-
`
`1001.
`
`
`
`501 Percentogesurvtvol 6
`L
`
`l
`O
`
`l
`4
`
`l
`8
`
`l'
`I2
`
`T
`'
`20
`16
`Months
`
`l
`24
`
`l’
`28
`
`'
`32
`
`‘l
`36
`
`l
`40
`
`FIG. 1. Percentage survival at different tumor estrogen receptor lev-
`els. Key: A 3-10 fmol/mg protein; A 11—30 fmol/mg protein; 0 3]—
`100 fmol/mg protein; 0 101-300 fmol/mg protein.
`
`tervals (Lee Desu, 9.72; dB; P< 0.025). The median
`durations of survival were 4.5 months (eight patients),
`14.4 months (24 patients), 19.3 months (11 patients)
`and 24.4 months (six patients). The ER intervals chosen
`for this analyses were 3—10, 11—30, 31—100, and 101—
`300 fmol/mg protein, respectively. Seven out of 24 pa-
`tients treated with schedule A (sequential
`tamoxifen
`alone) died while still receiving the first part of the pro-
`jected treatment schedule. During the subsequent period
`on CMF therapy, an additional seven patients died. Sim-
`ilarly, 13/28 patients died while receiving treatment B
`(concurrent CMF plus tamoxifen). These deaths account
`for 27 of the 34 deaths occurring on study. Fourteen
`patients received other chemotherapy and hormone
`therapies following withdrawal from the trial owing to
`treatment failure. These therapies included fluoxymes-
`terone (two patients), stilboestrol (one patient). vincris-
`tine + Adriamycin (seven patients) and mitomycin
`+ vinblastine + thiotepa (four patients). Five of the 14
`patients had a partial response following other treat-
`ments (fluoxymesterone three, chemotherapy two). These
`responses to second treatment schedules and patient sur-
`vivals with only symptomatic and supportive therapy,
`which ranged from 1—13 months, account for the dif-
`ferences between the median duration of remission and
`the median duration of survival.
`
`Discussion
`
`The results of the current study support the idea that
`a trial of hormone therapy for patients with ER positive
`disseminated breast cancer is both worthwhile and not
`
`disadvantageous in terms of subsequent survival. In this
`study there was no significant difference in patient sur-
`
`
`
`2750
`
`CANCER December 15
`
`1982
`
`Vol. 50
`
`vival in the sequential hormone therapy, chemotherapy
`arm and that of patients treated with concurrent che-
`motherapy and hormone treatment. Only a few studies
`have directly compared hormonal manipulation with
`chemotherapy and it must be remembered that in these
`studies no attempt was made to select out the patients
`with a poor chance of hormone response.“13 Although
`no previous randomised study similar to this one has
`been reported, there are indications from other studies
`that patients who have had prior hormone treatment
`respond to the same degree and for similar lengths of
`time as those patients who went straight onto chemo-
`therapy on diagnosis of metastatic involvement. Indeed
`one recent study tends to suggest that previous hormone
`responsiveness indicates a good and prolonged response
`to chemotherapy.3 While our study does not lend direct
`support to this notion, the results also do not conflict
`with this interpretation of the data. Certainly responses
`to chemotherapy were seen, both after cross over to
`CMF following the use of tamoxifen alone and to second
`line chemotherapy following the use of both CMF and
`tamoxifen. It might be asked why those patients with
`low level “positive” ER apparently gained no benefit
`from the addition of chemotherapy. Unfortunately the
`number of patients with low positive ER concentration
`was small and differences in response rate in this
`subgroup might have been missed. It is clear however
`that the major portion of the response duration in ER
`positive patients, particularly those with high level pos-
`itive ER is due to the use of hormone therapy. These
`findings differ from those of some authors who have
`found an inverse correlation between ER status and re-
`
`sponse to chemotherapy.M It has been suggested that
`some of the discrepant results could possibly be ex-
`plained by biologic differences in tumor behavior related
`to race. In our analysis no differences in either response
`rate or response duration were found when the data were
`analysed by race. This lack of any difference related to
`racial group is in accordance with our previous experi-
`ence.’5
`
`While there has been considerable controversy as to
`the relationship of ER levels, response to therapy, his-
`tologic type and natural history of the tumorm'l7 recent
`reports have indicated a longer survival for ER positive
`patients.18 In the current study although a cut off limit
`of 3 fmol/mg protein were used in order to define ER
`positivity, there was a progressive, significant survival
`advantage irrespective of type of treatment for patients
`with tumors with higher ER concentrations. Other
`known prognostic factors such as age, histologic type,
`and menstrual status were similarly distributed in both
`treatment groups and ER appeared to be an independent
`prognostic determinant. It is thus illogical to use out off
`points to determine “positivity” or “negativity” and it
`
`would be more valuable to define more accurately the
`
`given chances of success or failure at a determined level
`of ER concentration in order to be able to make ther-
`
`apeutic decision on the basis of ER levels. Thus, for the
`patient with a high level ER positive tumor, the high
`response rate and the low incidence and minor nature
`of the side effects from hormonal therapy justifies the
`use of initial hormone treatment alone. This form of
`
`treatment ensures maximum quality of survival without
`adversely affecting quantity of life.
`
`REFERENCES
`
`1. Cooper RG. Combination chemotherapy in hormone resistant
`breast cancer (Abstr). Proc Am Assoc Cancer Res 1969; 10:15.
`2. Henderson C, Cannellos GP. Cancer of the Breast: The past
`decade. N Engl J Med 1980; 302217—20, 78—90.
`3. Manni A, Trujillo JE. Pearson OH. Sequential use of endocrine
`therapy and chemotherapy for metastatic breast cancer: Effects on
`survival. Cancer Treat Rep 1980; 64:11 1—116.
`4. Stoll BA. Palliation by castration or hormone administration. In:
`Stoll BA, ed. Breast Cancer: Early and Late. Chicago: Year Book,
`1979; 133—146.
`5. Kennedy B]. Hormonal therapies in breast cancer. Semin Oncol
`1974; 1:119—130.
`6. McGuire WL, Horwitz KB, Pearson OH, Segaloff A. Current
`status of oestrogen and progesterone receptors in breast cancer. Cancer
`1977; 39:2934-2947.
`7. Legha SS, Carter SK. Anti-oestrogens in the treatment of breast
`cancer. Cancer Treat Rev 1976; 3:105—215.
`8. Kiang DT, Kennedy BJ. Tamoxifen (antioestrogen) therapy in
`advanced breast cancer. Ann Intern Med 1977', 87:687—690.
`9. Levin J, Kay G, Da Fonseca M, Lange M, De Moor NG, Savage
`N. Oestrogen receptors in tumors of breast cancer patients. S Afr Med
`J 1978; 53:577—578.
`10. McGuire WL, De La Garza M. Improved sensitivity in the
`measurement of oestrogen receptor in human breast cancer. Clin En-
`docrine! Metab 1973; 37:968—989.
`11. Hayward JL, Carbone PP, Heuson JC, Kumaoka S, Segaloff
`A, Reubens RD. Assessment of response to therapy in advanced breast
`cancer: A project of the panel on clinical oncology of the International
`Union Against Cancer, Geneva: Switzerland. Cancer 1977; 39:1289—
`1294.
`12. Nemoto T, Rosner D, Diaz R et al. Combination chemotherapy
`for metastatic breast cancer: Comparison of multiple drug therapy
`with 5-Fluorouracil, cytoxan, and prednisone with adriamycin or ad-
`renalectomy. Cancer 1978; 41:2073—2077.
`13. Priestman T, Baum M, Jones V, Forbes J. Treatment and sur-
`vival in advanced breast cancer. Br MedJ 1978; 2:1673—1674.
`14. Lippman ME, Allegra JC, Thompson EB et al. The relation
`between estrogen receptors and response rate to cytotoxic chemo-
`therapy in metastatic breast cancer. N Engl J Med 1978; 29821223—
`1228.
`15. Bezwoda WR, de Moor NG, Derman D. Lange M, Saner R,
`Dando R. Combination chemotherapy of metastatic breast cancer: A
`randomised trial comparing the use of adriamycin to that of vinblas-
`tine. Cancer 1979; 44:392—397.
`16. Fisher ER, Redmond CK, Lice H, Rockette H, Fisher B, and
`collaborating investigators. Correlation of estrogen receptor and patho-
`logic characteristic of invasive breast cancer. Cancer 1980; 45:349—
`353.
`17. Hill R, Feldstein ML, Gibson SL, Savlov ED. The relative im-
`portance of estrogen receptor analysis as a prognostic factor for re-
`currence or response to chemotherapy in women with breast cancer.
`Cancer 1980; 45:1993—2000.
`18. Fletcher WD, Loung BS, Davenport CE. The prognostic sig-
`nificance of estrogen receptors in human breast cancer. Am J Surg
`1978', 135:372—374,
`
`