The Contribution of Medicine to the Primary Treatment of ... · treatment of breast cancer but also because of its possible inhibition of carcinogenic influences. International Overview

(CANCER RESEARCH 48, 2314-2324, May I, 1988)

Perspectives in Cancer Research

The Contribution of Medicine to the Primary Treatment of Breast Cancer1

Gianni Bonadonna2 and Pinuccia Valagussa

Istituto Nazionale Tumori, Milan, Italy

About a quarter of a century ago, the world of clinicaloncology appeared rather well organized in compartments,almost like the old feudal society, with each of the protagonistsbeing confined to a given domain of action reflecting the prevailing strategic concepts of the time. Surgeons ranked aboveall clinicians. Endowed with uncanny faith in the anatomic andmechanistic dogma of tumor cell dissemination, they were veryproud of their skill to remove whatever could be radicallydissected when the disease was considered limited to local-regional areas. If the tumor was judged to be technically unre-sectable, then the patient was referred to radiation therapy,often considered a form of sophisticated palliation although,for given diseases or stages, it was already evident that thismodality could be curative. Only when the neoplasm appearedbeyond the feasibility of an effective radiation program, eitherbecause of a huge tumor mass or the presence of distantmÃ©tastases,did most clinicians agree that a therapeutic attemptcould be made by testing one of the growth-inhibiting compounds. At that time, medical treatment of human cancer wasin its infancy and altogether was considered a rather poorlydefined area of clinical activity. Within this field a few knowledgeable physicians, affected by the visionary concept of killingmalignant cells through drugs, were often considered helpfulmainly because they could pay attention to hopeless clinicalsituations. This rigid compartmental establishment was operationally peaceful, for surgeons and radiotherapists did not evendream of being directly engaged with chemotherapy, whilemedical oncologists did not dare interfere with the early management of cancer. Aside from a few coded pre- and postoperative irradiation programs, the modern strategy of the multi-

disciplinary approach had yet to be conceived.However, during the 1960s and the early 1970s, both in the

laboratory and at the patient bedside, there was a silent preparation for new biological concepts and strategies. Skipper,Schabel, Griswold, and coworkers (1, 2), studying unceasinglymurine leukemias and other transplantable tumors, providedthe biological and therapeutic response information that became very helpful in developing more effective drug treatmentsfor a variety of disseminated human neoplasms. The clinicalresults reported a few years later, particularly in acute leukemias, malignant lymphomas, and testicular cancer (3), wereactually the consequence of experimental models tested in thelaboratory. In rapid sequence, other important achievementsoccurred that allowed Medical Oncology to grow as a subspe-cialty of Internal Medicine (4). The major outcome of thisgrowth was a new, thorough clinical approach to all cancerpatients. Medical oncologists began to talk a common languagein terms of systematic staging and follow-up, objective assessment of response to treatment and pattern of disease recurrence,the concept of minimal residual disease, full-dose chemotherapyin selected situations, as well as effective supportive measures.

Received 8/28/87; revised 11/30/87; accepted 1/28/88.1Supported in part by Contract N01-CM-07338 with the Division of Cancer

Treatment, National Cancer Institute, NIH.*To whom requests for reprints should be addressed, at Division of Medical

Oncology, Istituto Nazionale Tumori, Via Venezian, 1, 20133 Milano, Italy.

In research centers, clinicians began to look at host-tumorrelationship as potential sources for developing new strategicapproaches. Therapeutic science in clinical oncology began tosurface also through the design of prospective randomized trialshaving biological end points.

During the same period of time, the Fisher brothers conducted a series of seminal studies which began to dismantle theconcept of anatomical importance in the spread of cancer. Theirlaboratory experiments for the first time revealed that regionallymph nodes were not an effective barrier to tumor cell dissemination as proposed by Virchow in 1863. Evidence indicatedthat regional nodes were of biological rather than anatomicalimportance in cancer, and the observed findings led to theconclusion that the lymphatic and blood vascular systems wereso interrelated that it was impractical to consider them asindependent routes of neoplastic cell dissemination (5). Thelaboratory studies provided a matrix upon which an alternativehypothesis, i.e., biological rather than anatomical and mechanistic, could be formulated. Bernard Fisher synthesized thathypothesis in 1968 (6) and elected breast cancer to test it (Table1).

The concepts underlying systemic adjuvant therapy wereintroduced into full-scale clinical trials in cancer in the 1970s.A forerunner of adjuvant therapy was the recognition of possiblecomplementary therapeutic action of different modalities whenchemotherapy was first introduced. Subsequently, the demonstration of the successful application of various treatment modalities directed against the early stages of Wilms' tumor led to

the concept of combined modality treatment. After many abortive attempts to utilize chemotherapy in the early stages of adultsolid tumors in the late 1950s and 1960s (3), the concept ofpostsurgical treatment of minimal residual disease emerged.Osteogenic sarcoma and breast cancer became the arena to testnot only the value of adjuvant chemotherapy but also thelingering difficulties in crossing medical compartments (7).

The first two publications on adjuvant therapy for node-positive breast cancer raised unexpected hopes and controversies. Contrary to the cautious conclusions written by the investigators (8, 9), many physicians and patients, as well as thepublic in general, derived the erroneous impression that theproblem of breast cancer therapy was practically solved andbegan to ask about the 5- to 15-year cure rate. When the results

of the midterm analyses were available and adjuvant chemotherapy (particularly CMF3) widely utilized, a number of sur

geons and radiation therapists began to look at the results froma schematic point of view (10-12). They concluded that chemotherapy was of no practical value in postmenopausal womenwhile in younger patients its effect was mediated through ovarian suppression (13, 14). A decade of countless debates madepeople aware that there was no single or simple solution for acomplex biological problem. They also led to the (hopefullyfalse) impression that the attempt to cross medical compart-

3The abbreviations used are: CMF, cyclophosphamide, methotrexate, fluo-rouracil; ER, estrogen receptor; PgR, progesterone receptor, NSABP, NationalSurgical Adjuvant Breast and Bowel Project.

2314

on May 22, 2020. © 1988 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

MEDICINE AND TREATMENT OF BREAST CANCER

Table 1 Hrcaxt cancer: biological hypotheses which led to the contemporarymultidisciplinar)/ treatment strategy"

There is no orderly pattern of tumor cell dissemination.Tumor cells traverse lymphatics by embolization challenging the merit of en

bloc dissection.The positive lymph nodes are an indicator of a host-tumor relationship that

permit development of mÃ©tastasesrather than the instigator of distant disease.

Regional lymph nodes are of biological importance but are ineffective as barriers to tumor cell spread.

The blood stream is of considerable importance in tumor dissemination.Complex host-tumor interrelationships affect every facet of the disease.Operable breast cancer is a systemic disease and variations in local-regional

therapy are unlikely to substantially affect survival.Only by distant disease control could there be an improvement in the outcome

of breast cancer patients.In homogeneous cell populations, the effective drug kill is best described by

first-order reaction kinetics.The efficacy of chemotherapy is dose dependent, related to the body burden at

the time of drug treatment and to the presence of primary resistant tumorcells.

There is an inverse relationship between tumor stem cell burden and curabilitywith chemotherapy, the possibility of resistant cell line(s) developing appearsclosely related to cell number. Therefore, drug therapy has more chance to beeffective when applied in full dosage and in the presence of limited number oftumor cells, i.e., when each micrometastasis contains 10* down to 1 neoplastic

cell.* Adapted from Refs. 1 and 6.

ments was indeed a more difficult task than killing cancer cellswith drugs.

The aim of this Perspective is to briefly survey some of themost important adjuvant studies in breast cancer and criticallyexamine the relative contribution of medical treatment to theprimary management of breast cancer.

Preclinical Basis

The experimental foundations of systemic adjuvant therapywere derived to a great extent from studies performed byinvestigators at the Southern Research Institute (1). Theirobservation of the survival of mice with transplantable solidtumors of varying age and size subjected to resection led themto postulate a great variability in metastatic body burden at thetime of surgery. Because of this, empirical considerations hadto be utilized in the selection of drugs, treatment schedule, andduration of treatment. Surgical cure rates in solid tumors ofmice also showed a relationship to the size of the tumor. Ofinterest was the "break" point in survival curves, after which

no further recurrences were noted. These break points appearedat similar times after surgery, although the percentage of survival was different in each case, indicating that the biology ofthe tumor was similar in the various animals, but the tumorcell burden accounted for the variability experienced. A corollary to this was that definitive conclusions about therapeuticeffectiveness must be withheld until this "break" point is

reached, since it is the patients with large tumor burdens whoare least likely to benefit from adjuvant chemotherapy and whomake up the first portion of the survival curves. As for selectionof drugs and treatment schedules, they must be effective toachieve a net reduction in cell burden by the next cycle; it isapparent that the optimal treatment intensity must generallybe identical to the intensity that yields the optimal results inadvanced disease (Table 1).

Martin (15) demonstrated that both drug combinations andsurgery were essential in making an optimal impact on thesurvival of mice bearing mammary tumors. However, since thetumor burden is generally lower if chemotherapy is started inadjuvant situations, the duration of treatment needed to assureeradication is likely to be briefer. The emergence of drugresistance is also likely to advance in parallel with the progress

of the disease and constitutes a major element in adjuvanttherapy failure. As we will discuss later in this review, withlower tumor burdens the emergence of primary resistant cellsoccurs to a lesser degree.

Starting from the mid 1970s, hormones began to receiverenewed attention as potential adjuvant treatments (16, 17). Incarcinogen-induced mammary tumor in mice, demonstrationthat the antiestrogen compound tamoxifen can both delay theappearance of tumors and decrease their overall frequency hasbeen of great interest not only for its relevance in adjuvanttreatment of breast cancer but also because of its possibleinhibition of carcinogenic influences.

International Overview on Breast Cancer

Individual opinions on the interpretation of results fromadjuvant treatments for resectable breast cancer have raisedinnumerable controversies during the past few years. Whilethere is a fair agreement on the ability of adjuvant chemotherapy and/or endocrine therapy to delay recurrence, at least ingiven patient subsets, the magnitude of the benefit in terms ofmortality reduction is still being questioned.

The first two randomized studies utilizing adjuvant chemotherapy in node-positive patients were activated in 1972 (8) and1973 (9), respectively, and reasonably sound data can be examined only for the first 10 years from surgery. There is nodoubt that the relapse-free survival benefit in favor of the treatedgroups was already evident during the first years and haspersisted throughout the entire observation period, albeit somewhat diminished over time. In contrast, the overall survivalbenefit, which was always slightly in favor of the treated groups,appears evident only after the fifth year from surgery.

The lack of a statistically significant difference in the overallsurvival has cast doubts on the efficacy of adjuvant chemotherapy. Admittedly, thousands of women would have been requiredto achieve statistical power and significance on small differenceswhich are of clinical importance. An overview of all availabledata from prospective randomized studies carried out thus farhas been strongly suggested to address the controversy. To bereliable, however, an overview (or meta-analysis) must be conducted using the same standards for quality that apply to anyclinical trial: the collection of data items with common definition should proceed with explicit quality control checks; theend point to evaluate treatment effect should be specified; andthe statistical procedure utilized, the inferences that can bemade, as well as the estimates that result should be clearlydescribed. Most important, analyses of treatment effect trendsmust maintain the requirement to compare like with like (18,19).

Such an effort was initiated by Richard Peto and colleagues(20) at the Oxford University Cancer Studies Unit in 1984,with the cooperation of more than 100 individual investigatorswho have contributed their data. A full report with details onthe methodology utilized has not yet been published but theresults of the analysis were presented at the NIH ConsensusConference held in September 1985 and reported in somepapers (19).

Fig. 1 summarizes the most important findings on more than14,000 women randomized to adjuvant chemotherapy versusno adjuvant chemotherapy. Briefly, when all drugs (i.e., singleagents as well as various combinations) were considered, therewas an overall reduction in the odds of death of 14 Â±3% (SD)(P< 0.001) in favor of the chemotherapy-treated groups. Whenonly CMF-type regimens were considered the reduction was

2315




All patients

Patient age<50yr

Patient age>SOyr

All drugs CMF- type regimens

Fig. 1. Overview estimate of reduction in odds of death during first 5 yearswith adjuvant chemotherapy. All drugs, all types of adjuvant chemotherapyincluded; CMF-type regimens, only regimens including CMF, CMF-prednisone,CMF-vincristine-prednisone, etc. (adapted from Ref. 19).

24% (P < 0.0001) for the total series, with a 34% reduction inpatients younger than 50 years old and a 17% reduction forolder women.

A similar analysis was carried out on more than 16,000women randomized to tamoxifen or no tamoxifen. There wasan overall reduction in the odds of death of 14 Â±3% during thefirst 5 years. Among women older than the age of SOyears, thereduction was 18%, while there was an insufficient number ofyounger patients to draw any sound conclusion.

The results of the overview must be interpreted cautiously,for overviews are not appropriate to answer all clinical questions. The process of combining different studies with differentpatient populations and different adjuvant programs was undertaken to test the null hypothesis of no treatment effect. Ifthere were no effect on mortality, then, and in spite of differences in the treatments applied, the summary measure wouldtend to demonstrate no effect. However, if there is an effect,the overview may not accurately reflect the real impact oftreatment. In addition, the "arithmetic construct" for estimat

ing the magnitude of the effect may no longer be valid and canlead to some bizarre conclusions (18).

Similarly, indirect comparisons from two separate overviewsmust be avoided because of the same difficulties and biases ofnonrandom comparisons in clinical trials. From the abovementioned overview we cannot conclude that CMF-like regi

mens and tamoxifen have the same effect on patients older thanSOyears simply because the reduction in the odds of death wasof the same magnitude. Such a conclusion could be valid onlyif derived from a direct comparison of like with like, while inan indirect comparison patients entered into a chemotherapy

program may differ from patients entered on endocrine therapytrials.

The results of the overview are important because they indicate that there is a reduction in the odds of death followingadjuvant systemic therapy and therefore this reduction substitutes facts for opinions. Nonetheless, we like to stress that thetransition from scientific processes to the application of therapyfor individual patients should still be made cautiously and thateven well-conducted overviews will never replace individual andwell-conducted clinical trials. These studies are indeed the onlyreliable methods to provide the detailed critical informationwhich should guide the choice of treatment for patients outsidethe context of trials.

First Chemotherapy Trials in Node-positive Tumors

The original scope of the first two adjuvant studies of themodern era was to determine whether single agent or combination chemotherapy could lengthen the disease-free intervalof all women having tumor-positive homolateral axillary lymphnodes (8, 9, 21, 22). During the past decade, the NSABP hascarried out a stepwise series of trials using as adjuvant therapyL-phenylalanine mustard alone or in combination with fluo-rouracil and methotrexate. The recent long-term results (23) ofthe above-mentioned trials have indicated that: (a) adjuvantchemotherapy has altered the natural history of node-positivebreast cancer by significantly prolonging disease-free and totalsurvival; (/>)the effect of adjuvant chemotherapy was not uniform in all patient subsets, underscoring the heterogeneity oftumor and/or hosts; (<â€¢)postmenopausal as well as premeno-

pausal patients demonstrated a significant improvement indisease-free and total survival from adjuvant chemotherapy; (</)there was a significant reduction in mortality from adjuvantchemotherapy with poorly differentiated tumors regardless ofage and the number of axillary lymph nodes which were tumorpositive.

The Milan Cancer Institute results of the first adjuvant trialwith combination chemotherapy utilizing CMF versus controlin node-positive women are illustrated in Figs. 2 and 3. Essentially, the findings indicate that the most evident benefit interms of relapse-free survival (top) from chemotherapy can beobserved during the first 3 years and the difference betweencontrol and treated patients remains important at the time oflatest analysis (12 years). In fact, the cumulative odds remainedgreater than unity throughout the years, indicating an advantagefor the CMF-treated patients. As far as total survival is concerned (bottom), a marginal advantage was evident during thefirst 7 years and only after that time was there an indication foran increased benefit in favor of the treated group. When theentire series was analyzed according to menopausa! subsets, astatistically significant difference was documented only in pre-menopausal women (Fig. 3). However, as shown in Table 2, inall age subsets except women >60 years, a benefit from adjuvantCMF was evident.

The role of adjuvant systemic chemotherapy for operablebreast cancer in inducing second malignancies has remainedcontroversial for many years, especially because of some anecdotal reports on second tumors following prolonged (in someinstances for more than 5 years) administration of alkylatingagents. The NSABP (24) has recently published the risk ofacute nonlymphocytic leukemia in a series of 8483 women withbreast cancer treated on seven different clinical trials activatedsince 1971. Following adjuvant chemotherapy with melphalan-containing regimens, the risk for acute nonlymphocytic leuke-

2316



50

Mgfi


TOTAL FAIL

CTR 179 124

CMF 207 115

5 7 9 Â»YEARS

100

so

&02

Mg l

PÃŒ

TOTAL FAIL

CTR 86 60

CMF 103 50

-CMF

P=00005

YEARS

5 Â»

TOTAL DEATH

106

101

= 0.10

1YEARS

Fig. 2. Comparative 12-year results of the first CMF adjuvant program (totalseries). Top, relapse-free survival and cumulative odds; bottom, total survival andcumulative odds. CTR, control.

mia was 1.3% at 10 years as compared to 0.06% followingsurgery alone and 1.4% following postoperative irradiation. Nocases of acute nonlymphocytic leukemia have been observedthus far in a series of 666 patients treated with adjuvant CMFat the Milan Cancer Institute between 1973 and 1978 (25). Inthis series of patients, the risk for second solid tumors was4.2% within 10 years as compared to 4.0% in 179 womensubjected to radical mastectomy alone. Median latency periods,types of second cancers, and survival from second malignancieswere not different between the two subsets of patients. Thisobservation would further suggest that second tumors documented thus far cannot be entirely ascribed to treatment withadjuvant chemotherapy.

Five other research groups have tested adjuvant chemotherapy, utilizing either i -phen> lalanine mustard or CMF-like regimens versus an untreated control group, and their long-termresults almost mirrored those reported by the NSABP andMilan investigators (21, 22).

Subsequent Chemotherapy Trials

Following the above-mentioned initial experiences, numerousprospective randomized trials were designed to answer specificclinical questions. The trials can be summarized in three groups:(a) combination versus single agent chemotherapy; (b) optimaladjuvant treatment duration; and (c) the addition of eitherpostoperative irradiation or immunotherapy or endocrine therapy to adjuvant chemotherapy. It is beyond the scope of thisreview to list or discuss in detail the countless studies undertaken during the past decade. We shall refer only to maturetrials which could be utilized in the interpretation of the overallresults.

TOTAL DEATH

52

U

P < 0.02

CMF

CTR

uÂ°2[

BHFig. 3. Comparative 12-year results of the first CMF adjuvant program in

premenopausal women. Top, relapse-free survival and cumulative odds; bottom,total survival and cumulative odds. CTR, control.

Table 2 First CMFstudy:<40yrControl%

relapse-freeAt3yrAt

5yrAt10yrAt12yr%

ofsurvivalAt3yrAtSyrAt

10yrAt12 yr(22)3023141468SO2715CMF(29)6547404083665248results

related to agegroup40-49

yrControl(52)5954363294815046CMF(66)776551488982645850-59yrControl(53)4342342872604944CMF(52)7961SO4594775847>60yrControl(52)6355353087734835CMF(60)5652362880634543

Available data also confirm that in the adjuvant therapy forbreast cancer, combination chemotherapy is superior to singleagent chemotherapy (21, 22). The only exception was reportedfrom a Danish study (26) testing low-dose CMF versus cyclo-phosphamide (130 mg/m2 daily for 2 consecutive weeks fol

lowed by 2 weeks of rest period) in a large series of premenopausal women. At 6 years, the outcome was no different betweenthe two treatment groups, and the lack of superiority of poh -versus monochemotherapy may be, in part, due to the administration of low-dose CMF.

A few studies have empirically addressed the problem ofoptimal treatment duration. The second Milan adjuvant trialtesting 12 versus 6 monthly cycles of CMF confirmed at the10-year analysis that there was no statistically significant difference in the relapse-free (12 cycles, 46%; 6 cycles, 53%) andtotal survival rates (12 cycles, 50%; 6 cycles, 60%). The 10-yearresults were comparable regardless of the receptor status of theprimary tumor (27). Also other trials (22, 28) testing various

2317




adjuvant combinations for relatively "long" versus "short" pe

riods of time failed to detect significant differences both in the5- to 10-year relapse-free and in the total survival rates.

Furthermore, regardless of patient subsets, there is no conclusive evidence that the addition of prednisone, vincristine,and Adriamycin, as given, to CMF have improved treatmentoutcome (19, 29). Limiting their study to postmenopausalwomen, investigators of Cancer Acute Leukemia Group B (30)have reported significant advantage of CMF-vincristine-pred-nisone over CMF. The issue of whether postoperative irradiation could further improve the relapse-free survival when combined with adjuvant chemotherapy was properly addressedthrough randomized studies by three research groups (22). Thereported findings were quite consistent, and the authors haveconcluded that the classic postoperative irradiation failed toincrease the relapse-free survival when combined with chemotherapy. Recently, the results of an overview of randomizedtrials of postoperative adjuvant radiotherapy alone in breastcancer have appeared in the medical literature (31). Thesefindings substantiate the uselessness of local-regional irradiation following radical or modified radical mastectomy in improving total survival. The possible role of adjuvant immuno-therapy combined with chemotherapy was prospectively testedin a number of trials, eight of which are presently Ã©valuableforthe 3- and 5-year relapse-free survival. As detailed in ourprevious review, there was no evidence that various immuno-stimulants (Bacillus Calmette-GuÃ©rin,methanol extraction residue, levamisole) used in conjunction with chemotherapy wereable to improve the results over adjuvant chemotherapy alone(22, 32).

Combination treatment with hormone manipulation and cy-totoxic drugs was considered a reasonable approach during themid-1970s, given that breast cancer is composed of a heterogeneous population of endocrine (or cytotoxic)-sensitive andendocrine (or cytotoxic)-resistant cells (33). Thus, a combinedchemoendocrine approach similar to that used in the treatmentof advanced breast cancer was tested by a few research groupsin an adjuvant situation, particularly in postmenopausalwomen. The largest experience achieved thus far comes fromthe study of the NSABP testing PF chemotherapy with orwithout tamoxifen in a total of 1858 Ã©valuablepatients. Theinvestigators concluded that no overall improvement was documented at 5 years from mastectomy by the addition of adjuvanttamoxifen. A significant prolongation of the relapse-free survival was almost entirely restricted to a small subset of women>50 years old with four or more involved axillary nodes, andwhen both ER and PgR receptors were positive (34). Cumulatively, present findings in postmenopausal node-positivewomen indicate that the relapse-free survival improvementfollowing the concomitant administration of adjuvant chemoendocrine therapy is either minimal or nonexistent over thatachieved with either modality alone (22). Recent reports fromNSABP and the Eastern Cooperative Oncology Group (35, 36)indicate that patients who benefit from tamoxifen given inconjunction with a chemotherapeutic regimen show an additional advantage when the tamoxifen administration is continued beyond cessation of the chemotherapy, i.e., beyond thesecond year.

Trials with Endocrine Therapy Alone

therapy dates back to many years ago when Schnizinger suggested in 1889 that ovariectomy be done before or at the timeof mastectomy. Subsequently, a number of attempts with adjuvant ovarian ablation followed, but poor study design andapparently minimal effects led to loss of interest in this modality. During the past decade, however, the discovery of hormonalreceptors together with the apparent failure to dramaticallyimprove total survival of postmenopausal women with adjuvantchemotherapy has led to renewed interest in adjuvant endocrinetherapy (17).

Let us briefly reexamine the findings reported during the past15 years with adjuvant castration. Many of these trials alsoincluded a subset of postmenopausal women and a mix of bothnode-negative and node-positive patients. A significant survivaladvantage was thus far observed in two series of premenopausalwomen. The Toronto trial (37) reported a survival benefit fromovarian irradiation limited to the subset older than 45 yearsand provided that low-dose prednisone, administered daily for5 years or more, was part of the treatment. In the Saskatchewanstudy (38) comprising 359 women, a survival benefit fromoophorectomy was detected only at the 10-year analysis andwas limited to a subset of 64 premenopausal patients with 1-3positive nodes. These favorable experiences, however, are somewhat in contrast to the results reported from the other series.In fact, the Oslo trial (39) indicated a benefit from ovariancastration in postmenopausal women, while in the Christie trial(40) a reduction in mortality was documented mainly in patientsyounger than 45 years.

Mainly because of its minimal toxicity, the antiestrogentamoxifen has in recent years most often been the drug ofchoice in trials of adjuvant endocrine therapy. Many prospectiverandomized trials with tamoxifen for 1 to 2 years versus no-treatment controls have been carried out (17). At present,almost all case series consist of postmenopausal women andthe observed findings have shown a relapse-free survival improvement only in patients with receptor-positive tumors. Thetwo randomized trials performed in Great Britain appear atpresent to be the exception to the above observation. In particular, the Nolvadex Adjuvant Trial Organization (41) reportedrelapse-free and total survival benefits from adjuvant tamoxifenadministered for 2 years regardless of the receptor status. Itshould be pointed out, however, that receptor assay were carriedout in only 46% of eligible patients. More recently, the Scottishtrial (42) observed highly significant delay in relapse-free andsurvival rates, favoring patients given adjuvant tamoxifen forat least 5 years. The findings appeared to be independent ofnodal and menopausa! status. Furthermore, in the subset (57%)of patients in whom estimates of ER content were made,treatment outcome did not differ significantly with estrogenreceptor level, although the greatest benefit in relapse-freesurvival was in women with levels 100 fmol/mg protein ormore. However, to explain the effectiveness of tamoxifen inestrogen receptor-negative tumors also, the British investigatorspostulate that this triphenylethylene derivative has multiplemodes of action and occasionally produces a response innonendocrine target tissue malignancies (43). A more simpleexplanation may, in part, be due to rigidity of receptor qualitycontrol since it is well known from quality assurance trials thatlarge interlaboratory variations exist for steroid receptor assays(44).

Many clinicians have always thought that the complex prob- Trials in Node-negative Tumors

lem of breast cancer could be interpreted, and therefore man- Until recently, node-negative patients represented only aaged, through hormonal mechanisms. Thus, adjuvant endocrine subgroup of larger studies concentrating on node-positive tu-

2318




mors. It should be first recalled that 15 to 20% of all node-negative patients relapse during the first 5 years after local-regional therapy, and in the subsequent 5 years another 5 to10% relapse. Therefore, the number of events analyzable arefew indeed unless specific criteria for selection of patients athigh risk of early relapse are utilized. This is the main reasonwhy no significant differences were detected in most patientseries treated thus far with postoperative adjuvant chemotherapy (19).

Among the more recent adjuvant studies, it is worth mentioning the prospective randomized trial begun in Milan on aselected group of patients having node-negative and estrogenreceptor-negative tumors (45). Following adequate local-re

gional therapy, patients having the above mentioned characteristics were assigned then to receive no further therapy or toreceive CMF i.v. every 3 weeks for a total of 12 courses. Theavailable 5-year results (Fig. 4) clearly indicated that the selected subset was associated with very poor prognosis followinglocal-regional therapy only and full-dose chemotherapy wasbeneficial in both menopausa! groups. Another randomizedstudy including a small group of women with node-negativeand estrogen receptor-negative tumors has been activated inCardiff. The 3-year results mirror those obtained in the Milanstudy. In fact, a statistically significant difference (/' = 0.02)favoring patients treated with a combination including Adria-mycin, vincristine, and prednisone (relapse-free, 83%) com

pared to the control group (71%) has been reported (46). Athird trial utilizing the same criteria for patient selection as inMilan and Cardiff is being performed by the NSABP group.More than 600 patients were accrued and randomized to receiveno further treatment or methotrexate-fluorouracil chemother-

1009030;*v?~^^^N^â€”-TMP\

TOTAL\

CTRV

CTRPREPOST-P

= 0.02CMF;

PRE1

1 > i 1 I POSTAS2718AS2718DEATH1275321I 2 3 A 5

YEARS

100so30;v^-%^\' CMFTOTAL\

CTRV

CTRPREPOST-P

= 0.02CMF;

PRE1

i i , i , ont.TAS2718AS2718DEATH127S32lI 2 3 A 5

YEARS

Fig. 4. Comparative 5-year results of the CMF adjuvant program in node-negative and ER-negative tumors. Top, relapse-free survival and cumulative odds;bottom, total survival and cumulative odds. CTR, control.

apy. At the time of this review, no results are available fromthis important study.

The Challenge of Interpretation

It is now clear that tumors, and breast cancer in particular.have a "societal" aspect which is changed or lost when the

number of one component is varied at the expense of another(47). Similarly, current adjuvant trials and their investigatorsalso reflect "societal" aspects which are mainly expressed by

patient selection (i.e., prognostic factors), treatment performance (e.g., drug combinations, doses, and schedules), assessmentof response, and methods of reporting data. Also, many research physicians appear to concentrate excessively only ontheir own results, and lack of internal criticism may adverselyinfluence the design of subsequent trials. Admittedly, the attempt to make a constructive critical interpretation of themultifaceted aspects of current adjuvant results is a real challenge. We first plan to examine the problem of tumor heterogeneity since it has a great influence on prognostic factors andprobably represents the major obstacle to drug therapy today.Then we shall try to redefine the actual effectiveness of systemicadjuvant therapy in various tumor and patient subsets. In doingthis, we shall attempt to identify the real contribution of medicine in the primary treatment of resectable breast cancer withsome comments on how biological principles should temperschematic clinical research.

Prognostic Factors. Full appreciation of variables influencingprognosis has proved clinically useful; this is the reason whysuch variables need constantly to be defined and redefined givenpositive or negative treatment results. In recent years, the searchfor new prognostic factors in breast cancer has become almostfashionable, in that many publications tend to conclude thatgiven morphological aspects or tests or age, subsets are, or maybe, of prognostic significance. How reliable or confounding areall these observations? It is difficult to draw a line between whatappears interesting but is actually premature and findings whichadd sound information to the basic biological principles oftherapeutic oncology and therefore could be utilized to developnew strategies. Such biological principles include tumor cellburden, primary cell resistance, and drug dosage intensity.

Skipper ( 1) often quotes V. T. De Vita's James Ewing Lecture

emphasizing the relationship between tumor mass and resistance to chemotherapy. De Vita noted that tumor mass negatively affects the outcome of cancer chemotherapy in a mannerquite different from the way in which it does surgery or radiotherapy: "Cancer chemotherapy fails because cells develop re

sistance to anticancer drugs... Based on the somatic mutationtheory, it now appears that resistant mutants arise spontaneously early in the natural history of cancers, and the likelihood of a resistant line developing appears closely related tocell number, such that one or more resistant lines are likelypresent before most human malignancies become clinically evident. The development of permanent resistance more preciselyaccounts for the invariable inverse relationship between cellnumber and curability by drugs and the greater effectiveness ofcombination chemotherapy over single agents" (48). As empha

sized in recent years by many investigators, but first by Goldieand Coldman (49-51), a substantial and growing body of evidence implicates the emergence of drug-resistant tumor cells asa major factor responsible for treatment failure in clinical cancerchemotherapy. The origin of these drug-resistant cells appearsto have a genetic basis and to be consistent with random andspontaneous mutations. Thus, the smaller the tumor cell pop-

2319




ulation, the less likely it is to contain significant numbers of with the number of histologically involved axillary nodes. Invarious classes of drug-resistant cells, and the chances for curewill always be greater when the tumor population is least, as inthe presence of mierometastases.

In resect able breast cancer, the prognostic importance oftumor mass is best expressed by the number of histologicallyinvolved lymph nodes (52,53). In fact, the probability of distantrelapse and mortality, also in women subjected to adjuvantchemotherapy, remains inversely related to the degree of nodalinvolvement, which may represent an indirect signal of tumorcell burden in distant micrometastases. As a consequence ofobservations made by the M. D. Anderson, Milan, and NSABPgroups (21, 54, 55), the NIH Consensus Development Conference (56) has defined four lymph node categories: negativenodes; one to three positive nodes; four to nine positive nodes;and ten or more positive nodes. During the past few years,several research groups began to retrospectively analyze a number of markers that could contribute to prediction of the outcome of breast cancer patients. The prognostic determinantsare represented by hormone receptor, tumor differentiation,and nuclear DNA content in women treated with and withoutsystemic adjuvant therapy. The complexity of reported findingswill be summarized below.

Starting with the initial observation of Knight et al. (57), inmost series treated only with local-regional therapy treatmentoutcome, at least during the first 5 years, was better in stage Iand II women with ER-positive than with ER-negative tumors(58, 59). Not all reported findings represented unambiguousand consistent information, probably reflecting different technical methods utilized in the receptor assay. An example of theclinical relevance of negative ER in node-negative patients canbe seen in Fig. 4. Subsequent findings have also indicated thatPgR levels are more important than the ER levels for predictingtime to recurrence (60), and the use of PgR in conjunction withER significantly added to the predictive value of the latter (61).Since PgR expression is related to tumor ER, it was thoughtthat PgR could also serve as a prognostic indicator through itsrole as a marker of tumor differentiation. The most significantpathological features related to a concordant-positive ER andPgR status were low (well-differentiated) nuclear and histolÃ³gica! grades, slight or absent tumor lymphoid infiltrate, slight orabsent necrosis, and moderate or marked elastica in decreasingorder of importance (62). Within the past 2 years, the resultsof three major studies have appeared in the medical literatureshowing the prognostic importance of tumor grade. In theevaluation of all three series, Cox regression analysis was utilized and findings based on a large patient population. In a firstpublication, the NSABP investigators have demonstrated thatpoorly differentiated tumors displayed a more favorable response to adjuvant chemotherapy (63). In a second report (61),they have clearly indicated that tumor nuclear grade, a majorcomponent of histolÃ³gica! grade, when combined with eitherER or PgR receptors or both, was a better predictor than eithermarker alone in node-positive breast cancer patients who received chemotherapy with or without tamoxifen. They concluded that not only did tumor nuclear grade significantly addto the predictive value of ER and/or PgR, but also moreaccurate assessment of treatment outcome can be obtainedwhen more than one marker is used. The Ludwig group (64)reported that in node-positive women histological tumor graderemained a statistically significant prognostic factor for relapse-free and total survival, while a French group (65) documentedthat the Scarff and Bloom histopathological grading represented the most valuable predictor of treatment outcome along

conclusion, from recent retrospective analyses, tumor gradeappears an important prognostic variable, either independentor strongly correlated with the number of axillary lymph nodes.

Other potential prognostic determinants explored in recentyears consisted of DNA measurements of proliferative tumorcell capacity. Thymidine labeling index techniques as utilizedby Meyer et al. (66), Tubiana et al. (67), and Silvestrini et al.(68) have shown that patients with high proliferative rate (highS phase) are at high risk of early relapse. When the relativeinfluence of tumor size, ER status, and labeling index ontreatment outcome were analyzed in a series of 215 node-negative women, the labeling index retained its independentprognostic significance (69). Ploidy measurements by DNAflow cytometry have also strongly suggested that aneuploidtumors are more likely to recur. The recent work of Dressler etal. (70) on 327 node-negative women demonstrated throughmult Â¡variateanalysis that ploidy status remained an importantindependent variable to predict for recurrence while patients(total, 215) whose tumors showed high S phase with DNA flowcytometry had a significantly shortened relapse-free survivalcompared to patients whose tumor was in a low S phase. Alsowhen nuclear DNA content was measured by static cytopho-tometry in smears from breast carcinomas, ploidy showed thestrongest correlation with survival and was a significant prognostic indicator even when other prognostic factors were takeninto account (71-73).

Tumor Heterogeneity. Among malignant neoplasms, breastcancer in particular is well known to be composed of diversecell populations that are heterogeneous for a variety of characteristics. The initial morphological observations of heterogeneity have been extended to include a wide variety of genetic,biochemical, enzymatic, immunological, biological, and otherproperties (74, 75). The two most sinister aspects of intraneo-plastic diversity are the genesis of clones with metastatic potential and the existence of drug-resistant variants in primarycancers and their mÃ©tastases.Since primary drug resistance haslong been identified as a major reason for therapy failure, thecritical problem is how many subpopulations are present in aheterogeneous tumor such as breast cancer. Considering thatthe mult Â¡factorialphenomenon of intraneoplastic diversity isdynamic and constantly changing with time, it is not unreasonable to hypothesize the existence of many thousands of clones,including several hundred drug-resistant variants in l g of solidtumor containing 1 billion cancer cells (76). As pointed out byGoldie and Coldman (51), the smaller the tumor cell population, the less likely it is that it will contain significant numbersof drug-resistant tumor cells. However, the number and fractionof drug (cytotoxic and hormonal)-sensitive versus resistant tumor cells fluctuate considerably within micrometastatic colonies of the same size, and therefore even when the tumorpopulation is smallest, we must face the possibility that a singlepatient with breast cancer may indeed have many diseases. Thefrightening possibility that there exist so many variables combined with pleiotropic drug resistance (77) seems to present aninsurmountable obstacle to the eradication of metastatic diseasein breast cancer. Certainly something different in the way ofapproaches and strategies is needed which would modify, overcome, or circumvent drug resistance. Potentially efficaciousstrategies can be envisioned and the laboratory experiencewould provide the data necessary for designing clinical trialswith radiosensitizers, activated macrophages, differentiatingagents, etc. (76).

Drug Dose Intensity. In animal model systems, outcome of2320




chemotherapy depends upon drug dose, tumor sensitivity, andtumor cell burden (1, 2). Animal models do not provide forreduced dosage or delay due to toxicity, and optimum treatmentcan be determined by treating beyond lethal toxicity. This isnot possible in humans. Instead, attempts to define optimumtreatments have resulted in a plethora of schedules, combinations, and schemes to reduce doses and delay treatments. Theseschemes have obscured dose-response relationship, which onlyrecently has been pointed out with adjuvant CMF (78, 79).Following the CMF retrospective analysis, there was a progressivewidespread interest in the subject of dose intensity in cancertreatment. Dose intensity is a concept developed by Hryniukand Bush (80) consisting of the amount of drug administeredper unit time. For a single drug regimen it may be expressedsimply as mg per square meter of body surface per week,regardless of the particular schedule used. Thus remission ratesin metastatic breast cancer, relapse-free survival in node-positive breast cancer, and remission rate and survival in advancedovarian cancer correlate with average relative dose intensity.Skipper (81) has concluded that, in retrospect, dose intensity isalso a key determinant of treatment outcome and toxicity inpreviously performed animal experiments.

The above-mentioned concept should be further expanded.Dose intensity, calculated after taking into account dosagereductions and treatment delays for toxicity, is designated asreceived dose intensity. Received dose intensity correlates muchmore closely with outcome than does dose intensity calculatedfrom the intended protocol (82). We have just completed ananalysis of received dose intensity in a previously reportedrandomized trial of 6 versus 12 months of treatment with CMFadjuvant chemotherapy and compared received dose intensitybetween the 6-month group and the first 6 complete cycles ofthe 12-month group. The results indicated that virtually theentire survival advantage accruing the 6-month treatment wasconcentrated in the postmenopausal women with greater than3 axillary lymph nodes involved, and this was the only subsetin which there was a difference in received dose intensity.Although it is tempting to conclude that the difference inreceived dose intensity produced the improved treatment outcome, other unrecognized factors may also have contributed(83). The concept of dose intensity must be separated from theconcept of total dose. It remains to be seen whether, and towhat extent, dose intensity correlates with outcome independently of total dose (82). The 10-year results of CMF 6 versus12 cycles would suggest that dose intensity is more importantthan total dose. Although the method of Hryniuk has beencriticized (84), the scientific hypothesis of dose intensity appears to be consistent with results from retrospective analysis.

The Contribution of Medicine

The contribution of medicine, or medical oncology, in theprimary management of breast cancer can be considered onvarious levels. The first, and probably the most important,contribution is the demonstration that often breast cancer isnever a localized disease in its clinical phase or for some timebefore this. The numerous trials performed during the past 15years with adjuvant chemotherapy and endocrine treatments inwomen with different age, tumor size, location, and nodal statushave eventually demonstrated through the systematic evaluation of treatment failure that clinical findings are consistentwith the initial scientific hypothesis that occult micrometastasesdo exist and represent the crucial aspect of clinical outcome.Never in the history of solid tumors was a disease so accurately

staged and restaged as breast cancer in an adjuvant situation.The recognition that the course of breast cancer is mainlydependent from its inception upon distant mÃ©tastasesand notupon tumor size and location within a quadrant has opened thecompartment in which this tumor was confined for decadeswhen it was considered primarily a surgical disease.

The second contribution of medical oncology concerns theobservation derived primarily from failures of drug treatments,continuing that breast cancer should not be considered at theclinical level as a single disease but as a protean group ofneoplasms the clinical behavior of which reflects tumor cellheterogeneity. Thus, drug resistance, a phenomenon that hastoo often defeated clinicians with depressing regularity, hasbeen most illuminating for scientists.

The third contribution deals with the limited but consistentsuccesses of systemic adjuvant therapy in altering the course ofresectable breast cancer. In spite of the strong possibility thattumor heterogeneity, and hence drug resistance, may also bepresent in micrometastatic foci, current combination chemotherapy (in particular CMF) and tamoxifen were able to alterthe course of high-risk resectable breast cancer during the first5 to 12 years from beginning of primary therapy, at least ingiven subsets.

The fourth contribution comes as a consequence of the abovethree considerations and refers to the complexity of new directions in which clinical research can improve treatment outcome.It appears obvious that studies on the technique(s) for optimallocal control will be useful for management but not influentialon survival. Indeed, since many mammary cancers at the timeof diagnosis are probably not a localized disease, clinical research efforts should not focus on less empirical administrationof growth-inhibiting compounds and on their optimal sequencewith surgery and radiotherapy. As already stressed by a numberof investigators (85-89), reversing the sequence of treatmentand thus beginning with short-term (e.g., 2-3 cycles) chemotherapy or endocrine therapy before local-regional therapy (so-called neoadjuvant therapy) may provide potentially useful clinical and biological information. In particular, the possible advantages of neoadjuvant therapy would be: (a) in vivomonitoring of drug response so guidance may be provided for subsequent agents to be used in the postoperative phase; (/>) todetermine the relation of drug response to the kinetic, biochemical, and histolÃ³gica!features of the tumor, measured in thepretreatment sample and, after treatment, in the mastectomyor tumorectomy specimen; (c) decreased local regional tumorcell burden with wider adoption of conservative operation alsoin tumor >4 cm; (d) in selected patients, the magnitude of drugresponse, i.e., pathological complete remission, may even eventuate into nonsurgical, nonradiotherapeutic treatment (85, 89).

These innovative studies may also allow a more appropriateselection of patients who will truly require postoperative (adjuvant) chemotherapy. Patient selection for adjuvant chemotherapy has been thus far exclusively based on the histolÃ³gica!evidence of axillary node involvement, and only a few studiesperformed in node-negative women have utilized the results ofhormonal receptors (45, 46). However, other prognostic variables, such as labeling index, S phase, and ploidy, should nowbe taken into consideration within the context of researchstudies, to separate "favorable" versus "unfavorable" subsets.

Furthermore, measurements of serum and morphological tumor markers (90) as well as measurements of mdr-l geneexpression (77) in human tumor specimens obtained at surgeryor by biopsy should be helpful in the design of treatmentprotocols and should more accurately predict which cancers

2321




will respond to certain drugs and which will not. The carefulevaluation of the new variables may enable complete redefinition of the prognostic factors within a multidisciplinary approach. By favorable subsets we mean those with slow-growingtumors which could be controlled through a minimum amountof therapy, i.e., surgery alone or radiotherapy alone, or short-term chemotherapy (or endocrine therapy) followed by tumo-rectomy. The unfavorable subsets would include the fast-growing neoplasms at high risk of developing clinical evidence ofdistant mÃ©tastases;they will require a more intensive and prolonged form of drug therapy. Thus, patients will not be classifiedas operable or inoperable because of the tumor size; rather type,intensity, and duration of treatment will be selected on the basisof their biological characteristics.

Breast cancer is the only major exception to the operationaldefinition for cure, defined as the disease-free survival plateau(3), because the risk of failure, while it decreases with time,continues beyond 5 and even 10 and IS years. Instead, cliniciansshould utilize the concept of clinical control, i.e., the time spentwithout evidence of tumor recurrence following primary treatment. If adjuvant therapy has a more variable effect resultingin a very prolonged survival prior to death from breast cancerin some patients, a very short survival improvement in others,and total eradication of tumor burden and cure in anothergroup of patients, then the effect of systemic therapy versuslocal-regional modality alone might be more accurately expressed by determining the difference in the mean or medianbetween the relapse-free or survival curves (19). These calculations will indicate the "average" prolongation of survival for all

treated patients. Unfortunately, even for the first two studies ofadjuvant chemotherapy (8, 9) started almost 15 years ago, suchcalculations cannot be readily made until either all patientshave died or the follow-up of all treated patients exceeds themedian survival of the total group.

To explore the validity of the new investigational approach,mechanisms other than prospective randomized trials shouldprobably be used, at least in the early phases. In fact, to studynew biological parameters will require scientific creativity inthe methodological organization of groups and evidence. Randomized clinical trials with their major accomplishmentsachieved thus far in the treatment of breast cancer could bereutilized in a later phase, i.e., when the formulation of hypotheses for biological mechanisms will need to be tested on alarge group of patients.

Concluding and Summary Remarks

To the young clinical oncologist, breast cancer today appearsas a vast and complex discipline involving almost all the techniques and knowledge of biology, epidemiology, and therapeutics. Fifteen years ago, to an objective critic the search toimprove the control of high-risk breast cancer by drug treatments could almost parallel Leonardo da Vinci's efforts to fly.

Today, in spite of the vast and intrinsic biological problems wecan appreciate the reality of the progress being achieved as wellas its limits.

Present chemotherapy and hormonal therapy are far frombeing optimal adjuvant treatments. However, their somewhatempirical administration has achieved prolonged, clinically useful results. The biological and strategic significance of thisobservation far exceeds the results themselves. Thus, cliniciansno longer have reason to believe that the primary therapy forresectable breast cancer should be confined only to the operatingtheater; investigators can realize that despite the expected cell

resistance to single or multiple drugs, early administration ofmedical treatment has produced some favorable long-term results and therefore renewed respect for the importance of clinical data. It is useful to stress at this point that the time requiredfrom the introduction of potentially curative treatments for agiven tumor to the recognition and confirmation that curativetreatment has, in fact, been delivered is substantial. For theadjuvant therapy of breast cancer, the time lag will be muchlonger compared to Hodgkin's disease and testicular cancer.

Also, additional time is required for the transfer of technologyto the community so that a decrease in national mortalitystatistics can be appreciated (3).

Competing forces have continuously played a role in determining the management of breast cancer (6). Knowing this,what is important is that clear biological principles direct newapproaches. At present, major creative challenges will probablybe in methodology rather than in hypothesis formulation. Oneattractive methodology will be to measure the degree of morphological and biological alterations in the primary tumorfollowing medical treatment administered in various forms,dosages, and duration. The search for new, potentially usefulprognostic factors should be performed, in our opinion, withoutbeing obsessed by the constraint of randomized trials. Until thesalient tumor variables are evaluated through controversies,dissensions, and doubts, randomization will remain the besttechnique in assuring that proposed treatments with new premises will be assigned without bias.

The multifaceted interests, including political and sociological ones, generated by the high frequency of breast cancer maytend to blur our research objectives. Therefore, a proper balancemust be achieved between research and medical care, and, mostimportantly, scientific objectivity should be maintained to avoidharsh and discouraging criticisms. Admittedly, technical andconceptual difficulties may arise with the army of physiciansinvolved in the clinical care of breast cancer, since many ofthem have adopted present findings as standard treatment andwill therefore be reluctant to refer patients for new studies.

Many actors are now crowding the stage of breast cancer.Some are investigating new avenues of research on treatmentthrough effective cooperation with basic scientists in an attemptto formulate meaningful hypotheses. Some are engaged inrepeating on a large scale studies which have previously showneither promising or controversial results, and their goal is torefÃ®nepresent state of the an. Some are busy in the organizationof new large cooperative groups, national or international,either to design protocols "easy to manage and which do nottake up a lot of time" or to evaluate questions that are notreally meaningful because of their concern that "what is goodfor science may not necessarily be good for the patient" (91):"O! there be players that I have seen play, and heard others

praise, and that highly, not to speak it profanely, that neitherhaving the accent of Christians nor the gait of Christian, pagan,nor man, have so strutted and bellowed " (W. Shakespeare,

Hamlet, Act II, Scene II). We are frightened by the possibilitythat these "players" will eventually occupy most of our time

and engender the greatest confusion.In conclusion, the contribution of medicine in the primary

management of breast cancer rests not only on the limited butconsistent reduction in the odds of death, i.e., on the possibilityof saving lives, but on revealing more about the biologicalcomplexity of this disease. Also it has stimulated basic scientistsand research clinicians to integrate their efforts. Further clinicalprogress is expected, but good action should depend on goodactors.

2322




References

1. Skipper, H. Â£.,and Schabel, F. M.. Jr. Tumor stem cell heterogeneity:implication with respect to the classification of cancers by chemotherapeuticeffect. Cancer Treat. Rep., 68: 43-61, 1984.

2. Griswold, D. P., Jr. Body burden of cancer in relationship to therapeuticoutcome: consideration of preclinical evidence. Cancer Treat. Rep., 70: 81-86, 1986.

3. Frei, E., III. Curative cancer chemotherapy. Cancer Res., 45: 6523-6537,1985.

4. Kennedy, B. J., Calabresi, P., Carbone, P. P., et al. Training program inmedical oncology. Ann. Intern. Med., 78: 127-130, 1973.

5. Fisher, B., and Fisher, E. R. The interrelationship of hematogenous andlymphatic tumor cell dissemination. Surg. Gynecol. Obstet., 122: 791-798,1966.

6. Fisher, B. The role of science in the evolution of breast cancer management.Heath Memorial Award Lecture. In: F. C. Ames, G. R. Blumenschein, andE. D. Montague (eds.), Current Controversies in Breast Cancer, pp. 1-21.Austin, TX: University of Texas Press, 1984.

7. De Vita, V. T., Jr. The evolution of therapeutic research in cancer. N. Engl.J. Med., 298: 907-909, 1978.

8. Fisher, B., Carbone, P., Economou, S. G., el al. L-Phenylalanine mustard (I.PAM) in the management of primary breast cancer a report of early findings.N. Engl. J. Med., 292: 117-122, 1975.

9. Bonadonna. G., Brusamolino, E., Valagussa, P., et al. Combination chemotherapy as an adjuvant treatment in operable breast cancer. N. Engl. J. Med.,294:405-410, 1976.

10. Pourquier, H. The results of adjuvant chemotherapy are predominantlycaused by the hormonal changes such therapy induces. In favor. In: M. B.Van Scoy-Mosher (ed.), Medical Oncology. Controversies in Cancer Treatment, pp. 83-99, 110-111. Boston: G. K. Hall & Co., 1981.

11. Bonadonna, G., Valagussa, P., and De Palo, G. The results of adjuvantchemotherapy are predominantly caused by the hormonal changes suchtherapy induces. Opposed. In: M. B. Van Scoy-Mosher (ed.), Medical Oncology. Controversies in Cancer Treatment, pp. 100-109, 112-115. Boston:G. K. Hall & Co., 1981.

12. Blarney, R. W. Adjuvant chemotherapy in breast cancer the case against. In:J. P. Delaney and R. L. Varco (eds.), Controversies in Surgery 2, pp. 94-103. Philadelphia: W. B. Saunders Company, 1983.

13. Padmanabhan, N., Howell, A., and Rubens, R. D. Mechanism of action ofadjuvant chemotherapy in early breast cancer. Lancet, 2:411-414, 1986.

14. Valagussa, P., and Bonadonna, G. Mechanism of action of adjuvant chemotherapy in early breast cancer. Letter to the Editor. Lancet, 2: 1035-1036,1986.

15. Martin, D. S., Fugman, R. A., Stolfi, R. L., and Hayworth, P. E. Solid tumoranimal model therapeutically predictive for human breast cancer. CancerChemother. Rep., 5:89-109, 1975.

16. Henderson, I. C. Adjuvant endocrine therapy. In: J. Allegra (ed.), Management of Breast Cancer through Endocrine Therapies, pp. 38-57. Amsterdam:Excerpta Medica, 1984.

17. Pritchard, K. I. Current status of adjuvant endocrine therapy for resectablebreast cancer. Semin. Oncol., 14: 23-33, 1987.

18. Gelber, R. D., and Goldhirsh, A. The concept of an overview of cancerclinical trials with special emphasis on early breast cancer. J. Clin. Oncol.,* 1696-1703, 1986.

19. Henderson, I. C. Adjuvant systemic therapy for early breast cancer. CUIT.Probi. Cancer, //: 125-207,1987.

20. Editorial. Review of mortality results in randomized trials in early breastcancer. Lancet, 2:1205, 1984.

21. Bonadonna, G., and Valagussa, P. Adjuvant systemic therapy for resectablebreast cancer. J. Clin. Oncol., 3: 259-275, 1985.

22. Bonadonna, G., and Valagussa, P. Current status of adjuvant chemotherapyfor breast cancer. Semin. Oncol., 14:8-22, 1987.

23. Fisher, B., Fisher, E. R., Redmond, C., and Participating NSABP Investigators. Ten-year results from the National Surgical Adjuvant Breast andBowel Project (NSABP) clinical trial evaluating the use of L-phenylalaninemustard (L-PAM) in the management of primary breast cancer. J. Clin.Oncol., * 929-941, 1986.

24. Fisher, B., Rockette, H., Fisher, E. R., et al. Leukemia in breast cancerpatients following adjuvant chemotherapy or postoperative radiotherapy: theNSABP experience. J. Clin. Oncol., 3: 1640-1658, 1985.

25. Valagussa, P., Tancini, G., and Bonadonna, G. Second malignancies afterCMF for resectable breast cancer. J. Clin. Oncol., 5:1138-1142, 1987.

26. Brincker, H., Mouridsen, H. T., and Andersen, K. W. Adjuvant chemotherapy with cyclophosphamide or CMF in premenopausa] women with stage IIbreast cancer. Breast Cancer Res. Treat., 3:91-95, 1983.

27. Bonadonna, G., Valagussa, P., Zambetti, M., et al. Milan adjuvant trials forstage I-11 breast cancer. In: S. E. Salmon (ed.), Adjuvant Therapy of CancerV, pp. 211-221. Orlando, FL: GruÃ±e& Stratton, 1987.

28. Henderson, I. C., Gelman, R., Harris, J., et al. Duration of therapy inadjuvant chemotherapy trials. Nati. Cancer Inst. Monogr., 1:95-98,1986.

29. Bonadonna, G., Valagussa, P., Moliterni, A., et al. Adjuvant chemotherapyfollowing quadrantectomy plus radiotherapy or mastectomy. Proc. Am. Soc.Clin. Oncol., 6:60, 1987.

30. Tormey, D. C., Weinberg, V. E., Holland, J. F., et al. A randomized trial offive and three drug chemotherapy and chemoimmunotherapy in women withoperable node positive breast cancer. J. Clin. Oncol., /: 138-145,1983.

31. Cuzick, J., Stewart, H., Peto, R., et al. Overview of randomized trials of

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

postoperative adjuvant radiotherapy in breast cancer. Cancer Treat. Rep., 71:15-29, 1987.Treumiet-Donker, A. D., Meischke-de Jongh, M. L., and Van Putten, W. L.J. Levamisole as adjuvant immunotherapy in breast cancer. Cancer (Phila.),59:1590-1593, 1987.Osborne, C. K. Heterogeneity in hormone receptor status in primary andmetastatic breast cancer. Semin. Oncol., 12: 317-326, 1985.Fisher, B., Redmond, C., Brown, A., et al. Adjuvant chemotherapy with andwithout tamoxifen in the treatment of primary breast cancer: 5-year resultsfrom the NSABP trial. J. Clin. Oncol., 4:459-471, 1986.Fisher, B., Brown, A., Wolmark, N., et al. Prolonging tamoxifen therapy forprimary breast cancer. Findings from the National Surgical Adjuvant Breastand Bowel Project clinical trial. Ann. Intern. Med., 106: 649-654, 1987.Tormey, D. C., Rasmussen, P., and Jordan, V. C. Long-term adjuvanttamoxifen study: clinical update. Breast Cancer Res. Treat., 9: 157-158,1987.Mi-akin, J. W., Allt, W. E. C., Beale, F. A., et al. Ovarian irradiation andprednisone following surgery and radiotherapy for carcinoma of the breast.Breast Cancer Res. Treat., 3:45-48, 1983.Bryant, A. J. S., and Weir, J. A. Prophylactic oophorectomy in operableinstances of carcinoma of the breast. Surg. Gynecol. Obstet., 153:660-664,1981.Nissen-Meyer, R. Ovarian suppression and its supplement by additive hormonal treatment. In: M. Namer and C. M. Lalanne (eds.), Hormones andBreast Cancer, Vol. 55, pp. 151-158. Paris: INSERM, 1975.Cole, M. P. A clinical trial of an artificial menopause in carcinoma of thebreast. In: M. Namer, C. M. Lalanne (eds.), Hormones and Breast Cancer,Vol. 55, pp. 143-150. Paris: INSERM, 1975.Nolvadex Adjuvant Trial Organization. Controlled trial of tamoxifen assingle adjuvant agent in management of early breast cancer. Analysis of sixyears. Lancet, /: 836-840, 1985.Breast Cancer Trials Committee. Adjuvant tamoxifen in the management ofoperable breast cancer: the Scottish Trial. Lancet, 2:171-175, 1987.Ebbs, S. R., Roberts, J. V., and Baum, M. Alternative mechanism of actionof "anti-oestrogens" in breast cancer. Letter to the Editor. Lancet, 2: 621,

1987.Borjesson, B. W., McGinley, R., Foo, T. M. S., et al. Estrogen and progesterone receptor assays in human breast cancer sources of variation betweenlaboratories. Eur. J. Cancer Clin. Oncol., 23: 999-1004, 1987.Bonadonna, G., Valagussa, P., Tancini, G., et al. Current status of Milanadjuvant chemotherapy trials for node-positive and node-negative breastcancer. Nati. Cancer Inst. Monogr., 1:65-69, 1986.Williams, C. J., Buchanan, R. B., Hall, H., et al. Adjuvant chemotherapy forTI 2, N,,, Mo estrogen receptor negative breast cancer: preliminary results ofa randomised trial. In: S. E. Salmon (ed.), Adjuvant Therapy of Cancer V,pp. 233-241. Orlando, FL: GruÃ±e& Stratton, 1987.Heppner, G. H. Tumor heterogeneity. Cancer Res., 44: 2259-2265, 1984.De Vita, V. T., Jr. The James Ewing Lecture. The relationship betweentumor mass and resistance to chemotherapy: implications for surgical adjuvant treatment of cancer. Cancer (Phila.), 51:1209-1220, 1983.Goldie, J. H., and Coldman, A. J. A mathematical model for relating thedrug sensitivity of tumors to their spontaneous mutation rate. Cancer Treat.Rep., 63: 1727-1733, 1979.Goldie, J. H., and Coldman, A. J. The genetic origin of drug resistance inneoplasms: implication for systemic therapy. Cancer Res., 44: 3643-3653,1984.Goldie, J. H., and Coldman, A. J. Genetic instability in the development ofdrug resistance. Semin. Oncol., 12: 222-230, 1985.Fisher, B., Slack, N., Katrych, D., and Wolmark, N. Ten-year follow-upresults of patients with carcinoma of the breast in a co-operative clinical trialevaluating surgical adjuvant chemotherapy. Surg. Gynecol. Obstet., 140:528-534, 1975.Valagussa, P., Bonadonna, G., and Veronesi, U. Patterns of relapse andsurvival following radical mastectomy. Analysis of 716 consecutive patients.Cancer (Phila.), 41:1170-1178, 1978.Budzar, A., Smith, T., Blumenschein, G., et al. Adjuvant chemotherapy withfluorouracil, doxorubicin, and cyclophosphamide (FAC) for stage II or IIIbreast cancer: 5-year results. In: S. E. Salmon and S. E. Jones (eds.). AdjuvantTherapy of Cancer III, pp. 419-426. Orlando, FL: GruÃ±e& Stratton, 1981.Fisher, B., Bauer, M., Wickerham, D. L., et al. Relation of number of positiveaxillary nodes to the prognosis of patients with primary breast cancer. AnNSABP update. Cancer (Phila.), 52: 1551-1557, 1983.Consensus Conference: Adjuvant chemotherapy for breast cancer. JAMA,254:3461-3463, 1985.Knight, W. A., Ill, Livingston, R. B., Gregory, E. J., et al. Estrogen receptoras an independent prognostic factor for early recurrence in breast cancer.Cancer Res., 37:4669-4671, 1977.Valagussa, P., Bignami, P., Buzzoni, R., et al. Are estrogen receptors alonea reliable prognostic factor in node negative breast cancer? In: S. E. Jonesand S. E. Salmon (eds.), Adjuvant Therapy of Cancer IV, pp. 407-415.Orlando, FL: GruÃ±e& Stratton, 1984.Bonadonna, G., and Valagussa, P. Chemotherapy of breast cancer: currentviews and results. Int. J. RadiÃ¢t.Oncol. Biol. Phys., 9: 279-297, 1983.Clark, G. M., McGuire, W. L., Hubay, C. A., et al. Progesterone receptor asa prognostic factor in stage II breast cancer. N. Engl. J. Med., 309: 1343-1347, 1983.Fisher, B., Fisher, E. R., Redmond, C., et al. Tumor nuclear grade, estrogenand progesterone receptors: their value alone or in combination as indicators

2323




of outcome following adjuvant therapy for breast cancer. Breast Cancer Res.Treat., 7:147-160, 1986.

62. Fisher, E. R., Sass, R., Fisher, B., and Collaborating NSABP Investigators.Pathologic findings from the National Surgical Adjuvant Breast and BowelProject. Correlations with concordant and discordant estrogen and progesterone receptors. Cancer (Phila.), 59:1554-1559, 1987.

63. Fisher, E. R., Redmond, ( '., Fisher, B., and Participating NSABP Investi

gators. Pathologic findings from the National Surgical Adjuvant BreastProject. VIII. Relationship of chemotherapeutic responsiveness to tumordifferentiation. Cancer (Phila.), 57:181-191, 1983.

64. Davis, B. W., Gelber, R. IX, Goldhirsh, A., et al. Prognostic significance oftumor grade in clinical trials of adjuvant therapy for breast cancer withaxillary lymph node metastasis. Cancer (Phila.), Sii: 2662-2670, 1986.

65. Contesso, G., Mouriesse, H., Friedman, S., et al. The importance of histologiegrade in long-term prognosis of breast cancer. A study of 1010 patients,uniformly treated, at the Institute Gustave Roussy. J. Clin. Oncol., 5:1378-1386, 1987.

66. Meyer, J. S., Friedman, E., McCrate, M. M., and Bauer, W. C. Predictionof early course of breast carcinoma by thymidine labeling. Cancer (Phila.),51:1879-1886, 1983.

67. Tubiana, M., Pejovic, M. II., Chavaudra, N., et al. The long-term prognosticsignificance of the thymidine labeling index in breast cancer. Int. J. Cancer,53:441-445, 1984.

68. Silvestrini, R., Daidone, M. G., and Gasparini, G. Cell kinetics as a persistentprognostic marker in node negative breast cancer. Cancer (Phila.), 56:1982-1987, 1985.

69. Silvestrini, R., Daidone, M. G., Di Fronzo, G., et al. Prognostic implicationof labeling index versus estrogen receptors and tumor size in node-negativebreast cancer. Breast Cancer Res. Treat., 7:161-169, 1986.

70. Dressier, L., Clark, G., Owens, M., et al. DNA flow cytometry predicts forrelapse in node negative breast cancer patients. Proc. Am. Soc. Clin. Oncol.,6: 57, 1987.

71. Hedley, D. W., Rugg, C. A., Ng, A. B. P., and Taylor, I. W. Influence ofcellular DNA content on disease-free survival of stage II breast cancerpatients. Cancer Res., 44: 5395-5398, 1984.

72. Stai, O., Klintenberg, C., KrÃ¤nzen.G., et al. A comparison of static cytofluo-rometry for thÃ©estimation of ploidy and DNA replication in human breastcancer. Breast Cancer Res. Treat., 7:15-22, 1986.

73. Harvey, J., De Klerk, N., Berryman, I., et al. Nuclear DNA content andprognosis in human breast cancer a static cytophotometric study. BreastCancer Res. Treat., 9:101-109, 1987.

74. Fidler, I. J., and Poste, G. The cellular heterogeneity of malignant neoplasms:

implications for adjuvant chemotherapy. Semin. Oncol., 12: 207-221, 1985.75. Nicholson, G. L. Tumor cell instability, diversification, and progression to

metastatic phenotype: from oncogene to oncofetal expression. Cancer Res.,47: 1473-1487, 1987.

76. Dexter, D. L., and Leith, J. T. Tumor heterogeneity and drug resistance. J.Clin. Oncol., 4: 244-257, 1986.

77. Pastan, I., and Gottesman, M. Multiple-drug resistance in human cancer. N.Engl. J. Med., 316:1388-1393, 1987.

78. Bonadonna, G., and Valagussa, P. Dose-response effect of adjuvant chemotherapy in breast cancer. N. Engl. J. Med., 304: 10-15, 1981.

79. Bonadonna, G., and Valagussa, P. Comment on "The mÃ©thodologiedilemma

in retrospectively correlating the amount of chemotherapy received in adjuvant therapy protocols with disease-free survival." Cancer Treat. Rep., 67:527-529, 1983.

80. Hryniuk, W., and Bush, H. The importance of dose intensity in chemotherapyof metastatic breast cancer. J. Clin. Oncol., 2:1281-1288, 1984.

81. Skipper, H. E. On optimum combination chemotherapy design. Does thedesign of a combination make a big difference in the end-results achievablein moderately and very advanced experimental leukemia? Booklet 13. Birmingham, AL: Southern Research Institute, 1987.

82. McGuire, W. L., Goldie, J., Hryniuk, W., and Tormey, D. C. Drug dosageintensityâ€”a panel discussion. Breast Cancer Res. Treat., 9:87-100, 1987.

83. Hryniuk, W. M., Bonadonna, G., and Valagussa, P. The effect of doseintensity in adjuvant chemotherapy. In: S. E. Salmon (ed.), Adjuvant Therapyof Cancer V, pp. 13-23. Orlando, FL: GruÃ±e& Stratton, 1987.

84. Gelman, R. S.. and Henderson, I. C. A reanalysis of dose intensity foradjuvant chemotherapy trials in stage II breast cancer. SAKK Bull., /: 10-12, 1987.

85. Tagnon, H. J. Some changing concepts of the natural history of humanmammary cancer and their effect on diagnosis and treatment. Eur. J. CancerClin. Oncol., 22: 123-128, 1986.

86. Ragaz, J., Band, P. R., and Goldie, J. H. (eds.). Preoperative (neoadjuvant)chemotherapy. Recent Results Cancer Res., 103: 1-162, 1986.

87. Jacquillat, C., Wiel, M., and Khayat, D. (eds.). Neo-Adjuvant Chemotherapy.London: John Libbey, 1986.

88. Wagener, D. J. T., Blijham, G. H., Smeets, J. B. E., and Wils, J. A. (eds.).Primary Chemotherapy in Cancer Medicine. New York: Alan R. Liss, Inc.,1985.

89. Forrest, A. P. M., Chetty, U., Miller, W. R., et al. A human tumour model.Lancet, 2:840-842, 1986.

90. Longo, D. L. (guest ed.). Tumor markers. Semin. Oncol., 14: 85-234, 1987.91. Fisher, B. Winds of change in clinical trials: from Daniel to Charlie Brown.

Contr. Clin. Trials, 4:65-73, 1983.

2324



1988;48:2314-2324. Cancer Res Gianni Bonadonna and Pinuccia Valagussa CancerThe Contribution of Medicine to the Primary Treatment of Breast

Updated version

http://cancerres.aacrjournals.org/content/48/9/2314

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/48/9/2314To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



Documents

The Contribution of Medicine to the Primary Treatment of ... · treatment of breast cancer but also because of its possible inhibition of carcinogenic influences. International Overview