Contribution of randomised controlled trials to understanding and management of early breast cancer

The randomised controlled trial has become the gold standard for evidence based medicine; through the unbiased comparison of competing treatments it is possible to accurately quantify the cost-benefits and harm of individual treatments. This allows clinicians to offer patients an informed choice and provides the data on which purchasing authorities can make financial decisions. We, of course, subscribe to this view but also recognise this as a gross oversimplification of the power of the randomised controlled trial. The randomised controlled trial is the expression of deductive science in clinical medicine. Not only is it the most powerful tool we have for subjecting therapeutic hypotheses to the hazard of refutation¹ but also the biological fallout from such trials should allow clinical scientists to refine biological hypotheses. Trials of treatments for breast cancer have, at least twice, contributed substantially to a paradigm shift in our understanding of the disease.²

Summary points

• Clinical trials allow clinicians to accurately inform patients with breast cancer of the benefits and harm of different treatments

• Breast conserving techniques produce equivalent survival outcomes as more radical operations, but without the anticipated improvement in psychosocial morbidity

• The introduction of adjuvant systemic treatments has been associated with a significant fall in mortality from breast cancer in all age groups antedating the start of the national breast screening programme

• Counterintuitive results from clinical trials are being incorporated into an emerging conceptual model of the disease

Trials of local therapy

Postoperative radiotherapy

The first randomised trial in the management of early breast cancer can be credited to the Christie Hospital in Manchester. Patients undergoing radical mastectomy were randomised to receive postoperative radiotherapy or not.³ The study found no difference in survival, although the morbidity of the combined procedure was substantial, with 30% of patients who received radiotherapy suffering lymphoedema. It took nearly two decades for the biological importance of those observations to be appreciated and for two trials to seriously challenge the prevailing belief of the centrifugal, mechanistic spread of breast cancer.⁴,⁵ The more radical treatments in these trials (surgery plus radiotherapy) were associated with a reduced rate of local relapse, but failure to treat the axillary nodes either by surgery or radiotherapy left the long term survival unchanged.

More mature follow up of the early radiotherapy trials, together with later meta-analysis, provided another curious and unexpected result that might be considered part of the biological fallout of the deductive process. An excess mortality from cardiovascular events, particularly in those patients with left sided breast cancer, compensated for a modest reduction in mortality from breast cancer.⁶ Two recent studies of postoperative radiotherapy for patients with poor prognosis support these observations, which challenge the contemporary belief of biological predeterminism⁷,⁸ (more of this later).

Surgery

The first randomised controlled trial of breast conserving surgery compared with classical radical mastectomy was conducted by Sir Hedley Atkins and colleagues at Guy’s Hospital.⁹ This trial has been criticised for the inadequate dose of postoperative radiotherapy, but the long term results demonstrated an equivalence in outcome for node negative disease but poorer survival in the group with node positive disease treated by breast conserving surgery. Subsequently, larger and perhaps better conducted trials addressed the same issue; all have demonstrated an equivalence in survival between the conservative and more radically treated groups.¹⁰ Inevitably, the penalty of breast conservation was a modest but significantly higher rate of local relapse, and perhaps these events, or their anticipation, prevented the expected improvement in psychological morbidity with breast conserving surgery¹¹

Conclusions

Forty years of controlled trials in the local treatment of breast cancer have provided us with some reasonably robust conclusions: the more radical the treatments the better the local control, but also the greater physical morbidity. Breast conserving techniques—which usually involve wide local excision, axillary node dissection, and radiotherapy—do not increase the risk of distant relapse and death but have had no demonstrable effect on psychosocial or psychosexual morbidity.¹² The untreated axilla might be regarded as an index of systemic disease. Even though up to 30% of axillas contain occult metastases, failure to treat them is associated with a remarkably low incidence of progression and uncontrolled regional recurrence.¹³,¹⁴ Finally, there is the tantalising suggestion that, if the excess deaths from cardiovascular events could be excluded, postoperative radiotherapy for patients with poor prognosis might be associated with a 5% reduction in late distant relapse and death.⁶,¹⁵ These observations all have to be incorporated in the evolving paradigm and further evaluated by clinical trials.

Trials of adjuvant systemic therapy

Endocrine manipulation

Ovarian ablation

It is just over 100 years since George Beatson described the dramatic responses of three women with advanced breast cancer to surgical castration.¹⁶ This approach became standard treatment for premenopausal women with advanced breast cancer, and provided the first approach to trials of adjuvant systemic therapy by pioneers Dr Cole (Manchester)¹⁷ and Dr Nissen-Meyer (Oslo).¹⁸ These early trials of ovarian ablation produced provocative hints that the natural course of breast cancer could be perturbed by a systemic approach. Unfortunately, they were underpowered statistically, and, although they suggested that relapse free survival could be prolonged, it was not until the publication of the first world overview (see below) that we could confidently conclude that such an approach improved absolute survival.¹⁹

Endocrine ablation has fallen out of fashion, but we predict a renewed interest after the presentation of early results from modern trials using luteinising hormone releasing hormone agonists as a reversible method of ovarian suppression.²⁰ However, a concern about this approach must be the possible increased incidence of osteoporosis and heart disease among women with an early artificial menopause.

SUE SHARPLES

Tamoxifen

The next endocrine approach was the use of tamoxifen, developed as an “oral antioestrogen” in the late 1960s.²¹ Its clinical efficacy in advanced breast cancer was demonstrated in the early 1970s, and it was introduced into trials of adjuvant systemic therapy in the late ’70s. A survival advantage from two years of treatment with tamoxifen was reported in 1983,²² and many other trials reported similar results. The first world overview of data available in 1985 demonstrated unequivocally that, for unselected patients, tamoxifen was associated with a highly significant improvement in relapse free and overall survival.¹⁹ Since then trials have addressed the issue of combinations of chemotherapy and endocrine therapy²³ and have looked at the optimum duration of tamoxifen treatment.²⁴,²⁵

Fascinating biological fallout has emerged. Perhaps most notable was the first report in 1985 that two years of tamoxifen treatment significantly reduced the incidence of contralateral breast cancers,²⁶ which in part contributed to the development of trials for prevention of breast cancer among women deemed to be at high risk. Furthermore, the simplistic assumption that tamoxifen was of benefit only in women who had oestrogen receptor positive tumours has been refuted. There are now convincing data that tamoxifen given in the absence of chemotherapy to postmenopausal women has substantial benefits for patients whose tumours are oestrogen receptor negative.²⁷ A biological rationale for these observations has evolved with the realisation that tamoxifen has alternative mechanisms of action independent of its binding capacity for oestrogen receptor. This may involve the induction of transforming growth factor β from stromal fibroblasts, the down regulation of insulin-like growth factor, and perhaps a previously unrecognised effect on angiogenesis.

Adjuvant chemotherapy

Short term chemotherapy

One of the first trials of adjuvant chemotherapy was carried out by Roar Nissen-Meyer (one of the great unsung heroes of the subject), who studied the effect of five days’ perioperative treatment with cyclophosphamide.²⁸ This approach was based on the belief, current at that time, that prognosis might be in part determined by cancer cells shed at the time of surgery. His initial results were encouraging, suggesting an absolute improvement in 10 year survival of about 10%. Unfortunately, the survival benefit could not be corroborated by a large scale trial mounted by the Cancer Research Campaign Breast Cancer Trials Group.²⁹ Other trials, typically employing six cycles of adjuvant chemotherapy, have effectively refuted the suggestion that a short course of perioperative chemotherapy might be of value.²³

Long term chemotherapy

The seminal studies that demonstrated that long term chemotherapy could have an important survival advantage emerged from the national surgical adjuvant breast and bowel project, led by Professor Bernard Fisher,³⁰ and the historic trial of cyclophosphamide, methotrexate, and fluorouracil led by Dr Gianni Bonadonna in Milan.³¹ It was probably this second trial more than any other that established the role of adjuvant chemotherapy in the management of early breast cancer. Many subsequent trials have attempted to fine tune or better select the optimum duration and combination of cytotoxic drugs.

It became clear from the early results that premenopausal women gain the greatest benefit from chemotherapy, and to some extent the benefit could be predicted by the onset of amenorrhoea. This observation led to a continuing debate as to whether a component of the action of adjuvant chemotherapy might be related to chemical ovarian ablation.³² However, with longer follow up and meta-analysis of trials, there is unequivocal evidence that adjuvant chemotherapy is also of modest value in postmenopausal women.³³ Research continues in this area, including studies addressing possible synergistic or additive effects of endocrine therapy and chemotherapy.

Several studies have also investigated the use of primary (neoadjuvant) chemotherapy.^34–36 These have shown that, in spite of treatment causing dramatic down staging of the original tumour, overall survival is unchanged.

High dose chemotherapy

Frustrated by the limitations of conventional chemotherapy, many medical oncologists, led by groups in the United States, embarked on programmes of high dose chemotherapy with bone marrow transplant or stem cell rescue. Even before the results of randomised controlled trials became available, such treatments were demanded (off trial) by medical oncologists and their patients, leading to some high profile court cases in the United States. However, at the meeting of the American Society of Clinical Oncology in Atlanta (May 1999) it was reported that three trials of high dose chemotherapy had found no advantage. In our view high dose chemotherapy is the reductio ad absurdum of a conceptual belief no longer tenable. It echoes the frustrations of doctors in the 1960s, when extended radical mastectomies became fashionable among the surgical zealots.

Contribution of the world overview

By 1985 there were some tens of thousands of women already randomised into trials comparing different local regional regimens or comparing adjuvant systemic therapy with an untreated control group. It had become too complicated for any individual to make sense from even the most systematic of reviews of the literature. Furthermore, as many of these trials were seriously underpowered to detect small but clinically important differences, the scene was set for the first world overview.¹⁹

This was undertaken with the support of the United Kingdom Coordinating Committee of Cancer Research and masterminded by Professor Richard Peto’s group at Oxford University. This overview demonstrated with enormous statistical confidence the advantages of adjuvant chemotherapy for premenopausal women, tamoxifen for postmenopausal women, and the surprise finding that ovarian ablation could produce benefits of the same order as that achieved by 12 cycles of treatment with cyclophosphamide, methotrexate, and fluorouracil. Over the next 10 years, tens of thousands of women were entered into further randomised trials that allowed us to extend and refine the observations of the first world overview. Subsequent analyses were carried out in 1990 and 1995 and published two to three years later (see box).²³,²⁷,³³

Key results from the overview of breast cancer trials

• Prophylactic radiotherapy after mastectomy is associated with a significant reduction in local relapse but no major net effect on survival. Mortality from breast cancer is slightly reduced, but this is to some extent counterbalanced by a small excess in cardiac mortality

• Breast conserving regimens produce equivalent survival outcomes to more radical approaches

• Ovarian ablation, either by surgical castration or radiation induced menopause, is associated with roughly 25% relative risk reduction in mortality from breast cancer at 15 years

• Adjuvant chemotherapy using multidrug regimens for six cycles or more is associated with roughly 30% relative risk reduction in mortality from breast cancer in premenopausal women and about half this in postmenopausal women

• Adjuvant tamoxifen given for two years or more is of benefit in all subgroups of patients except premenopausal women with oestrogen receptor negative tumours. Indirect comparisons suggest that five years’ treatment might be better than two years’ and that it is associated with almost 50% reduction in the risk of contralateral breast cancer

Since the publication of the first world overview there has been a significant reduction in mortality from breast cancer among all age groups, both in the United States and the United Kingdom, which closely parallels that which might be expected by the widespread adoption of systemic treatment.³⁷

The future

Over the past few years the rate of progress has slowed, and, apart from fine tuning the duration of tamoxifen treatment and the optimum selection of cytotoxic drugs, the important new results are negative ones. Primary chemotherapy has so far failed to demonstrate survival advantages, and the recent announcement of negative results for at least three of the trials of high dose chemotherapy with stem cell rescue perhaps mark the boundary of the contemporary paradigm.

The next leap forward will require better understanding of the molecular basis of breast cancer, but, in planning future strategies, it will also be necessary to learn from the lessons of the past and to incorporate the clinical inconsistencies within the contemporary model. We have reached an exciting time in the treatment of early breast cancer, when we can build on the successes of the past, attempt to understand the failures, and incorporate the clinical observations and our new biological understanding into a revolutionary model of disease that will direct therapeutic innovations for the new millennium.