Breast cancer management: The road to today
When Cancer began publishing in 1948, 1 of the first articles to appear was a review of sex hormones and advanced carcinoma of the breast. Because little was then known about the biology of breast cancer, standard treatment was characterized by radical and disfiguring surgery, often with only a limited effect on long-term outcomes. Several pivotal developments in the ensuing 60 years changed this picture dramatically. The large tumors that were common at initial presentation in 1948 became increasingly uncommon with the growing use of screening mammography. It was soon suggested that these smaller tumors could be successfully treated with more conservative surgery, especially with the addition of multidrug chemotherapy and hormonal therapy as adjuvant treatments. These revolutionary developments, which have allowed many women to remain free of disease for extended periods of time, were chronicled in landmark papers that appeared in Cancer: the first clinical trial to determine whether screening mammography would improve outcomes, the early small trials comparing less invasive surgery with Halsted's radical mastectomy, the initial National Surgical Adjuvant Breast Project trial testing the efficacy of triethylenethiophosphoramide in combination with radical surgery, and the first antiestrogen trials. These articles are extraordinary not only for breaking new ground in their respective technical areas, but also for the keen insights shown by the authors into what would become important in the future. Cancer 2008;113(7 suppl):1844–9. © 2008 American Cancer Society.
The last half of the 20th century transformed the management of breast cancer. When the first edition of Cancer was published in 1948, the majority of breast tumors were discovered by accident, typically when they were relatively large. Radical and disfiguring surgery was the standard treatment for these large tumors, often with minimal long-term success. Today, most women with breast cancer are diagnosed very early, receive minimally invasive surgery, have a variety of adjuvant treatment options, and stand a good chance of remaining disease free for extended periods of time.
Several key developments were integral to this evolutionary process. The introduction and widespread use of screening mammography resulted in a sharp reduction in the average size of tumors when first discovered. The small size and the associated reduced risk of lymph node involvement meant that less invasive surgical techniques could be used for local treatment. The growing recognition of breast cancer as a systemic disease led to the use of chemotherapy or hormonal therapy for the majority of patients, with the continuous introduction of new drug regimens leading to small but steady improvements in patient survival. The beginning stages of these developments were described early on in landmark articles that were published in Cancer.1–4
The use of radiography to detect tumors in the breast was first demonstrated in 1913 by Albert Salomon, who performed radiography of mastectomy specimens to show the spread of tumor to the axillary lymph nodes.5 For the next 50 years, the technology needed to use mammography as a screening tool slowly developed, as clinicians in the US and Europe fine-tuned their abilities to produce and interpret images in a way that would be easily reproducible and clinically relevant.6–8
The idea of using mammography as a screening tool for asymptomatic women was suggested by several researchers in the early 20th century,6, 8 but it did not really gain any traction until the publication of results of the first randomized controlled trial of periodic screening with mammography and physical examination. Initial results from this study, conducted from 1963 to 1966 by Strax et al under the auspices of the Health Insurance Plan of New York, were published in Cancer in 1967.1
The question addressed in that study was whether periodic screening for breast cancer using mammography and clinical examination would lead to a lowering of the mortality rate from breast cancer. The study cohort consisted of 30,000 randomly chosen women between the ages of 40 and 64 years, with a randomly chosen control group of similar size. The study cohort was offered an initial and 3 annual evaluations consisting of medical history, clinical examination, and mammography. Mammographic technique involved low kilovoltage, various target-object distance, close coning, and industrial fine-grain film. Cephalocaudal and mediolateral views were taken.
The initial screening examinations detected 55 asymptomatic cancers (2.72 per 1000 women), whereas the control group had 1.55 cancers per 1000 women detected during the first year. This indicated an average duration of 21 months between when mammary cancer was detectable through the screening program and when it ordinarily would be detected. For the 55 cancers in the screened women, 65% had no evidence of axillary lymph node involvement, compared with only 41% with no lymph node involvement in the control group. These observations suggested that the screening program would indeed lower mortality rates from breast cancer. This turned out to be the case, with a significant decrease in mortality in the screened group compared with the control group, continuing through 16 years of follow-up.9
The study was significant as the first prospective, controlled trial to examine the benefits of screening. As such, it provided the foundation for the worldwide spread of screening mammography. Despite technology that would seem primitive by present-day digital standards, the authors were able to observe that: cancers in very small breasts were difficult to diagnose by mammography; cancers found by mammography alone differed in location from cancers found only on clinical examination; the ability to observe microcalcifications was an important asset in cancer detection; and mammography was able to detect ductal carcinoma in situ more effectively than clinical examination. Finally, they predicted that it would eventually be possible to identify high-risk groups of women based on demographic characteristics, medical history, clinical findings, mammographic breast types, and benign conditions of the breast found on mammography.
Radical Versus Conservative Surgery
Radical mastectomy was the treatment of choice for breast cancer from the early 1900s until the late 1970s. As developed by William Halsted at Johns Hopkins School of Medicine, this extreme and disfiguring surgery removed the affected breast, the axillary lymph nodes, and the chest wall muscles on the affected side. Because of the belief that breast cancer was a localized disease that could be cured with adequate surgery, virtually every woman with breast cancer received a radical mastectomy, regardless of the extent of the disease.
After World War II, the value of radical surgery for all patients was questioned by clinicians in many countries, including Robert McWhirter in Edinburgh, Sakari Mustakallio in Finland, George Crile, Jr. at the Cleveland Clinic in the US, and R.A. Dixon at the South Metropolitan Cancer Registry in southeast England. The questioning resulted in a diversity of opinions regarding the optimal treatment for patients with operable breast cancer, and established the need for controlled clinical trials to resolve this issue. Some of the earliest of these controlled trials were conducted in Copenhagen and England,10–12 with results that set the stage for the practice-changing National Surgical Adjuvant Breast Project (NSABP) B04 trial that was initiated in 1971. In 1974, John Hayward from Guy's Hospital, London, reviewed the results of the early trials in Cancer (Table 1).2
Table 1. Results of Early Trials Comparing Radical Mastectomy With Less Invasive Surgery for the Treatment of Patients With Operable Breast Cancer
|Radium Centre of Copenhagen, 1951-1957 (Kaae & Johansen 1969)10||n=425 patients (290 with clinical stage I, and 135 with operable non–stage I)||SM + XRT vs RM (including supraclavicular and internal mammary lymph nodes) with no XRT||No significant difference in 5-y LR, DR, or OS in either stage I or non–stage I patients|
|Addenbrooke's Hospital, Cambridge, 1958-1965 (Brinkley & Haybittle 1966)11||n=204 stage II patients (lymph node–positive)||RM + XRT vs SM (including accessible lymph nodes) + XRT||No significant difference in 5-y OS|
|Guy's Hospital, London, 1961-1970 (Atkins 1972)12||n=370 stage I and II patients||RM + XRT to axilla, supraclavicular fossa, and mediastinum vs wide excision + XRT (as above) +XRT to other breast||Stage I: no significant difference in DR or 5-y OS; LR was significantly higher with wide excision vs RM. Stage II: LR and DR were significantly higher with wide excision vs RM; trend toward decreased 5-y OS with wide excision vs RM.|
Superficially at least, the results of these 3 trials do not do much to clarify the issue of whether less invasive surgery is appropriate for some cases of breast cancer. Some insights can be drawn from them, however. There is a clear division between patients who have a heavily involved axilla and those who have marginal or no involvement. The Copenhagen and Guy's trials suggested that surgical interference with the axilla in patients with clinically stage I tumors was most likely unnecessary. Conversely, leaving behind heavily involved axillary lymph nodes would almost certainly have a significant effect on long-term outcomes.
On the basis of the results of these trials, Hayword provided some insights into how treatment would change in the future. First, treatment makes a difference in survival of the patient. Major changes in treatment produce major changes in recurrence and survival. He further suggested that minor changes in treatment regimens would also have their effect, something that has been verified in ongoing reports from the Early Breast Cancer Trialists Collaborative Group overview.13, 14 Second, not all patients with early breast cancer should be treated in the same way. He predicted that a growing roster of clinical, histologic, immunologic, and other factors would eventually enable us to tailor a treatment precisely for the individual patient. Third, he predicted that breast conservation would become an accepted treatment option. Although it would have no effect on patient survival, it would provide freedom from mutilation and improve quality of life. Finally, he foresaw that the availability of less invasive surgery for earlier-stage disease would provide the impetus to convince women to undergo screening mammography. He correctly predicted that this synergism would result in a decrease in the average size of tumors at presentation, resulting in an improved survival rate.
Systemic Chemotherapy for a Systemic Disease
The idea of treating cancer systemically has had appeal throughout recorded history. In the 2nd century, Galen reported compounding naturopathic remedies that included stinging nettle, honey, and powdered falcon feces. By the time of the Renaissance, Paraclesus was experimenting with mercury, lead, sulfur, and arsenical compounds to treat his patients. The first modern trial of chemotherapy came during World War II, when nitrogen mustard was used for patients with advanced lymphoma. By the early 1960s, combination chemotherapy was being used for pediatric acute lymphocytic leukemia.
The concept of using systemic chemotherapy in conjunction with surgery to decrease the recurrence rate and enhance survival in patients with breast cancer was put to the test with a series of clinical trials involving the cooperative efforts of 45 institutions that was begun in 1957 and became known as the NSABP. In an article published in Cancer in 1969, Fisher reported on the results of the initial NSABP trial.3
This trial involved 826 patients with operable breast cancer who were randomized in double-blind fashion between 2 treatment groups. In 1 group, patients received a conventional Halsted radical mastectomy and adjuvant treatment with triethylenethiophosphoramide (thio-TEPA). In the second control group, patients received radical mastectomy with or without a placebo. Patients in the drug treatment arm received thio-TEPA intravenously in a dose of either 0.8 or 0.6 mg/kg of body weight—1 dose of 0.2 or 0.4 mg/kg at the time of surgery and 1 dose of 0.2 mg/kg on each of the first 2 postoperative days.3
Recurrence and survival data were grouped according to lymph node and menopausal status. Lymph node–positive patients were further subdivided into those having 1 to 3 or ≥4 positive lymph nodes. After 5 years, there was no significant difference noted with regard to local recurrence in patients receiving thio-TEPA compared with control patients in any of the 6 groups.3 There was, however, a noteworthy difference in disease-free interval: 50% of patients in the placebo arm had recurrences by the 13th month of follow-up, compared with the 45th month of follow-up for patients receiving thio-TEPA.3
At 5 years of follow-up, survival rates for the thio-TEPA and placebo groups were not significantly different, except in premenopausal patients with ≥4 positive lymph nodes, who demonstrated a 33% higher survival rate associated with thio-TEPA.3
When accrual to this study was terminated in 1961, a second study was begun comparing thio-TEPA with 5-fluorouracil (5-FU). The results from thio-TEPA were similar to those noted in the first study, and the results with 5-FU were not promising, in addition to being associated with a high incidence of complications.
There are some interesting observations to be made from these trials. These protocols were extremely abbreviated by current standards, involving only 3 treatments at and immediately after surgery. Whereas the use of single-agent chemotherapy was largely unproductive with these protocols, combination chemotherapy was starting to be introduced at this time for hematologic malignancies; by the early 1970s, 5-FU would once again come to the forefront in combination with cyclophosphamide and methotrexate as an effective chemotherapy regimen.
Despite the disappointing results of these early trials, Fisher was far from discouraged, believing (correctly) that the use of systemic therapy as an adjunct to surgery would provide the most likely means to overcome the treatment plateau for the disease, which had existed for > 30 years.3 He foresaw that prolonging the treatment time and using multiple drugs would be a fruitful direction to follow, and also predicted that chemotherapy would likely not be recommended for the treatment of early-stage patients with negative lymph nodes. The successful completion of these trials demonstrated that large-scale studies could be successfully performed by cooperative groups.
An important conclusion reached in this article, which shaped future clinical trials for better or for worse, was that controlled clinical trials would supply the most meaningful results in the shortest period of time if they recruited only high-risk, lymph node–positive patients. Because of the growing percentage of patients with early-stage disease at the time of presentation, and because of the likelihood that late-stage disease may involve cell populations that are quite different biologically from those found in early-stage disease, this conclusion is being revisited.
A Revolution in Hormonal Therapy
Since the earliest description of breast cancer, it was clear that something unique to the female body made it an especially fertile ground for this disease. The finding that the majority of breast cancers were nourished by estrogen was not to be elucidated for many years, but the higher frequency of occurrence in women than in men was noted and studied. In the 17th century, breast cancer had become known as the nun's disease, because of the high incidence in Catholic nuns. Although many theories were put forward for this high incidence (binding of the breasts, exposure to toxic chemicals in inks used for manuscript illumination, etc.), it appears likely that a major contributing cause was the finding that these women were largely nulliparous, and had never born or nursed babies.
By the early 20th century, the connection between breast cancer and estrogen was well accepted if not completely understood. One of the first articles regarding breast cancer published in Cancer was a review of the effect of sex hormones on advanced carcinoma of the breast.15 At that time, available evidence was largely in the form of case studies and small series. There was some reported success in patients treated with androgen therapy, ovarian ablation, and the estrogenic compounds diethyl stilbestrol and ethinyl estradiol. Not surprisingly, treatment with the estrogenic compounds also resulted in an apparent acceleration of tumor growth in some patients, especially those aged < 60 years.
By the 1970s, antiestrogens were being actively investigated. In 1977, Tagnon published a review of what was then known about these compounds and their effect on advanced breast cancer.4 Estrogen receptors (ERs) in breast tissue had been described, and the concept that receptor measurement might be important in determining treatment type had been introduced. Although Tagnon reviewed small clinical studies using both nafoxidine and tamoxifen, tamoxifen came to be the treatment of choice because of excellent tolerance and ease of administration.
Tagnon analyzed results from 3 small uncontrolled trials in which patients with advanced breast cancer were treated with 10 to 40 mg/day of tamoxifen16, 17 (Table 2). These trials demonstrated regression rates (≥50% reduction in tumor size) ranging from 23% to 38%. These regression rates were better than those observed with the use of androgens, and approximated the results noted with the large ablative procedures. On the basis of these studies, it appeared that a dose of 20 mg/day, the current dose recommendation, was acceptable, although Tagnon recommended a dose of 40 mg/day based on the results reported by Ward.17
Table 2. Outcomes From Tamoxifen Treatment in Patients With Advanced Breast Cancer: Results of Early Trials
|Cole et al 197216||96||10-20 mg/day||23%, compared to 25% for stilbestrol(historic control)|
|Brewin (unpublished data)||98||10 mg bid||25%|
|Ward 197317||68||10 mg bid or 20 mg bid||38%|
Tagnon made 2 related observations based on the slim data available to him: first, that the proportion of cases responding to hormonal manipulations never exceeded 30% to 35%, and second, that ERs can be found in approximately 80% of the cases of breast cancer.4 The finding that more ER-positive patients do not respond to hormonal therapy speaks to issues that are just beginning to be fully understood today. It was recognized as far back as the 1970s when Tagnon's review was written that there are different levels of ER positivity, and that low levels of ERs are associated with a decreased response to tamoxifen. Accurate measurement and quantification of ER status remain important considerations today. In addition, much is now known regarding the clinical pharmacology of tamoxifen response. Most of the efficacy of tamoxifen is actually because of the clinical activity of its principal metabolic breakdown product, endoxifen. The enzyme responsible for this breakdown, cytochrome P450-2D6, is present in various allelic forms in different individuals, affecting its ability to metabolize tamoxifen to its active form.18
Tagnon included an important discussion of a problem that continues to vex clinicians today: how to accurately assess the efficacy of a therapeutic agent by monitoring clinical response. He pointed out that a decrease of at least 50% in clinical tumor size actually represents a very significant outcome.4 This is because, for some rapidly growing tumors, killed cells occupy nearly as much space as living cells, and the metabolism of these cells may contribute to the growth and proliferation of the surviving cells. Thus, even a modest regression in tumor size can represent a major cell kill, and a decrease of 50% may indicate that >95% of all cells have been killed. The accurate measurement of clinical response has presented a continuing problem in the design of neoadjuvant chemotherapy trials and, more recently, in the development of noninvasive ablative techniques such as radiofrequency ablation. Improvement in functional imaging techniques that accurately measure the presence of living cancer cells may finally provide an answer to this problem that was so clearly outlined ≥30 years ago.
These early articles are noteworthy for several reasons. First, they remind us of the relatively primitive standard of care that was available for our mothers and grandmothers if they were unlucky enough to develop breast cancer. More important, however, especially given the state of the technology and the paucity of good data, were the amazing insights these early clinicians had into what would be important in the future development of breast cancer management. It reminds us of the 2-fold nature of scientific progress, requiring both the painstaking accumulation of new data and the synthesis of that data by visionary thinkers.