A 100 years ago, breast cancer struck 1 woman in 20. Since then, breast cancer incidence has increased steadily, not only because diagnostic modalities have become more sensitive and women and physicians are more aware of the disease, but because risk factors for the disease have become more pervasive. These include proliferation of the Western-style diet with its high proportion of animal fats, earlier menarche, delayed menopause, delayed first pregnancy, and decline of breastfeeding. These factors mainly affect developed countries but are increasingly evident in emerging countries, where, unfortunately, locally advanced breast cancers are usually diagnosed. At the end of the last century, 1 woman in 10 contracted breast cancer; now the figure is 1 in 8 in developed countries.1

Notwithstanding this alarming increase in breast cancer incidence (3 million new cases a year worldwide), we have seen a decrease in mortality for the disease in developed countries, and treatments have become much less aggressive, placing greater emphasis on the patient’s quality of life.2

Italy, and the city of Milan with its 2 distinguished Cancer Institutes, has been at the forefront of breast cancer treatment and research for many years. The first pivotal clinical trials on less aggressive surgical approaches to the disease, which had a worldwide impact, were performed in the city. These trials marked the beginning of the shift in paradigm from maximum tolerated treatment to minimum effective treatment; from aggressive surgery and radiotherapy to targeted conservative treatments; from an anatomic concept of cancer spread toward a biologic concept. Today, breast cancer treatments are much more tolerable, personalized, and effective, encouraging women to present early for breast cancer screening so that the disease is diagnosed earlier and treatments are safer. More recently, the Milan Institutes have carried out important research in pharmaco-prevention, results of which are now applied to populations of women most exposed to breast cancer risk factors.

This chapter traces the history of the landmark studies on breast cancer carried out in Milan over the last 40 years. These studies were initiated in the 1970s by our team at the National Cancer Institute and have continued at the European Institute of Oncology (Italian: Istituto Europeo di Oncologia [IEO]) of Milan.

Up to the end of the 19th century, breast cancer was considered invariably fatal. However, the American surgeon W.S. Halsted became convinced that breast cancer spread by direct extension into muscle and skin and through lymphatic ducts to regional lymph nodes, which held the cancer cells in check prior to widespread dissemination. If this were true, it meant that breast cancer was a mainly locoregional disease that could be cured definitively by timely and radical surgery.

In 1898, Halsted presented 76 mastectomized cases to the American Surgical Association3 illustrating his thesis that breast cancer could be cured. Thereafter, radical mastectomy became the accepted treatment for breast cancer. In fact, in Western countries up to the 1970s, the Halsted mastectomy was the standard treatment even for small breast cancers. And although it undoubtedly saved the lives of many women, the treatment left an ugly scar, a depression beneath the clavicle, protruding ribs, and often required a skin graft.

Two currents of thought developed slowly from Halsted’s ideas. The first sought to convince ordinary people and physicians that if diagnosed early, breast cancer was curable; it implied the development of means to diagnose early disease and also wide dissemination of the message that early diagnosis was vital.

The second current of thought has been concerned with reducing the extent of surgical and subsequent treatment, and emerged as understanding of the natural history of breast cancer increased and the mode of its initial presentation changed. At the beginning of the century, breast cancer was usually diagnosed when locally advanced: when the mass was often large and the skin was involved and sometimes ulcerated. Today, small palpable lesions are the normal presentation, and thanks to mammography and other imaging techniques, preclinical nonpalpable lesions are increasingly identified.

As with many types of malignant tumor, surgeons became interested in the use of conservative surgery to treat small-size breast cancers in the 1970s. At that time, the disease came to be viewed as one generally involving only part of the breast and not necessarily affecting the whole gland.

The first sporadic attempts to preserve the breast in an aesthetically acceptable way took place in the 1920s. Joseph Hirsch, a Frankfurt gynecologist, was probably the first to treat a substantial series of patients by simple breast resection followed by interstitial radiotherapy with encouraging results.4 Numerous small-scale studies using limited surgery, and more extensive trials with radiotherapy, sought to find a substitute for the Halsted mastectomy. And although the overall trend of the results was encouraging, little notice was taken because the era of evidence-based medicine was at hand, and only a controlled, randomized study could provide convincing evidence as to whether conservative surgery could achieve an acceptable cure rate for breast cancer.

By the 1960s, it had become clear that very aggressive surgical and radiotherapeutic approaches (dissection of the internal mammary lymph nodes and high-dose radiotherapy to regional lymph nodes) were not effective. The first randomized controlled study of breast-conserving treatment was conducted at Guy’s Hospital in the 1960s and early 1970s.5 It compared the Halsted mastectomy plus radiotherapy at 32 Gy to the regional lymph nodes with a wide resection of the tumor followed by radiotherapy to the breast and supraclavicular, axillary, and internal mammary lymph nodes, again at the dose of 32 Gy. The results, published in 1972, revealed that the percentage of local recurrences was much higher in the group treated conservatively. However, almost 20% of these recurrences were in the axilla, whereas such recurrences were rare in the mastectomized group—as expected because the Halsted operation includes removal of all axillary lymph nodes. Initial results also indicated that survival was significantly inferior among the women treated by the conservative approach. However a long-term survival analysis, conducted later, showed that only in women with T1 tumors was the Halsted mastectomy superior to breast conservation.6 This finding suggests that in most patients with tumor larger than 2 cm, occult metastases are already present in other body areas, so that the extent of local treatment in this group has little influence on long-term outcome. By contrast, for small tumors that were more often confined to the breast, the more radical local surgery was more likely to affect a cure.

An important aspect of the Guy’s study is that the conservative treatment adopted would not be considered adequate today. First, because axillary dissection was not performed, even when there was clinical evidence of metastatic involvement, and second, because the radiotherapy dose (32 Gy) was too low to eradicate any residual local disease.

The results of the Guy’s study were not encouraging for breast conservation, and when, in 1968 we proposed a randomized trial study to compare Halsted mastectomy with a new conservative approach, the World Health Organization (WHO) Expert Committee in Geneva rejected our proposal. Fortunately, my perseverance and conviction that breast conservation should be further investigated won through against opposition, and the study was accepted by the WHO Committee in December 1969. The novelty of our approach was that the conservation procedure aimed to the radical, that is, to achieve effective locoregional control of the disease. It consisted of a wide local resection of the lesion (quadrantectomy), complete axillary dissection, and high-dose radiotherapy (50 Gy) to the breast (but not to the axilla) with an additional 10-Gy boost to the tumor bed. This treatment, called quadrantectomy plus axillary dissection and radiotherapy (QUART), was to prove so successful that it remains in widespread use today, although several modifications have evolved.

Following WHO approval, what was to be known as the Milan I trial began in 1973. Seven hundred and one patients with breast cancer were recruited: 349 were randomized to the standard treatment (Halsted mastectomy) and 352 were assigned to QUART. Eligible patients had infiltrating carcinoma up to 2 cm, without clinically suspect axillary lymph nodes (T1N0). Recruitment closed in 1980. The 2 groups of patients were closely comparable thanks to the strict randomization protocol. Patients were randomized in the operating room after excisional biopsy had confirmed the histology and size of the tumor. Patients were told of the 2 possible outcomes, breast removal or breast conservation, and only if they accepted these 2 possibilities were they admitted to the study. In the first 2 years, many patients refused to enter and specifically chose mastectomy, whereas in the final years of recruitment many patients refused mastectomy and were excluded because they chose quadrantectomy. Even among surgeons, attitudes changed over this period, and the scepticism that had been widespread gave way to enthusiasm as the initial results were published in the New England Journal of Medicine in 1981.7

The most recent analysis of the Milan I trial, after more than 20 years of follow-up8 showed indistinguishable survival curves for the 2 study arms (Fig. 43-1). The latest data are that 28 of the patients (7.9%) treated conservatively and 8 (2.3%) in the Halsted group developed local recurrences. These events had no impact on survival.

The satisfyingly low rate of unfavorable events in the QUART group is in part attributable to the quality of the quadrantectomy operation. In developing this operation our aim had been to achieve secure local control. Earlier studies had indicated that intraductal spread was relatively frequent in breast cancer, and therefore it was necessary to excise the entire portion of the ductal tree (right up to the nipple) that was involved by the carcinoma. The ductal system of the breast is made up 10 to 15 relatively independent sets of branching ducts and lobules called lactiferous units, or mammary lobes. The term quadrantectomy seemed appropriate because the operation maintained the radicality that characterized mastectomy (removal of skin, subcutaneous tissue, gland, and fascia of the pectoralis muscle) but was limited to more or less 1 breast lobe. Furthermore, quadrantectomy indicates the position of the tumor to the surgeon.

Subsequent studies confirmed that local recurrence is often due to residual disease within the ductal system,9 and Holland showed that invasive and in situ breast cancer was often limited to a single breast lobe and commented that breast cancer could be considered a “disease of the quadrant.”10

Following analysis of the initial results of the Milan I trial, our group set about critically examining their experience. First, it was important to define a local failure. The practical attitude was adopted that any glandular, subcutaneous, or cutaneous lesion arising within 3 cm of the scar should be defined as local failure, and a neoplastic lesion within the same breast larger than 3 cm from the quadrantectomy or in a different breast quadrant was considered a second primary tumor. Next, it was important to determine whether the good results were mainly due to surgery or to radiotherapy. Thus we began a second randomized trial in 1985 to compare quadrantectomy with tumorectomy combined with complete axillary dissection and radiotherapy (TART). Tumorectomy removed the tumor mass with only a limited margin of surrounding tissue, so that the task of ensuring local control was entrusted mainly to radiotherapy (dose 46 Gy to the whole breast with 15 Gy to the tumor bed by interstitial 192Ir). This second study (Milan II) concluded in December 1989, after recruiting 705 patients. Inclusion criteria were similar to those in Milan I except that tumor size could be up to 2.5 cm. As with Milan I, the comparability of the 2 groups was excellent. After 10 years of follow-up, the 2 groups differed significantly in terms of the frequency of local failures: 25 (9.3%) in the QUART group and 63 (23%) in the TART group, although overall survival was identical, again indicating that local failure is not a negative factor for survival. An additional finding of Milan II regarded the prognostic significance of the histopathologic entity known as extensive intraductal component (EIC). The presence of EIC was associated in both groups with increased risk of local failure; however, this risk was much higher in the TART than the QUART group. The reason for this was that when EIC was present, simple tumorectomy inevitably left residual local tumor. It was also shown that breast cancer with EIC is less radiosensitive. This finding had not emerged from the Milan I trial, in which only 3 local failures with EIC were observed, almost certainly because of the greater local extent of QUART. The conclusion of Milan II was that a wide resection such as quadrantectomy was necessary to reduce the risk of local failure to acceptable levels.11