Age

Age is the single most important risk factor for the development of breast cancer (Colditz and Rosner 2000). The incidence of breast cancer increases with age, doubling every 10 years to the menopause, and then the rate of increase slows, levelling to a plateau after 80 years (Anderson et al 2005). This creates a point of inflection in the age-specific incidence curve known as ‘Clemmensen’s hook’ (Clemmensen 1948). The incidence of oestrogen-receptor-alpha (ERα)-negative tumours rises rapidly to 50 years and then flattens out/decreases, whereas the incidence of ERα-positive tumours is similar up to the age of 50 years but then continues to increase, albeit at a slower pace. As a result, ERα-negative tumours tend to occur earlier in life and ERα-positive tumours are more common in older women. The peak age of onset for these two tumour phenotypes is 50 years and 70 years, respectively.

Ethnicity changes the effect of age on breast cancer risk. African-American women under 50 years of age have a higher age-specific incidence of breast cancer than their Caucasian counterparts (Vogel 1998). However, the incidence of breast cancer is higher in Caucasian women after 50 years of age. The difference in age-adjusted breast cancer mortality rates between African-American and Caucasian women in the 1980s was probably due to the greater incidence of ERα-negative tumours in African-American women, who will not have benefited from the introduction of adjuvant hormonal therapy (Jatoi et al 2005).

Geographical variation

Worldwide, more than 1 million women are diagnosed with breast cancer every year, accounting for 10% of all cancer cases and 23% of all female cancer cases (Ferlay et al 2004). Incidence rates for breast cancer are six times greater in the Western world than in underdeveloped countries. Approximately 430,000 new cases of breast cancer occur each year in Europe and an estimated 212,920 in the USA. The lowest rates in Europe are found in Romania and Latvia, and the highest rates are found in Northern and Western Europe. The incidence of breast cancer in American Hispanic women is 40–50% that of non-Hispanic White women.

Migrants from low-risk countries to high-risk countries have been shown to acquire the risk of the ‘host nation’ within two generations (Tominaga 1985, Ziegler et al 1993). Japanese migrants to the USA acquire an increased breast cancer risk compared with the population in Japan, and there is increasing evidence that the earlier in life a woman takes up residence in a ‘high-risk’ country, the higher the risk of breast cancer compared with the country of origin (Shimizu et al 1991, Ziegler et al 1993).

Ovulatory cycles

Events in a woman’s life that alter her lifetime ovulatory cycles appear to correlate with the risk of breast cancer. Women who start menstruating early or have a late menopause have a 30–50% increase in risk of developing breast cancer. Similarly, late menarche and an early menopause lead to an equivalent reduction in breast cancer risk. The risk of developing breast cancer is doubled in those women who have a natural menopause after 55 years of age compared with those who experience the menopause before 45 years of age. A menopause induced before 40 years of age reduces the risk of breast cancer by almost two-thirds (Hankinson et al 2004).

Age at first pregnancy

Late age at first birth and nulliparity increase the lifetime risk of breast cancer. Nulliparity has been a well-known risk factor for developing breast cancer since Ramizzini described horrendis mammarium canceris in Catholic nuns (Ramazzini 1713).The risk of breast cancer in women who have their first child after 30 years of age is double that of women who have their first child before 20 years of age. The group at highest risk are those women who have their first child after 35 years of age. These women have a greater risk than nulliparous women. An early age of birth of a second child confers a reduced risk. The protective effects of an early full-term pregnancy have been observed in a number of ethnic groups and geographic locations, suggesting that the parity-induced protection results from biological changes in the breast rather than environmental factors.

Benign breast disease

Prospective and retrospective studies have shown a relative risk of breast cancer of 1.5–1.6 for women with benign breast disease compared with the general population (Hartmann et al 2005). This increased risk has been shown to persist for 25 years after biopsy. The relative risk of proliferative changes with atypia was 4.24 compared with a relative risk of 1.88 for those without atypia and 1.27 for non-proliferative lesions. The age at diagnosis of the benign breast disease appears to modify the risks related to the histological appearance of benign breast disease. The presence of atypia in women under 45 years of age conveys twice the risk observed among women over 55 years of age. The Breast Cancer Detection and Demonstration Project showed that the risk of breast cancer among premenopausal women with atypia was elevated by a factor of 12.0, compared with 3.3 for postmenopausal women with aytpia (London et al 1992, Dupont et al 1993). An increase in breast cancers has been demonstrated in the same breast during the first 5 years of follow-up following a diagnosis of benign breast disease, particularly in women with atypia.

Ionizing radiation exposure

The understanding of radiation-related breast cancer in women derives from epidemiological studies of patients exposed to diagnostic or therapeutic medical radiation (mantle irradiation for lymphoma) and of the Japanese atomic bomb survivors (United Nations Scientific Committee on the Effects of Atomic Radiation 2000). The breast tissue of young women is highly sensitive to the carcinogenic action of ionizing radiation. Only the bone marrow and the infant thyroid gland are more sensitive to the cancer-causing effects of radiation. Ionizing radiation is an established breast cancer risk factor and this risk has been shown to increase linearly with dose. Age at exposure directly affects radiation-related breast cancer risk, with the greatest risk seen in women exposed before 20 years of age and a significantly reduced risk seen for postmenopausal women. Periods of enhanced cell proliferation, namely in utero, puberty and pregnancy, have been proposed to represent windows of increased susceptibility for mammary carcinogenesis (Ronckers et al 2005).

A two-view mammogram results in a breast tissue dose of approximately 0.3 cGy (annual exposure from all sources of natural ‘background’ radiation being approximately 0.1 cGy). Therefore, with a woman receiving 10 mammograms as a young woman, the total dose would be approximately 3 cGy to each breast. Bearing in mind that epidemiological studies have not detected statistically significant increases below a dose of approximately 20 cGy, and knowing that 100 cGy increases risk by approximately 40%, one can estimate that 3 cGy from periodic mammography screenings would increase the risk by approximately 1.2% or a relative risk of 1.012. Such low risks are not detectable in human studies. Nonetheless, all unnecessary radiation should be avoided and although the presumed risk is very small, it should be clear that the benefit from the medical exposure would far outweigh it.

Oral contraceptive pill

Use of the oral contraceptive pill slightly increases the risk of breast cancer in current and recent users (1–4 years following cessation) with relative risks of 1.24 and 1.16, respectively (Collaborative Group on Hormonal Factors in Breast Cancer 2002). This risk diminishes after discontinuing use and returns to normal after 10 years.

These estimates are based on the Collaborative Group on Hormonal Factors in Breast Cancer study; a collaborative meta-analysis of 54 studies in 25 countries with data on over 50,000 women with breast cancer (Anonymous 1996). Cancers diagnosed in women who have used the oral contraceptive pill tend to be less clinically advanced than those detected in women who have never used it. Users of the oral contraceptive pill are generally younger women whose breast cancer risk is comparatively low, so the small excess in current users will result in a relatively small number of additional cases. Other findings of the study were:

A large case–controlled study of 4575 women aged 35–64 years showed that current or former use of the oral contraceptive pill was not associated with a significantly increased risk of breast cancer (Marchbanks et al 2002).

Hormone replacement therapy

Hormone replacement therapy (HRT) use increases the risk of breast cancer and reduces the sensitivity of mammography. The risk of breast cancer for current or recent users of HRT increases by 2.3% per year of use. For women who have used HRT for at least 5 years (average 11 years), the increased risk was 35% (Anonymous 1997a). The effect is substantially greater for oestrogen–progesterone combinations than for oestrogen-only HRT. Risk increases with duration of use; the risk for current users of oestrogen–progesterone combinations for more than 10 years was 2.31 compared with 1.74 for 1–4 years of use. Risk decreases with cessation of use; past users (>5 years following cessation) have a similar risk to those who have never taken HRT. One recent study reported that current users of combined HRT had a 2.7-fold elevated risk of lobular cancer and a 3.3-fold risk of ductal cancer. The risk of lobular cancer was only raised in women who had used HRT for 3 years or more.

In the UK over the past 10 years, it is estimated that 20,000 extra breast cancer cases have occurred among women aged 50–64 years as a result of HRT use, and three-quarters (15,000) of these additional breast cancers are due to the use of oestrogen–progesterone HRT. HRT is used by over 20 million women in Western countries for the alleviation of perimenopausal and postmenopausal symptoms, and is an important source of exogenous oestrogen and, in some cases, progesterone exposure. A recent review concluded that the excess incidence of breast cancer, stroke and pulmonary embolism in postmenopausal women who use HRT for 5 years was greater than the reduction in incidence of colorectal cancer and hip fracture (Beral et al 2002). The risks and benefits for treating menopausal symptoms should be evaluated on an individual basis.

A decrease in the incidence of breast cancer among postmenopausal women in the UK may be linked to a decrease in the use of HRT. Researchers calculated that between 1999 and 2005, the risk of the disease fell by 14% in women in their 50s, representing 1400 fewer cases in 2005, and 3300 fewer over the period (Parkin 2009). They suggest that the decrease in breast cancer since 1999 in women aged 50–59 years and since 2003 in women aged 60–64 years is a consequence of the reduced use of HRT. The use of HRT began to fall after the Million Women Study in the UK and the Women’s Health Initiative Study in the USA showed that HRT could increase the risk of breast cancer. In women aged 45–69 years, the use of HRT increased from 1992 to reach a peak of 25% in 2000–2001, before falling to approximately half that by 2006.

Hormone therapy after breast cancer is becoming an increasingly relevant problem as more women survive breast cancer. The HABITS (Hormonal Replacement Therapy After Breast Cancer — Is It Safe?) trial was designed to confirm the efficacy of HRT given to women after treatment for breast cancer (Holmberg and Anderson 2004). The trial confirmed an unacceptably high risk in women allocated to receive HRT compared with those receiving best symptomatic treatment, and as a result was terminated. Analysis of women in the HABITS trial after a median follow-up of 4 years showed that 17.6% in the HRT treatment arm had developed breast cancer recurrence or a new breast cancer, compared with 7.7% in the control arm. The estimated 5-year cumulative rate for disease recurrence was 22.2% in the HRT arm and 9.5% in the control arm.

Endogenous hormones

Higher levels of endogenous hormones have been hypothesized to increase breast cancer risk. A pooled analysis of nine prospective cohort studies found a statistically significant increased risk of breast cancer in postmenopausal women with higher levels of sex hormones (Key et al 2002). The risk for women whose oestradiol levels were in the top quintile was approximately twice that compared with women whose oestradiol levels were in the bottom quintile. Evidence in premenopausal women was inconclusive.

Body weight

Increased weight and obesity, as measured by high body mass index (BMI), has been shown to moderately increase the risk of postmenopausal breast cancer and is one of the few modifiable risk factors. It is thought that between 7% and 15% of breast cancers in developed countries may be attributable to obesity. Two large studies — the EPIC study (European Prospective Investigation into Cancer and Nutrition) and the Million Women Study — have found that obese women have a 30% higher risk of postmenopausal breast cancer than women of a healthy weight. That means that if the average lifetime risk of breast cancer is one in nine, an obese woman’s lifetime risk is one in seven. An increase in weight over time can also increase the risk of breast cancer.

Diet

A high-fat diet has been positively associated with breast cancer in a number of animal and case–control studies. Pooled analysis of cohort studies found no significant association between fat intake and breast cancer risk, whilst a meta-analysis found an association between higher total and saturated fat intake and an increased risk of breast cancer. A recent prospective study confirmed a significant association between saturated fat and breast cancer risk (Bingham et al 2003).

Alcohol intake

A significant association between alcohol intake and breast cancer has been found, with an increased risk of 7% for each 10 g alcohol/day (Hamajima et al 2002, Baan et al 2007). Approximately 4% of breast cancers in women from developed countries may be attributable to alcohol. In a recent large cohort of middle-aged women in the UK, a population with low-to-moderate alcohol consumption [10 g alcohol (one drink)/day] has shown a statistically increased risk of breast cancer; they also showed an increase in cancers of the oral cavity, pharynx, oesophagus, larynx, rectum and liver (Allen et al 2009). They calculated that the increased risk accounted for an additional 11 breast cancers per 1000 women up to 75 years of age. Although alcohol and tobacco are closely related social habits, there is no direct association between tobacco and breast cancer.

Height

Taller women have an increased risk of breast cancer (Hunter and Willett 1993). A pooled analysis estimated that the relative risk for women over 1.75 m tall compared with women under 1.6 m tall was 1.22 for all women and 1.28 for postmenopausal women (van den Brandt et al 2000).

Physical activity

A report from the International Agency for Research on Cancer concluded that physical activity has a preventive effect on breast cancer. This may be an indirect effect as exercise lowers BMI, or a direct effect on hormonal and growth factor levels. This varies between studies, with one showing a 30–40% reduction in the risk of breast cancer with a few hours of vigorous activity per week versus none (Monninkhof et al 2007).

Socioeconomic status

Breast cancer has a higher incidence in higher social classes. The age-standardized incidence rate is 115.1 per 100,000 women in the least deprived quintile compared with 97.3 per 100,000 women in the most deprived quintile. This is likely to be a reflection of other factors including reproductive history and early nutritional intake.

Mammographic density

There is extensive evidence that mammographic density is a risk factor for breast cancer, independent of other risk factors, and is associated with large relative and attributable risks for the disease. Women with increasingly dense breasts have two to six times the risk of breast cancer compared with women with less dense breasts (Boyd et al 1995). Parity, menopause and other risk factors only explain 20–30% of the variance in mammographic density (Boyd et al 2002). Early studies of mother–daughter sets and small twin studies suggested that genetic factors might explain a proportion of the variation of breast tissue patterns within a given population (Boyd et al 2002, Stone et al 2006). The genetic factors that influence mammographic density may also explain variations in breast tissue involution.

Previous history of breast cancer

If a woman has had breast cancer previously, the risk of developing a second primary breast cancer is two to six times greater compared with the the risk of the general population of developing a primary breast cancer (Chen and Thompson 1999).

Family history

A family history of breast cancer, particularly in first-degree relatives, is a well-known risk factor for the development of breast cancer. The past few years have seen a significant increase in knowledge of inherited breast cancer. Approximately 80% of breast cancers are sporadic, and the other 20% are familial, occurring within the context of a positive family history. Twin studies have shown that inherited factors may account for up to 25–30% of all breast cancers (Lalloo et al 2003). These cancers are the result of a combination of genetic and environmental factors that cause acquired genetic mutations over time (Vogel and Bevers 2003). A woman with one first-degree relative has approximately double the risk of breast cancer of a woman with no family history of the disease. The risk is greater if two (or more) relatives are affected. Multiple primary cancers in one individual or related early-onset cancers in a family pedigree are highly suggestive of a predisposing gene.

The first two major susceptibility genes for breast cancer, BRCA1 and BRCA2, were identified in 1994 and 1995, respectively. Mutations and genomic rearrangements within the BRCA1 and BRCA2 genes have been clearly associated with breast cancer. The inherited mutation of these genes represents the first hit of Knudson’s two-hit model of tumorigenesis. BRCA gene mutations interfere with the DNA-repair function of the normal gene, resulting in the accumulation of chromosomal abnormalities and an increased susceptibility to develop malignancy.

BRCA1 and BRCA2 account for the largest proportion of familial breast cancer cases. It is thought that 20–25% of familial breast cancer cases are due to mutations or genomic rearrangements within these genes. The frequency of these mutations is rare, occurring in 0.1–0.5% of the general population, compared with 2% in Ashkenazi Jews (Rebbeck et al 2002). The breast cancer risk attributable to BRCA1 and BRCA2 mutations in Ashkenazi Jews is as high as 15–30% (FitzGerald et al 1996, Abeliovich et al 1997, Struewing et al 1997, Metcalfe et al 2004). A recent analysis of 22 studies showed that carrying a deleterious BRCA1 or BRCA2 mutation confers an estimated lifetime risk for developing breast cancer of 65% and 45%, respectively (Antoniou et al 2003). By the age of 40 years, carrying a deleterious BRCA1 mutation confers a 20% chance of developing breast cancer, and the risk increases with age (Kauff et al 2002). Mutations in BRCA1 are strongly associated with ovarian and fallopian tube cancers (Antoniou et al 2003). The risk of ovarian carcinoma for BRCA1 carriers is 17%, 39% and 54% at 40, 70 and 80 years, respectively (Antoniou et al 2003). Penetrance of BRCA1 with respect to ovarian, fallopian and breast cancer is greater than that for BRCA2.

Tumours associated with BRCA1 carriers are more frequently grade 3, and ERα-negative and ovarian carcinomas, which occur in BRCA1/2 families, are mostly non-mucinous epithelial cancers (Lakhani et al 2002). No single technique is able to detect all mutations, and even by sequencing the entire gene, the detection rate is only 85% (Evans et al 2003). Once a mutation has been identified in a family, definitive genetic testing can inform women more accurately of their risks and give them an informed choice of their options. Mutational analysis of an unaffected individual is problematic without checking an affected relative. Identification of a mutation will confirm an increased risk; however, the absence of a mutation does not exclude the possibility of a refractory mutation of another gene. Where there is not a dominant family history or a BRCA1/2 mutation is not identified, risk estimation gives a 1.5–3 fold relative risk with a family history of a single first-degree relative affected (Newman et al 1988, Claus et al 1994).

The Gail model

The Gail model quantifies a woman’s lifetime risk of developing breast cancer by incorporating patient age, age at menarche, age at first birth/nulliparity, number of breast biopsies, ethnic origin and history of breast cancer in first-degree relatives (Gail et al 1989, Costantino et al 1999). It does not, however, consider the ages of first-degree relatives with breast cancer, and overlooks a family history of bilateral breast disease, second-degree relatives with breast cancer and a family history of ovarian cancer. Also, the model does not account for a personal history of atypical hyperplasia and lobular carcinoma in situ, previous radiation therapy to the chest for the treatment of Hodgkin’s lymphoma, or recent migration from a region of low breast cancer risk (e.g. rural China) to a high-risk region. The Gail model remains the most frequently used in defining eligibility for risk reduction trials.

The Claus model

The Claus model provides breast cancer risk estimates based on which relatives were diagnosed with breast cancer and at what age their diagnosis was made. Initially, only data from mothers and sisters were included, but second-degree relatives were included subsequently (Claus et al 1994).

BRCAPRO

BRCAPRO was developed by statisticians at Duke University to calculate age-specific probabilities of developing breast and ovarian cancer, based on the probability that the individual carries a mutation on one of the BRCA genes (Berry et al 1997). The model uses the observed incidences of breast and ovarian cancer among BRCA gene mutation carriers and non-carriers to calculate the probability that a given individual is a mutation carrier based on his/her family history.

The Tyrer–Cuzick model

The Tyrer–Cuzick model identifies the risk of breast cancer in unaffected women by taking into account their probability of carrying genetic risk factors, namely a BRCA1/BRCA2 mutation; a notional common, low-penetrance dominant susceptibility allele that stands for all other genetic risk factors; and a number of other individual factors known to influence risk, such as age at menopause and menarche, weight, height, age, use of HRT, previous benign breast biopsies and parity (Tyrer et al 2004).

Future developments

The Breast Cancer Prevention Collaborative Group has advised that the following risk factors should be further examined by multivariate analysis in future studies: mammographic density, plasma hormone levels, bone density and fracture history, history of weight gain, BMI and hip–waist ratio (Santen et al 2007). It is thought that the addition of these risk factors will lead to a more accurate quantitative risk assessment model for use in future breast cancer prevention trials. Whilst breast cancer risk assessment using mathematical models based on epidemiological data is valid, no one model integrates benign breast disease, oestrogen exposure and family history in a comprehensive fashion. Therefore, it is important to use a variety of models in a specialized risk assessment clinic.

Surgical management

Prophylactic/risk-reducing mastectomy (RRM) is an option for breast cancer risk reduction in high-risk women. The significant psychological and physical burden associated with RRM is reserved for those women whose lifetime risk of developing breast cancer is high. High-risk women are classified as those with a lifetime risk above 25%. This risk is the equivalent of having one first-degree relative with breast cancer diagnosed below the age of 50 years, or three affected relatives (first-degree) diagnosed under the age of 60 years. The efficacy of RRM is controversial and depends on the amount of residual tissue following the procedure. A retrospective study of 639 women with a family history of breast cancer suggests that RRM is associated with a 90% reduction in risk (Hartmann et al 1999). One small prospective study investigating the efficacy of prophylactic mastectomy in BRCA1/2 carriers with a mean follow-up of 3 years showed that those women undergoing mastectomy had a significant reduction in the incidence of breast cancer (0 of 76 women) compared with the surveillance group (eight of 63 women) (Meijers-Heijboer et al 2001).

The use of RRM is increasing, with women often choosing to have breast reconstruction surgery either immediately or delayed. The timing of the reconstruction is controversial, with immediate reconstruction appropriate in the majority of women. It is now felt that the psychological benefit of emerging from a mastectomy with a breast mound far outweighs the need for a waiting period (Kronowitz 2007). The process from first consultation to surgery takes 6–12 months; this delay deliberately allows women enough time for the decision-making process.

Bilateral prophylactic oophorectomy

Studies have shown a significant reduction in the incidence of breast cancer in women with BRCA1/2 mutations that have undergone prophylactic bilateral oophorectomy (Rebbeck et al 2002). Before undergoing the procedure, the patient should take into account how long she wishes to maintain her fertility, and should receive counselling about the risks and benefits of prophylactic oophorectomy. Opinion is divided on the use of HRT following prophylactic oophorectomy, with some centres routinely recommending HRT for all patients up to 50 years of age; the decision to use oestrogens should be based on a consideration of symptoms affecting future health and quality of life.

Chemoprevention

Chemoprevention provides a non-invasive option for breast cancer risk reduction for many high-risk women. Tamoxifen, a selective oestrogen receptor modulator (SERM) well known for its antioestrogenic effects in breast tissue, was the first agent used in breast cancer risk reduction after it was found to reduce the incidence of all breast cancers by 38% and ER-positive tumours by 48% (Cuzick et al 2003).

In-situ carcinoma

These are preinvasive carcinomas. The proliferation of epithelial cells remains confined by an intact basement membrane, with no invasion into the surrounding stroma.

Ductal carcinoma in situ

Breast screening has resulted in a marked increase in the detection of ductal carcinoma in situ (DCIS), accounting for 25–30% of all screen-detected tumours (Schwartz et al 2000). Over 90% of DCIS is impalpable, asymptomatic and often detected by screening as microcalcifications (Figure 47.1). The remaining 10% are symptomatic, presenting with nipple discharge, a palpable mass or Paget’s disease of the breast. When diagnosed clinically, DCIS is often extensive or associated with a concurrent invasive tumour. Postmortem studies have shown prevalence of DCIS from 0.2% to 14% (Bartow et al 1987, Nielsen et al 1987). Risk factors for DCIS include older age at first childbirth, nulliparity and a family history of breast cancer (Raktovich 2000). Retrospective studies of cases of low-grade DCIS misdiagnosed as benign showed that 20 years after local excision, approximately 33% had developed invasive disease (Page et al 1995). DCIS is classified into two major subtypes according to the presence/absence of comedo necrosis (atypical cells with abundant luminal necrosis that fill at least one duct) (Silverstein et al 1996). A system of low, intermediate and high nuclear grade is used to classify DCIS.

• Low-grade DCIS: evenly spaced cells with centrally placed small nuclei, few mitoses and the nucleoli are not well seen.

• High-grade DCIS: pleomorphic irregularly spaced cells, with large irregular nuclei, frequent mitoses and prominent nucleoli. Often solid with comedo necrosis and calcification.

• Intermediate-grade DCIS: features between high- and low-grade DCIS.

Figure 47.1 Ductal carcinoma in situ. Intermediate-grade ductal carcinoma in situ (DCIS) exhibiting classical comedo-type casting calcification within the ducts. The linear branching pattern is very typical of DCIS. Small arrows demonstrate the casting branching pattern of DCIS. The more density clustered area of calcification is associated with a local area of invasive ductal carcinoma (arrow head).

Most cases of DCIS are unicentric, with only 1% showing multicentric disease (Holland et al 1990b).

Multicentric tumour is defined as separate foci of tumour found in more than one breast quadrant or more than 5 cm from the primary tumour. Multifocal tumour is defined as more than one tumour foci in the same quadrant (Anonymous 1998c).

The spread of DCIS locally is along the branching ducts that form the glandular breast, and explains why most DCIS recurrences occur at or near the site of the initial tumour (Holland et al 1998). Micro-invasion is uncommon in DCIS, but confers a 2% risk of lymph node metastasis (van Dongen et al 1989). Studies have shown that poorly differentiated, high-grade comedo DCIS has low ER expression, high rates of cell proliferation, and overexpression of c-erbB2 (HER-2/neu) and epidermal growth factor receptor (EGFR) (Millis et al 1996). Low-grade lesions have high ER expression, lower rates of cell proliferation and rarely express HER-2 (Millis et al 1996, Boland et al 2002). Small, localized areas of DCIS (<4 cm) should be treated with breast-conserving surgery (BCS) with or without radiotherapy, and larger lesions need to be treated with mastectomy (Schwartz et al 2000).

The incidence of macroscopic nodal involvement in DCIS is less than 1%; however, nodal micrometastases have been reported in 5–14% of patients (Schuh et al 1986, Kitchen et al 2000). Two small, single-institution studies have reported a 3.1% rate of sentinel lymph node involvement in 223 patients with pure DCIS. Sentinel lymph node biopsy (SNB) for DCIS is not currently indicated but is under investigation, and may have a role in mastectomy for DCIS (Intra et al 2003). SNB should only be considered in those patients with a higher likelihood of underlying, undetected invasive disease (i.e. younger patients, DCIS >4 cm, high-grade DCIS and patients undergoing mastectomy). Twenty-five percent of cases recur within 8 years of BCS alone, 50% of which will present with invasive disease (Fisher et al 1999, Julien et al 2000, Ottesen et al 2000, Chan et al 2001); the recurrence rate for DCIS following mastectomy is 1% (Silverstein et al 1995). The key factor for recurrence is a clear margin at the time of surgery, and poor prognostic indicators include younger age at diagnosis (<40 years), poorly differentiated/high-grade tumours, presence of comedo necrosis, ER negativity and HER-2 positivity (Yen et al 2005).

The NSABP-B17, EORTC 10853 and UK/ANZ DCIS trials evaluated the value of radiotherapy following BCS for DCIS (Fisher et al 1999, Julien et al 2000, Houghton et al 2003). All of the trials reported a significant reduction in ipsilateral recurrence following radiotherapy. The EORTC trial showed a reduction in recurrence of DCIS from 8% to 5%, and a reduction in invasive recurrence from 8% to 4% at 5 years.

Invasive carcinoma

Invasive lobular carcinoma

Invasive lobular carcinoma (ILC) is the second most common type of invasive breast cancer accounting for 8–14% of all invasive breast cancers (Arpino et al 2004). ILC clinically presents as a poorly defined thickening of the breast rather than a dominant mass. This makes the extent of the disease difficult to estimate on clinical examination, and difficult to visualize on mammography. Ultrasonography is more sensitive than mammography in detecting ILC, but may significantly underestimate the size of the lesions (Pritt et al 2004, Selinko et al 2004). Magnetic resonance imaging (MRI) is more accurate than either mammography or ultrasonography in defining the extent of the disease but is less widely available (Weinstein et al 2001). Due to the infiltrative growth pattern, there is a higher incidence of resection margin involvement than for IDC and a higher conversion rate to mastectomy (Yeatman et al 1995). ILC has a higher incidence of being bilateral, multifocal and multicentric than IDC (Figure 47.2) (Ashikari et al 1973). Axillary lymph nodes in ILC may remain impalpable even when extensively involved (Grube et al 2002). A large study from the National Cancer Database to examine treatment and outcomes in patients with ILC showed a similar 5-year survival rate between patients with ILC and IDC, and between those ILC patients receiving breast-conserving therapy and those receiving mastectomy (Winchester et al 1998).

Figure 47.2 Multifocal breast cancer. Most commonly seen in invasive lobular cancer. The left mammogram images [mediolateral oblique (MLO) and craniocaudal (CC)] have dense parenchyma and only show a single subtle mass with slight distortion (arrowhead) on the MLO view. Ultrasound shows two subtle lesions (arrowed) close to each other (LT BREAST 7.2 and 7.4). Contrast-enhanced subtraction magnetic resonance shows multiple enhancing lesions (arrow heads) in the left breast on different slices.