Breast Cancer Risk Factors

Many risk factors for breast cancer have been identified, including genetic, environmental, and lifestyle factors. Some are modifiable and others are not. To estimate an individual patient's risk for breast cancer, the physician needs to elicit a detailed family history and personal past medical and breast health history. 

Once the level of risk has been established, physician and patient can discuss the best screening and management, which may involve measures such as addressing modifiable risk factors or genetic counseling. Screening is important, because early detection may lead to better outcomes.

Breast cancer is the most common type of cancer diagnosed in women, comprising 31% of all women’s cancer diagnoses in the United States. ^[1] The American Cancer Society estimates that 297,790 new cases of breast cancer will be diagnosed in women in 2023 (along with about 2800 cases in men). After lung cancer, breast cancer is the second leading cause of cancer-related death in women, accounting for 15% of cancer-related deaths. ^[1]

The incidence of breast cancer has increased by approximately 0.5% per year since the mid-2000s. ^[1]  Of note, the incidence of invasive breast cancers decreased between 1999 and 2004, which coincides with and is possibly attributable to better adherence to screening mammography recommendations for the general population of women, as well as decreasing use of menopausal hormone replacement therapy (HRT). ^{[2, 3]}

Worldwide, breast cancer is the leading cause of cancer death in women. Although the United States and Western Europe have a five-fold higher incidence of new cases of breast cancer compared with Africa and Asia, ^[4]  since 1990, the death rate of breast cancer has declined by 24% in the United States (as well as other countries in Western Europe). This may be due to increased use of screening mammography and of adjuvant chemotherapy. ^[4]

Given the high incidence and mortality of breast cancer, defining the risk factors for breast cancer has significant clinical value. Physicians can use this information to work with patients to minimize modifiable factors, and to determine appropriate screening procedures.

Factors associated with the highest risk for development of breast cancer (relative risk [RR] > 4.0) are as follows ^{[5, 6, 7]} :

• Advanced age (65 years and older)
• Atypical hyperplasia of breast (biopsy proven)
• Certain inherited genetic mutations ( BRCA1, BRCA2, TP53, ATM, CDH1); RR 4-8
• Ductal or lobular carcinoma in situ (DCIS/LCIS); RR 8-10
• Family history of early ovarian cancer (age < 50 years)
• Multiple first-degree relatives with breast cancer
• Thoracic ionizing radiation exposure before age 30 (RR 22-40)
• Personal history of early breast cancer (age < 40)

Factors associated with RR 2.1-4.0 for breast cancer are as follows ^{[5, 6, 7]} :

• High endogenous estrogen or testosterone level (postmenopausal)
• First full-term pregnancy after age 35 years
• Very dense breasts (> 50%, compared with 11-25% mammographically)
• One first-degree relative with breast cancer
• Proliferative breast diseases (eg, atypical ductal hyperplasia)
• Certain inherited genetic mutations (eg, CHEK2, PTEN)

Factors associated with RR 1.1-2.0 for breast cancer are as follows ^{[5, 6, 7]} :

• Alcohol consumption
• Age 30-35 at first full-term pregnancy
• Diethylstilbestrol exposure in utero
• Early menarche (age < 12 years)
• Height (> 5 feet 3 inches) ^[5]
• High socioeconomic class
• Ashkenazi Jewish heritage
• Personal history of breast cancer (age of onset > 40)
• Dense breasts (25-50%, compared with 11-25% mammographically)
• Benign breast conditions: Non-atypical ductal hyperplasia, fibroadenoma, sclerosing adenosis, microglandular adenosis, papillomatosis, radial scar
• Never breastfed a child
• Nulliparity (no full-term pregnancies)
• Late menopause (age > 55)
• Type II diabetes mellitus
• Obesity (post-menopausal)
• Personal history of uterine, ovarian, or colon cancer
• Recent and long-term use of hormone replacement therapy (HRT) containing estrogen and progestin
• Recent oral contraceptive use ^[8]
• Occupation: night shift
• Tobacco abuse
• Sedentary lifestyle
• Inferior cardiovascular health
• High bone mineral density ^[9]

Factors that reduce risk of breast cancer (RR < 1) include the following ^[7] :

• Asian, Hispanic, or Pacific islander race
• Breastfeeding
• Age < 20 at first pregnancy
• Tamoxifen use
• Prior risk-reduction breast surgery
• History of cervical cancer
• History of oophorectomy
• Exercise/active lifestyle
• Low bone mineral density

Age is the most significant risk factor for breast cancer. The disease is rare in women younger than 25 years, and the incidence increases with increasing age, reaching a plateau in women aged 50-69 years. In 2019, 50% of all new cases of invasive breast cancer occurred in women 50 to 69 years of age. ^[5] See Tables 1 and 2, below.

Table 1. Estimated New Female Breast Cancer Cases and Death by Age in the United States, 2019 (Open Table in a new window)

Table 2. 10-, 20-, and 30-Year Risk of Developing Breast Cancer (Open Table in a new window)

A family history of breast cancer in a first-degree relative is the most widely recognized breast cancer risk factor, but only 5-10% of women diagnosed with breast cancer have a known genetic predisposition. Women with a family history of breast cancer in a mother or sister have a 1.5-3 fold increase in the risk of developing breast cancer. 

Family history of breast cancer is a heterogeneous risk factor that depends on the number of family members affected and the age at diagnosis, as well as the number of unaffected women in the pedigree. Even in the absence of a known genetic risk factor, the presence of a family history may suggest the presence of an unknown genetic risk, or a shared environmental risk.

A family history of ovarian cancer in a first-degree relative, especially if the disease occurred at an early age (< 50 y), has been associated with an increased risk of breast cancer risk. ^[11]   

The family history characteristics that suggest increased risk of cancer are summarized as follows ^[12] :

• One or more relatives with breast or ovarian cancer
• Breast cancer occurring in an affected relative younger than 50 years
• Male relatives with breast cancer
• BRCA1 and BRCA2 mutations
• Ataxia-telangiectasia heterozygotes (4 times’ increased risk)
• Ashkenazi Jewish descent (2 times’ greater risk; independent of BRCA positivity

Although 20-30% of women with breast cancer have at least one relative with a history of breast cancer, only 5-10% of women with breast cancer have an identifiable hereditary predisposition. BRCA1 and BRCA2 mutations are responsible for 3-8% of all cases of breast cancer and 15-20% of familial cases. Rare mutations include PTEN, TP53, MLH1, MLH2, and STK11 genes, as well as ATM, BRIP1, CDH1, CHEK2, MRE11A, NBN, PALB2, RAD50, RAD51C, and SEC23B.

The BRCA1 and BRCA2 gene mutations, on chromosomes 17 and 13, respectively, account for the majority of autosomal dominant inherited breast cancers. Both genes are believed to be tumor suppressor genes whose transcribed protein products are involved with maintaining DNA integrity and transcriptional regulation.

Prevalence rates of these mutations vary by ethnic and racial groups. For BRCA1 mutations, the highest rates occur among Ashkenazi Jewish women (8.3%), followed by Hispanic women (3.5%), non-Hispanic white women (2.2%), black women (1.3%), and Asian women (0.5%). Moreover, 95% of Ashkenazi Jews with a BRCA gene mutation will have 1 of the 3 founder mutations (185delAG, 538insC in BRCA1; 6174delT in BRCA2). Women who inherit a mutation in the BRCA1 or BRCA2 gene have an estimated 50-80% lifetime risk of developing breast cancer. 

BRCA1 mutations are seen in 7% of families with multiple breast cancers and 40% of families with breast and ovarian cancer. Women with a BRCA1 mutation have a 40% lifetime risk of developing ovarian cancer. Breast cancers that develop in BRCA1 mutation carriers are more likely to be high grade, as well as estrogen receptor (ER) negative, progesterone receptor (PR) negative, and HER2-negative (triple negative) or basal-like subtype. BRCA1 mutations are also associated with a higher risk of colon, pancreatic, and prostate cancer. ^[13]

BRCA2 mutations are identified in 10-20% of families at high risk for breast and ovarian cancers but in only 2.7% of women with early-onset breast cancer. Women with a BRCA2 mutation have an approximately 10% lifetime risk of ovarian cancer. BRCA2 mutation carriers who develop breast cancer are more likely to have a high-grade, ER-positive, PR-positive, and HER2-negative cancer (luminal type). BRCA2 is also a risk factor for male breast cancer. Other cancers associated with BRCA2 mutations include the following ^[14] :

• Prostate
• Pancreas
• Gallbladder and bile duct
• Stomach
• Malignant melanoma

The US Preventive Services Task Force (USPSTF) recommends that primary care clinicians assess women with a personal or family history of breast, ovarian, tubal, or peritoneal cancer or who have an ancestry associated with BRCA1/2 gene mutations with an appropriate brief familial risk assessment tool. Women with a positive result on the risk assessment tool should receive genetic counseling and, if indicated after counseling, genetic testing. ^[15]

Li-Fraumeni syndrome, caused by TP53 mutations, is responsible for approximately 1% of cases of familial breast cancer. Bilateral breast cancer is noted in up to 25% of patients. Li-Fraumeni syndrome is also associated with multiple cancers, including the SBLLA syndrome (sarcoma, breast and brain tumors, leukemia, and laryngeal and lung cancer). Cancer susceptibility is transmitted in an autosomal dominant pattern, with a 90% lifetime risk of breast cancer (and a 56% risk by age 45), necessitating earlier screening in TP53 mutation carriers.) ^[16]

Cowden disease is a rare genetic syndrome caused by PTEN mutations. It is associated with intestinal hamartoma, cutaneous lesions, and thyroid cancer. The lifetime risk of breast cancer ranges from 25-50%. Benign mammary abnormalities (eg, fibroadenomas, fibrocystic breast disease, ductal epithelial hyperplasia, and nipple malformations) are also common. ^[16]

Other rare genetic disorders, such as Peutz-Jeghers syndrome and hereditary nonpolyposis colorectal carcinoma (HNPCC), are associated with an increased risk of breast cancer.

Hereditary syndromes may also affect the response to treatment in breast cancer. A study by Mangoni et al found an association between MSH2 and MSH3 genetic variants and the development of radiosensitivity in patients with breast cancer. The authors propose a hypothesis that mismatch repair mechanisms may be involved in the cellular response to radiotherapy and that genetic polymorphisms warrant further study as candidates for predicting acute radiosensitivity. ^[17]

The table below lists the most prevalent genetically determined breast cancer syndromes.

Table 3. Genetic Breast Cancer Syndromes (Open Table in a new window)

White and black women have a higher incidence of breast cancer than Hispanic, Asian/Pacific Islander, and Native American women. From 2015-2020, rates per 100,000 women were as follows ^[18] :

• White - 133.7
• Black - 127.8
• Hispanic - 99.2
• Asian/Pacific Islander - 101.3
• American Indian/Alaska Native - 111.3

 

Neoplastic conditions that increase the risk of breast cancer include the following:

• Previous breast cancer
• Ovarian cancer
• Endometrial cancer
• Ductal carcinoma in situ (DCIS)
• Lobular carcinoma in situ (LCIS)

Benign breast conditions that slightly increase the risk of breast cancer include the following ^[19] :

• Hyperplasia (unless mild)
• Complex fibroadenoma
• Radial scar
• Papillomatosis
• Sclerosing adenosis
• Microglandular adenosis

A meta-analysis of 386,590 women who underwent mammography found a two-fold increased breast cancer risk for those with extremely dense breast tissue (BI-RADS density D) compared with women having scattered dense breast tissue (BI-RADS density B). ^[20]  

Interestingly, a personal history of cervical cancer is associated with a lower incidence of developing breast cancer. ^[21]

One of the most widely studied risk factors in breast cancer is the use of exogenous hormones in the form of oral contraceptives (OCs) and hormone replacement therapy (HRT).

The overall evidence suggests a modestly increased risk in current users of oral contraceptives. In 1996, the Collaborative Group on Hormonal Factors in Breast Cance Risk reviewed data on 53,297 women with breast cancer and 100,239 women without breast cancer from 54 studies conducted in 25 countries,  and concluded that risk is increased 1.24 times for 10 years’ use. However, 10 years after discontinuation, this risk regresses to normal. ^[22] The progestin (synthetic progesterone)-only pill (“mini pill”) does not seem to be associated with an increased breast cancer risk, ^[23] and is commonly prescribed to women who experienced adverse effects from combination OCs or who have thrombotic risk, such as smokers or those with sickle cell disease.

In 2017, A Danish study of 1.8 million women who were followed on average for 10.9 years reported a higher risk of breast cancer with current or recent use of contemporary hormonal contraceptives. Compared with women who had never used hormonal contraception, the relative risk of breast cancer was 1.20 (95% confidence interval [CI], 1.14-1.26) in current and recent users of hormonal contraception. This risk increased from 1.09 with less than 1 year of use to 1.38 with more than 10 years of use (P=0.002). ^[24]

After discontinuation of hormonal contraception, breast cancer risk remained higher in women who had used hormonal contraceptives for 5 years or more than in those who had not used hormonal contraceptives. Estimated risk associated with various oral combination (estrogen–progestin) contraceptives varied between 1.0 and 1.6. ^[24]

Consistent epidemiologic data support an increased breast cancer risk with the use of postmenopausal HRT. Risk is directly associated with length of exposure, with the greatest risk observed for the development of hormonally responsive lobular, mixed ductal-lobular, and tubular cancers. ^[25]

Studies, including the Women’s Health Initiative (WHI) trial, have shown that the incidence of breast cancer was greater in women taking combination estrogen plus progestin formulations than in those taking estrogen-only formulations, and the cancers in women taking combination HRT were more commonly advanced or node positive. This risk seems to achieve demonstrable significance at 3 or more years of exposure. Combination HRT also appears to be associated with increased mortality. ^[26]

Published results of the WHI of estrogen-only and combination-HRT for the prevention of chronic disease indicate that the adverse outcomes associated with long-term use outweigh the potential disease prevention benefits, particularly for women older than 65 years.

A 2019 meta-analysis by the Collaborative Group on Hormonal Factors in Breast Cancer of 58 international studies that included 143,887 postmenopausal women with invasive breast cancer and 424,972 without breast cancer concluded the following about menopausal HRT and breast cancer risk ^[27] :

• Estrogen plus daily progestin, used for 5 years starting at age 50 years, would increase 20-year breast cancer risks at ages 50–69 years from 6.3% to 8.3%, an absolute increase of 2.0 per 100 women (one in every 50 users).
• Estrogen plus intermittent progestin, used for 5 years, would increase the 20-year risk from 6.3% to 7.7%, an absolute increase of 1.4 per 100 women (one in 70 users).
• Estrogen-only menopausal HRT would increase the 20-year risk from 6.3% to 6.8%, an absolute increase of 0.5 per 100 women (one in 200 users), especially in lean women, with little excess risk in obese women.
• For 10 years of use, the 20-year increases in incidence would be about twice as great as for 5 years of use.

In contrast, long-term data from two WHI trials in more than 10,000 women showed lasting decreases in breast cancer incidence and death in those receiving estrogen-only HRT with conjugated equine estrogen. At over 16 years of cumulative follow-up, breast cancer diagnosis was 23% lower in women with prior hysterectomy randomized to estrogen-only HRT compared with those assigned to placebo (hazard ratio [HR] 0.77, 95% CI 0.62-0.92); most of the reduction was due to fewer diagnoses of estrogen receptor–positive/progesterone receptor–negative disease. In addition, breast cancer deaths were 44% lower with estrogen-only HRT (HR 0.56, 95% CI 0.34-0.92). As in other studies, menopausal HRT with estrogen plus progestin was associated with persistent increases in breast cancer incidence and death. ^[28]

Factors that increase the number of menstrual cycles also increase the risk of breast cancer, probably due to increased endogenous estrogen exposure. Such factors include the following:

• Menarche when younger than 13 years (2 times the risk)
• Nulliparity
• First full pregnancy when older than 30 years
• Not breastfeeding
• Menopause when older than 50 years

Conversely, late menarche, anovulation, and early menopause (spontaneous or induced) are protective, owing to their effect on lowering endogenous estrogen levels or shortening the duration of estrogenic exposure.

Other factors affecting the risk of breast cancer include the following:

• Tobacco exposure
• Diabetes/insulin resistance
• Diethylstilbestrol exposure in utero ^[29]
• Alcohol consumption
• Ionizing radiation exposure
• Exposure to dichlorodiphenyldichloroethylene (DDE), a metabolite of the insecticide dichlorodiphenyltrichloroethane (DDT)
• Socioeconomic class
• Night shift work
• Poor cardiovascular health
• High bone density
• Hair product use
• Processed foods

Tobacco smoking

Tobacco abuse portends a 24% higher risk of developing invasive breast cancer. Former smokers carry a 13% increased risk. Starting smoking at an earlier age has a profound impact. Compared with never smoking, beginning tobacco use prior to menarche increases breast cancer risk by 61%, and beginning tobacco use 11 or more years prior to parity carries a 45% increased risk. ^[30]

Diabetes/insulin resistance

Postmenopausal women with type 2 diabetes have a 17% higher risk of developing breast cancer. In addition, diabetic women with a new breast cancer have a higher incidence of being diagnosed at a more advanced stage. Diabetics are diagnosed with stage I cancer at about 49% of the frequency as non-diabetics; they are 21% and 16% more likely to be diagnosed with stage III or stage IV disease, respectively. This elevated risk was noted specifically in postmenopausal women with estrogen receptor-positive disease. ^[31]

Diethylstilbestrol exposure in utero

Diethylstilbestrol, known as DES, was used clinically to prevent complications of pregnancy. In the late 1960s, an unusual occurrence of a rare cancer of the vagina among young women, called clear cell adenocarcinoma (CCA), was observed and subsequently linked to their intrauterine exposure to DES and in 1971, DES was removed from the market by the US Food and Drug Administration. However,  5 to10 million pregnant women and babies had been exposed to the drug. DES daughters have been found to have an increased risk for late breast cancer (~1.8-fol).

Alcohol consumption

A study by Chen et al found that low levels of alcohol consumption were associated with a small increase in breast cancer risk; cumulative alcohol intake throughout adult life was the most consistent measure. Alcohol intake that occurred early and late in adult life was independently associated with risk. ^[32] The mechanism, though unclear, likely is mediated via increasing estrogen levels.

A meta-analysis showed that for every 10 grams of alcohol consumed per day, there is a 7% increase in the risk of breast cancer. Compared with women who never drank, women who drank 35-44 g of alcohol (or, roughly 2-3 alcoholic beverages) per day had a 32% increased risk. ^[9]

Irradiation

Radiation, particularly to the chest or in the first decade of life, profoundly increases the risk of developing breast cancer. Studies involving patients who had received thoracic mantle radiation for Hodgkin lymphoma have repeatedly corroborated this risk.

Socioeconomic class

The incidence of breast cancer is increased in individuals in higher socioeconomic classes. However, breast cancer survival rates are lower in women from lower socioeconomic classes. This underscores the benefits of mammography in early detection, which likely contributes to the better outcomes seen in people with the means to diligently follow cancer screening recommendations.

Night shift work

The International Agency for Research on Cancer and the World Health Organization (IARC/WHO) have recognized night shift work as a probable carcinogen. A 2005 meta-analysis (including 13 studies of airline cabin attendants and nighttime shift workers ) explored the relationship between night work and breast cancer risk and reported a relative risk (RR) of 1.48; they noted similar risk levels in female air cabin crew and female night shift workers (RR of 1.51 vs 1.44, respectively). ^[33]

Poor cardiovascular health

A review of the American Heart Association cardiovascular health (CVH) score in 161,809 Women's Health Initiative participants followed from 1993 through 2010 found that, compared with women with the highest (best) CVH scores, those with the lowest (worst) CVH scores had a 52% greater risk of incident cancer, including breast cancer. Fewer minority women and less-educated women achieved an ideal CVH score. The CVH score is based on smoking, body mass index, physical activity, diet, total cholesterol, blood pressure, and fasting glucose. ^[34]

High bone density

High bone density is associated with higher endogenous estrogen levels, which are associated with a higher risk of breast cancer. Women in the highest hip bone mineral density (BMD) category are 62% more likely to develop breast cancer than are women in the lowest BMD category. Of note, women of east Asian ethnicity have lower bone density and have a lower incidence of breast cancer than white and black Americans. ^{[35, 36, 37]}

Long-term use of calcium channel blockers

Several studies had previously shown a correlation between use of calcium channel blockers and breast cancer incidence. ^{[38, 39]} More powerful data, however, has emerged recently, showing that calcium channel blockers do not increase the risk of breast cancer. ^[40] Given that the newer data is derived from a larger study, current evidence does not implicate calcium channel blockers as a significant risk factor.

Hair product use

A study of 4,285 women found a significant increase in breast cancer risk in black women who used dark shades of hair dye and in white women who used chemical hair relaxers. The risk of breast cancer was 51% higher in black women who reported using dark hair dye, compared with those who did not, and was 74% higher in white women who reported using chemical relaxers. Risk was more than doubled in whites who used both relaxers and hair dyes (odds ratio [OR] = 2.40, 95% confidence index [CI],1.35–4.27). ^[41]

In white women, the use of dark hair dyes was associated with increased estrogen receptor (ER)–positive disease (OR = 1.54; 95% CI, 1.01–2.33). Relaxer use was associated with increased ER-negative disease (OR = 2.56; 95% CI, 1.06–6.16). ^[41]

Similarly, data from 46,709 women in the Sister Study showed an increased risk for breast cancer in women who frequently use permanent hair dye and chemical hair straighteners. Women who regularly used permanent hair dye in the year prior to enrolling in the study were 9% more likely to develop breast cancer, compared with women who did not use hair dye. In African-American women, using permanent dyes every 5 to 8 weeks or more often was associated with a 60% increased risk of breast cancer as compared with an 8% increased risk for white women. Use of semi-permanent or temporary hair dye was associated with little to no increase in breast cancer risk. ^[42]

Women who used hair straighteners at least every 5 to 8 weeks were about 30% more likely to develop breast cancer. The association between straightener use and breast cancer was similar in African-American and white women, but straightener use was much more common in African-American women. ^[42]

Milk consumption

In prospective cohort studies, milk consumption has not been associated with a greater risk of breast cancer. ^[43]  In contrast, a study of 52,795 North American women found that higher intakes of dairy milk were associated with greater risk of breast cancer, when adjusted for soy intake. ^[44]  Consuming as little as 1/4 to 1/3 cup of dairy milk per day was associated with a 30% increase in risk of breast cancer; for those drinking 1 cup per day, risk rose to 50%, and for those drinking 2 to 3 cups per day, the risk increased further to 70% to 80%. However, because the study was observational design, it cannot prove that cow's milk causes breast cancer; in addition, the results were significant only for postmenopausal women, and only for hormone receptor-positive cancers. ^[45] In premenopausal women, dairy consumption has been tied to decreased risk for breast cancer. ^[46]

Processed foods

Epidemiologic studies have shown a correlation between breast cancer risk and consumption of advanced glycation end products (AGEs), which are reactive metabolites found at high levels in many processed foods, and in foods high in sugar, protein, and fat. ^{[47, 48]}  To establish a possible direct cause-and-effect relationship, Krisanits et al studied mammary gland development in mice fed a regular, low AGE, or high AGE diet during puberty. Mammary glands in the mice that were fed a diet high in AGEs showed disrupted ductal morphogenesis and atypical hyperplasia, with changes in both epithelial and stromal cells. The mammary glands showed many features of preneoplastic lesions, suggesting that a high AGE diet may influence tumorigenesis. ^[49]

Genetic counseling about BRCA mutation testing may be done by trained health professionals, including trained primary care providers. Several professional organizations describe the skills and training necessary to provide comprehensive genetic counseling. The process of genetic counseling includes the following:

• Detailed kindred analysis and risk assessment for potentially harmful BRCA mutations
• Education about the possible results of testing and their implications
• Identification of affected family members who may be preferred candidates for testing
• Outlining options for screening, risk-reducing medications, or surgery for eligible patients
• Follow-up counseling for interpretation of test results

Current genetic sequencing tests can accurately detect BRCA mutations. Initial testing of a family member who has breast or ovarian cancer is the preferred strategy in most cases, but it is reasonable to test the patient if no affected relative is available.

Several organizations have published guidelines concerning the use of genetic testing for BRCA mutations. ^{[50, 51, 52, 53, 54]} Their recommendations include the following:

• Testing should be done only when an individual has personal or family history that suggests an inherited cancer susceptibility.
• Testing of an individual without a cancer diagnosis should be considered onlly when an appropriate affected family member is unavailable for testing.
• Testing must be voluntary.
• Informed consent (by the patient or a legal proxy) is necessary and should encompass pretest education and counseling about the risks, benefits, and limitations of genetic testing, including the implications of both positive and negative results.
• The patient should have access to a health professional who is trained to provide genetic counseling and interpret test results
• The patient should receive pretest and posttest counseling.
• Testing should be done only if the results will aid in diagnosis or medical management of the patient or family member who has hereditary risk for cancer.
• Patients found to be mutation carriers should be encouraged to advise close family members to obtain genetic counseling.

The American Congress of Obstetricians and Gynecologists and the Society of Gynecologic Oncologists recommend genetic risk assessment for women who have more than a 20% to 25% risk for an inherited predisposition to breast and ovarian cancer, and suggest that it may be helpful for patients with more than a 5% to 10% risk. ^[50]

The United States Preventive Services Task Force (USPSTF) recommends that women who have family members with breast, ovarian, tubal, or peritoneal cancer undergo screening with a risk assessment tool designed to identify increased risk for carriage of BRCA mutations. The USPSTF recommends that women with positive screening results receive genetic counseling and, if indicated after counseling, BRCA testing. ^[15]

However, the USPSTF recognizes that each risk assessment tool has limitations. The USPSTF evaluated the Ontario Family History Assessment Tool, Manchester Scoring System, Referral Screening Tool, Pedigree Assessment Tool, and FHS-7, and found insufficient comparative evidence to recommend one tool over another. ^[15]

The USPSTF advises that general breast cancer risk assessment models (eg, the National Cancer Institute Breast Cancer Risk Assessment Tool, which is based on the Gail model) are not designed to identify women who should receive genetic counseling or BRCA testing. The USPSTF also found insufficient evidence to support a specific risk threshold for referral for testing. ^[15]

The National Comprehensive Cancer Network (NCCN) recommends genetic testing not only for BRCA, but also for other high-penetrance breast cancer susceptibility genes, including CDH1, PALB2, PTEN, STK11, and TP53. Testing is indicated for women with a personal history of breast cancer at age 50 years or younger. In individuals at any age with a personal history of breast cancer, other indications are male breast cancer or Ashkenazi Jewish ancestry. ^[54]

Treatment indications for testing at any age include the following ^[54] :

• To aid in systemic treatment decisions using PARP inhibitors for breast cancer in the metastatic setting
• To aid in adjuvant treatment decisions with olaparib for high-risk, HER2-negative breast cancer

NCCN also recommends testing at any age for patient with the following histology results ^[54] :

• Triple-negative breast cancer
• Multiple primary breast cancers (synchronous or metachronous)
• Lobular breast cancer with personal or family history of diffuse gastric cancer

Additionally, the NCCN recommends testing for anyone with a personal history of breast cancer and a family history of three or more diagnoses of breast and/or prostate cancer (any grade) on the same side of the family, including the patient with breast cancer or one or more close blood relatives with any of the following ^[54] :

• Breast cancer at age ≤50 years
• Male breast cancer
• Ovarian cancer
• Pancreatic cancer
• Metastatic or high- or very-high-risk prostate cancer

Risk assessment tools are designed to quantify a patient's risk for breast cancer, and thus can assist in the decision on whether or not to refer a patient for genetics counselling and/or whether or not to pursue a workup for inherited cancer syndromes. Several risk assessment tools exist. In general, these tools elicit information about factors that epidemiologic studies have identified as associated with increased risk, including the likelihood of BRCA mutations, such as the following:

• Breast cancer diagnosis before age 50 years
• Bilateral breast cancer
• Presence of breast and ovarian cancer
• Presence of breast cancer in 1 or more male family members
• Multiple cases of breast cancer in the family
• One or more family members with 2 primary types of BRCA-related cancer
• Ashkenazi Jewish ethnicity

Gail model

The Gail Model is a statistical breast cancer risk assessment algorithm that was developed in 1989 by Dr. Mitchell Gail and colleagues with the Biostatistics Branch of the National Cancer Institute’s Division of Cancer Epidemiology and Genetics. It was derived from a huge screening study of 280,000 women 35 to 74 years of age. The Gail model has proved to be a reasonable tool for estimating breast cancer risk in white women, and other researchers have subsequently supplemented the model to provide accurate risk assessments for African-American, Hispanic, and Asian women.

However, the Gail model underestimates the breast cancer risk for women with a significant family history. Consequently, it should not be used for women suspected to have a hereditary syndrome associated with increased risk of breast cancer. ^[15]

The Gail Model looked at a woman’s personal medical history, familial history, and reproductive history. These variables were then adjusted according to age and associated higher risk for older women. The Gail Model is a risk prediction tool that is designed to derive individual risk estimates for the development of breast cancer over time. It was developed to estimate the probability of developing breast cancer over a defined age interval; it was also intended to improve screening guidelines.

However, the Gail model did not take into account racial or ethnic differences or the risk of women with atypical hyperplasia on a breast biopsy (atypia), BRCA genetic variants, or tamoxifen use. In addition, it excluded women who had already had a confirmed diagnosis of either ductal or lobular breast carcinoma in situ.

In 2008, the accuracy of the Gail Model for women with a history of atypia was reported. ^[55] Women with atypia were identified from the Mayo Benign Breast Disease (BBD) cohort (1967 to 1991). Their risk factors for breast cancer were obtained, and the Gail Model was used to predict 5-year– and follow-up–specific risks for each woman. The predicted and observed numbers of breast cancers were compared, and the concordance between individual risk levels and outcomes was computed. Of the 9,376 women in the BBD cohort, 331 women had atypia (3.5%). At a mean follow-up of 13.7 years, 58 of 331 (17.5%) patients had developed invasive breast cancer, 1.66 times more than the 34.9 predicted by the Gail model (95% confidence index [CI], 1.29 to 2.15; P < 0.001).

For individual women, the concordance between predicted and observed outcomes was low, with a concordance statistic of 0.50 (95% CI, 0.44 to 0.55). The C-statistic (sometimes called the “concordance” statistic or C-index) is a measure of goodness of fit for binary outcomes in a logistic regression model. In clinical studies, the C-statistic gives the probability a randomly selected patient who experienced an event (eg, a disease or condition) had a higher risk score than a patient who had not experienced the event. A concordance statistic of 0.5 means that the predication ability of the Gail Model was no better than chance for women with abnormal biopsy results in the past. ^[56]

However, the model was subsequently revised (Gail Model 2) and validated to predict risk of invasive breast cancer, including information on the history of first-degree affected family members. The Gail Model 2 has been used extensively in clinical practice and has served as the basis for eligibility for a number of the breast cancer prevention trials.

The US Food and Drug Administration (FDA) guidelines use the National Surgical Adjuvant Breast and Bowel Project’s (NSABP) modified Gail model as the basis for eligibility for the prophylactic use of tamoxifen. Tamoxifen, a selective estrogen receptor (SERM), is approved for women aged 35 years and older who have a 5-year modified Gail risk of breast cancer of 1.67% or more. The Gail Model 2 also forms the basis of the National Cancer Institute’s Breast Cancer Risk Assessment Tool.

The Gail Model 2 is most accurate for non-Hispanic white women who receive annual mammograms, but the model tends to overestimate risk in younger women who do not receive annual mammograms. The model also demonstrates reduced accuracy in populations with demographics (ie, age, race, extent of screening) that differ from the population on which it was built. At the individual level, the model lacks adequate discrimination in predicting risk and has been challenged on its generalizability across populations. The updated Gail Model Calculator (which incorporates information for women of other races and ethnicities), also known as the Breast Cancer Risk Assessment Tool, is available online at the National Cancer Institute website: Breast Cancer Risk Assessment Tool.

To address concerns regarding applicability of the modified Gail model to black women, Gail and colleagues derived a model using data from a large case-control study of black women participating in the Women’s Contraceptive and Reproductive Experiences (CARE) study. The CARE model demonstrated high concordance between the number of breast cancers predicted and the number of breast cancers observed among black women when validated in the WHI cohort. The CARE model better estimates the risk in black women (whereas the Gail model underestimates breast cancer risk in them) and can be additive to other factors (such as family history, genomics, and environmental factors) when assessing risk and providing counsel for black women. ^[56]

Breast Cancer Surveillance Consortium (BCSC) Risk Calculator

The Breast Cancer Surveillance Consortium (BCSC) Risk Calculator was developed and validated in 1.1 million women undergoing mammography across the United States, of whom 18,000 were diagnosed with invasive breast cancer. The BCSC Risk Calculator has been externally validated in the Mayo Mammography Health Study. In 2015, the BCSC risk calculator was updated to include benign breast disease diagnoses and to estimate both 5-year and 10-year breast cancer risk.

BRCAPRO

BRCAPRO calculates a person’s probability of carrying a deleterious mutation of BRCA1, BRCA2, or both on the basis of the person's cancer status and the history of breast and ovarian cancer in first- and second-degree relatives. BRCAPRO is especially accurate in predicting testing results when the probability of the woman being a carrier of a deleterious mutation is less than 70%. ^[57]

BRCAPRO results can help patients decide whether to undergo genetic testing. Testing is generally most useful for women whose BRCAPRO results indicate that they have intermediate probability of being a BRCA carrier. If BRCAPRO results indicate that the patient's probability of being a carrier is very low, genetic testing will almost certainly yield negative results. ^[57] For patients with a very high pretest probability of BRCA carriage, genetic testing results can help guide screening of other family members.

Before using BRCAPRO, the clinician should determine whether the patient is prepared to make decisions on the basis of her level of risk. If she is not prepared to make lifestyle or healthcare changes to reduce moderate risk, or perhaps consider prophylactic surgery for extremely high risk, then the information provided by BRCAPRO has no value. ^[57]

Tyrer–Cuzick model

The Tyrer-Cuzick model, or IBIS tool, is used to calculate a person’s likelihood of carrying the BRCA1 or BRCA2 mutations. It estimates the likelihood of a woman developing breast cancer in 10 years and over the course of her lifetime. The tool is used to help inform a person’s decision-making about genetic counselling and testing. If the model predicts a 10% or greater chance that the woman has a mutation in BRCA1, BRCA2, or both, genetic counseling is advised. ^[58]

The tool estimates breast cancer risk on the basis of the following risk factors ^[58] :

• Age
• Body mass index
• Age at menarche
• Obstetric history
• Age at menopause (if applicable)
• History of a benign breast condition that increases breast cancer risk (hyperplasia, atypical hyperplasia, LCIS)
• History of ovarian cancer
• Use of hormone replacement therapy
• Family history (including breast and ovarian cancer, Ashkenazi inheritance, genetic testing if done)

A comparison study of breast cancer risk models concluded that the Tyrer–Cuzick model is the most consistently accurate, whereas the Gail, Claus, and Ford models all significantly underestimate risk. ^[59] However, a study by Boughey et al found that the Tyrer-Cuzick model significantly overestimated the risk of breast cancer in women with atypical hyperplasia. ^[60]

Mirai model

The Mirai model is an artificial intelligence (AI)–based approach to calculation of breast cancer risk. It estimates risk of breast cancer across multiple future time points, based on analysis of a woman’s mammogram and clinical risk factors (eg, age, family history, hormonal factors). If clinical risk factors are not available, Mirai can predict those factors directly from the mammogram.

Mirai works as follows:

1. An image encoder receives and processes all standard mammogram views.
2. An image aggregator module combines the image information across all views to create a representation of the mammogram.
3. If necessary, a risk-factor predictor module predicts the patient's risk factors from the mammogram.
4. An additive-hazard module predicts the patient's risk each year over the next 5 years, based on analysis of the mammogram and the patient's risk factors.

Yala et al performed a validation study of Mirai, using 128,793 mammograms from 62,185 patients at seven sites (in the United States, Israel, Sweden, Taiwan, and Brazil), of which 3,815 were followed by a cancer diagnosis within 5 years. The study confirmed Mirai’s accuracy across globally diverse test sets, with receiver operating characteristic area under the curve values ranging from 0.78-0.90 at 1 year to 0.75-0.82 at 5 years. ^[61]