Genetic Factors: Hereditary Cancer Predisposition Syndromes

Clinical Features

Although only about 18,000 cases of breast cancer each year are associated with an obvious hereditary predisposition, primary cancers developed in more than 200,000 breast cancer survivors in the United States as a result of a hereditary predisposition, and these survivors remain at risk for secondary cancers. Genetic testing has emerged as one of the most important indicators of risk factors, pointing to a need for intensified screening for breast cancer. When detected at an early stage, more than 90% of breast cancers are curable. These statistics underscore the rationale for the use of genetics in clinical oncology. We have previously reviewed in detail the management of women and men who are at hereditary risk for breast cancer, and this review is summarized and updated here.

From 1 in 150 to 1 in 800 individuals in the population carry a genetic susceptibility to breast cancer, and the prevalence is much higher in certain ethnic groups. Syndromes of breast cancer susceptibility are linked to mutations of BRCA1 and BRCA2, in addition to a smaller number of cases with germline mutations of ATM, PALB2, PTEN, p53, CHEK2 , STK11 , CDH1, and rarer syndromes ( Tables 13.4 and 13.5 ). Cowden syndrome (CS) was initially described as a dominant inheritance of multiple hamartomatous lesions, including papillomas of the lips and mucous membranes and acral keratoses of the skin. This disease was ultimately linked to germline mutations of PTEN and is discussed separately later. In persons with Li-Fraumeni syndrome, early-onset breast cancer occurs with soft tissue sarcomas, osteosarcoma, leukemia (particularly hypodiploid acute lymphoblastic leukemia), brain tumors, adrenal cortical tumors, and other cancers. Rarely, atypical breast-ovarian kindred may be found to have a germline p53 mutation. Although rare, women with Peutz-Jeghers syndrome, associated with germline mutations in the STK11 gene, are at increased risk for breast cancer as well as a variety of other malignancies. More recently, mutations in PALB2, partner and localizer of BRCA2, have been implicated in both familial breast and pancreas cancer risk.

Early linkage studies suggested that in about 50% of breast cancer kindreds, the cancer was linked to BRCA1 ; in 30%, it was linked to BRCA2 ; and in the remainder, it was linked to other either identified or unidentified genes. In up to two-thirds of families with both male and female breast cancer, the cancers were due to BRCA2, whereas more than 80% of families with both breast and ovarian cancer harbored mutations in BRCA1. BRCA1 -linked breast cancers are associated with a basal-like subtype and higher mitotic indices. They are more likely to be of higher grade, are more frequently estrogen and progesterone receptor negative, and demonstrate a basal-like phenotype. They display a more “aggressive” phenotype, including a higher proportion of cells in S phase, and other indices.

Multiple moderate-penetrance breast cancer susceptibility genes have also been identified (see Table 13.4 ). In Northern European families, specific mutations of CHEK2 are associated with familial breast cancer, with pathogenic CHEK2 variants being present in ~1% of Caucasians of European descent. Although the common European mutation in CHEK2 is rare in persons in North America, analysis of women for CHEK2 founder mutations stratified for family history of breast cancer demonstrated that carriers with a positive family history had a greater than 25% lifetime risk for breast cancer. Important to note, degree of breast cancer risk in women with mutations in CHEK2 appears to be dependent on both the nature of the specific mutation (truncating versus missense) and the strength of the family breast cancer history. Carriers of mutations in ATM, another moderate-penetrance breast cancer susceptibility gene, have an approximately twofold elevated breast cancer risk.

Multiple, common, lower-penetrance genes are likely to account for a significant component of currently unexplained familial breast cancer risk. A host of putative lower-penetrance gene mutations have been identified through whole-genome association studies, although the clinical relevance of these associations remains unclear. In a large cohort of breast cancer patients referred for genetic testing, approximately 10% had a related germline mutation.

Compared with the 10% breast cancer risk for women in the general population, estimates of the breast cancer risk that is conferred by a common high-penetrance susceptibility gene ranged from 67% to 69% by age 70 years based on epidemiologic analyses. This was confirmed in a prospective study that defined breast cancer risk by age 80 of 72% and 69% for BRCA1 and BRCA2 carriers, respectively. This study also reported a contralateral breast cancer risk, up to 20 years after diagnosis, of 40% and 26% for BRCA1 and BRCA2, respectively. In the setting of a mutation in PALB2, the breast cancer risk to age 80 is approximately 45%, whereas for those with mutations in ATM and CHEK2 it is less than 30%.

Epithelial ovarian carcinoma can be delineated into six distinct histologic subtypes—high-grade serous, carcinosarcoma, clear cell, endometrioid (International Federation of Gynecology and Obstetrics [FIGO] grades 1–3), mucinous, and low-grade serous—with differing manifestations, prognoses, and molecular and immunohistochemistry profiles. The high-grade serous subtype comprises the majority of cases of ovarian cancer. Mutations in BRCA1 and BRCA2 show a strong association with the high-grade serous carcinoma of ovary and fallopian tube or peritoneal origin, such that germline testing for BRCA1 and BRCA2 together can have up to a 25% detection rate in this histologic subtype. Germline testing for BRCA1 and BRCA2 is standard of care for all patients with a high-grade serous ovarian cancer diagnosis. Mutations in the mismatch repair (MMR) genes are also associated with ovarian cancer risk, usually endometrioid histology; this is further discussed in the later section on Lynch syndrome (LS ). Similar to breast cancer, a number of more moderate-penetrance ovarian cancer susceptibility genes have also been identified. BRIP1, RAD51C, and RAD51D are associated with approximately a 5-fold to 12-fold risk of ovarian cancer (see Table 13.5 ). Earlier data suggested an association with BRIP1 and breast cancer, but a large population-based study demonstrated that truncating mutations in this gene were not associated with a substantial increase in breast cancer risk.

In a study of close to 2000 women with ovarian cancer, 18% of patients had a germline mutation in a gene associated with ovarian cancer risk. The majority of patients in this study had high-grade serous histology (78%), but all histologic subtypes were represented. As expected, mutations in BRCA1 and BRCA2 accounted for the bulk of inherited risk, with mutations in BRIP1, RAD51C, RAD51D, and the MMR genes accounting for 3% cumulatively. Compared with a population risk for ovarian cancer of approximately 1.5%, mutations in BRCA1 and BRCA2 are associated with a risk of 44% and 17% by age 80, respectively. For BRIP1, RAD51C, and RAD51D, the cumulative lifetime risk by age 80 is 4% to 14%.

The role of ascertainment and other possible biases in deriving cancer risk estimates, as well as risks for cancer of the prostate, colon, pancreas, and other sites in BRCA mutation carriers has been reviewed. In addition to Fanconi anemia, persons with compound BRCA2 mutations may experience childhood medulloblastomas. Heterozygous mutations in BRCA1 and BRCA2 harbored by pediatric cancer patients have been detected through tumor normal sequencing; however, it is not clear at this point if these are cancers associated with the familial syndrome or simply true and unrelated, and this is an active area of research.

Genetics

BRCA1 and BRCA2 are both important for DNA repair, in particular homologous recombination, which enables repair of double-stranded DNA breaks. BRCA1 is a large gene, spanning more than 100,000 bases of genomic DNA with 22 coding and 2 noncoding exons. BRCA2 is also large, consisting of 27 exons across 70 kb of genomic DNA. Both genes, by coincidence, have a large exon 11. An update of reported mutations is accessible through the Internet at http://research.nhgri.nih.gov/bic/ .

Most BRCA1 mutations cause premature truncation of the peptide by frameshift or nonsense sequence changes. Large germline rearrangements also occur in BRCA1 and BRCA2. Five percent to 10% of BRCA mutations that are missense are problematic because they are of unknown clinical significance. The proportion of these variants that are of unknown significance was as high as 10% to 23% in some series, posing counseling challenges. The role of BRCA1 and BRCA2 in DNA damage response and homologous recombination is reviewed elsewhere.

Founder BRCA mutations have been documented in genetically isolated populations. In North American families, the most common founder mutations occur in persons of Ashkenazi Jewish origin. These mutations include a two-base-pair deletion in codon 23 of BRCA1, termed 185delAG (c.68_69delAG); another mutation in BRCA1, 5382insC (c.5266dupC); and the 6174delT (c.5946delT) mutation in BRCA2. About 1 in 40 Ashkenazi Jews harbor one of the common BRCA1 or BRCA2 mutations, a relatively high carrier frequency for an inherited cancer predisposition syndrome. Other mutations in BRCA1 and BRCA2 occur in the Ashkenazim; 16 of 737 Ashkenazi Jews who were tested in a clinic-based ascertainment had a nonfounder mutation (2%). In another study of Ashkenazi individuals with a personal history of breast or ovarian cancer who had previously been shown not to have a founder mutation, 3 of 70 (4.3%) had a deleterious nonfounder mutation. This very low nonfounder BRCA rate in Ashkenazim was also confirmed again in a 2017 study. Founder mutations in populations other than the Ashkenazim have also been observed.

Other Genes

Protein truncating variants in PALB2, CHEK2, and ATM have been associated with increased breast cancer risk as outlined earlier. The risk associated varies per study and variant. A Finnish founder mutation in PALB2 c.1592delT is associated with higher than population risk of breast cancer, but risk estimates are lower than those reported with other PALB2 mutations. Most of the data for CHEK2 and increased breast cancer risk relates to the truncating c.1100delC variant found in Northern Europeans. The missense variant CHEK2 I157T is associated with a very modest increased risk (RR, 1.58; 95% confidence interval [CI], 1.42–1.75), with some laboratories not reporting this variant as being pathogenic. The autosomal recessive neurologic disorder ataxia telangiectasia is associated with compound heterozygote mutations in ATM, a gene involved with DNA double-strand break repair and cell cycle control . Although this is a rare event, monoallelic ATM mutations are relatively common, occurring in approximately 3% of Caucasians. Women with heterozygous ATM mutations have a moderate, approximately double population risk of breast cancer as noted earlier. BRIP1, BRCA1 interacting protein C-terminal helicase 1 gene, is part of the DNA repair machinery. A founder mutation, BRIP1, c.2040_2041insTT was associated with increased risk of ovarian cancer in the Icelandic population. Subsequently, case-control series reported increased ovarian cancer risk for other truncating BRIP1 mutations. RAD51C and RAD51D are part of the Fanconi- BRCA homologous recombination repair pathway and have been associated with a modest increased risk of ovarian cancer.

Clinical Management

Three elements of breast surveillance that are recommended to women with increased risk of breast cancer are self-examination, clinician examination, and imaging. The evidence base underlying these recommendations has been reviewed, and updated guidelines are available.

Increasingly, breast cancer screening including mammography and magnetic resonance imaging (MRI) together has been shown to have greatest sensitivity. For women who are at the highest hereditary risk for breast cancer, whose breasts are difficult to examine, or who have had biopsy results showing atypia, it might be appropriate to discuss the option of removing the healthy breasts as a preventive measure (i.e., prophylactic mastectomy). Retrospective studies and reviews document the well-established risk-reducing role of surgery in high-risk patients. Prospective and combined consortium studies have also shown the efficacy of this approach.

Because of their antiestrogen properties, tamoxifen, raloxifene, and aromatase inhibitors have been shown to decrease breast cancer rates in persons who are at increased risk. Two studies on the impact of tamoxifen on subsequent breast cancer risk in BRCA1 and BRCA2 mutation carriers have shown conflicting conclusions, although only one of these studies was sufficiently powered to reach significant results. Tamoxifen was confirmed to decrease contralateral breast cancer risk in a follow-up of one of these studies of BRCA mutation carriers. Providers are encouraged to discuss breast cancer chemoprevention with women who are at increased risk of breast cancer. There are no studies demonstrating a mortality benefit for this approach, and a careful discussion of potential benefit versus toxicity is required.

Small trials have demonstrated the ability of ultrasound with Doppler and CA125 to reveal early-stage ovarian cancers in BRCA mutation carriers. However, the efficacy of this approach has not been proven in large, prospective studies, and the US Food and Drug Administration (FDA) has recommended against screening for ovarian cancer in all women, including those at high risk. Therefore prophylactic removal of the ovaries and fallopian tubes is recommended to women with strong family histories of ovarian cancer, in families linked to BRCA1 or BRCA2, or to women who are considering hysterectomy in the setting of a germline mutation associated with LS. Studies examining specimens from prophylactic salpingo-oophorectomy in BRCA1 and BRCA2 mutation carriers have implicated the fimbriated part of the fallopian tube as harboring precursor lesions to high-grade serous epithelial ovarian cancer. Thus prophylactic salpingo-oophorectomy is recommended because such surgeries not only decrease the incidence of subsequent breast and ovarian cancer but also reveal occult early-stage pelvic neoplasms, decreasing mortality. These studies have also confirmed that high-grade serous carcinoma can still occur after prophylactic oophorectomy, albeit at much reduced rates. For women with high-penetrance risk genes, risk-reducing prophylactic oophorectomy is recommended when childbearing is complete or at approximately age 40 . Given the lack of an adequate screening modality, difficulty in diagnosis, and resultant late stage at presentation and the associated mortality with invasive ovarian cancer, risk-reducing surgery is also recommended for carriers of the more modest-risk mutations in genes BRIP1, RAD51C, and RAD51D. However, given the later age of onset of ovarian cancer compared with BRCA1 and BRCA2, risk-reducing surgery may be delayed until patients in this group are perimenopausal or postmenopausal.

Combination oral contraceptives that contain estrogen and high-dose progestin result in a time-dependent, protective effect against ovarian cancer in some but not all studies of BRCA mutation carriers. There remains the concern of a small increased risk of breast cancer due to oral contraceptives in this group, particularly in persons with BRCA1 mutations.

Germline BRCA1 and BRCA2 mutation–associated tumors have defective DNA repair mechanisms because of the loss of function of the remaining normal copy of the gene. This inability of the tumor to repair its DNA effectively can be exploited by therapies that disrupt alternative DNA repair pathways, such that the tumor cell accumulates so much DNA damage that it can no longer survive. Inhibitors of the base excision repair enzyme poly (ADP-ribose) polymerase (PARP inhibitors) use this principle of synthetic lethality and are FDA approved for women with recurrent ovarian cancer in a number of clinical settings. Although benefit has been demonstrated in all patients with ovarian cancer, it is substantially greater in patients with a germline or somatic BRCA1/2 mutation. A recent randomized trial has also demonstrated that patients with metastatic breast cancer and germline BRCA1/2 mutations had significant progression-free survival benefit with olaparib.

Clinical Features

CS is an autosomal dominant disorder characterized by multiple hamartomas with a high risk of benign and malignant tumors of the thyroid, breast, and endometrium. Consensus criteria for CS establish three diagnostic categories.

Pathognomonic criteria include adult Lhermitte-Duclos disease (LDD) (defined as presence of a cerebellar dysplastic gangliocytoma) and mucocutaneous lesions, facial trichilemmomas, acral keratoses, papillomatous lesions, and mucosal lesions. Major criteria include breast cancer, thyroid cancer (especially follicular histology), macrocephaly, and endometrial carcinoma. Minor criteria include other thyroid lesions (e.g., goiter), mental retardation, hamartomatous intestinal polyps, fibrocystic breast disease, lipomas, fibromas, and genitourinary tumors (e.g., uterine fibroids and renal cell carcinoma) or genitourinary malformation. The operational diagnosis of CS is made if an individual meets any one of the following criteria: pathognomonic mucocutaneous lesions alone if there are six or more facial papules, of which three or more must be trichilemmoma, or cutaneous facial papules and oral mucosal papillomatosis; or oral mucosal papillomatosis and acral keratoses; or six or more palmoplantar keratoses. Alternatively, the individual may fulfill two major criteria, but one must include either macrocephaly or LDD. Alternatively, the individual may have one major and three minor criteria or four minor criteria.

The palmar and plantar hyperkeratotic pits usually become evident later in childhood. Subcutaneous lipomas and cutaneous hemangiomas are seen in persons with CS with low frequency. An increased risk of early-onset male breast cancer has been noted in mutation carriers.

Cumulative lifetime risks for female breast cancer are 81% to 85.2% ; for LDD, 32% (CI, 19%–49%) ; for thyroid cancer, 21%–35.2%, ; for endometrial cancer, 19%–28.2% ; and for renal cancer, 15%–33.6% (CI, 6%–32%). The lifetime risks for colorectal cancer (CRC) are 9% to 16%, and the risk for melanoma is 6%.

Genetics

The CS-linked PTEN was mapped to 10q22-23. Bannayan-Riley-Ruvalcaba syndrome, characterized by macrocephaly, intestinal polyps, lipomas, and pigmented penile macules, is also caused by germline mutations in PTEN .

PTEN acts as a tumor suppressor by mediating cell cycle arrest, apoptosis, or both. Full sequencing and deletion/duplication analysis are available clinically, and promoter analysis is available on a research basis. Heterozygous germline mutations in PTEN cause most cases of CS. Nonsense, missense, and frameshift mutations that are predicted to disrupt normal PTEN function have been identified in some families, including mutations that disrupt the protein tyrosine/dual-specificity phosphatase domain. Initially it was shown that PTEN mutations can be detected in about 80% of patients with CS. More recently, the mutation detection rate in persons meeting CS criteria was found to be 34%, raising the possibility that current aspects of testing criteria are too broad.

Risk Management Recommendations

Persons with known germline PTEN mutations should undergo appropriate cancer screening. In view of recent work outlining cancer risks in persons with CS, some investigators have suggested that National Comprehensive Cancer Network (NCCN) management guidelines be modified to include renal cancer screening using biannual renal imaging from age 40 years or 5 years earlier than the earliest kidney cancer diagnosis in the family, and endometrial sampling in adulthood or 5 years earlier than the earliest endometrial cancer diagnosis in the family. Female patients with CS should have breast self-examination training and education starting at age 18 years. Beginning at age 25 years, clinical breast examinations should be performed every 6 to 12 months, and annual mammography and MRI screening should start at age 30 years or 5 years before the earliest age of breast cancer diagnosis in the family. Men should perform monthly breast self-examination. Female patients should receive endometrial cancer screening beginning around age 30 years or 5 years before the earliest age of endometrial cancer diagnosis in the family. Persons with CS should undergo biannual colonoscopy from age 40 years. Both men and women should have a comprehensive annual physical examination starting at diagnosis with screening for skin and thyroid lesions, including a baseline and annual thyroid ultrasound and dermatologic examination. Finally, preimplantation genetic testing for CS can be performed if a mutation is described in a parent.

Lynch Syndrome

A constellation of colon and endometrial cancers became known as Lynch syndrome, formerly referred to as hereditary nonpolyposis colon cancer (HNPCC). LS is an autosomal dominant syndrome caused by mutations in one of four MMR genes (MLH1, MSh2, MSH6, PMS2), or a gene located nearby (EPCAM), with resultant errors in DNA replication yielding a microsatellite instability (MSI) phenotype. LS accounts for 3% of all CRCs.

Polyposis Syndromes