The incidence of ovarian carcinoma is about 15 per 100,000 females in the UK; the lifetime risk to age 70 years is about 1.5 % in the UK and North America, slightly less in Southern Europe, making it the fifth most common malignancy in females in Western Europe and North America. Ovarian cancer has a very high mortality and is the most frequent cause of death from gynecological malignancies in the Western world. Worldwide, one of the highest incidence for ovarian cancer is seen in Israeli Jewish women who were born in North America or Europe – these are highly likely to be Ashkenazi Jews (13.5 per 100,000, cumulative incidence, 1.55 % to 74 years of age). Interestingly, in Israeli non-Jews, these figures are 3.0 per 100,000 and 0.32, respectively. These observations fit with the finding that up to 30 % of unselected Ashkenazi Jewish women with ovarian cancer carry mutations (Moslehi et al. 2000). This is a much higher figure than that observed in other groups. Early data suggested that for BRCA1 in unselected non-Ashkenazi women with ovarian cancer, between 4 and 9 % can be expected to carry a BRCA1 mutation, and between 0 and 4 % will carry BRCA2 mutations, although this figure will be higher in certain founder populations, such as Iceland. The most recent data, using more comprehensive mutation analysis, has suggested that in population unselected for ethnicity, about 15 % of women with ovarian cancer carry a mutation in BRCA1 (10 %) or BRCA2 (5 %). Notably, about 50 % of BRCA2 carriers are diagnosed at over 60 years of age, whereas 90 % of BRCA1 carriers are diagnosed below 60 years’ age.
Irrespective of family history, between 20 and 25 % of women diagnosed with ovarian cancer between the ages of 41 and 50 will carry a BRCA1/2 mutation (Pal et al. 2005). These findings have important implications for referral guidelines to cancer genetics services. Among the Ashkenazim, these percentages are much higher, although some of the studies include selected cases. Overall, 25–30 % of all ovarian cancer in Ashkenazi Jewish women is attributable to one of the three founder mutations, with BRCA1 mutation predominating at younger ages and BRCA2 at older ages. Ovarian cancer is diagnosed at a significantly younger age in BRCA1 carriers than in the general population. BRCA2-related ovarian cancers tend to be diagnosed at an older age than both the aforementioned groups.
Women with ovarian cancer (carcinoma) are more likely than expected to have a family history of ovarian neoplasia, and first-degree female relatives of patients with ovarian cancer are at increased risk. A collection of studies from around the world are summarized in Table 6.4; as can be seen, the odds ratios for ovarian cancer in association with a positive family history of ovarian cancer vary substantially, but are usually greater if the index case was diagnosed at a young age. The risk of death from ovarian cancer is very much greater if two first-degree relatives are affected. In this study, relatives of ovarian cancer cases also had significantly increased mortality from cancers of the stomach and rectum, but interestingly, the observed increased mortality from colon cancer, breast cancer, and pancreatic cancer failed to reach statistical significance.
Families in which there appears to be an autosomal, dominant susceptibility to ovarian cancer, with a high penetrance of the cancer in predisposed females have been recognized for years, and the identification of BRCA1 and BRCA2 has permitted more precise estimates of age-dependent cumulative incidence (i.e., penetrance). An analysis of 22 studies suggested that the risk of ovarian cancer among BRCA1/2 carriers who were not selected on the basis of family history of breast or ovarian cancer is very low before age 40 (below 2 % for both genes), but rises substantially with age in the case of BRCA1, and is above 1 % per year from age 45 onwards, reaching 2.5 % per year by age 65. For BRCA2, the risks are lower, start to rise at later ages, and never reach 1 % per year (Antoniou et al. 2003).
The ovarian Tumours arising in BRCA1/2 carriers cases are carcinomas, with a predominance of serous papillary cystadenocarcinoma, and mucinous and borderline ovarian carcinoma are underrepresented.
The frequency of BRCA mutations in the general population is estimated to be about 1 in 800 for BRCA1 and somewhat less for BRCA2, but it can vary significantly among some ethnic groups or geographic regions. Thus, the prevalence of the three BRCA1/2 founder mutations among the Ashkenazim is approximately 1 in 50. The Icelandic population carries the founder BRCA2 999del5 mutation at a frequency of 0.4 %. The frequency of BRCA1 and BRCA2 mutations in unselected series of women with ovarian carcinoma has been extensively studied, particularly in the so-called founder populations. A founder effect can occur when a relatively small group is genetically isolated from the rest of the population, because of geographic conditions or religious belief. If an individual in that isolated population carries a rare genetic alteration, the frequency of this allele in the next generations could increase in the absence of selection. Specific BRCA1 and BRCA2 mutations have been identified in diverse populations, such as in Ashkenazi Jewish, Icelandic, Swedish, Norwegian, Austrian, Dutch, British, Belgian, Russian, Hungarian, French Canadian, and Polish families (summarized in Fackenthal and Olopade 2007). The knowledge of well-characterized founder mutations in individuals of particular ethnic origins can simplify genetic counseling and testing, as the mutation screening can be limited to specific panels of mutations.
The founder mutations (often referred to as BRCA1 185delAG and 5382insC in BRCA1, BRCA2 6174delT) have been identified in the Ashkenazi Jewish families of Eastern European ancestry. These mutations are carried by about 2.5 % of the Ashkenazi Jewish population. These founder mutations are particularly common in Ashkenazi Jewish women with ovarian cancer, even without a family history of breast/ovarian cancer (Table 1).
Table 1 Risk of ovarian cancer in women with a family history of OC: studies prior to BRCA1/2
|Relatives studied||Country||Age group (years)||Cases||Controls||Odds ratio (95 % CI)||References|
|Any||USA||All||150||300||“No positive association”||Wynder et al. (1969)|
|First + second degree||USA||<50||150||150||15.7 (0.9–278)||Casagrande et al. (1979)|
|First degree||USA||45–74||62||1,068||18.2 (4.8–69)||Hildreth et al. (1981)|
|First degree||USA||18–80||215||215||11.3 (0.6–211)||Cramer et al. (1983)|
|First degree||Greece||All||146||243||∞ (3.4–∞)||Tzonou et al. (1984)|
|First degree||Japan||N/A||110||220||∞ (0.1–∞)||Mori et al. (1988)|
|First degree||USA||20–54||493||2,465||3.6 (1.8–7.1)||Schildkraut and Thompson (1988)|
|Second degree||2.9 (1.6–5.3)|
|First degree||USA||20–79||296||343||3.3 (1.1–9.4)||Hartge et al. (1989)|
|First degree + aunts||Canada||All||197||210||2.5 (0.7–11.1)||Koch et al. (1989)|
|First degree||Italy||25–74||755||2,023||1.9 (1.1–3.6)||Parazzini et al. (1992)|
|First degree||USA||N/A||883||Population||2.1 (1.0–3.4)||Goldgar et al. (1994)|
|First degree||USA||<65||441||2,065||8.2 (3.0–23)||Rosenberg et al. (1994)|
|First degree||USA||All||662||2,647||4.3 (2.4–7.9)||Kerber and Slattery et al. (1995)|
These results show that among women with ovarian cancer, BRCA1 and BRCA2 mutations are at least 3 times more likely to be found in Ashkenazi Jewish women than in non-Ashkenazi women. Interestingly, among otherwise unselected, very young-onset cases (under 30 years of age at diagnosis) BRCA1/2 mutations have not been observed, even though from epidemiological studies, mutations in genes are likely to be playing an important role in susceptibility.
There is evidence for genotype-phenotype relationships for ovarian cancer in BRCA1/2 mutation carriers. Mutations in the 3‘ portion of the BRCA1 gene (exons 13–24) were initially associated with a higher frequency of breast cancer relative to ovarian cancer in a series of 32 European families. This observation has not been confirmed by most larger studies (Stoppa- Lyonnet et al. 1997; Ford et al. 1998), although one study provided nonsignificant evidence in favor of the original finding (Moslehi et al. 2000). In a series of 25 English breast/ovarian cancer families, ovarian cancer was more prevalent than breast cancer when BRCA2 truncating mutations were located in a region of approximately 3.3 kb in exon 11 (the ovarian cancer cluster region (OCCR), nucleotides 3035–6629). Additional data from 45 BRCA2 families ascertained outside the UK provided support for this clustering. The analysis of 164 families with BRCA2 mutations, 67 of whom had mutations in the OCCR, has been reported (Thompson and Easton 2001).
The odds ratio for ovarian versus breast cancer in families with mutations in the OCCR, relative to non-OCCR mutations was 3.9 (P < 0.0001), confirming the importance of the OCCR in terms of ovarian cancer risk, but the effect was lost when Ashkenazi Jewish cases were omitted (6174delT is an exon 11 mutation). Case reports of multiple-case ovarian cancer families with mutations outside the OCCR temper enthusiasm for the clinical utility of
these observations (Al Saffar and Foulkes 2002).
The existence of a premalignant lesion for ovarian carcinoma is uncertain (Scully 2000). Careful histopathological analysis of prophylactic oophorectomy specimens among high-risk women, either because they have been identified as BRCA1/2 mutation carriers or based on their family history, gave conflicting results regarding the presence of histological alterations that could evolve towards invasive carcinoma. A tiny carcinoma in situ was identified in an otherwise normal prophylactic oophorectomy specimen from a woman with a BRCA1 mutation (Werness et al. 2000), but carcinoma in situ is very rarely seen in ovarian tissue, and this finding has not been replicated.
Perhaps the Tumours arose from ectopic fallopian tube epithelium that had been implanted in the ovary. More recently, attention has become focused on the hypothesis that the origin for clinically diagnosed BRCA1/2-related ovarian carcinoma is the distal fallopian tube.
The incidence of primary serous carcinoma of the peritoneum among BRCA1/2 carriers after oophorectomy is thought to be 0.2 % per year (Finch et al. 2006). The protective effect of oophorectomy against high-grade serous cancer in this study was approximately 80 %, which was in agreement with a population-based study from Israel (Rutter et al. 2003). These two studies suggested substantially less protection than was seen in selected series (Rebbeck et al. 2002), implying that among women who have undergone prophylactic surgery, this cancer may become one of the major threats to health among BRCA1/2 mutation carriers in their later years. Peritoneal cancer is indistinguishable histologically or macroscopically from ovarian cancer occurring among BRCA1/2 mutation carriers and represents a major challenge in terms of prevention of cancer in mutation carriers. The potential increased risk of malignant transformation of the entire peritoneal surface is thought to reflect the common origin of the ovarian epithelium and peritoneum from embryonic mesoderm. However, the peritoneum on the surface of the ovary may be particularly vulnerable to malignant transformation as a result of repeated injury following ovulation and/or high levels of local estrogen exposure – or perhaps this leads to implantation of fimbrial epithelium that when internalized within the ovary, can undergo malignant change. Some peritoneal carcinomas may arise multifocally, particularly in the context of BRCA1 mutations, and there may be a unique molecular pathogenesis of BRCA1-related papillary serous carcinoma of the peritoneum (Schorge et al. 2000). It seems implausible that these cancers arise from the fallopian tube.
The natural history of ovarian cancer in BRCA mutation carriers is thought to be different from sporadic ovarian cancer. Forty-three serous ovarian adenocarcinomas (81 % of the total) had an actuarial median survival of 77 months, compared with 29 months for the age, stage, and histological type-matched control group who were believed not to have mutations in BRCA1 on the basis of family history (P < 0.001). This good prognosis was attributed partly to the relative youth of the patients (mean age 48 years) but was also thought to be directly related to the presence of a BRCA1 mutation.
This study was criticized on methodological grounds, but a second study using a historical cohort approach gave similar results (Boyd et al. 2000). Furthermore, the better survival in hereditary cases was particularly noted for those women receiving platinum-containing chemotherapy. Other more recent studies have supported the notion that the prognosis is better for BRCA1/2-related ovarian cancer, but the durability of this survival advantage (most evident at 5 years following diagnosis) remains in question.
Mutations in other genes can cause ovarian cancer (Table 6.3). Lynch syndrome is one of the most common autosomal conditions predisposing to cancer, accounting for 3–4 % of all colorectal cancers (Moreira et al. 2012). Mutations in MLH1 and MSH2 are rare in ovarian cancers not selected on the basis of family history of cancer. The risk of developing ovarian cancer in MMR mutation carriers is about 8 %, with the highest risks in MLH1 and MSH2 mutation carriers and the lowest risk in MSH6 mutation carriers (Walsh et al. 2011). Most symptomatic ovarian cancers (77–81 %) in Lynch syndrome are diagnosed at an early stage (FIGO stage I and II) and may have a relatively good prognosis (reviewed in Vasen et al. 2013). There may be an increased risk for ovarian cancer in Carney Complex (Stratakis et al. 2000), but the numbers are too small for accurate risk assessment.
A growing list of cancer predisposition alleles in the genes associated with the Fanconi anemia genetic pathway (Fanconi anemia-like) genes – BRIP1 (Rafnar et al. 2011), RAD51C (Meindl et al. 2010), and RAD51D (Loveday et al. 2011) – have been shown to predispose to ovarian cancer. In addition, rare mutations in a number of genes involved in homologous repair, including BARD1, CHEK1, MRE11A, NBN, and RAD50 have also been found in patients with ovarian cancer (Walsh et al. 2011).
Identifying patients with these mutations may have clinical importance because Tumours with defects in the Fanconi anemia and homologous repair pathways are sensitive to platinum agents and PARP inhibition. The therapeutic effectiveness of PARP inhibitors for treating ovarian cancer in women with germline BRCA mutations has been acclaimed, but more recent trials have failed to confirm an overall survival benefit with such treatment (Ledermann et al. 2012), possibly because of the development of resistance (Barber et al. 2013).
Prevention of hereditary ovarian cancer is a major topic and is beyond the scope of this book; suffice to say that medically, the oral contraceptive pill is likely to offer significant (40 % of more) risk reduction (Narod et al. 1998;
Whittemore et al. 2004a). Surgically, ovarian removal (Rutter et al. 2003) is an option. Prophylactic salpingo-oophorectomy is often offered to women with germline BRCA1 or BRCA2 mutations, as the evidence is that it very much reduces the risk of ovarian cancer and also reduces the risk of breast cancer to 50 % in premenopausal women (Rutter et al. 2003; Domchek et al. 2010). It may also be offered to women with Lynch syndrome (Schmeler et al. 2006). While surgery clearly is effective in preventing serous papillary cancers arising from the ovary, the peritoneum remains at risk (as above).
The early detection of carcinoma at a stage in which it might be surgically curable, or more amenable to chemotherapeutic agents, would have a significant impact on prognosis. The traditional screening methods for
ovarian carcinoma are the measurement of the serum Tumours marker, CA-125, transvaginal ultrasonography, and clinical examination. None of these methods alone achieve the required level of sensitivity; for example, CA-125 detects only about two-thirds of Stage 1 ovarian Tumours. In the general population, clinical examination, transabdominal, transvaginal (van Nagell et al. 2000) ultrasound, and serum CA-125 screening tests either alone or in combination have all been assessed as potential screening tests in various settings. However, no test or combination of tests has yet proved to be effective for population screening. In a longer-term follow-up of their previous study, Jacobs and colleagues showed that although the median survival in the screened group was significantly better than for the control group, the number of deaths from an index cancer did not differ significantly between the two groups, with 18 deaths in 10 977 controls versus 9 in 10 958 screened women, RR 2.0 (95 % CI 0.78–5.13) (Jacobs et al. 1999). Long-term follow-up of this study is awaited.
For women at high risk of ovarian carcinoma, it has been recommended that they have an annual pelvic examination, vaginal ultrasonography, and serum CA-125 measurement every 6- or 12-month intervals from their mid- the 20s or 5 years less than the earliest age of onset of ovarian cancer in the family. However, data showing the clear benefit of such screening remains unconvincing. In the absence of good supportive evidence, it is appropriate to restrict these tests to a research setting. A worrying feature of hereditary ovarian cancer is the possibility that high-grade cancers can arise in very small foci that would not be detectable by ultrasound. A case report of a case of ultimately fatal advanced ovarian cancer diagnosed in a woman referred for a prophylactic oophorectomy was an early suggestion that delaying surgery could be dangerous (Rose and Hunter 1994). Studies among proven BRCA1/2 carriers have supported this finding. Three of 33 women with BRCA1/2 mutations were found to have early ovarian cancers lesions on examination of prophylactic oophorectomy specimens. Notably, two of the three cases were bilateral at diagnosis, and both were only noted on histopathological review and not in the operating room (Lu et al. 2000).
Similarly, high-risk lesions or frank cancers can be found in the fallopian tubes of BRCA1/2 carriers at preventive oophorectomy (Leeper et al. 2002; Carcangiu et al. 2004; McEwen et al. 2004). In the latter study, 5 of 60 consecutive BRCA1/2-positive women undergoing preventive oophorectomy were found to have occult ovarian or fallopian cancer: one death had occurred at 4 years follow-up. In a similar study from the USA, 5 of 30 BRCA1/2 carriers had cancer at surgery: one was a primary peritoneal cancer, three had tubal cancer, and one had ovarian adenofibroma with adjacent areas of low malignant potential carcinoma (Leeper et al. 2002). Taken together, these studies suggest that screening for ovarian cancer by ultrasound and CA-125 will prove to be insufficiently sensitive for routine use in BRCA1/2 carriers, but definitive studies are awaited, and until such times as definitive benefit of screening has been demonstrated, it is appropriate to advocate consideration of prophylactic oophorectomy after childbearing age in high-risk women.