The prevention of breast CA, the most common malignancy affecting U.S. women, is paramount to diminish the number of women who succumb to this disease. As a result of the work done by Beatson [2] and Lacassagne [3], in conjunction with the identification of the ER, SERMs emerged as the first class of therapeutic agents considered in breast CA prevention trials. Tamoxifen, a SERM approved for use in patients with advanced, as well as early-stage, breast CA, was compared to placebo in the National Surgical Adjuvant Breast and Bowel Project P-1:BCPT (of the NSABP) which accrued 13, 388 women at increased risk of breast CA between 1992 and 1997. Increased risk was defined as a 5-year predicted risk of breast CA > 1.66% according to modified Gail model criteria [4]. Tamoxifen reduced the risk of invasive breast CA by 49% and noninvasive breast CA by 50%. Only ER positive breast CAs were affected, showing a rate reduction of 69% with tamoxifen versus placebo [5]. The greatest benefit was observed in women at high risk secondary to a history of lobular carcinoma in situ (LCIS) or atypical hyperplasia, with a 56% and an 87% reduction in risk, respectively. The major toxicities associated with tamoxifen were EC (relative risk RR = 2.53, tamoxifen versus placebo); and thromboembolic disease, including pulmonary emboli (RR = 3.01), deep vein thrombosis (RR = 1.60) and stroke (RR = 1.59, non-significant increase). Secondary outcomes included a reduction in hip, radius and spine fractures. Three similar but smaller randomized trials of tamoxifen versus placebo were completed, only one of which demonstrated a statistically significant reduction in breast CA. The two other trials were underpowered, and as well, had important methodologic differences (Fig. 1) [6], [7], [8] and [9]. The U.S. Food and Drug Administration approved tamoxifen for risk reduction of breast CA in high risk women [10].

Raloxifene, another SERM, was compared to tamoxifen in an equivalency study carried out by the NSABP P-2:BCPT (STAR Trial). This study included 19,747 postmenopausal women over age 35 with a 5-year breast CA risk of 1.66% utilizing the Gail model or lobular carcinoma in situ [11]. After a median follow-up of 3.9 years, no difference was observed in the incidence of invasive breast CA, with both SERMs reducing the risk by 50%. Raloxifene did not provide protection against noninvasive breast CA. The EC rate was 38% lower in the raloxifene group; the number of strokes was similar; but a statistically significant lower number of thromboembolic events occurred in the raloxifene group.

Are the women who would benefit from tamoxifen as a chemoprevention agent against breast CA actually being prescribed the medication? Oddly, there has been limited published data regarding its actual use. The acceptance of tamoxifen chemoprevention by physicians and women at risk remains largely unknown. Extrapolating from the year 2000 National Health Interview Survey data, Freedman et al. [12] estimated that approximately 10.2 million women were eligible for tamoxifen using the National Surgical Adjuvant Breast and Bowel Project (NSABP) P-1 trial criteria. By applying a risk/benefit ratio analysis to the estimates derived from the National Health Interview Survey, approximately 2.4 million women were estimated to derive a net benefit from tamoxifen [13]. In a recent study by Tchou et al., only 42% of 137 women offered tamoxifen for breast CA chemoprevention actually accepted it [14]. The recognized barriers against the utilization of this SERM include the identification of appropriate candidates, the potential life-threatening toxicities associated with tamoxifen usage (DVT, PE and endometrial carcinoma) and finally, the question of survival benefit. Even with available prospective placebo-controlled studies, the debate continues about who should receive tamoxifen for the prevention of breast CA, although reasonable candidates include women with atypical hyperplasia or lobular carcinoma in situ, as well as those with a strong family history. Women with a BRCA1 or BRCA2 mutation represent a particularly important subgroup of patients. Prophylactic oophorectomy in this patient population is known to reduce breast cancer risk by at least 50%, which suggests that SERM therapy could be a particularly important intervention [15]. However, in that 80% of breast CAs in BRCA1 positive women are ER negative, as discussed by Honrado et al., the question is raised whether SERM therapy would be of actual benefit [16]. King et al., in a subgroup analysis of the NSABP Breast Cancer Prevention Trial, found that tamoxifin reduced breast CA incidence among healthy BRCA2 carriers, but the same beneficial effect was not observed among healthy BRCA1 carriers [17]. Further randomized, placebo-controlled trials are needed in order to establish a clear benefit for SERM therapy in this patient population.

Aromatase inhibitors (AIs), an alternative class of anti-estrogenic agents being evaluated for breast CA prevention, block aromatase, the rate-limiting enzyme in estrogen synthesis. Thus the conversion of androstenedione to estrone and testosterone to estradiol cannot occur, and in postmenopausal women, this effectively depletes the body of circulating estrogen. Thus the anti-estrogenic effect of AIs is unaffected by the type and distribution of ER subtype. Investigators have established a clear role for AIs in the treatment of women with metastatic or early stage, ER-positive breast CA. Several Phase III trials are under way assessing AIs in the prevention of this same disease. Drawing on the benefit to the contralateral breast CA risk demonstrated in the Anastrozole, Tamoxifen Alone or in Combination (ATAC) Trial in women with early-stage breast CA [18], the IBIS-II breast CA prevention trial includes one branch with 4000 women with ductal carcinoma in situ randomly assigned to anastrozole versus tamoxifen for 5 years. The other branch will randomly assign 6000 high-risk women to anastrozole versus placebo for 5 years. AIs have been tested exclusively in postmenopausal women because of their limited ability to obliterate aromatase activity in premenopausal women. The major unresolved issues with AIs include potency, selectivity and toxicity due to the elimination of desirable pro-estrogenic activity in the bone.

Prophylactic oophorectomy is an important “hormonal intervention” that has been shown to decrease the risk of developing breast CA, not only in the general population when performed at an early age, but more importantly, in women with BRCA1 and BRCA2 gene mutations. These patients have a 65–80% lifetime risk of developing breast CA. Prophylactic oophorectomy was assessed in a case–control study, evaluating 1439 patients with breast CA and 1866 matched controls derived from a registry of BRCA1 and BRCA2 carriers. The odds ratio (OR) for breast CA after bilateral oophorectomy was established using conditional logistic regression, and matched for parity and OC usage [19]. A previous history of oophorectomy significantly reduced the breast CA risk in BRCA1 carriers (OR = 0.44; 95% C.I., 0.29–0.66). In women with a BRCA2 mutation, statistic significance was not reached, but a trend was identified (OR = 0.57; 95% C.I., 0.28–1.15). The risk reduction for breast CA was greatest if the bilateral oophorectomy was performed before age 40. As noted above, this represents an inconsistent picture, in that 80% of BRCA 1 breast CAs are ER negative.

Endometrial carcinoma (EC) is the most common gynecologic malignancy affecting U.S. women, with approximately 42,000 new cases annually. Distinguishing between EC subtypes is critical when considering prevention, and they are broadly classified into two major groups. Type II, which are non-estrogen-dependent cancers, are mostly high-grade serous papillary, clear cell or adenosquamous carcinomas that have a poor prognosis [20]. Type I, or estrogen-dependent cancers, are predominantly low-grade endometrioid tumors, which are associated with a significantly better prognosis, and are by far the more common. A type I precancerous lesion does exist, atypical endometrial hyperplasia. The excessive estrogen exposure that “drives” type I EC most commonly results from either unopposed exogenous estrogen or obesity, where there are higher circulating levels of estrogen. The association between excessive estrogen exposure and the development of Type I EC provides us with significant opportunity for EC chemoprevention.

Oral Contraceptives (OCs) have been shown to decrease the incidence of Type 1 EC. The risk reduction is greatest in women taking OCs for more than 1 year. A 20% reduction in EC risk is seen in women who take OCs for 1 year, and an approximately 80% reduction is observed with usage greater than 10 years. The protective effect occurs in low- and high-dose progestin preparations, but is most prominent in women of low parity. After discontinuation of OCs, a patient can expect to enjoy this reduction for up to 20 years [21] and [22]. The Royal College of General Practitioners' oral contraceptive study, which began in 1968, recruited approximately 23,000 women who were using OCs and 23,000 women who had never used them. This inception cohort study, with 339,000 women-years of observation of never-users, compared with 744,000 women-years of ever-users, revealed a relative risk (RR) of developing EC with ever-use equal to 0.58 with a 95% confidence interval of 0.42–0.79. While the study's limitations include the basic weaknesses of any cohort study, a major strength was the ability to include more than a million women-years of observation, accumulated over 36 years. Virtually all of the women in this study were postmenopausal when the study was published in 2007, allowing for a large number of events for analysis [23].

A plethora of data have shown that obesity increases the risk of EC, with up to 40% of EC cases being attributable to excessive body mass index (BMI) [24]. Excessive adipose tissue raises the levels of endogenous estrogen, which is accompanied by enhanced proliferation of endometrial epithelial cells and subsequent higher risk of EC [25]. Obese women are also more likely to have increased levels of bioavailable estrogen, secondary to decreased sex hormone binding globulin (SHBG). Anovulatory cycles and decreased levels of progesterone also contribute to the overall picture of estrogen excess. Importantly, not only is there a higher incidence of EC in obese women, but in the Cancer Prevention Study II, Calle EE et al. established that the RR of death in women with EC, when comparing those with a BMI greater than 40 to the reference group (BMI 18.5 to 24.9), was 6.25 with 16 years of follow-up (Fig. 2) [26].

 

There is a suggestion that the rates of EC are starting to rise and, one likely explanation for any such increase is the rise in prevalence of obesity [27]. Importantly, EC is predominantly a disease of older age, whereas obesity is currently more common among middle-aged women, supporting the premise that the full effects of the obesity epidemic on EC rates have yet to be appreciated. Preventing obesity should prevent a large proportion of ECs. The challenge is enormous due to the high-calorie diet and sedentary lifestyle which is firmly entrenched in the lives of many American women.

A more relevant question for U.S. women who are already obese is whether losing weight will reduce their EC risk, or if this is predetermined by their earlier body size. It is believed that the effects of obesity are largely mediated by the associated increases in exposure to endogenous estrogen. Thus it is likely that obesity acts at a late stage as a cancer promoter, leading one to believe that weight loss will reduce a woman's EC risk. However, despite the evidence that weight gain and current obesity are strongly associated with EC risk, there is currently very limited information regarding the effects of weight loss. In a study from the Netherlands, women whose BMI at the start of follow-up was lower than their BMI at age 20 (12% of the cohort) had a 50% lower risk of developing EC than those whose BMI had remained constant or increased slightly [28]. Similarly, an investigation by Trentham-Dietz et al., evaluated the effects of weight loss on EC risk by comparing women's maximum and minimum weights during adulthood. Women who had sustained some degree of weight loss (the amount was not specified) for 5 years or more had a 25% lower risk of developing EC than those who had not lost weight [29]. There continues to be insufficient evidence to firmly conclude that weight loss decreases the risk of EC. Further studies that pay particular attention to the effects of sustained weight loss among obese individuals are necessary to firmly establish this relationship. However, the data are compelling that avoiding excessive weight gain correlates with a lower risk of developing Type I EC.

Endometrial hyperplasia is recognized as a precursor to Type 1 EC. The treatment of this precancerous lesion, logically, would diminish a woman's risk of developing EC. While oral progestin therapy is an accepted therapeutic modality in the treatment of endometrial hyperplasia, these agents are associated with poor compliance and systemic side effects that may limit overall efficacy [30], [31] and [32]. Interest in the levonorgestrel-releasing intrauterine system (LNG-IUS) is significant. Varma R et al. prospectively evaluated 105 women diagnosed with endometrial hyperplasia and treated with LNG-IUS between 1999 and 2004. Baseline characteristics and outpatient Pipelle sampling were undertaken at 3 and 6 months, and thereafter at 6-month intervals in all cases. LNG-IUS achieved endometrial regression in 90% of cases (94/105) by 2 years, with a significant proportion, 96% (90/94) achieving this within 1 year. Regression occurred in 92% (88/96) of nonatypical and 67% (6/9) of atypical hyperplasias, and in all 22 cases of endometrial hyperplasia associated with HRT. Regression rates did not differ between histological types of hyperplasia. Twenty-three women (22%) underwent hysterectomy, of which 13 were indicated and 10 were performed at patient request despite regressed endometrium. Two cases of cancer (one uterine and one ovarian) were identified [33]. Randomized, controlled trials are needed to evaluate the effectiveness of LNG-IUS in the treatment of women with endometrial hyperplasia. Prospective studies also are needed to assess the role of LNG-IUS in the prevention of EC in women deemed high risk on the basis of BMI.

Preventing EC by screening with endometrial biopsy in women with high levels of unopposed estrogen on the basis of obesity is a logical consideration, albeit not a hormonal intervention per se. Cost-effectiveness is an obvious concern when considering prevention strategies for EC. Kwon, J. and Lu, K. utilized a Markov decision-analytic model to evaluate four prevention strategies for EC including: no prevention, OCs, annual screening with endometrial biopsy from age 30, and biennial screening from age 30 [34]. Net health benefit was life expectancy and primary outcome was the incremental cost-effectiveness ratio. Baseline and transition probabilities were obtained from published literature and the Surveillance Epidemiology and End Results database, and costs were from the U.S. Department of Health and Human Services and Agency for Healthcare Research and Quality. Sensitivity analyses were performed for uncertainty around various measures. None of the strategies had an incremental cost-effectiveness ratio less than $50,000 per year of life saved, relative to the next-best strategy. EC risk in obese women had to be 13 times greater than the general population risk before OCs were a cost-effective intervention. The authors concluded that OCs and current screening methods are not cost-effective EC prevention strategies for obese women. The authors point out that this analysis is Level III data only, and concluded that larger prospective studies are needed to assess interventions that reduce EC, as well as provide additional health benefits.

Finally, Hereditary Nonpolyposis Colorectal Cancer (HNPCC), also known as Lynch Syndrome, is a hereditary syndrome associated with a cumulative lifetime risk for colorectal cancer of approximately 50%. The lifetime cumulative risk of EC for women with Lynch syndrome equals or exceeds their risk of colorectal cancer (40 to 60 %) [35]. The current gynecologic cancer screening guidelines for women with Lynch Syndrome include annual endometrial sampling and transvaginal ultrasonography (TVUS) beginning at the age range of 30 to 35 years [36].

Epithelial ovarian cancer (EOC) has the highest mortality rate of all gynecologic cancers in the United States. The American Cancer Society estimated that 21,550 new cases of EOC would be diagnosed in 2009 and 14,600 of these women would die of the disease [37]. The most consistently observed influences on the increased risk of sporadic ovarian cancer are infertility and low parity [38], [39] and [40]. A family history of EOC or breast cancer in a first-degree relative approximately triples the risk. This risk is particularly high among carriers of the BRCA1 or BRCA2 gene mutations, where the lifetime risk is 40–45% and 15–20%, respectively [41]. This subset of women who are at high risk because of genetic mutations constitutes approximately 8–10% of all women who are diagnosed with EOC. The high mortality rate observed in women with this malignancy, in part, results from the lack of early symptoms. Greater than 70% of women diagnosed with EOC are diagnosed when the disease has progressed to stage III or IV. It is estimated that approximately $2.2 billion is spent in the United States each year on treatment of ovarian cancer. In 2007, the NCI spent $96.9 million on ovarian cancer research. While this review is focused on prevention strategies, it is impossible, in the case of EOC, to ignore the issue of screening. The resources directed toward effective screening continue to be significant, recognizing that early detection, which is associated with improved survival, depends on screening strategies that are both medically effective and cost-effective. At this time, routine EOC screening of the general population is not recommended by any professional society (Table 1) [42]. However, women who are high risk for EOC on the basis of a BRCA mutation or who have a family history of breast or ovarian CA warrant a different “screening” approach. While there is no evidence demonstrating that measurement of CA125 and transvaginal ultrasonography every 6 months provides a survival benefit, the National Comprehensive Cancer Network does recommend this strategy for these high-risk patients. In summary, for women with or without a genetic predisposition to EOC, the development of prevention strategies against this disease is vitally important, particularly in the face of no clearly effective screening paradigm and its associated high mortality rate.

 

Several studies have reported that the use of OCs decreases the risk of developing EOC. The Cancer and Steroid Hormone Study investigated the association between OC usage and the risk of developing EOC in a case-controlled study [43]. From 1980 to 1982, 546 women with EOC, ranging in age from 20 to 54, were enrolled from eight population-based cancer registries. The controls were 4228 women selected from the same geographic areas. Women who had used oral OCs had a risk of EOC of 0.6 (95% confidence interval, 0.5 to 0.7), as compared with those who had never used them. The protection was seen in women who had used OCs for as little as 3 to 6 months, and it continued for 15 years after use ended. The protection was independent of the specific oral-contraceptive formulation and of the histologic type of EOC. More recently, The Royal College of General Practitioners' Oral Contraception Study firmly substantiated the protective effect of OCs against EOC. This study was initiated in 1968, and follow-up ceased in 1996. Over a 14-month period, 1400 general practitioners in the United Kingdom recruited 23,000 women using OCs (744,000 women-years of observation), and approximately the same number who did not (339,000 women-years of observation). The RR for developing EOC in ever-users of OCs was 0.54 with a CI 95% = 0.40–0.71. In order to more fully evaluate the eventual public-health effects of reducing the incidence of EOC by OC usage, specifically how long the protection lasts after use ceases, the Collaborative Group on Epidemiological Studies of Ovarian Cancer [44] analyzed centrally individual data for 23,257 women with EOC (cases) and 87,303 without EOC (controls) from 45 epidemiological studies in 21 countries. The RR of EOC in relation to OC use was estimated, stratifying by study, age, parity and hysterectomy. Overall 7308 cases (31%) and 32,717 (37%) controls had ever used OCs, for average durations of 4.4 and 5.0 years, respectively. The longer women had used OCs, the greater the reduction in EOC risk (p < 0.0001). This risk reduction persisted for more than 30 years after OC use, but lessened over time. The proportional risk reductions per 5 years of use were 29% (95% CI 23–34%) for use that had ceased less than 10 years previously, 19% (14–24%) for use that had ceased 10–19 years previously, and 15% (9–21%) for use that had ceased 20–29 years previously. Use during the 1960s, 1970s and 1980s was associated with similar proportional risk reductions, although typical estrogen doses in the 1960s were more than double those in the 1980s. The authors concluded that OC usage provides long-term protection against EOC. They estimate that approximately 200,000 cases of EOC and 100,000 deaths from the disease were prevented in the general population.

Women known to carry a BRCA 1 or 2 gene mutation have a significantly increased risk of breast CA (65–80%), as well as an increased risk of EOC (45% and 20%, respectively), as described above. Prophylactic salpingo-oophorectomy (BSO) (a hormonal intervention) has been shown to be an effective prevention strategy in the general population, as well as women with a BRCA gene mutation. Finch A et al. evaluated the reduction in incidence of ovarian, fallopian tube and primary peritoneal cancer with surgical intervention in women who carry a deleterious mutation in either BRCA1 or BRCA2 [45]. Women known to carry either of these mutations were identified from an international registry between 1992 and 2003. A total of 1828 carriers at 1 of 32 centers in Canada, the United States, Europe and Israel completed questionnaires at baseline and follow-up. Participants were observed from the date of study entry until: diagnosis of ovarian, fallopian tube, or peritoneal cancer; death; or the date of the most recent follow-up. Incidence was determined by a time-dependent survival analysis that was adjusted for covariates to assess the risk reduction associated with prophylactic BSO. With a mean follow-up of 3.5 years, 50 incident ovarian, fallopian tube and peritoneal cancer cases were reported in the cohort. Of the 1828 women, 555 (30%) underwent a prophylactic BSO prior to study entry, 490 (27%) underwent the procedure after entering the study, and 783 (43%) did not undergo the procedure. There were 32 incident cancers diagnosed in women with intact ovaries (1,015/100,000 per year). Eleven cancer cases were identified at the time of prophylactic BSO, and 7 were diagnosed following prophylactic BSO (217/100,000 per year). The estimated cumulative incidence of peritoneal cancer is 4.3% at 20 years after BSO. The overall (adjusted) reduction in cancer risk associated with prophylactic BSO is 80% (multivariate hazard ratio = 0.20; 95% confidence interval, 0.07–0.58; P = .003). The authors concluded that bilateral BSO is associated with reduced risk of ovarian and fallopian tube cancers in high-risk women, although there is a substantial residual risk for peritoneal cancer in BRCA1 and BRCA2 mutation carriers following prophylactic BSO.

The diagnosis of EOC in women who are BRCA gene carriers occurs at a significantly younger age than that of women who are not gene carriers. While prophylactic BSO affords protection as described above, is there a chemoprevention strategy available that would provide protection for these women until they have completed childbearing? To evaluate the potential benefit of OC use in women at high risk for EOC, Narod SA et al. studied 207 patients with BRCA1 or BRCA2 mutations and EOC, and 161 of their sisters, who served as controls. The adjusted OR for EOC associated with any past use of OCs was 0.5 (95% C.I., 0.3 to 0.8). The risk decreased with increased duration of use (P for trend, < 0.001); use for six or more years was associated with a 60% reduction in risk. OC use protected against EOC both for carriers of the BRCA1 mutation (OR, 0.5; 95% C.I., 0.3 to 0.9) and for carriers of the BRCA2 mutation (OR, 0.4; 95% C.I. 0.2 to 1.1). The authors concluded that OC use may reduce the risk of EOC in women with pathogenic mutations in the BRCA1 or BRCA2 gene [46]. While in general it is recommended that premenopausal BRCA carriers remain on OCs, the possible benefit must be weighed against the slight increased risk of breast CA seen in women with BRCA mutations.

Cancer prevention for the women under our medical care continues to be the ultimate goal for all gynecologic oncologists. Prospective, randomized trials, designed to control for all variables known to exist, are mandatory to fully assess the potential for hormonal chemoprevention in breast, endometrial and ovarian cancers.

The author declares that there are no conflicts of interest.