Introduction
According to statistics regarding cancer incidence and mortality, breast cancer is the most common malignancy diagnosed in females, accounting for nearly one in three cancers [
1]. Breast cancer is also the second leading cause of cancer-related death after lung cancer among women [
2]. With the development of the early screening, detection, and systemic treatment of breast cancer, significant improvements in breast cancer survival outcomes have been made. Thus, women with breast cancer can achieve prolonged survival times and better life expectancies. However, a higher risk of developing a second primary female genital cancer among breast cancer survivors may translate into an important health problem in their lifetimes [
3]. Within 10 years after the initial diagnosis, approximately 10% of breast cancer patients develop a subsequent primary cancer, including second primary endometrial cancer and second primary ovarian cancer [
4].
Previous studies have suggested a consensus in the increased risk of second primary female genital cancers among breast cancer survivors. A large cohort study reported that women with breast cancer had a 30% excess risk for second primary cancers, particularly for endometrial cancer and ovarian cancer [
5]. The problem of second primary female genital cancers among breast cancer survivors may be related to treatment side effects or to etiological associations for multiple cancers. With the growing interest in identifying possible relationships between second female genital malignancies and first cancers, studies have focused on multiple factors, such as lifestyle, environment, treatment side effects, and hormonal and/or genetic factors. A European study reported that risk factors of second female genital malignancies included the body mass index and smoking status of breast cancer patients [
5]. Moreover, estrogen receptor (ER)-positive breast cancer patients treated with tamoxifen have an increased risk of subsequent endometrial cancer [
6,
7]. However, one study has shown that patients with breast cancer have a higher risk of subsequent endometrial cancer regardless of ER or progesterone receptor (PR) status [
8].
The aims of our study were to systemically evaluate the incidence of second primary female genital cancers in a large cohort of breast cancer patients and to identify risk factors for second primary corpus uteri cancer and ovarian cancer. Using the Surveillance, Epidemiology, and End Results (SEER) database, we calculated the standardized incidence ratio (SIR) for second primary cancers after breast cancer diagnosed between 2000 and 2014. Meanwhile, we used multivariable analysis to examine the risk factors for the development of a second primary female genital cancer after breast cancer.
Discussion
In this study, we found a significantly decreased risk of second cervical cancer in patients with first breast cancer. Viral carcinogenesis is the main risk factor for cervical cancer. Moreover, a previous study showed the induction of aromatase expression in cervical carcinomas and opened the possibility that the use of aromatase inhibitors (AIs) for primary breast cancer may be a potential favorable factor in the development of subsequent cervical cancer [
11]. Additionally, the cervix is targeted by estrogen through estrogen receptor alpha and beta (ER-α and ER-β). ER-α expression occurs mainly in the normal cervical epithelium but is lost in cervical cancer. However, ER-β is expressed in both the normal cervix and cervical cancer. One study has suggested that estrogen may affect ER-β, which is a suppressor of ER-α in carcinogenesis [
12]. Thus, AI therapy for postmenopausal breast cancer patients and ER-β expression might be the causes of the decreased risk of second cervical cancer among breast cancer survivors.
Our study showed that the SIR for second primary corpus uteri cancer was significantly increased in patients with first breast cancer regardless of ER and PR status. Previous studies have confirmed an increased risk of second corpus uteri cancer after ER-positive breast cancer in patients with tamoxifen therapy compared to the general population [
13,
14]. Interestingly, ER−PR− breast cancer patients who were less likely to receive tamoxifen therapy also had a significantly increased SIR for second corpus uteri cancer, and this result suggests that tamoxifen therapy may not be the only potential reason for second corpus uteri cancer after initial breast cancer. The results of a recent study were consistent with our study, which showed that breast cancer patients with different hormone receptor statuses had increased risks of subsequent corpus uteri cancer [
8]. In addition to tamoxifen therapy, shared BRCA1/2 gene mutations may contribute to a higher risk of corpus uteri cancer among breast cancer patients, compared to the general population [
15]. Our study suggested a trend in which the SIR of second corpus uteri cancer was declining more in recent years. However, this trend may be merely a reflection of the shorter latency period for breast cancer survivors more recently. On the one hand, Table
1 shows that the risk of second uterine corpus cancer increases with the latency period. On the other hand, a large case-control study found that patients using tamoxifen for less than 2 years failed to show a statistically increased risk of developing uterine corpus cancer [
16]. Our study continued until December 31, 2014, and revealed that numerous breast cancer patients diagnosed from 2010 to 2014 may have undergone tamoxifen treatment for less than 2 years, so the effect of tamoxifen on second uterine corpus cancer development among these patients was underestimated.
Our study suggested that survivors had a higher risk of developing second primary ovarian cancer after the first breast cancer than the general population, which was in line with previous studies [
17,
18]. The results further revealed that the young breast cancer survivors had a higher risk of second primary ovarian cancer than young women in the general population, and the peak of this risk was particularly notable in the 30–39 age group but lowest in the 70+ age group. This may be associated with a common germline mutation (BRCA1) and similar hormonal exposure. To some extent, BRCA1 mutations expressly cause breast cancer at a young age [
19]. In addition, oophorectomy, a treatment for breast cancer, can lead to an incidental finding of early ovarian cancer. The SIRs for second primary ovarian cancer were significantly increased after ER-negative breast cancer. This increase may be related to the common BRCA1 gene for both early-onset ER-negative breast cancer and ovarian cancer [
20]. Surprisingly, for both corpus uteri and ovarian sites, women with ER−PR+ breast cancer appeared to have the highest risk. However, progesterone signaling should not be viewed as a risk factor for both uterine corpus and ovarian cancers. Progesterone is a favorable prognostic factor and inhibits cancer cell growth and metastasis development in both uterine corpus cancer and ovarian cancer [
21]. ER−PR+ breast cancer is rare because PRs are ER-induced genes and because PR positivity is driven by an active ER. Thus, the expression of ER−PR+ may be induced by other signaling. This alternative signaling may explain how ER−PR+ breast cancers are associated with earlier recurrence times and poorer overall survival rates than ER+PR+ breast cancers [
22]. Patients with ER−PR+ breast cancers still receive tamoxifen therapy and are influenced by the side effects of tamoxifen, but these patients have poorer outcomes than ER+PR+ patients.
In contrast to previous studies, we had extensive information on cancer risk factors for second female genital cancers, including the year of breast cancer diagnosis, age at breast cancer diagnosis, race, and hormone receptor status. For breast cancer survivors, in the development of second primary female genital cancer, older age (45+ years), black race, and PR+ breast cancer tended to be associated with a higher risk of second corpus uteri cancer. Meanwhile, older age (56+ years), white race, and ER- breast cancer needed heightened vigilance for potential second ovarian cancer. Elevated risks were noted in patients who were perimenopausal or postmenopausal, which may indicate an age-related preference in the risk of second primary female genital cancer among breast cancer patients. However, ER+ breast cancer patients had a decreased risk of developing second ovarian cancer compared to ER− patients. This decreased risk may be due to hormone therapy for first breast cancer. Nevertheless, the SEER database does not contain complete treatment information to evaluate the impact of hormone therapy.
Our study had strengths and weaknesses. This study was based on a large, well-established, and standardized population database. However, the heterogeneous population and retrospective analysis of the SEER data were the main limitations. In addition, patients with first primary breast cancer tend to be given more thorough attention than the general population, and a surveillance bias exists for patients with second primary cancer. To address these issues, we analyzed data from a 15-year period and data from 18 registries, making these systemic errors non-differential. Potential pre-existing or synchronous cancers were another concern. Therefore, patients with second primary cancers diagnosed within 6 months of breast cancer diagnosis were excluded from our cohort. Moreover, due to the composition of the dataset, the numbers of patients for second female genital cancers in some race and age groups of breast cancer patients were small. These relevant findings would need further confirmation.
In conclusion, this large population-based study suggests that women with first primary breast cancer have higher risks of developing second primary corpus uteri cancer and second primary ovarian cancer than the general population. In contrast, breast cancer survivors have a decreased risk of developing second primary cervical cancer. These increased and decreased risks of developing second primary cancers suggest associations with breast tumor biology and shared genetic susceptibility. Meanwhile, several risk factors were associated with an increase (age at first cancer, race, PR status) or a decrease (year of first cancer) of second primary corpus uteri cancer and with an increase (year of first cancer, age at first cancer) or a decrease (race, ER status) of second primary ovarian cancer. These findings are useful for health planning, including screening and the development of specific guidelines for both the general population and breast cancer patients. Further work is needed to explore possible reasons for this association and the survival outcomes of these patients.