Borderline Ovarian Tumors Share Familial Risks with Themselves and Invasive Cancers

Borderline ovarian tumors (BOTs) are a subgroup of ovarian malignancies with low malignant potential (1). The common histologic types of BOTs are serous and mucinous BOTs; papillary type has been considered as a distinctive variant of serous BOTs (2). In the 2014 WHO Classification of Tumours of the Female Genital Organs, the main refinements concerned serous BOTs, which was described to share molecular and genetic alterations with low-grade serous carcinomas, specifically for micropapillary variant of serous BOTs (2). BOTs display epithelial proliferation higher than that seen in benign tissue and variable nuclear atypia. Compared with ovarian carcinomas, BOTs usually occur in young women, have no destructive stromal invasion, and have better prognosis (3). However, the treatment for BOT is usually surgery, which may affect the fertility of women (4).

Risk factors for BOTs are largely similar to those reported for invasive ovarian tumors, including protective effects of pregnancy and lactation (5, 6). However, some studies on reproductive and hormone-related factors showed distinct risks between BOTs and ovarian cancer (7–9). For example, risks for BOTs were less reduced among women who had used oral contraceptives and more elevated among women with a history of infertility compared with women with ovarian cancer (7). In agreement with ovarian cancer, smoking is a risk for mucinous BOT (RR = 2.25) but not for serous BOT (10). Pelvic inflammatory disease has been reported as a modest risk factor (11).

Family history of ovarian cancer is one of the strongest known risk factors for ovarian cancer (12). Yet family history data on BOTs are sparse. No significant risk was observed for concordant BOTs, that is, BOT in two family members; however, an association of BOT with invasive ovarian cancer was found among sisters but not among mother–sister pairs (13, 14). Very few studies reported BOTs are related to germline mutations in BRCA1/2 (15, 16), MSH2 (15), and CHEK2 (17). Low-risk genetic loci were described for serous BOT alone or in combination with low-grade serous ovarian cancer in a large genome-wide association study (18). We examine here familial risks for patients with BOT considering both BOT and any invasive cancers in family members based on the nationwide Swedish Family-Cancer Database (FCD), the largest dataset of its kind in the world, to resolve whether BOTs share familial risk with themselves and invasive cancers.

FCD includes all Swedish people born since 1932 (offspring generation) with their biological parents (parental generation; ref. 19). The recently updated version of FCD contains 16.1 million individuals, among which almost 2.0 million were patients with cancer recorded to the end of 2015. The 3-digital codes of 7th revision of the International Classification of Diseases (ICD-7) were used to identify BOTs, 35 most common primary cancers, and cancer of unknown primary. BOTs are classified as in situ tumors in the Swedish Cancer Registry. Histologic subtypes of BOTs were classified according to Systemized Nomenclature of Medicine (SNOMED) codes since 1993, into serous, papillary, and mucinous types. The follow-up for cancer in the offspring generation was started from the beginning of 1958 (for histologic analysis, it began from 1993), the birth year, or the immigration year, whichever came latest. The follow-up was terminated when a person was diagnosed with BOT or cancer, or when he emigrated or died, or at the end of 2015, whichever came first. Having first-degree relatives (including parents and/or siblings) who were affected with cancer was considered as family history.

For familial risk analysis, the incidence in those with a family history of BOTs or cancer (affected relatives) was compared with the incidence in persons (reference group) without a family history (unaffected relatives). A two-way comparison was employed to estimate familial RRs for (a histology-specific) BOTs in daughters when a first-degree relative was diagnosed with any invasive cancer, for example, brother with cancer X (Fig. 1, left), and conversely familial risk for invasive cancer X in offspring when mothers or sisters were diagnosed with (a histology-specific) BOT (Fig. 1, right). For parents and offspring relation (large majority of familial cases), the two comparisons are independent, but for siblings, the pairs of cases are the same.

The Poisson regression model was employed to estimate RRs and corresponding confidence intervals (CIs) for 5%, 1%, and 0.1% significance levels. We use here different significance levels to differentiate the likely true associations from likely chance findings, and a single 95% CI is not very informative in the context of multiple comparisons. Potential confounders, including age group (17 groups with 5-year gap), sex, calendar period, residential area, and socioeconomic status, were added to the model as covariates. SAS version 9.4 was used to perform the statistical analysis.

The study was approved by the Ethical Committee of Lund University (Lund, Sweden), and the study was conducted in accordance with the approved guidelines.

The data that support the findings of this study are available from Lund University, but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available.

A total of 4,729 BOT cases were found in our database and of these, 4,199 cases were in the offspring generation diagnosed at ages 0 to 83 years, for which RRs were calculated. Among all BOT cases in the offspring generation, there were 34 (0.8%) cases with a first-degree relative affected by BOT, of which seven cases were serous, seven were papillary, and nine were mucinous (Table 1). Individuals have increased risk of being diagnosed with BOT when first-degree relatives were affected by any histologic type of BOTs (n = 34, RR = 2.20, P < 0.001), and by papillary BOT (n = 7, RR = 4.02, P < 0.001). Risk for papillary BOT was elevated with a family history of any histologic type of BOTs (n = 7, RR = 2.40) and papillary BOT (n = 4, RR = 10.10, P < 0.001) in first-degree relatives. Increased risk for mucinous BOT was observed in families with any histologic type of patients with BOT (n = 9, RR = 2.06).

Among all BOT cases in the offspring generation, a total of 2,489 (59.3%) cases had first-degree relatives diagnosed with invasive cancers. Significant familial associations of all BOTs with invasive cancers were found for pancreatic (1.39, P < 0.01), lung (1.20, P < 0.01), and bone (2.27, P < 0.01) cancers and leukemia (1.23; Table 2). In the reverse order, RRs for colorectal and ovarian cancers were 1.18 and 1.59 (P < 0.01) when a family member was diagnosed with BOT. The RRs of BOT were 1.14 in families with patients diagnosed with any cancer including or excluding ovarian cancer. On the contrary, a family history of BOT was associated with the increased risk of any cancers (1.06, P < 0.01), and the risk slightly decreased (1.05) when only considering any cancers other than ovarian cancer. Only cancer sites with at least 10 familial cases (column n) in either of the two-way comparison were included in Table 2.

Results from detailed analyses of familial risks in histology-specific BOT with invasive cancers are shown in Table 3. However, included are only invasive cancers with at least one significant association with histology-specific BOT; in addition, ovarian cancer was included, although it showed no significant associations. The risk for serous BOT was not associated with a family of invasive cancers. Increased risk of papillary BOT was associated with family history of lung (1.39) and endocrine gland (1.68) cancers and with all cancers (1.25, P < 0.01). Mucinous BOT risk was elevated when first-degree relatives were diagnosed with anal (2.69) and pancreatic (3.30, P < 0.01) cancers. In the reverse analysis, elevated risks for pancreatic (3.03, P < 0.001) and kidney (2.87, P < 0.001) cancers were found in family with papillary BOT patients thyroid gland cancer in family with serous BOT patients.

BOTs are one subset of epithelial neoplasms with low malignant potential that affect women in the reproductive age group and show excellent prognosis with 5-year survival rates of 95% to 97% (20). The current study on the relation between BOT and invasive cancers is mainly about the personal risk of invasive cancer after diagnosis of BOT, especially risk of ovarian cancer, as the histopathologic findings strongly suggest a continuum from benign presentation to borderline and to invasive ovarian tumors (21). The vast majority of serous carcinoma arise de novo, whereas only 5% to 10% of them are derived from the precursor lesion serous BOT (22). High heterogeneity (benign, borderline, intraepithelial carcinoma, microinvasion, and invasive carcinoma) is presented in mucinous tumor (23). There are a few studies focusing on the familial association of BOT with invasive cancers (24). The current and the previous studies provide insight into the etiology of BOT that helps in the management of patients with BOT and counselling of their family members.

On the basis of the FCD, we found the familial risk of BOT with a family history of BOT was 2.20, higher than the risk with a family history of any invasive cancers (1.14). However, BOT cases with a family history of BOT only accounted for a small proportion (0.8%) of all the patients with BOT compared with the large amount of the patients with BOT with the first-degree relative diagnosed with invasive cancers (59.3%). The familial risk for BOT (RR = 2.20) was also higher than the association of BOT with ovarian cancer (RRs = 1.17 and 1.59) in two-way analyses of which only the latter was significant. These data probably suggest that some risk factors responsible for the development of familial BOTs do not predispose to development of invasive ovarian cancer, even though the two diseases have common risk factors such as parity and sex hormone levels (7, 9–11, 25). As far as we know, risk of BOT with a family history of BOT has never been reported before. Furthermore, in histology-specific analysis, only concordant family history of papillary BOT was significant, which may suggest that this histology has a unique pathophysiology, probably in accordance with the recent WHO classification in which papillary variant of serous BOT is mentioned as a distinct subtype of serous BOT (2). Of cautionary note, all the results of the familial associations of BOT were based on small numbers of cases.

We used two-way comparisons in the search of familial associations of BOT with invasive cancers to find internal support to the findings. However, incidence rates and diagnostic age ranges differ between BOT and many cancers; thus, a lacking two-way association is no strong evidence against an association. We found significant results including or excluding ovarian cancer. We did this because ovarian cancer shares many common risk factors with BOT that can drive the associations between BOT and other cancers. However, the magnitude of effect remained similar after excluding ovarian cancer, indicating that BOT shares common risk factors with other cancers.

RRs for BOT were found to be increased in families with pancreatic, lung and bone cancers, and leukemia. In the reverse analyses, RRs for colorectal and ovarian cancers were increased in families with BOT. No associations were significant in both of the two-way analyses, although for lung and ovarian cancer, one RR was significant at 5% level and the other one was of borderline significance. The association with pancreatic cancer was earlier reported from a population-based study in Finland (24). The associations between BOT with invasive cancers found here may be related to shared common genetic or environmental factors. The association of BOT with ovarian may be contributed by germline mutations in BRCA1/2 and MSH2 as well as by some reproductive and hormone‐related factors. Association with pancreatic cancer may probably be due to germline mutations in BRCA1/2 (26), colorectal cancer due to MSH2 (27), and bone cancer due to CHEK2 (28). Furthermore, smoking may contribute to the associations with lung and pancreatic cancers.

The analyses of familial risk of BOTs stratified by histology revealed associations with some cancers, although those associations were only significant in the one-way analysis. Smoking is a risk factor for mucinous BOT, and we found familial associations of mucinous BOT with smoking-related cancers, including anal, pancreatic, and lung cancers, although the latter was marginally significant (10, 29, 30). Serous BOT is thought to be a precursor to low-grade serous ovarian carcinoma (18); no familial associations between these two were observed in the two-way comparison, yet the results were both moderately significant. Papillary BOT was associated with most number of cancers, including pancreatic, lung, kidney, and endocrine gland cancers. Papillary BOT is characterized by a frequent association with extraovarian implants (particularly invasive implants), which confer prognostic information of increased risk (31).

To the best of our knowledge, this is so far the largest study focusing on familial aggregation of BOTs. As the FCD is based on registered resources with practically complete nationwide coverage of medically diagnosed cancers, it provides a reliable estimation of familial risks. However, there are limitations in our study. First, identifiable histology was diagnosed only after 1993 when the ICD-O/2 classification was taken to use in the cancer registry. A 22-year follow-up, although the longest yet reported, is still short for intergenerational study considering risks of both the parental and offspring generations. Information on possible confounders, such as smoking and use of oral contraceptive use, was not available, yet we adjusted for socio-economic and demographic factors to reduce possibility of confounding.

In summary, we report for the first time that family history contributes to the risk of concordant BOT. BOTs showed familial associations with some invasive cancers, most consistently with colorectal, pancreatic, ovarian, and bone cancers as well as with leukemia. These results provide insight into familial associations of BOT, which may also help in surveillance and genetic counseling of patients with BOT. Furthermore, the familial clustering of BOT with invasive cancers suggests that some preventive counselling for invasive cancer, such as smoking cessation, can also be applied to BOT. However, due to small numbers of cases in this study, larger studies with detailed lifestyle information are needed to validate the present familial associations between BOT and invasive cancers.

No potential conflicts of interest were disclosed.

Conception and design: K. Hemminki

Development of methodology: G. Zheng

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): K. Sundquist, K. Hemminki

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): G. Zheng, H. Yu, K. Sundquist, K. Hemminki

Writing, review, and/or revision of the manuscript: G. Zheng, A. Kanerva, A. Forsti, K. Sundquist, K. Hemminki

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): K. Sundquist

Study supervision: K. Sundquist, K. Hemminki

This work was funded by the German Cancer Aid, the EU Transcan funding by the German Federal Ministry of Education and Research, the Swedish Research Council for Health, Working Life and Welfare (FORTE; Reg. no. 2013-1836), FORTE (Reg. no. 2014-0804), and the Swedish Research Council (2012-2378 and 2014-10134), and ALF funding from Region Skåne as well as by the China Scholarship Council (201606100057, to G. Zheng).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.