By Robert L. Coleman, MD, Professor, University of Texas; M.D. Anderson Cancer Center, Houston. Dr. Coleman reports no financial relationships relevant to this field of study.
Synopsis: Self-reported personal history of endometriosis was associated with an increased risk of ovarian cancer. Further, it was differentially associated with clear cell, endometrioid, and low-grade serous ovarian carcinoma in this pooled analysis. No relationship appeared between endometriosis and high-grade serous or mucinous ovarian cancer, or borderline variants of these two histologies. The results suggest further work is necessary to understand whether endometriosis plays a strategic precursor role in certain ovarian cancer histological subtypes.
Source: Pearce CL, et al. Association between endometriosis and risk of histological subtypes of ovarian cancer: A pooled analysis of case-control studies. Lancet Oncol 2012;13:385-394.
Several studies have implicated endometriosis as a risk factor for the subsequent development of invasive epithelial ovarian cancer. However, while some histological subtypes, such as clear cell and endometrioid, have been associated, little is known of the other subtypes (e.g., serous and mucinous). To address the hypothesis, members of the Ovarian Cancer Association Consortium (OCAC) pooled data from 13 predominately population-based case-control studies conducted in Australia (1), Europe (3), and the United States (9), assessing risk factors for the development of ovarian malignancy. The data cohort included 23,144 women; 7911 (34%) with invasive epithelial ovarian cancer, 1907 (8%) with borderline tumors, and 13,326 (58%) controls. Controlled confounding variables were age, ethnic origin, oral contraceptive use, parity, breastfeeding, weight, height, body mass index, tubal ligation, and family history. Serous histology was re-classified by WHO grade with grade I representing "low-grade" serous (LGSOC) and all others as "high-grade" serous (HGSOC). A history of endometriosis was found in 9.3% of women diagnosed with invasive ovarian cancer compared to 6.2% of controls (odds ratio [OR] 1.46, 95% confidence interval [CI] 1.31-1.63); in addition, among the five histological subtypes considered, the association was strongest for clear cell carcinoma (OR 3.05, 95% CI 2.43-3.84), compared to endometrioid (OR 2.04, 95% CI 1.67-2.48) or LGSOC (OR 2.11, 95% CI 1.39-3.20). No relationship was observed between endometriosis and invasive HGSOC or mucinous cancer, or of borderline serous and mucinous histology. The effects were upheld in a sensitivity analysis, which considered the time between endometriosis diagnosis and the diagnosis of cancer. The authors conclude that the association should raise awareness in treating clinicians and spark investigation into mechanistic processes driving malignant progression.
This article is by far the largest collection of retrospective data to address the relationship between endometriosis and ovarian cancer. The strength of the pooled analysis is that it provides the ability to delve further into histological association, which has been done only to a limited degree and for only certain subtypes.1 The provocative data do raise the hypothesis that, albeit uncommon, a malignant ovarian phenotype occurs to a greater degree in women with a personal history of endometriosis than those without, particularly for clear cell and endometrioid tumors (previously known) and LGSOC (previously unknown). The relationships are compelling, but as is the case with retrospective studies, cannot be inferred as causal, and must consider several confounding issues and procedural assumptions that could affect the strength of association.
First, all the data regarding the diagnosis of endometriosis in this study is self-reported. There was no attempt to confirm that, indeed, patients had the disease (e.g., a site audit). In two sites where central registries are maintained, the diagnosis was confirmed by review of discharge summaries, but even in these cases it is unknown if these are histologically based or an inferred diagnosis for pelvic symptoms. The authors acknowledge this limitation but state it would unlikely affect the differential association among the subtypes. However, it is plausible that since women with clear cell, endometrioid, and LGSOC are diagnosed at a younger age, bona fide endometriosis may be differentially recognized by a higher frequency of surgical procedures and confirmatory histology. Second, the histological criteria by which clear cell carcinoma and LGSOC is made are not clearly provided. This has been a significant issue in treatment studies where adjudication of the pathology must be made prior to registration. For instance, it is unclear how cases of mixed clear cell and serous carcinoma were handled. Third, ethnic differences were considered in the confounding variables, which is appropriate given the substantially higher frequency of clear cell carcinoma in Asian-Pacific Islanders; however, it is not clear that endometriosis is similarly increased in this cohort.2 The higher prevalence of clear cell cancer in this population could significantly attenuate the association. Fourth, the one truly novel finding of the study, the association of endometriosis to LGSOC, was based on reclassification of the WHO grading criteria and not on pathological review. A recent study has found that this methodology misclassifies about 5% of cases, predominately WHO grade 2 (HGSOC) patients being more appropriately identified with LGSOC.3 Since the frequency of each WHO grade was not given and pathology not reviewed for consistency, it is difficult to know how to interpret this association.
Nevertheless, the recent discovery of ARID1A mutations in clear cell and endometrioid cancers and the association of these tumors to endometriosis provide sufficient rationale to investigate the underlying biology. In addition, microarray interrogation of HGSOC has identified an inflammatory-like signature for some cases.4 It would be of interest to know if these cases too had an association with endometriosis, which is well recognized to induce stromal alterations and an inflammatory response.
1. Gates MA, et al. Risk factors for epithelial ovarian cancer by histologic subtype. Am J Epidemiol 2010;171:45-53.
2. Haruta S, et al. Molecular genetics and epidemiology of epithelial ovarian cancer (Review). Oncol Rep 2011;26:1347-1356.
3. Bodurka DC, et al. Reclassification of serous ovarian carcinoma by a 2-tier system: A Gynecologic Oncology Group Study. Cancer 2012;118:3087-3094.
4. Tothill RW, et al. Novel molecular subtypes of serous and endometrioid ovarian cancer linked to clinical outcome. Clin Cancer Res 2008;14:5198-5208.