Background
Several studies have suggested that endometriosis is associated with an increased risk of ovarian cancer, especially for the clear-cell and endometrioid subtypes. However, previous studies lack sufficient power or diagnostic certainty.
Objective
The objective of the study was to assess the association between histologically proven endometriosis and ovarian cancer in a large population-based cohort study.
Study Design
We identified 131,450 women with a histological diagnosis of endometriosis between 1990 and 2015 from the Dutch nationwide registry of histopathology and cytopathology (PALGA). For the control cohort 132,654 women with a benign dermal nevus were matched on age and inclusion year with the endometriosis cases. Histological diagnoses of ovarian, fallopian tubes, and peritoneal cancers between January 1990 and July 2017 were retrieved. Incidence rate ratios were estimated for ovarian cancer and its subtypes for the whole follow-up period as well as for women with more than 1 person-year at risk.
Results
We found a crude incidence rate ratio of 4.79 (95% confidence interval, 4.33–5.31) and an age-adjusted incidence rate ratio of 7.18 (95% confidence interval, 6.17–8.36) for ovarian cancer overall. Endometrioid and clear-cell ovarian cancer had the highest age-adjusted incidence rate ratio of 29.06 (95% confidence interval, 20.66–40.87) and 21.34 (95% confidence interval, 14.01–32.51), respectively. Median age at ovarian cancer diagnosis was 56 years (interquartile range, 49–63) for the endometriosis cohort and 60 years (interquartile range, 53–67) for the nevus cohort, ( P < .05). After excluding women with less than 1 person-year at risk following an endometriosis diagnosis, we found a crude incidence rate ratio of 1.04 (95% confidence interval, 0.91–1.19) and an age-adjusted incidence rate ratio of 1.08 (95% confidence interval, 0.87–1.35) for ovarian cancer overall. However, statistically significant age-adjusted incidence rate ratios of 2.29 (95% confidence interval, 1.24–4.20) for clear-cell ovarian cancer and 2.56 (95% confidence interval, 1.47–4.47) for endometrioid ovarian cancer were found.
Conclusion
A significantly higher incidence of clear-cell and endometrioid ovarian cancer was found in women with histologically proven endometriosis. Additionally, we found an increased incidence of all ovarian cancer subtypes in histologically proven endometriosis; however, in many of these women, endometriosis and ovarian cancer were diagnosed synchronously after the average menopausal age, which may suggest that the risk of ovarian cancer in endometriosis patients remains, even when clinical endometriosis symptoms are no longer present.
Ovarian cancer has the highest mortality rate of all gynecological cancers, with an estimated 295,400 new cases of ovarian cancer and 184,800 cancer deaths in 2018. This makes ovarian cancer the ninth most common type of cancer in women worldwide and the eighth most common type regarding cancer mortality in women.
What is the purpose of this study?
To compare the incidence of ovarian cancer subtypes between 131,450 women with histologically proven endometriosis and women with a benign dermal nevus.
Key findings
A significantly higher incidence of clear-cell and endometrioid ovarian cancer was found in women with histologically proven endometriosis. Additionally, we found an increased incidence of all ovarian cancer subtypes in histologically proven endometriosis; however, in many of these women, endometriosis and ovarian cancer were diagnosed synchronously after the average menopausal age.
What does this add to what is known?
This is the largest study so far and, to our knowledge, this is the first large epidemiological study to report on the high frequency of synchronously diagnosed endometriosis and ovarian cancer after the average menopausal age. This could suggest that the risk of ovarian cancer in endometriosis patients remains, even when clinical symptoms of endometriosis are no longer present.
Ovarian cancer often presents in an advanced stage because of the aspecific symptoms in early stage disease. The 5 year survival rate for all stages of ovarian cancer together is less than 50%. The mortality of this disease has not improved much over the last decades, and early diagnosis by screening of high-risk groups, such as women with germline mutations, has failed to improve prognosis. , To improve the prognosis, efforts should be directed toward the prevention of ovarian cancer. Identification of high-risk groups is a first step in this process.
Several studies have shown that endometriosis is associated with an increased risk of ovarian cancer, but not all subtypes of ovarian cancer have been associated with this increased risk. Endometriosis patients are more frequently diagnosed with clear-cell and endometrioid subtypes and less often with serous or mucinous ovarian cancer.
Endometriosis is a common disease with an estimated 176 million women worldwide being affected. The disease is characterized by endometrial-like tissue outside the endometrium and is commonly associated with pelvic pain and infertility. A contributing problem in previous studies is the definition of endometriosis diagnosis, with large variation between the definitions used in different studies, varying from self-reported endometriosis to histological proven endometriosis. However, laparoscopy with histological confirmation is still considered the gold standard of endometriosis diagnosis.
Given the lack of diagnostic certainty of endometriosis diagnosis in previous studies on the association between endometriosis and ovarian cancer, as well as the lack of sufficient power, larger studies are needed with inclusion of histological proven endometriosis patients. The aim of this study was to determine whether there is an association between histological confirmed endometriosis and ovarian cancer and whether this risk concerns specific histological subgroups.
Material and Methods
Study population and design
All women with a histological diagnosis of endometriosis between Jan. 1, 1990, and Dec. 31, 2015, were identified from the Dutch nationwide registry of histopathology and cytopathology (PALGA, Houten, The Netherlands). Women registered in the database with a benign dermal nevus but who never had histological diagnosed endometriosis were used as the control cohort.
We aimed to select a control cohort with an equal number of women of the same age. Because we used a pathology database, we had to select a group with a tissue examination registered in this database. We choose a benign dermal nevus because this is a common diagnosis in women of all ages. A benign dermal nevus is a melanocytic tumor that can either be congenital or acquired. Women with both endometriosis and a benign dermal nevus were included in the endometriosis cohort. From the total cohort of women with a benign dermal nevus, random-frequency, 1-to1 matching was performed with the endometriosis cohort by age and year of diagnosis.
Histology reports for cancers of the ovaries, fallopian tubes, and peritoneum between January 1990 and July 2017 were retrieved and merged as ovarian cancer cases. Borderline tumors were not included in this study. All ovarian cancers that were found were classified according to the ovarian cancer subtype. All questionable cancer diagnoses were discussed with a second reviewer, and consensus was reached.
Women were censored after ovarian cancer diagnosis, as well as after a bilateral salpingo-oophorectomy (BSO), or autopsy. Other than an autopsy report, there were no data available on possible death. Women with a BSO or ovarian cancer diagnosed more than half a year before the endometriosis or benign dermal nevus diagnosis were considered not eligible for this study.
Because we used a pathology-based database, there were no data available on the type of surgical procedure, the endometriosis stage, or other possible risk factors for ovarian cancer.
Statistical analysis
First, we analyzed the full follow-up period. Second, we excluded all women with less than 1 person-year at risk (synchronous group) to account for detection bias. The remaining cases were classified as the metachronous group. The synchronous group was defined as ovarian cancer, BSO, or autopsy within a half a year before or 1 year after the histological diagnosis of endometriosis or nevus.
We determined the number of ovarian cancers for both the endometriosis and nevus cohort and calculated the person-years at risk based on censoring as described above. We calculated incidence rates (IR) per 100,000 person-years for the endometriosis and nevus cohort by using Poisson regression by including person-years as an offset in the model. Crude and age-adjusted incidence rate ratios (IRRs) were calculated, comparing women with an endometriosis diagnosis with the matched controls.
Age adjustment was done by using the age at endometriosis or nevus diagnosis and was adjusted to the distribution of the Dutch female population in the median inclusion year of this study to make the results easier to compare with the general population. A χ 2 test was used to compare the endometriosis cohort and the nevus cohort per ovarian cancer subtype divided into serous, mucinous, clear-cell, endometrioid type, and adenocarcinoma not otherwise specified. All statistical analyses were performed with STATA version 15.1 (StataCorp LLC, College Station, TX).
Ethical approval
This study is exempt from medical ethical approval because there is no physical involvement of the women in the study, and data were not traceable to individual participants. The study was approved by the scientific committee of PALGA (lzv2017-104).
Results
Cohort characteristics
We identified 131,902 women with histologically proven endometriosis between 1990 and 2015 and 547,924 women with a histologically proven benign dermal nevus. Frequency matching resulted in a total of 133,398 women in the nevus cohort. There were 452 women in the endometriosis cohort and 744 women in the nevus cohort with a censoring date more than half a year before the diagnosis of endometriosis or nevus. This resulted in 131,450 and 132,654 women, respectively, in the endometriosis and nevus cohort ( Figure ), who were included for further analysis.
The average number of histological endometriosis diagnoses per year was 5131. The median age at endometriosis or nevus diagnosis was 45 years, with an interquartile range (IQR) of 38–50. Patient characteristics at inclusion are listed in the Supplemental Table . Median follow-up for the endometriosis cohort was 14 years (range 0–27 years) and 16 years (range 0–27 years) for the nevus cohort, resulting in a total of 1,836,582 and 2,029,538 person-years, respectively.
In the endometriosis cohort, 12,940 women had person-years between minus a half year and plus 1 year (synchronous group); 11,305 of these women had a BSO without ovarian cancer, five had an autopsy, and the remaining 1630 women had ovarian cancer. This left 118,510 women in the endometriosis cohort with more than a year of follow-up (metachronous group) for the secondary analysis.
In the nevus cohort, 174 women had a censoring date synchronous to the nevus date, of which 142 women had BSO without ovarian cancer and 32 women had ovarian cancer. This left 132,480 women in the metachronous group for the nevus cohort ( Figure 1 ).
Ovarian cancer risk
As described in the Materials and Methods section, we first analyzed the full follow-up period. Second, we analyzed the women with at least 1 person-year at risk to account for detection bias (metachronous group).
Analysis of the full follow-up period
We detected a total of 2043 women with ovarian cancer in the endometriosis cohort and a total of 471 women with ovarian cancer in the nevus cohort ( Table 1 ). Median age at ovarian cancer diagnosis was 56 years (IQR, 49–63) and 60 years (IQR, 53–67), in the endometriosis cohort and nevus cohort, respectively ( P < .05) ( Table 2 ).
| Variables | Endometriosis | Nevus | ||||
|---|---|---|---|---|---|---|
| Observed number | Percentage of cohort total, % | Person-years | Observed number | Percentage of cohort total, % | Person-years | |
| Total group | 2043 | 1.6 | 1,836,582 | 471 | 0.4 | 2,029,538 |
| Metachronous group a | 413 | 0.3 | 1,836,305 | 439 | 0.3 | 2,029,483 |
a Metachronous group is defined as endometriosis or nevus diagnosis at least a year before censoring date (autopsy, bilateral salpingo-oophorectomy, or ovarian cancer).
| Variables | Endometriosis | Nevus | P value |
|---|---|---|---|
| Total group | 56 (49–63) | 60 (53–67) | < .05 |
| Metachronous group a | 58 (51–65) | 60 (54–67) | < .05 |
a Metachronous group is defined as endometriosis or nevus diagnosis at least a year before censoring date (autopsy, bilateral salpingo-oophorectomy, or ovarian cancer).
Table 3 shows the observed number of ovarian cancers, estimated incidence rate per 100,000 person-years, crude incidence rate ratios, and age-adjusted incidence rate ratios of ovarian cancer of women with endometriosis, compared with women with a benign dermal nevus, per ovarian cancer subtype and overall.
| Variables | Total group | Metachronous group a | ||||||
|---|---|---|---|---|---|---|---|---|
| Observed number | Incidence rate per 100,000 person-years (95% CI) | Crude incidence rate ratio (95% CI) | Age-adjusted incidence rate ratio (95% CI) | Observed number | Incidence rate per 100,000 person-years (95% CI) | Crude incidence rate ratio (95% CI) | Age-adjusted incidence rate ratio (95% CI) | |
| Clear-cell | ||||||||
| Endometriosis | 346 | 18.37 (16.96–20.93) | 11.25 (7.89–16.50) | 21.34 (14.01–32.51) | 54 | 2.94 (2.25–3.84) | 1.76 (1.12–2.78) | 2.29 (1.24–4.20) |
| Nevus | 34 | 1.67 (1.20–2.34) | 34 | 1.68 (1.20–2.34) | ||||
| Endometrioid | ||||||||
| Endometriosis | 545 | 29.68 (27.29–32.27) | 12.55 (9.33–17.23) | 29.06 (20.66–40.87) | 75 | 4.08 (3.26–5.12) | 2.02 (1.36–3.03) | 2.56 (1.47–4.47) |
| Nevus | 48 | 2.37 (1.78–3.14) | 41 | 2.02 (1.49–2.74) | ||||
| Serous | ||||||||
| Endometriosis | 702 | 38.22 (35.50–41.16) | 2.80 (2.43–3.23) | 4.19 (3.41–5.15) | 198 | 10.78 (9.38–12.39) | 0.82 (0.68–0.98) | 0.78 (0.59–1.04) |
| Nevus | 277 | 13.65 (12.13–15.35) | 268 | 13.21 (11.72–14.88) | ||||
| Mucinous | ||||||||
| Endometriosis | 239 | 13.01 (11.46–14.77) | 5.62 (4.10–7.86) | 5.87 (3.58–9.61) | 30 | 1.63 (1.14–2.34) | 0.85 (0.51–1.40) | 0.70 (0.31–1.60) |
| Nevus | 47 | 2.32 (1.74–3.08) | 39 | 1.92 (1.40–2.63) | ||||
| Adenocarcinoma NOS | ||||||||
| Endometriosis | 211 | 11.49 (10.04–13.15) | 3.59 (2.70–4.81) | 5.87 (3.89–8.86) | 56 | 3.05 (2.35–3.96) | 1.09 (0.74–1.60) | 1.77 (1.04–3.01) |
| Nevus | 65 | 3.20 (2.51–4.08) | 57 | 2.81 (2.17–3.64) | ||||
| All ovarian cancers | ||||||||
| Endometriosis | 2043 | 111.24 (106.52–116.17) | 4.79 (4.33–5.31) | 7.18 (6.17–8.36) | 413 | 22.49 (20.42–24.77) | 1.04 (0.91–1.19) | 1.08 (0.87–1.35) |
| Nevus | 471 | 23.21 (21.20–25.40) | 439 | 21.63 (19.70–23.75) | ||||
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
