Endometrial Hyperplasia in Epithelial Ovarian Cancer
Endometrial Hyperplasia in Epithelial Ovarian Cancer
In this study, an endometrial (pre)malignancy was identified in one-third of patients diagnosed with EOC. Atypical hyperplasia was found in half of endometrioid ovarian carcinomas and in 7% to 33% of other histology types. Endometrial intraepithelial carcinoma was found limited to serous EOC. Synchronous endometrial carcinomas were identified in patients with endometrioid and serous ovarian cancer in 11% and 3%, respectively.
The prevalence of synchronous endometrial carcinomas is comparable to other studies reporting on endometrial carcinomas in patients with EOC. Synchronous atypical hyperplasia was common in patients with EOC, and similar to the prevalence of synchronous endometrial carcinoma, atypical hyperplasia had the highest prevalence in patients with endometrioid EOC. Atypical hyperplasia is known as a premalignancy of endometrioid endometrial carcinoma. It is thought that endometrial atypical hyperplasia progresses to endometrial carcinoma in 23% of patients, and if so, progression occurs over a mean duration of 4 years.
The development of atypical hyperplasia is thought to be influenced by unopposed estrogens. In the present study, BMI was the only clinical variable that was significantly higher in women diagnosed with endometrial atypical hyperplasia than in those diagnosed with a normal endometrium. Other variables, including neoadjuvant chemotherapy, did not differ between both groups of women.
Atypical hyperplasia could represent an adverse effect of increased estrogen production by the EOC. Normal ovarian surface epithelial cells and/or ovarian tumors are capable of the production of steroids, and occasionally EOC has been reported to produce high estrogen levels. However, this is a retrospective study, and we were not informed of estrogen levels of the included cases. Furthermore, most reported estrogen-producing ovarian tumors are of endometrioid histology, and estrogen production is only very rarely described for the more common serous EOC. Still, in the current study, atypical hyperplasia was also identified in one-fourth of patients with serous EOC.
All but two (4%) patients with EOC diagnosed with concurrent atypical hyperplasia did not report signs of abnormal vaginal bleeding at presentation. Eleven patients diagnosed with atypical hyperplasia were premenopausal, and none showed specific signs of endometrial disease at presentation.
Preoperative ultrasound on endometrial thickness could not always be retrieved from the medical records of included cases. If an ovarian carcinoma was suspected, less focus was given to describing the endometrial thickness. Furthermore, neoadjuvant chemotherapy often had started before preoperative ultrasound was performed at the RUNMC. Other limitations of the current study are that clinical variables could not always be retrieved from the medical records, and no follow-up was available on the endometrial pathology of patients without a hysterectomy. Genetic counseling for a BRCA mutation or Lynch syndrome was offered to women with a family history at risk of hereditary cancer syndrome. However, results were not always available in the medical record, or genetic counseling was not or had not been performed because of refusal by the patient or rapid progression of the disease. Furthermore, the cohorts of nonserous EOC were relatively small, and the endometrium was embedded according to routine protocol. Therefore, only a part of the endometrium was available for histologic assessment.
Endometrial intraepithelial carcinomas were limited to the serous subtype of EOC and identified in 5% of this group. It is known as a precursor of serous endometrial carcinoma and defined as noninvasive serous endometrial carcinoma. Recently, the endometrium was also suggested as a potential source of precursor lesions for a limited proportion of serous EOC. Serous tubal intraepithelial carcinoma (STIC) is thought to be an intraepithelial carcinoma of serous ovarian carcinoma and can be found in the fallopian tubes of 19% to 61% of women diagnosed with serous EOC. STIC has also been suggested as a putative precursor of tubal and primary peritoneal serous carcinoma and rarely of serous endometrial carcinoma
WT1 staining is thought to be able to discriminate between primary serous EOC and serous endometrial carcinoma. Half of the analyzed cases showed strong expression for WT1 in both the serous EOC and concurrent EIC, which is suggestive of a primary serous EOC. One patient showed negative expression in both lesions, suggestive of primary endometrial carcinoma, although in two patients, the WT1 staining was suggestive of a primary serous EOC as well as a primary endometrial carcinoma. In the latter, the EIC showed no expression of WT1, whereas the concurrent serous EOC showed strong and diffuse WT1 staining.
Immunohistochemical staining showed comparable expression patterns for EIC and concurrent serous EOC. The expression pattern of atypical endometrial hyperplasia was more distinct. It showed strong exhibition of ER and PR, as well as negative to moderate Ki-67 and p53 staining. This expression pattern of atypical hyperplasia is comparable to the expression pattern known for endometrioid endometrial carcinoma. Therefore, atypical hyperplasia and concurrent serous EOC are most likely two independent lesions.
Prevalence of endometrial pathology seemed significantly higher concurrent to endometrioid EOC in comparison with other histologic subtypes. To confirm this finding in a larger cohort, we selected patients with endometrioid EOC operated on at the Canisius Wilhelmina Hospital from 1996 through 2010 (unpublished data). Twenty-two women diagnosed with endometrioid EOC could be identified, and sections were retrieved from the pathology archives to review the endometrium and confirm tumor histology. In six (27%) patients, a benign endometrium was found, whereas 11 (50%) had atypical hyperplasia and five (23%) had endometrioid endometrial carcinoma. The high prevalence of atypical hyperplasia concurrent to endometrioid EOC was confirmed in these additionally selected cases and did not differ from findings in the initial cohort (nine [47%] of 19 vs 11 [50%] of 22).
In patients diagnosed with EOC, the presence of a concurrent endometrial premalignancy has to be taken into account. However, the clinical importance of these often asymptomatic endometrial lesions has not yet been clarified. Most likely, concurrent atypical hyperplasia represents a premalignant stage of synchronous endometrioid endometrial carcinoma. Lynch syndrome in families with hereditary nonpolyposis colorectal cancer could explain the frequent presence of an endometrial carcinoma as well as ovarian carcinoma. However, the chance that these women develop endometrial carcinoma or colorectal cancer as their first malignancy is much higher compared with ovarian carcinoma, with a 40% to 60% chance of developing endometrial carcinoma first.
Occasionally, conservative surgery is performed in women diagnosed with early stage EOC, sparing the uterus and contralateral ovary. Studies investigating recurrence risk in these women showed that a recurrence was mostly found in the contralateral ovary, and in only 0% to 3% of cases did carcinoma recur in the endometrium. However, all women included in these studies were aged 40 years or younger. Women investigated in the current study had a median age of 60 years. This is comparable to the median age in which EOC is diagnosed. Therefore, the cohorts included in studies analyzing conservative surgery are not comparable to the current study. However, we support the recommendation of performing endometrial sampling at the time of conservative surgery, since endometrial atypical hyperplasia was also found in patients with early stage disease. In the current study, eight patients with a FIGO stage I EOC were diagnosed with atypical hyperplasia, and three (38%) were premenopausal.
In conclusion, in 31% of women diagnosed with EOC, the endometrium showed a (pre)malignant endometrial lesion. In patients with endometrioid EOC, atypical hyperplasia occurred in half. Identified endometrial premalignancies are vital when considering treatment options for these women. If conservative surgery is considered with preservation of the uterus, attention should be given to the possibility of concurrent endometrial pathology.
Discussion
In this study, an endometrial (pre)malignancy was identified in one-third of patients diagnosed with EOC. Atypical hyperplasia was found in half of endometrioid ovarian carcinomas and in 7% to 33% of other histology types. Endometrial intraepithelial carcinoma was found limited to serous EOC. Synchronous endometrial carcinomas were identified in patients with endometrioid and serous ovarian cancer in 11% and 3%, respectively.
The prevalence of synchronous endometrial carcinomas is comparable to other studies reporting on endometrial carcinomas in patients with EOC. Synchronous atypical hyperplasia was common in patients with EOC, and similar to the prevalence of synchronous endometrial carcinoma, atypical hyperplasia had the highest prevalence in patients with endometrioid EOC. Atypical hyperplasia is known as a premalignancy of endometrioid endometrial carcinoma. It is thought that endometrial atypical hyperplasia progresses to endometrial carcinoma in 23% of patients, and if so, progression occurs over a mean duration of 4 years.
The development of atypical hyperplasia is thought to be influenced by unopposed estrogens. In the present study, BMI was the only clinical variable that was significantly higher in women diagnosed with endometrial atypical hyperplasia than in those diagnosed with a normal endometrium. Other variables, including neoadjuvant chemotherapy, did not differ between both groups of women.
Atypical hyperplasia could represent an adverse effect of increased estrogen production by the EOC. Normal ovarian surface epithelial cells and/or ovarian tumors are capable of the production of steroids, and occasionally EOC has been reported to produce high estrogen levels. However, this is a retrospective study, and we were not informed of estrogen levels of the included cases. Furthermore, most reported estrogen-producing ovarian tumors are of endometrioid histology, and estrogen production is only very rarely described for the more common serous EOC. Still, in the current study, atypical hyperplasia was also identified in one-fourth of patients with serous EOC.
All but two (4%) patients with EOC diagnosed with concurrent atypical hyperplasia did not report signs of abnormal vaginal bleeding at presentation. Eleven patients diagnosed with atypical hyperplasia were premenopausal, and none showed specific signs of endometrial disease at presentation.
Preoperative ultrasound on endometrial thickness could not always be retrieved from the medical records of included cases. If an ovarian carcinoma was suspected, less focus was given to describing the endometrial thickness. Furthermore, neoadjuvant chemotherapy often had started before preoperative ultrasound was performed at the RUNMC. Other limitations of the current study are that clinical variables could not always be retrieved from the medical records, and no follow-up was available on the endometrial pathology of patients without a hysterectomy. Genetic counseling for a BRCA mutation or Lynch syndrome was offered to women with a family history at risk of hereditary cancer syndrome. However, results were not always available in the medical record, or genetic counseling was not or had not been performed because of refusal by the patient or rapid progression of the disease. Furthermore, the cohorts of nonserous EOC were relatively small, and the endometrium was embedded according to routine protocol. Therefore, only a part of the endometrium was available for histologic assessment.
Endometrial intraepithelial carcinomas were limited to the serous subtype of EOC and identified in 5% of this group. It is known as a precursor of serous endometrial carcinoma and defined as noninvasive serous endometrial carcinoma. Recently, the endometrium was also suggested as a potential source of precursor lesions for a limited proportion of serous EOC. Serous tubal intraepithelial carcinoma (STIC) is thought to be an intraepithelial carcinoma of serous ovarian carcinoma and can be found in the fallopian tubes of 19% to 61% of women diagnosed with serous EOC. STIC has also been suggested as a putative precursor of tubal and primary peritoneal serous carcinoma and rarely of serous endometrial carcinoma
WT1 staining is thought to be able to discriminate between primary serous EOC and serous endometrial carcinoma. Half of the analyzed cases showed strong expression for WT1 in both the serous EOC and concurrent EIC, which is suggestive of a primary serous EOC. One patient showed negative expression in both lesions, suggestive of primary endometrial carcinoma, although in two patients, the WT1 staining was suggestive of a primary serous EOC as well as a primary endometrial carcinoma. In the latter, the EIC showed no expression of WT1, whereas the concurrent serous EOC showed strong and diffuse WT1 staining.
Immunohistochemical staining showed comparable expression patterns for EIC and concurrent serous EOC. The expression pattern of atypical endometrial hyperplasia was more distinct. It showed strong exhibition of ER and PR, as well as negative to moderate Ki-67 and p53 staining. This expression pattern of atypical hyperplasia is comparable to the expression pattern known for endometrioid endometrial carcinoma. Therefore, atypical hyperplasia and concurrent serous EOC are most likely two independent lesions.
Prevalence of endometrial pathology seemed significantly higher concurrent to endometrioid EOC in comparison with other histologic subtypes. To confirm this finding in a larger cohort, we selected patients with endometrioid EOC operated on at the Canisius Wilhelmina Hospital from 1996 through 2010 (unpublished data). Twenty-two women diagnosed with endometrioid EOC could be identified, and sections were retrieved from the pathology archives to review the endometrium and confirm tumor histology. In six (27%) patients, a benign endometrium was found, whereas 11 (50%) had atypical hyperplasia and five (23%) had endometrioid endometrial carcinoma. The high prevalence of atypical hyperplasia concurrent to endometrioid EOC was confirmed in these additionally selected cases and did not differ from findings in the initial cohort (nine [47%] of 19 vs 11 [50%] of 22).
In patients diagnosed with EOC, the presence of a concurrent endometrial premalignancy has to be taken into account. However, the clinical importance of these often asymptomatic endometrial lesions has not yet been clarified. Most likely, concurrent atypical hyperplasia represents a premalignant stage of synchronous endometrioid endometrial carcinoma. Lynch syndrome in families with hereditary nonpolyposis colorectal cancer could explain the frequent presence of an endometrial carcinoma as well as ovarian carcinoma. However, the chance that these women develop endometrial carcinoma or colorectal cancer as their first malignancy is much higher compared with ovarian carcinoma, with a 40% to 60% chance of developing endometrial carcinoma first.
Occasionally, conservative surgery is performed in women diagnosed with early stage EOC, sparing the uterus and contralateral ovary. Studies investigating recurrence risk in these women showed that a recurrence was mostly found in the contralateral ovary, and in only 0% to 3% of cases did carcinoma recur in the endometrium. However, all women included in these studies were aged 40 years or younger. Women investigated in the current study had a median age of 60 years. This is comparable to the median age in which EOC is diagnosed. Therefore, the cohorts included in studies analyzing conservative surgery are not comparable to the current study. However, we support the recommendation of performing endometrial sampling at the time of conservative surgery, since endometrial atypical hyperplasia was also found in patients with early stage disease. In the current study, eight patients with a FIGO stage I EOC were diagnosed with atypical hyperplasia, and three (38%) were premenopausal.
In conclusion, in 31% of women diagnosed with EOC, the endometrium showed a (pre)malignant endometrial lesion. In patients with endometrioid EOC, atypical hyperplasia occurred in half. Identified endometrial premalignancies are vital when considering treatment options for these women. If conservative surgery is considered with preservation of the uterus, attention should be given to the possibility of concurrent endometrial pathology.
