An analysis of clinical characteristics and prognosis of endometrioid ovarian cancer based on the SEER database and two centers in China

Endometrioid ovarian cancer (EOVC) accounts for approximately 10% ~ 13% of epithelial ovarian cancer (EOC) and has clinical and biological difference compared with other pathological subtypes [1‐3]. The incidence of EOVC in non-Hispanic, Hispanic and Asian populations is about 8%, 9.6% and 11%, respectively [4]. Domestic and foreign research generally believed that endometriosis, genes mutation, imbalance of female reproductive tract microenvironment, delayed menopause and menopausal hormone replacement, obesity and other factors can increase the risk of EOVC [5‐7]. EOVC is one of endometriosis-associated ovarian cancer, the risk of which may increase 2.32 times if a patient has endometriosis [8]. Despite that, origin and pathogenesis of EOVC have not been revealed in detail, we still need lots of basic and clinical studies to explore.

Most patients with EOVC are diagnosed at an early stage, and the age of onset tends to be younger, with a better prognosis than that of high-grade serous cancer (HGSOC) and clear cell carcinoma (OCCC) [9]. CA125 and HE4 are tumor markers commonly used in clinical diagnosis and identification of EOC, but they all have a certain false positive rate. Moreover, in patients with EOVC, CA125 elevation is not as significant as in HGSOC, and about a quarter of patients had no abnormalities in CA 125 levels [10]. Study showed that HE4 has high sensitivity and specificity in diagnosis for EOC, with great potential in early patient diagnosis [11], but HE4 whether EOVC diagnostic and therapeutic efficacy monitoring is more advantageous still lack corresponding clinical evidence.

The Surveillance, Epidemiology, and End Results (SEER) database is currently the largest publicly available cancer statistics database that includes approximately 30% of the United States population [12]. In this study, we hope to evaluate clinical characteristics and explore risk factors affecting the prognosis of EOVC patients registered in the SEER database and two centers in China in order to provide some reference for EOVC diagnosis, decision making and prognosis assessment.

EOVC, as a special subtype of EOC, has significant geographical and ethnic differences in incidence. Higher rates of EOVC in Asian women have been documented in the United States and Eastern Asia, although reasons remain unknown [3, 9, 14]. Previously, two single-center studies from China had confirmed that patients with EOVC were significantly younger compared to other histological types [15, 16]. In our study, we found that the average age of EOVC patients in two centers was approximately 8 years younger than that of patients in SEER database (47.14 years vs 55.77 years). However, due to the diversity of races in SEER database, we further analyzed and found that the average age of EOVC of Asian/Pacific Islander was younger than that of whites and blacks. In terms of age stratification, Asian/Pacific Islander had the largest number of cases in the 41 ~ 50 age group, which may partly reflect that EOVC may occur at a younger age in Asians, including Chinese people. Meanwhile, the results from SEER database also found that age is an independent risk factor affecting the prognosis, but same conclusion was not drawn according to 87 EOVC patients from China. This may be related to the small samples we obtained, followed by differences in morbidity characteristics because of different populations, regions and health care measures, therefore, multicentric clinical studies are needed to analyze the distribution characteristics and incidence of EOVC patients in China in the future.

Endometriosis is thought to play a role in the pathogenesis of ovarian cancer, it’s estimated that patients with endometriosis have a fourfold and 2.32-fold increased risk of ovarian cancer and EOVC, respectively [8, 17]. In our study, we found that endometriosis occurred in approximately 27.6% of patients and was not an independent prognostic factor, which is in line with published articles [18, 19]. However, two large cohort studies based on the PALGA database and the Netherlands Cancer Registry revealed that a significantly higher incidence of EOVC was observed in women with endometriosis and a better survival presented in women with endometriosis who had stage I ~ II EOVC [20, 21]. In a study by Giovanna et al. [22], unilateral ovarian involvement was more frequently found in OCCC with endometriosis, but similar results in EOVC have not been reported. In our cohort, we observed that unilateral ovarian involvement was more common in EOVC with endometriosis than without (87.5% vs 57.1%, P = 0.044). If the role of endometriosis in the pathogenesis of EOVC is further elucidated, a proper risk model for patients with endometriosis can be developed to prevent the occurrence of EOVC.

Surgical combined with chemotherapy is the most-common strategy used to treatment EOVC. The NCCN guidelines recommend patients with high grade should undergo adjuvant chemotherapy after surgery regardless of FIGO staging. A study assessing the effect of surgical staging and adjuvant chemotherapy on survival in stage I, low grade EOVC patients found that patients with G2 had a significantly high recurrence rate, but adjuvant chemotherapy and staging lymphadenectomy didn’t improve survival [23]. In contrast, another retrospective study by Dimitrios et al. showed that patients with G2 stage I tumors could benefit from adjuvant chemotherapy, which isn’t recommended for G1 tumor [24]. On the other hand, research from Oseledchyk et al. found that the benefit of adjuvant chemotherapy is limited to patients with stage IC and G3 EOVC [25]. Moreover, a multicenter study from Europe demonstrated disease grade wasn’t of prognostic significance if restricted to early-stage disease [26]. The frequency of G1 ~ 2 and G3 in patients with EOVC was 84% ~ 97% and 3% ~ 16%, respectively [19, 27, 28]. Although the proportion of poorly differentiated patient is low, some of these are more likely to develop chemotherapy resistance when recurrence occurred, then both treatment and prognosis face great challenges. In our two centers, a total of 12 EOVC patients experienced disease progression or recurrence during follow-up and there were seven cases with advanced stage, eventually 10 patients died within 3 years of relapse.

LN are a common site of metastasis in patients with EOC and the therapeutic value of lymphadenectomy during debulking surgery is still under debate [29]. As latest NCCN guidelines described, systematic lymph dissection for early-stage patients can confer clinical benefits in staging and follow-up treatment, the removal of all visible lesions for advanced-stage patients is recommendation in surgery, including enlarged or suspected LN. According to previous research data, the rate of LN metastasis was about 6.1% ~ 29.6% in patients with FIGO stage I ~ II ovarian cancer, and more than 50% in advanced patients. The probability of LN metastasis occurring in early-stage EOVC patients is about 2.1% ~ 6.5% [30, 31], but there are various views as to whether LN resection can provide survival benefits. A Chinese single-center retrospective study showed that LN resection was an independent protective factor for recurrence after operation in patients with FIGO stage I EOVC [15]. Meanwhile, another large-scale clinical study from US also confirmed that lymphadenectomy is associated with favorable survival [31]. However, a study from Italy didn’t reach the same conclusion [32]. In the study, the rates of LN metastasis were 12.6% and 5.7% in two Chinese centers and SEER database and, moreover, patients with FIGO stage III ~ IV EOVC are more often lymph-metastasized. Notably, LN metastases in the SEER database were related to poor prognosis. However, analysis for two Chinese centers didn’t reach same results (P = 0.887).

We found 12 patients diagnosed with endometrioid carcinoma of the ovary and uterus. However, these patients were excluded from our study because of the uncertainty about tumor origin. Synchronous diagnosis of endometrioid ovarian carcinoma and endometrioid endometrial carcinoma (SEO-EEC) is largely documented. Patients with SEO-EEC can be classified into three groups: endometrial cancer with metastasis to the ovary, ovarian cancer with metastasis to the endometrium, or synchronous primary cancers [33, 34]. Women with synchronous primary cancers have better survival rates than those with single cancers with metastases [35, 36]. The accurate distinction between metastases and independent primary carcinoma mainly depends on pathologic features [37], such as stage, whether there is fallopian tube and myometrial invasion, unilateral or bilateral ovarian involvement, presence of atypical endometrial hyperplasia, presence of ovarian endometriosis, and follow-up status of patients, which is important. Moreover, molecular profiling is also helpful in the evaluation of primary or metastatic tumors, and can identify Lynch syndrome. Unfortunately, due to the limitations of basic healthcare insurance in China, molecular profiling wasn’t done as most patients were not willing to pay out-of-pocket for, thus it was difficult to identify metastasis from independent primary carcinoma. An accurate distinction is of great importance for prognostic and therapeutic assessment.

The study has a few limitations that should be noted. Firstly, the satisfaction of surgery and the sensitivity of adjuvant chemotherapy after surgery affect the overall therapeutic effect of EOVC patients to a large extent. Regrettably, the SEER database didn’t contain detailed postoperative chemotherapy information. The nomogram model constructed based on SEER database may have certain deviation in prognostic prediction on account of the lack of important adjuvant chemotherapy indicators, which can be further improved in subsequent prospective multi-center studies and include more effective indicators in order to build a prognostic prediction model for Chinese patients with EOVC. Secondly, the SEER database provided no information on BMI, HE4, CA19-9, cancer recurrence, and history of endometriosis, it wasn’t possible to compare more detailed differences in clinical characteristics between the two cohorts. Thirdly, the tolerance and severity of adverse reactions of patients in our centers to chemotherapy will affect the cycle and dose of chemotherapy and the therapeutic effect. Due to the limitation of the retrospective study, we did not include more factors affecting chemotherapy for prognostic evaluation, which may cause some offset in the results of the final analysis. We expect to conduct more clinical studies in subsequent studies on patients with ovarian cancer who are not sensitive to advanced chemotherapy. In addition, BMI was an independent prognostic factor in the analysis of data from the two centers in China. Owing to the nature of our study, we did not have information on the hip circumference or the waist-to-hip ratio, which might to be better predictors of all-cause mortality. CA125 and HE4 are important indictors for diagnosis and prognosis evaluation of EOC, it was found that different from HGSOC, HE4 may have greater predictive value for EOVC. Of course, this conclusion was only based on small sample size data, which lacks representation of the whole population, and needs to be verified by multi-center and large-sample clinical studies.

In conclusion, most EOVCs were diagnosed at an early stage, low grade and had better prognosis. The age of onset for EOVC may be younger in Asian/Pacific Islander and Chinese. The elder, high grade, advanced FIGO stage, only unilateral salpingo-oophorectomy, and BMI < 19.34 kg/m² may be unfavorable risk factors for prognosis. In addition, HE4 and CA125, as prognostic indicator, has certain guiding significance. A prognostic nomogram was developed and validated to assist clinicians in evaluating prognosis of EOVC patients. In the future, more studies are needed to further explore the molecular pathogenesis and treatment strategies of EOVC, a special pathological type, so as to provide new directions and ideas for early screening and early diagnosis of EOVC and improvement of advanced poor prognosis.