Clinical outcomes of patients with endometrioid epithelial ovarian cancer following surgical treatment

We identified a total of 265 patients with EEOC who were treated at MSKCC between January 2002 and September 2017. Upon review, 11 patients were diagnosed with borderline tumors on final pathology, 11 were diagnosed with advanced synchronous or metachronous malignancies (with the exception of synchronous endometrial cancer), 24 presented for treatment of recurrent disease, and 1 had no central MSKCC pathology review. Six patients presented to MSKCC for treatment more than 3 months after their initial surgery. The above-mentioned patients were excluded, resulting in a final cohort of 212 patients. The clinical characteristics of the 212 patients meeting the study inclusion criteria are listed in Table 1 . The median age at diagnosis was 52 years (range, 20–88 years). A majority were Caucasian (86.1%). Disease stage distribution was as follows: stage I, n=145 (68%); stage II, n=47 (22%); stage III/IV, n=20 (9%). FIGO grade disease distribution was as follows: grade 1, 127 patients (60%); grade 2, 66 patients (31%); grade 3, 17 patients (8%); unknown, 2 patients (1%). Endometriosis was seen on pathology in 128 patients (60%). Borderline tumors were seen in 34.9%, with a majority (77.8%) showing endometrioid histology. Seventy-five patients (35%) had a synchronous endometrioid endometrial cancer, of which 80% were stage IA. Sixty percent of patients received chemotherapy after surgery, and 27.6% received no further treatment. Radiation was administered with or without chemotherapy in 9.8% of cases. Four patients received hormonal therapy after surgery. All patients underwent primary surgery. Comprehensive surgical staging was performed in 101 (48%) patients ( Table 1 ). Eight of the 101 (5%) patients had positive pelvic lymph nodes (LNs) and 6 (4%) had positive paraaortic LNs. All 6 patients (4%) with positive para-aortic nodes had FIGO grade 2 or 3 histology. Two patients with grade 1 histology had positive pelvic LNs, and both had a concomitant endometrial cancer. Nearly all patients (n=176; 97%) had a documented complete gross resection (CGR) at time of surgery. Three (1.4%) and 2 (0.9%) patients had optimal and sub-optimal resections, respectively. Postoperative chemotherapy was administered in 123 (60%) patients. Fifty-six (28%) patients had no additional treatment. Three patients (1.4%) received radiation therapy alone, and 17 (8%) received radiation with chemotherapy. The median follow-up time was 63.9 months (range, 0.7–192 months). The 5-year PFS was 83% (95% CI: 76.6–87.8%) and the 5-year OS was 92.7% (95%CI: 87.7–95.8%). On univariate analysis, age, stage, and comprehensive surgical staging were associated with improved 5-year PFS, but only younger age at diagnosis was associated with improved 5-year OS (p<0.001) ( Table 2 ). On multivariate analysis, age, stage and comprehensive surgical staging were associated with improved 5-year PFS ( Table 3 ). Compared with observation alone, chemotherapy did not improve 5-year PFS for patients with stage IA or IB disease but did improve survival in patients with stage IC disease (HR 1.01; 95% CI 0.22–4.59; p=0.99; HR 0.17; 95% CI 0.04–0.7; p=0.006) ( Table 4 ; Figure 1 ).

This study was approved by the Institutional Review Board at Memorial Sloan Kettering Cancer Center (MSKCC). We identified all patients treated for EEOC at our institution between January 2002 and September 2017. We included patients who had their initial surgery at MSKCC or who presented to MSKCC within 3 months of their initial surgery. The medical records of all patients meeting the inclusion criteria were reviewed. Abstracted data included patient’s age, date of initial diagnosis, tumor histology and grade, residual tumor after all operations, any post-resection therapy, dates and sites of all recurrences, and disease status at last follow-up. Comprehensive surgical staging was defined as, at minimum, unilateral salpingo-oophorectomy, pelvic lymphadenectomy, para-aortic lymphadenectomy, peritoneal biopsies and collection of pelvic washings. The amount of residual disease was determined based on operative reports. All cases were re-reviewed by an expert gynecologic pathologist. Progression-free survival (PFS) was defined from the initial surgery date to recurrence/progression date/death date/last follow-up date, whichever came first. OS was calculated from the initial surgery date to death date/last follow-up date, whichever came first. P-values were obtained using the Log-Rank test for categorical variables and Cox Proportional-Hazards Models for continuous variables. Survival was estimated using the Kaplan-Meier method, and estimates were compared using the Log-Rank test. All statistical tests were performed using SPSS 25.0 software.

EEOCs comprise only 5–10% of all epithelial ovarian cancers and are thought to arise from endometriosis. These tumors commonly co-occur with synchronous endometrial cancer and generally have a better prognosis than other histologic subtypes. Because of their rarity, optimal intraoperative and postoperative management remains poorly defined. In this study we examined a large cohort of patients with EEOCs and found favorable outcomes, with age, stage, and comprehensive surgical staging associated with improved 5-year PFS. In our cohort, 60.3% of patients had findings of endometriosis on final pathology. In these patients, the presence of endometriosis was associated with a trend toward better PFS and OS—which, although clinically interesting, was not statistically significant. Mangili et al., reviewed 65 patients diagnosed with EEOC and found that 21/65 (32%) cases were associated with endometriosis [ 9 ]. In their cohort, patients with endometriosis were significantly younger and less likely to have high-grade tumors. However, no difference in PFS or OS was seen between those with and without endometriosis. The 5-year survival for patients with and without endometriosis was 66% and 60%, respectively. Though the findings are similar, the actual percent survival seen in our cohort of patients is considerably higher: 95.3% and 89.4% 5-year survival in patients with and without endometriosis, respectively. This difference may reflect the fact that there were more patients with stage III/IV disease in the Mangili study. The NCCN recommends observation for patients with surgically staged IA or IB, grade 1 endometrioid carcinomas [ 7 ]. In our cohort, chemotherapy did not improve 5-year PFS for patients with stage IA or IB disease (all grades) compared with observation alone. However, the cohort was too small for analysis of results based on stage and grade. Due to low event rates in progression, our data should be interpreted with caution. Nevertheless, it is reassuring that our findings are consistent with the analysis by Oseledchyk et al., which showed that chemotherapy improved OS in stage IC grade 3 tumors, but not in stage I grade 1 or 2 tumors [ 8 ]. Of the 60% of patients in our study who received chemotherapy, the majority (87%) received ≥4 cycles of platinum-based chemotherapy. Despite previous data showing that 3 cycles is sufficient, and despite data showing that 6 cycles increases toxicity without altering rates of recurrence in early-stage EOC, most patients received 6 cycles nevertheless [ 14 , 15 ]. If observation is considered for stage IA or IB grade 1 or 2 tumors, a surgical staging procedure is recommended in all such cases [ 7 ]. In our cohort, 47.6% of patients had comprehensive surgical staging, which was associated with improved PFS on multivariate analysis (HR 0.39, 0.2–0.76). In surgically staged patients, 8 (5%) had positive pelvic LNs and 6 (4%) had positive paraaortic LNs. Only patients with grade 2 or 3 EEOC had positive paraaortic nodes. Two patients with grade 1 EEOC had positive pelvic LNs, and both had a concomitant endometrial cancer. This highlights the fact that, in the absence of an endometrial pathology, low-grade EEOCs are unlikely to involve pelvic or paraaortic LNs. The current study does have limitations. EEOCs demonstrate genomic heterogenicity (as do endometrioid endometrial cancers). Unfortunately, due to the age of our cohort, testing for mismatch repair deficiency and Lynch syndrome was performed in only a small number of patients. Genomic heterogeneity may be the main driver of survival, rather than surgical staging or adjuvant chemotherapy [ 2 , 16 ]. As noted above, it has been proposed that endometrioid ovarian carcinomas that arise from endometriosis are molecularly distinct from those not associated with endometriosis [ 17 , 18 ]. Genes that aid in the regulation of autoimmune response and inflammation appear to be key in the etiology of both endometriosis and associated EEOCs, which could explain the differences in survival seen in other studies. Despite its limitations, however, our study adds to the literature on this rare malignancy. Additional studies are needed to explore the genomic basis of EEOC and survival, and to develop more tumor-specific treatment algorithms.

In summary, EEOC is a rare tumor associated with relatively favorable outcomes. In this study we found that age, stage, and comprehensive surgical staging impacted 5-year PFS. Chemotherapy improved OS in the setting of stage IC grade 3 tumors, but not in stage I/grade 1 or 2 tumors. Chemotherapy did not improve 5-year PFS for patients with stage IA or IB disease (of any grade) compared with observation alone. We found that, in the absence of an endometrial pathology, low-grade EEOCs were unlikely to involve pelvic or paraaortic LNs. Genes associated with the regulation of autoimmune and inflammatory responses appear to be key in the etiology of both endometriosis and associated EEOCs. Further studies exploring the relationship between the genomic landscape of EEOCs and survival will be required in order to develop treatment specific to this tumor type.

Endometrioid epithelial ovarian cancers (EEOCs) comprise 5–10% of all epithelial ovarian cancers (EOCs) and are thought to arise from atypical endometriosis. They commonly co-occur with a synchronous endometrial carcinoma. Paradoxically, patients with concurrent EEOCs and uterine cancers have better survival than those with ovarian tumors alone [ 1 ]. EEOCs are distinct from high-grade serous carcinomas (HGSOC), both clinically and molecularly; molecularly, in fact, EEOCs bear more similarity to endometrioid endometrial cancers than to HGSOCs [ 2 ]. Age is younger and stage lower at presentation in EEOCs compared with HGSOCs. EEOCs are more amenable to optimal debulking. Stage for stage, EEOCs tend to have a better prognosis than other EOC histologies [ 3 , 4 ]. Compared with HGSOCs, EEOCs less frequently harbor TP53 mutations but more frequently display KRAS , PIK3CA , PIK3R1 , PTEN and CTNNB1 mutations [ 5 , 6 ]. According to the National Comprehensive Cancer Network (NCCN), adjuvant chemotherapy is indicated for grade 1 EOCs stage ≥II, and grade 2–3 EOCs stage ≥IC [ 7 ]. A recent Surveillance, Epidemiology, and End Results (SEER) analysis of 3552 stage I EEOC patients showed no survival benefit to chemotherapy for patients with grade 1 or 2 tumors [ 8 ]. Only patients with stage IC, grade 3 disease had an improved overall survival (OS) (5-yr OS 81% vs. 62%). In this single-institution study, we evaluated characteristics, surgical management, adjuvant therapies, and oncologic outcomes in a cohort of patients diagnosed with primary EEOC over a 15-year period.