Analysis of prognostic factors in ovarian clear cell carcinoma: a retrospective cohort study

Ovarian cancer (OC) is one of the three major gynecological malignancies, characterized by high malignancy, insidious onset, and a higher fatality rate compared to other gynecological cancers. Epithelial ovarian cancer (EOC) accounts for approximately 90% of OC cases, with ovarian clear cell carcinoma (OCCC) comprising approximately 6% of all EOCs ( 1 , 2 ). OCCC presents distinct biological characteristics and prognosis compared to other OCs ( 3 ). The incidence of OCCC exhibits significant ethnic and geographical variation ( 4 ). In Asia, Japan has the highest incidence, with rates ranging from 15% to 25% ( 5 ). In China, OCCC accounts for 9.7% of all EOCs, according to a study by Takano et al. ( 6 ). The median onset age of patients with OCCC is 55 years, compared to 64 years for patients with ovarian serous carcinoma ( 7 ). Approximately 90% of OCCC cases are confined to a unilateral ovary, while only 19% to 22% involve bilateral ovarian involvement ( 8 , 9 ). Although most OCCC cases are diagnosed at an early stage, treatment typically involves standard staging surgery or tumor cytoreduction when feasible. Advanced-stage patients may undergo intermediate tumor cytoreduction following neoadjuvant chemotherapy. Most patients receive platinum-based chemotherapy regimens as first-line postoperative treatment. However, OCCC exhibits a high recurrence rate following chemotherapy ( 10 ). Several prognostic factors for OCCC have been identified, including demographic characteristics, disease stage, the presence of endometriosis (EM), molecular features, surgical methods, postoperative tumor residuals, adjuvant therapy, and venous thromboembolism (VTE). Previous studies have debated whether EM-associated OCCC is associated with a better prognosis due to earlier stage at diagnosis or distinct biological factors ( 11 ). This study aims to further explore the relationship between EM and OCCC prognosis. Additionally, while carbohydrate antigen 125 (CA-125) is a sensitive biomarker for OCCC, there is limited research on the dynamic changes in CA-125 levels and their correlation with prognosis. This study investigated the prognostic factors of OCCC, focusing on the association between EM, initial CA-125 levels before treatment, the final CA-125 levels post-treatment, and the rate of CA-125 normalization. We present this article in accordance with the STROBE reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1643/rc ).

This is a retrospective cohort study. After preliminary screening by the medical record department of Fujian Provincial Cancer Hospital, a total of 117 patients with OCCC diagnosed between January 2007 and December 2018 were identified. Excluding 10 patients who did not receive any treatment and 5 patients who were found to have recurrence at the time of their visit, with unavailable prior treatment history, 102 patients remained in the sample. Detailed clinical data were collected from medical records and clinical follow-up information, including age, menstrual status, presenting symptoms, CA-125 levels, International Federation of Gynecology and Obstetrics (FIGO) stage, tumor size and location, type of surgery, preoperative or intraoperative tumor rupture, ascites, stage, presence of EM (ovarian and/or extraovarian), treatment response, and recurrence. Inclusion criteria were as follows: (I) histologically confirmed pure OCCC; (II) surgery performed; (III) standardized and completed postoperative follow-up. Exclusion criteria were as follows: (I) patients with recurrence or metastasis; (II) patients who did not receive treatment. All included patients provided written informed consent for the use of their clinical data. The study was approved by the Ethics Committee of Fujian Cancer Hospital (No. K2023-045-01). All methods were conducted in compliance with relevant guidelines and regulations. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. Chemotherapy resistance was defined as definitive tumor progression or recurrence within six months of the last chemotherapy. Standard treatment was defined as undergoing OC tumor cell reduction surgery followed by at least three cycles of chemotherapy. Postoperative survival data were obtained through outpatient visits or telephone follow-ups, with the final follow-up period extending to March 2022. Patients without follow-up information for the preceding six months were classified as lost to follow-up. Survival analysis included overall survival (OS) and progression-free survival (PFS). OS was defined as the time from the first operation to death or the last follow-up. PFS was defined as the time from the initial surgery to the occurrence of tumor progression or recurrence. Statistical analysis was performed using the SPSS software (Version 25.0, SPSS Inc., Chicago, IL, USA). Multivariate Cox regression analysis was applied to assess clinical prognostic factors. A P value <0.05 was considered statistically significant. For the multivariate Cox regression model, the number of covariates was based on the “Application Conditions of the Cox Regression Model” (2024), where one covariate required 5 events per variable (EPV) to avoid overfitting. Risk factors with a P value <0.10 were selected for inclusion in the multivariate Cox regression analysis, and differences were considered statistically significant.

A total of 102 patients with OCCC were included in the study. Among these, 64 cases (62.7%) were diagnosed at stage I, 18 cases (17.6%) at stage II, 16 cases (15.7%) at stage III, and 4 cases (3.9%) at stage IV. The patients’ ages ranged from 31 to 72 years, with a mean age of 49.40±7.94 years, and a median age of 49 years. Symptoms at presentation included abdominal mass in 65 cases (63.7%), abdominal pain in 22 cases (21.5%), abdominal distension in 8 cases (7.8%), and postmenopausal vaginal bleeding in 2 cases (1.9%). One patient had concurrent endometrial carcinoma at stage Ia, and 5 patients (4.8%) presented with other symptoms, including menstrual disorders (2 cases), frequent micturition (2 cases), and dysuria (1 case). Among the 102 patients, 6 (6.0%) had additional malignancies, and 40 patients (39.2%) had EM. Of the 102 patients, 95 (93.1%) received first-line platinum-based chemotherapy with a mean of 5.38±2.51 cycles, while 7 (6.9%) did not receive chemotherapy following surgery. Chemotherapy resistance was observed in 34 patients (36.2%). Of the 43 patients who experienced recurrence, 38 (88.3%) patients died, accounting for 86.4% (38/44) of all deaths in the cohort. The OS time was 60.67±39.04 months, with a median survival time of 52.84 months. The disease-free progression survival time was 52.63±42.43 months, with a median of 45.10 months ( Table 1 ). CA-125, carbohydrate antigen 125; EP, etoposide plus platinum; TP, paclitaxel plus platinum. All 102 patients were followed for a minimum of 3 years, with a median follow-up time of 89.62 months (range, 40.1 to 182.7 months). The median OS was 52.84 months, and the median PFS was 45.10 months. The 3-, 5-, and 10-year OS rates were 71.6%, 59.2%, and 44.4%, respectively, while the 3-, 5-, and 10-year PFS rates were 56.9%, 50.0%, and 38.9%, respectively. Comparison of OS and PFS rates between early-stage (stages I–II) and advanced-stage (stages III–IV) patients is shown in Figure 1 . The median OS and PFS for early-stage patients were not reached, whereas the median OS and PFS for advanced-stage patients were 20.84 and 10.45 months, respectively, demonstrating a statistically significant difference (P<0.001) ( Figure 1 ). Kaplan-Meier survival curves for patients at different stages. (A) OS; (B) PFS. OS, overall survival; PFS, progression-free survival. Univariate analysis of the Cox proportional hazards regression model indicated that OS was associated with stage, concurrent EM, chemotherapy, initial lymph node metastasis, platinum chemoresistance, CA-125 normalization, final CA-125 ≤20 U/mL, and venous thrombosis. Among these factors, initial CA-125, final CA-125, and CA-125 normalization had a significant number of missing values, and therefore, these covariates were excluded from the multivariate Cox regression model. Multivariate analysis revealed that stage [hazard ratio (HR) =2.780, 95% confidence interval (CI): 1.265–6.110, P=0.01], chemotherapy (HR =0.211, 95% CI: 0.075–0.593, P=0.003), and platinum resistance (HR =8.233, 95% CI: 3.617–18.743, P<0.001) were independent risk factors for OS ( Table 2 ). Variables with substantial missing data (CA-125 normalization, initial and final CA-125) were excluded from multivariate analysis. CA-125, carbohydrate antigen 125; CI, confidence interval; EP, etoposide plus platinum; HR, hazard ratio; TP, paclitaxel plus platinum. Univariate analysis of the Cox proportional hazards regression model indicated that PFS was associated with stage, concurrent EM, chemotherapy, CA-125 normalization, and final CA-125 ≤20 U/mL. Multivariate analysis revealed that concurrent EM (HR =0.385, 95% CI: 0.189–0.784, P=0.009) and stage (HR =4.507, 95% CI: 2.346–8.660, P<0.001) were independent risk factors for PFS ( Table 3 ). Variables with substantial missing data (CA-125 normalization, initial and final CA-125) were excluded from multivariate analysis. CA-125, carbohydrate antigen 125; CI, confidence interval; EP, etoposide plus platinum; HR, hazard ratio; TP, paclitaxel plus platinum. Univariate Cox regression analysis revealed that stage, concurrent EM, chemotherapy, CA-125 normalization, and the last CA-125 ≤20 U/mL were associated with both OS and PFS (P<0.05). Lymph node metastasis, platinum resistance, and venous thrombosis were only associated with OS (P<0.05). Multivariate analysis identified stage (HR =2.780, 95% CI: 1.265–6.110, P=0.01), chemotherapy (HR =0.211, 95% CI: 0.075–0.593, P=0.003), and platinum resistance (HR =8.233, 95% CI: 3.617–18.743, P<0.001) as independent prognostic factors for OS. Concurrent EM (HR =0.385, 95% CI: 0.189–0.784, P=0.009) and stage (HR =4.507, 95% CI: 2.346–8.660, P<0.001) were independent prognostic factors for PFS. Survival curve comparisons between OS and PFS were performed using the log-rank test. To assess whether differences in OS and PFS existed among the three groups of patients with different CA-125 normalization timelines, the results indicated significant differences in both OS and PFS among the early normalized, late normalized, and non-normalized groups (P<0.001) ( Figure 2 ). Kaplan-Meier survival curves for patients with different CA-125 normalization timelines. (A) OS; (B) PFS. CA-125, carbohydrate antigen 125; OS, overall survival; PFS, progression-free survival. An analysis was conducted to compare the OS and PFS between the 102 patients who completed standard treatment and those who did not. The results showed a significant difference in OS between the two groups (P<0.001), while no significant difference in PFS was observed (P=0.15) ( Figure 3 ). Kaplan-Meier survival curves for patients who completed standardized treatment and those who did not. (A) OS; (B) PFS. OS, overall survival; PFS, progression-free survival. A comparison of OS between patients with and without platinum chemoresistance revealed a significant difference, with patients exhibiting platinum resistance showing worse OS (P<0.001) ( Figure 4 ). Kaplan-Meier survival curves of OS for patients with and without platinum chemoresistance. OS, overall survival.

Patients with stage I OCCC accounted for 62.7% of the total OCCC cases during the same period, slightly higher than the previously reported proportion. This is comparable to the 66% reported by Okamoto et al. ( 8 ). The CA-125 levels in clear cell carcinoma patients vary significantly before treatment, with more than half of the patients showing a notable increase, which aids in early tumor detection and diagnosis. However, some patients may have normal CA-125 levels before treatment and during recurrence or metastasis, which can lead to diagnostic challenges. While CA-125 is the most commonly used marker for gynecological malignancies, its diagnostic and prognostic value in patients with clear cell carcinoma remains debated. In this study, 10.87% of patients had normal CA-125 levels at diagnosis, and the CA-125 level at diagnosis was not associated with prognosis (P>0.05) ( 9 ). However, other studies have suggested that the timing of CA-125 normalization during chemotherapy correlates with prognosis. Specifically, patients whose CA-125 levels normalized earlier during treatment had better outcomes compared to those whose levels did not normalize ( 10 ). Furthermore, the final CA-125 value, when using a cut-off of 20 U/mL, was significantly associated with prognosis. Patients with a final CA-125 value ≤20 U/mL had longer PFS and OS. Thus, while serum CA-125 levels may not be helpful for diagnosing OCCC, they can be valuable for adjusting treatment plans and identifying high-risk patients prone to recurrence and metastasis during follow-up ( 12 ). Twelve studies identified biomarkers with sensitivity and specificity exceeding 80%. A biomarker panel consisting of IMP3, napsin A, and hepatocyte nuclear factor 1 beta holds potential as therapeutic targets ( 13 ). Most OCCC occurrences arise from the same precursor lesion as endometrioid carcinoma of the ovary, namely EM. In this study, three cases of OCCC complicated by endometrial carcinoma were associated with intraovarian EM or adenomyosis, further supporting the relationship between these conditions. Previous studies have reported that 50% to 74% of OCCCs are associated with EM, with ovarian EM being the most common type ( 14 ). In our study, the proportion of OCCC cases with EM was 39.2%, which is lower than the rates previously reported. Prognostic analysis showed that OCCCs with EM had a better prognosis, consistent with previous case studies. Moreover, the symptoms of EM may prompt earlier medical consultations, aiding in the early detection of OCCC ( 11 , 15 ). In our cohort, 74.3% of patients with OCCC and EM were diagnosed at stage I, which is higher than the 62.7% of stage I patients in the overall study cohort (39.2%). Previous studies have indicated that EM-associated OC has a lower incidence rate but a better early prognosis, although the survival rate in late stages is significantly lower compared to high-grade serous ovarian carcinoma ( 16 ). Clinical stage is a critical prognostic factor for OCCC (P<0.001), with stage II/III lymph node metastasis at initial diagnosis being most closely associated with prognosis (P=0.004) ( 17 ). The 3-, 5-, and 10-year OS rates for patients with OCCC were 81.7%, 68.9%, and 50.0% for stages I–II, respectively, and 30.0%, 20.0%, and 0.0% for stages III–IV. These results indicate that late-stage patients have a significantly lower survival rate compared to early-stage patients, highlighting the importance of early detection and treatment ( 18 ). Although multivariate analysis in this study showed that lymph node metastasis at initial diagnosis was not an independent prognostic factor for OS and PFS, this may be due to the limited number of cases. For example, a study by Bennett et al. ( 19 ), which used the log-rank test and Kaplan-Meier analysis to evaluate prognostic factors of OCCC, concluded that staging and lymph node status were the only parameters with statistical significance (P<0.0001). Regarding treatment, surgery remains the primary approach for managing OC. Takano et al. reported that patients with EOC who underwent satisfactory tumor reduction surgery (residual lesions <1 cm) had a significant survival advantage compared to those who did not undergo such surgery ( 20 ). However, Sugiyama et al. ( 21 ) showed that residual tumor size did not significantly affect survival rates. In this study, data analysis indicated no significant difference in survival rates based on the effect of surgery (P=0.495), except for two patients with stage IV. Most patients underwent satisfactory tumor reduction, with few negative data or bias, but the accuracy of these results cannot be guaranteed. Chemotherapy plays a critical role in the treatment of OCCC, with standard postoperative regimens typically involving platinum-based combinations, such as TP (paclitaxel + cisplatin) or TC (paclitaxel + carboplatin). The efficacy of standard platinum-based chemotherapy in OC is approximately 30%, significantly lower than that observed in high-grade serous carcinoma. Takano et al. ( 22 ) reported that the response rate to paclitaxel and carboplatin ranged from 22% to 56%. In this study, paclitaxel chemotherapy (P=0.046) significantly impacted survival in patients with OCCC. Completion of surgery and more than three cycles of chemotherapy were considered as potential contributing factors. The study also revealed a platinum resistance rate of 36.2%, with platinum resistance identified as an independent risk factor for OS (P<0.001). The prognosis of patients with advanced OCCC improves with the addition of immunotherapy. In a non-randomized clinical trial, patients with advanced OCCC who received a combination of nivolumab and ipilimumab demonstrated an objective response rate (ORR) of 55% (95% CI: 35–73%). This suggests that anti-PD-1/CTLA-4 blockade shows promising activity with a high rate of durable responses in patients with advanced OCCC ( 23 ). In recent years, the tumor microenvironment (TME) of OCCC has garnered significant attention. Devlin et al. found notable differences in immune cell populations, collagen matrix composition, and cytokine expression between ARID1A wild-type (ARID1Awt) and ARID1A mutant (ARID1Amut) OCCCs, which may reflect distinct tumorigenesis pathways and their association with EM. Increased infiltration of CD8 + T cells in the malignant cell area (MCA) and increased CD4 + T cell infiltration in the leading edge (LE) and stroma were significantly associated with reduced OS ( 24 ). However, the expression of PD-L1 and PD-1 in tumor and stromal cells was not significantly linked to prognosis. The presence of CD8 + T cells showed significant potential for predicting the prognosis of patients with OCCC ( 25 ). This study is a retrospective, single-center analysis, and several limitations are inherent in this design. First, the retrospective nature of the study inherently introduces limitations. Second, the study’s single-center design may introduce bias. Third, the absence of a central pathology review is a notable limitation, particularly since most of the patients were diagnosed and treated before the routine implementation of molecular testing (e.g., ARID1A, PIK3CA status). A larger, prospective, multi-center study is needed to further validate these findings.

This study identified stage, EM, platinum resistance, and platinum-based chemotherapy as key prognostic factors in patients with OCCC. Additionally, CA-125 levels, positive lymph nodes, and venous thrombosis were found to be associated with survival outcomes.

The article’s supplementary files as