{"paper_id":"6b4ae8a4-446b-4532-9bb2-582fe0a4511a","body_text":"Epithelial ovarian cancer (EOC) is a highly lethal carcinoma, the global 5‐year overall survival rate of which has stayed at 30%–40% for decades. The standard treatment recommended by the National Comprehensive Cancer Network guidelines is comprehensive surgical staging followed by platinum‐based chemotherapy. \n 1 \n ,  \n 2 \n  However, disease recurrence and chemoresistance pose major challenges to therapeutic efficacy, and contribute to the high mortality rate. \n 3 \n ,  \n 4 \n  Ovarian clear cell carcinoma (OCCC) is an infrequent histological subtype of EOC, which accounts for a variable 5%–25% of all cases of EOC. It has different biologic behavior and clinicopathological characteristics from those of other subtypes of EOC. \n 5 \n ,  \n 6 \n  To be exact, most OCCC are negative with WT1 transcription factor and estrogen receptors. Conversely, high‐grade serous carcinoma is usually positive. OCCC shows frequent phosphatidylinositol‐4,5‐bisphosphate 3‐kinase catalytic subunit alpha mutations and PTEN inactivation, while tumor protein 53 and breast cancer 1/2 mutations are ubiquitous in high‐grade serous carcinoma. \n 7 \n  Furthermore, OCCC may arise from ovarian endometriosis and usually presents at a younger age and earlier stage in comparison with high‐grade serous carcinoma. Early OCCC tends to be associated with a more favorable prognosis than late OCCC. \n 8 \n ,  \n 9 \n  However, OCCC is considered to be a high‐grade tumor, and advanced stage OCCC is associated with a poor prognosis due to its intrinsic resistance. \n 10 \n ,  \n 11 \n  Although postoperative treatment is the standard treatment regimen for OCCC, \n 12 \n  its overall efficacy for early stage disease remains unclear. \n 13 \n  In addition to the insensitivity of OCCC to chemotherapy, one major cause is that OCCC is frequently analyzed concomitantly with other histological subtypes due to its rarity. \n 14 \n  For that reason, detailed and integrated information regarding the treatment and prognosis of early OCCC is difficult to acquire and, therefore, scarce. \n 15 \n  Subtype‐specific data are thus required if therapeutic strategies for early OCCC are to be optimized in the future.\nThe present study provided specific data from the US Surveillance, Epidemiology and End Results (SEER) database, which is devoted to improving data quality by performing rigorous quality control studies and various data assessments. Patients with stage IA OCCC were incorporated into our study cohort. All patients received surgery and their follow‐up data were retrospectively analyzed. This study aimed to determine whether adjuvant therapy should be used for the treatment of stage IA OCCC.\n\nData were extracted from the SEER database.\nPatients with primary OCCC (ICD‐O‐3, 8310/3) between 2004 and 2015 were identified and all subjects underwent surgery ( n  = 6231). Data from a total of 1038 patients with stage IA OCCC (Derived AJCC Stage Group, 6th ed) were collected for the present study. The following variables were included and presented as numbers and percentages: Patient identity card, Age at diagnosis, Ethnicity record, Tumor stage (Derived AJCC Stage Group, 6th ed), Laterality (origin of primary), CS (collaborative staging) tumor size, Regional lymph nodes removed, Primary surgical site (ovary), Surgical type, Chemotherapy recode, Survival months, Vital status (dead or alive) and SEER cause‐specific death classification. The subjects were classified into three age groups: <50, 50–60 and > 60 years. Ethnicity was divided into Caucasian and non‐Caucasian. Tumor laterality was grouped into whether the tumor originated from the left or the right ovary. Tumor size was classified into the following three groups: ≤2 cm, >2 cm and unknown. The lymph nodes removed were categorized into ≤3 and ≥4. In the extracted dataset, chemotherapy recode was divided into yes and no/unknown. The effect of chemotherapy on the 5‐year overall survival was assessed for the subjects (Figure  1 ).\nFlow chart.\nThe follow‐up data of each subject was integrated. Pearson's  χ \n 2 ‐tests were used to assess the univariate associations. Two‐sided test was conducted for all statistical tests.  p  < 0.05 was considered statistically significant. Statistical analysis was performed using R and SPSS software.\nSelection bias, which is common in retrospective studies, existed in our research and might affect the results of the present study. To minimize the influence of selection bias, PSM and logistic regression models were used to match age, ethnicity, tumor laterality, tumor size and number of regional lymph nodes removed between the chemotherapy and non‐chemotherapy groups.\nThe Kaplan–Meier method was applied to perform survival analysis. The hazard ratios (HRs) and 95% confidence intervals (CIs) were evaluated using the Cox proportional hazards model.\n\nA total of 1038 patients (346 who had not received chemotherapy and 692 who had) met the study inclusion criteria. Among these patients, 26.2% were aged <50 years, 43.0% were aged 50–60 years and 30.8% were aged ≥60 years. The majority of patients were Caucasian ( n  = 823; 79.3%). An equal number of left‐ ( n  = 519; 50.0%) and right‐sided ( n  = 519; 50.0%) cases were observed in terms of tumor laterality. Only 8.0% of the patients had tumors measuring ≤2 cm in size, whereas 79.5% of the patients had tumors measuring >2 cm. In total, ≤3 lymph nodes were removed from 28.4% of the patients, whilst ≥4 lymph nodes were removed from 71.6% of the patients (Table  1 ).\nDemographics\nData are expressed as  n  (%).\nLaterality: the side of ovaries on which the primary tumor originated.\nCompared with patients in the non‐chemotherapy groups, patients in the chemotherapy groups were younger prior to PSM (<50 years, 24.4 vs. 29.8%, respectively; 50–60 years, 46.8 vs. 35.3%, respectively; and > 60 years, 28.8 vs. 35.0%, respectively;  p  < 0.05). The patients were more likely to be Caucasian (79.6 vs. 78.6%, respectively;  p  = 0.705). There was also a higher number of patients in whom the left ovary was the primary origin in the chemotherapy groups (50.3 vs. 49.4%, respectively;  p  = 0.792). Fewer patients had tumors measuring ≤2 cm in the chemotherapy groups (7.8 vs. 8.4%, respectively;  p  = 0.090). Finally, there were fewer patients with ≤3 lymph nodes removed in the chemotherapy groups compared with the non‐chemotherapy groups (24.3 vs. 36.7%, respectively;  p  < 0.05). PSM was performed to eliminate the non‐uniformity between the two groups and equalized the number of patients ( n  = 346). The logistic regression analysis demonstrated that the covariates were well balanced and markedly decreased (Table  2 ).\nComparison of univariate covariates\nAbbreviation: PSM, propensity score matching.\nThe potential relevance between chemotherapy and 5‐year overall survival was analyzed. No statistically significances were identified between the chemotherapy and non‐chemotherapy groups (54.6 vs. 55.8%, respectively;  p  = 0.760). Therefore, patients with stage IA OCCC did not benefit from chemotherapy in the field of 5‐year overall survival (Table  3 ).\nRelevance of chemotherapy with 5‐year overall survival\nA univariate analysis was conducted in the matched patients to study the prognostic effects of the clinical factors (Table  4  and Figure  2 ). No significant differences in the 5‐year overall survival rates were identified between the chemotherapy and non‐chemotherapy groups (HR, 0.934; 95% CI, 0.764–1.142;  p  = 0.506). Older age was a risk factor for 5‐year overall survival (50–60 years, HR, 0.713; 95% CI, 0.555–0.915; >60 years, HR, 0.750; 95% CI, 0.584–0.963;  p  < 0.05) rates. Tumors >2 cm in size were associated with a lower 5‐year overall survival (HR, 1.545; 95% CI, 0.976–2.446;  p  < 0.05) rates.\nUnivariate analysis and multivariate cox regression analysis for 5‐year overall survival\nAbbreviations: CI, confidence intervals; HR, hazard ratios; Ref: reference.\nKaplan–Meier survival curves for 5‐year overall survival in patients with stage IA OCCC after propensity score matching.  p  = 0.497.\nThe correlation between variables and survival is extensively investigated using the Cox proportional hazards model. The Kaplan–Meier method and log‐rank tests are used to analyze patient survival according to one factor in isolation, without including the effect of other factors. They are only applicable for categorical variables, not for quantitative or parametric variables. The Cox proportional hazards model is suitable for both categorical and quantitative variables. In addition, the Cox proportional hazards model allows for the simultaneous assessment of the effects of several variables on survival.\nTo investigate how the variables collectively affect survival, the present study incorporated these variables into the multivariate Cox regression analysis (Table  4 ). The results suggested that elderly patients (age, ≥50 years) had a higher mortality and a lower 5‐year overall survival (50–60 years, HR, 1.052; 95% CI, 0.826–1.339; >60 years, HR, 1.403; 95% CI, 1.093–1.801;  p  < 0.05) rates. After eliminating the effect of miscellaneous factors, however, chemotherapy remained statistically insignificant in the field of the 5‐year overall survival ( p  = 0.524).\n\nOCCC accounts for 3%–11% of all EOC cases in the United States. \n 16 \n ,  \n 17 \n  Staging surgery or cytoreductive surgery combined with chemotherapy is the current standardized treatment for OCCC, and clinical guidelines recommend standard intravenous platinum‐based chemotherapy for patients with stage I‐IV OCCC, \n 18 \n  though it is characterized as inherent chemoresistance and poor prognosis. \n 19 \n ,  \n 20 \n  However, the effect of adjuvant chemotherapy in early OCCC remains the subject of intense debate on account of the rarity of the disease and the lack of large sample clinical trials. \n 21\nThe strength of the present study was that it provided specific and powerful evidence on the efficacy of adjuvant chemotherapy for treating stage IA OCCC because the cohort was based on a great and distinct population. However, all SEER database analyses have inherent limitations, which may have resulted in important biases in the present study. \n 22 \n  Precise surgical data were arduous to acquire for the subjects. In addition, this dataset lacked information regarding the details of adjuvant chemotherapy, including the number of cycles, agents and dosage. \n 23 \n  To weaken selection bias, the dataset was randomized and causal arguments were strengthened using PSM. The Cox proportional‐hazards models were performed to analyze the correlation between variables and survival.\nNo significant differences in the 5‐year overall survival were discovered between the chemotherapy and non‐chemotherapy groups ( p  = 0.497). The multivariate Cox regression analysis showed that elderly patients (age, ≥50 years) had a higher mortality and a lower 5‐year overall survival ( p  < 0.05). After eliminating the effect of miscellaneous factors, it was observed that the difference between chemotherapy and the 5‐year overall survival was not statistically significant ( p  = 0.524).\nThe present findings were consistent with those of previous studies. A retrospective investigation suggested that adjuvant chemotherapy had little effect on the survival of stage I OCCC patients ( n  = 219), and no statistical difference between progression‐free and overall survival was identified. \n 24 \n  Similarly, Mizuno et al found that postoperative adjuvant chemotherapy was not helpful in improving the prognosis of patients with stage IA OCCC ( p  = 134). There was no significant prognostic factor for either recurrence or survival. \n 25 \n  Equally, Takada et al revealed that postoperative adjuvant chemotherapy did not influence progression‐free and overall survival in patients with stage IA CCC ( n  = 20). \n 26 \n  By the same token, Oseledchyk et al evaluated the impact of adjuvant chemotherapy on overall survival in patients with stage I OCCC using the SEER database. Their analysis indicated that chemotherapy was not associated with improved overall survival ( n  = 1995). \n 27 \n  Another retrospective study demonstrated that postoperative adjuvant chemotherapy was not associated with overall survival, and could be safely omitted in patients with stage IA and IB OCCC ( n  = 45). \n 28 \n  In addition, a meta‐analysis revealed that adjuvant chemotherapy did not affect disease‐free survival and overall survival in patients with stage IA and IB OCCC ( n  = 2809). \n 29 \n  However, the present findings were inconsistent with those of a study by Hogen et al, which supported the use of adjuvant chemotherapy for surgical stage I OCCC ( n  = 60). \n 30 \n  Furthermore, a cohort of postoperative patients with stage I OCCC (1298 patients with stage IA, 35 patients with stage IB, and 1007 patients with stage IC) was drawn from the National Cancer Database. That study indicated that adjuvant chemotherapy was associated with a better overall survival. \n 31 \n  The discordance in findings could be ascribed to several factors, including differences in patient population, surgical protocols, chemotherapy paradigm, tumor misclassifications, or tumor heterogeneity.\nIn conclusion, the present investigation suggested that the adoption of adjuvant chemotherapy does not improve the outcomes of patients with stage IA OCCC after surgery has been performed. However, convincing evidences from large‐scale clinical trials are needed to confirm this encouraging outcome. Further study is required to provide guidance on the optimal management of patients with early OCCC.\n\nShuqing Li:  Conceptualization (lead); data curation (lead); formal analysis (lead); investigation (lead); methodology (lead); project administration (lead); resources (lead); software (lead); writing – original draft (lead); writing – review and editing (lead).\n\nThe authors declare that they have no conflicts of interest.","source_license":"CC-BY-4.0","license_restricted":false}