In this study, the ED rates of EOC at 1, 3, and 6 months post-operation were 9.98%, 18.73%, and 33.34%, respectively. Identifying patients at risk of ED is essential to reduce the burden on patients. Our results show that ED from ovarian cancer is mainly related the clinical factors AJCC stage, residual lesion size, chemotherapy, serum CA125 level, tumor size, number of lymph nodes examined, surgery of primary site, and age. Therefore, we constructed a nomogram that integrates these factors for predicting ED in ovarian cancer patients.
Comparing the ROCs and time-dependent AUCs shows that the nomogram had higher predictive power than the AJCC staging system. AJCC staging is commonly used to clinically evaluate ovarian cancer prognosis but is limited in that it cannot give individualized predictions. Additional clinical factors, such as those included in our nomogram, are often ignored. Internal validation showed that the ED rate predicted with the nomogram is consistent with the actual ED rate. The C-index of the nomogram was 0.787, which represents a better degree of differentiation and ability to provide individualized prediction compared with the AJCC stage. In addition, both the training and validation sets showed good consistency with observed values.
Nomograms have been widely used to evaluate the prognosis and death risk in malignant tumor patients. Chen et al.11 studied all-cause and specific mortalities of ovarian clear cell carcinoma and constructed a nomogram with the prognostic factors age, laterality, organ metastasis, AJCC stage, number of lymphadenectomies, and chemotherapy. Yuan et al.12 assessed lung metastasis incidence in ovarian cancer and proposed that stage, liver, bone, and brain metastases and TN stage are predictors of lung metastasis. The nomogram constructed by Mosgaard et al.6 showed that the survival rate was higher for patients with chemotherapy, smaller residual cancer and tumor size, younger age, lower CA125 level, higher differentiation, debulking surgery, and more lymph node resections. In the present study, analysis and internal validation using ROC and DCA curves showed that the nomogram has good discrimination and calibration. Thus, the nomogram may be an effective tool to predict ED in ovarian cancer patients, guide their individualized treatment, and improve their hospice care and quality of life.
Cancer stage is closely related to cancer survival and ED rates. In our nomogram, stage occupied the entire 100-point scale, with stage I being scored 0 and stage IV being scored 100. This implies that stage is an important independent influencing factor. This is consistent with the results of previous studies on prognostic evaluation models for pancreatic cancer and uterine sarcoma13,14. The prognosis of stage III and IV patients is generally poor and mainly depends on the size of intraperitoneal metastases15. As stage progresses, the risk of death increases16,17. A large retrospective cohort study based on the SEER database suggested that mortality in patients with stage III-IV ovarian cancer is as high as 10% in 1 year18.
After stage, the residual size of cancer foci was the second most influential factor in this study. The larger the residual cancer foci, the higher the chances of ED, which is consistent with the findings of previous studies. Surgery is essential for ovarian cancer treatment, and postoperative residual tumor is one of the most relevant clinical prognostic factors19–21. Surgery aimed at minimizing tumor cells can lead to better outcomes. Complete tumor resection is considered a major predictor of survival22. Approximately 20% of patients with advanced ovarian cancer survived for more than 12 years after treatment and were eventually effectively cured. Debulking surgery is performed to eliminate cancer cells and cancer foci as much as possible, preferably without significant residue. An article previously suggested that recovery depends on whether the combination of surgery and chemotherapy can effectively eliminate all cancer cells23.
As an important adjuvant treatment for ovarian cancer, chemotherapy is commonly used to kill residual cancer foci and control or treat recurrent foci. In our study, chemotherapy was significantly correlated with prognosis, which was of great value in improving survival outcomes. Chemotherapy can reduce tumors and create conditions for surgery. Major tumor debulking surgery and platinum chemotherapy remain as the standard treatments for patients with stage III-IV EOC. Some patients with International Federation of Obstetrics and Gynecology stage III-IV ovarian cancer may benefit from neoadjuvant chemotherapy24. EOC patients (especially high-grade serous cancer) respond well to initial chemotherapy3, with approximately 80% responding to neoadjuvant chemotherapy as an alternative treatment. Poly ADP-ribose polymerase inhibitors are one of the most studied, most effective, and least toxic drugs, and have therefore become one of the best targeted therapeutic options for treating recurrent EOC, especially in cases of platinum-sensitive recurrent ovarian cancer 25,26.
Age and stage are independent risk factors for ovarian cancer prognosis6. In our study, age, lymph node examined, and tumor size were stratified, with cutoff values of 62 and 72 years for age, 2 for number of lymph nodes examined, and 90 mm for tumor size. In general, older patients are at higher risk for ovarian cancer and are more likely to have poor survival outcomes due to lower immune responses27. However, we observed that age was not highly significant, which may be related to the relatively conservative surgeries and pre- and postoperative chemotherapies given to young patients. In clinical practice, we often explore the pelvic and abdominal cavities of ovarian cancer patients to histologically examine suspected lesions and sites prone to metastasis and to clear the pelvic and abdominal para-aortic lymph nodes. Operation scope is determined according to the results of intraoperative exploration and frozen pathology examinations. The thoroughness of the first operation of EOC is closely related to the prognosis. Lymph node metastasis has an important effect on EOC prognosis. Some scholars suggest that lymph node dissection in advanced EOC may be used as a treatment. The number of lymph nodes examined and the resection of paraaortic lymph nodes may also be helpful28. In stage III, the subcategories IIIA and IIIB are based on the presence or absence of gross external pelvic peritoneal metastasis. However, this method is unable to distinguish the prognosis of patients with different numbers of lymphadenectomy in the same pathological stage. Therefore, we used X-tile to analyze the optimal cutoff value for the number of lymph nodes, which was subsequently included in the prediction after univariate and multivariate analysis.
Serum CA125 level is a high-sensitivity index for disease monitoring. Among EOC patients, Serum CA125 level are higher than the normal value, with more than 90% being consistent with the remission or deterioration of the disease. Increasing Serum CA125 level is considered an important predictor of death29 and is therefore important for predicting prognosis.
The primary ovarian focus in stage I patients are larger than those in stage III patients15. In addition, the ovarian foci in early ovarian cancer are more than twice as large as those in advanced ovarian cancer30. These support the fact that early and advanced ovarian cancer are two separate disease processes. Early tumors grow locally and do not spread, while advanced tumors that are relatively small are prone to spreading. It has been suggested that there may be a key substance differentiating the two processes; that is, the tumor in patients with advanced disease produces a substance that allows early-stage transmission. Without this substance, the tumor only grows locally.
The standard process for determining treatment in early EOC is clinical/surgical staging, which includes hysterectomy, bilateral ovariectomy, omentum resection, abdominal irrigation, and pelvic and paraaortic lymph node biopsy. Preserving the reproductive function means preserving the uterus and at least one side of the ovary. A study based on the SEER database found that fertility-sparing surgery was associated with an increased risk of death in women with advanced serous EOC31. However, some studies have found that the effect of fertility-sparing surgery on survival in stage I ovarian cancer is no worse than that of radical surgery. This suggests that specific histological subtypes have a greater effect on tumor prognosis than the retention of reproductive function. Radical surgery is unlikely to reduce the risk of recurrence of certain histological subtypes32. In a previous study, there was no significant difference in overall survival between stage I and radical surgery in EOC33. Prognosis may be more related to the natural history of the disease and the cancer type rather than to the specific type of surgery34. The nomogram we constructed refined the prognostic prediction by classifying the surgical modalities. And based on that, it is obviously that the score the difference of surgical type is non-significant.
This study has several limitations. First, our model did not include molecular markers that elucidate ovarian cancer mechanisms as these were not part of the SEER database. Many of these markers have been used to build predictive models for ovarian cancer35, including five genes related to glucose metabolism36 and 11 genes related to lipid metabolism37. MRPS12 may be a promising candidate for prognosis38. Second, several factors were also unavailable from the SEER database, including patients’ family history, underlying pre-operative diseases, BMI, types of anesthetics, induction time, blood pressure, blood oxygen, heart rate fluctuations, cancer cell detection in pre- and postoperative ascites, thrombosis and surgical incision infection, specific preoperative and postoperative chemotherapy, chemotherapy times, and chemotherapeutic agents. Third, we did not analyze humanistic and sociological factors such as income and insurance, which have an important influence on the psychological and physiological aspects of ovarian cancer8. We also excluded economic status, education level, and follow-up by gynecologic oncologists, which are considered closely related to ovarian cancer prognosis. Lastly, our study is retrospective and has potential for selection bias since the data were extracted from the SEER database. Without external data validation, a more comprehensive prediction is impossible. Further studies combining our research data with those of others are needed for better prediction.
In conclusion, using a large cohort study, we identified several factors associated with ED in ovarian cancer and constructed a nomogram with better prognostic performance than the AJCC staging system. Our nomogram can be used in future clinical work as a more effective tool for screening high-risk patients. It may also play an important role in predicting ED from ovarian cancer and providing relatively reliable and individualized postoperative treatment advice to improve the quality of life of ovarian cancer patients.