Prognostic factor analysis and nomogram construction for elderly patients with stages III and IV epithelial ovarian cancer: a study based on the SEER database

Abstract

Background: Epithelial ovarian cancer (EOC), one of the most fatal diseases affecting the elderly women. Advanced stages EOC (stage III and stage IV) presents significant challenges in prognosis and treatment due to factors such as poor treatment tolerance, comorbidities, and immune dysfunction. There is a lack of reliable prognostic tools for elderly EOC patients. This study aimed to develop two nomograms to predict overall survival (OS) and cancer-specific survival (CSS) in elderly patients with advanced-stage EOC using Surveillance, Epidemiology, and End Results (SEER) database, providing a tool for more personalized treatment decisions.

Methods: Data about patients diagnosed with ovarian cancer at stages III and IV from 2010 to 2015 were extracted from the SEER database. Participants were randomly assigned to a training set and a validation set in a 7:3 ratio with OS and CSS as outcome events. Independent prognostic indicators determined in the multivariable analysis were employed in nomograms for predicting 1-, 3-, and 5-year OS and CSS for elderly EOC patients. The predictive performance and clinical utility were assessed using the concordance index (C-index), receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA).

Results: The majority of included participants were in stage III (71.38%), while 28.62% were in stage IV. In the OS training set, identified independent prognostic factors included age, race, marital status, tumor grade, T stage, American Joint Committee on Cancer (AJCC) stage, laterality, surgical method, chemotherapy, and cancer antigen 125 (CA-125). In the CSS training set, all these factors were retained except for the variable 'race'. The area under the ROC curve (AUC) for OS in the training set was 0.77 (0.75, 0.80) for 1-year, 0.68 (0.66, 0.70) for 3-year, and 0.66 (0.63, 0.68) for 5-year; in the validation set, the AUCs were 0.74 (0.70, 0.79), 0.69 (0.66, 0.72), and 0.70 (0.67, 0.73), respectively. For CSS in the training set, the AUCs were 0.77 (0.74, 0.79), 0.68 (0.66, 0.70), and 0.67 (0.64, 0.69) for 1, 3, and 5 years; in the validation set, the AUCs were 0.76 (0.71, 0.81), 0.66 (0.63, 0.70), and 0.67 (0.63, 0.70). These results indicate that the developed nomograms possess robust discriminative ability in predicting patients' OS and CSS.

Conclusions: This study establishes clinically relevant nomograms for elderly patients with advanced ovarian cancer, demonstrating significant diagnostic value in predicting OS and CSS.

Keywords: Elderly women; Surveillance, Epidemiology, and End Results database (SEER database); epithelial ovarian cancer (EOC); nomogram; prognostic factors.

Figure 1

Flow chart of stage III and IV elderly EOC patients based on the inclusion and exclusion criteria. AJCC, American Joint Committee on Cancer; CA-125, cancer antigen 125; EOC, epithelial ovarian cancer; N, node; NA, not available; SEER, Surveillance, Epidemiology, and End Results; T, tumor.

Figure 2

Kaplan-Meier survival analysis curve for comparing OS (A) and CSS (B) between different chemotherapy groups. CSS, cancer-specific survival; OS, overall survival.

Figure 3

Kaplan-Meier survival analysis curve for comparing OS (A) and CSS (B) between different surgery methods. CSS, cancer-specific survival; OS, overall survival.

Figure 4

Nomograms to predict 1-, 3-, and 5-year OS (A) and CSS (B) for stage III and IV elderly patients with EOC. AJCC, American Joint Committee on Cancer; CA-125, cancer antigen 125; CSS, cancer-specific survival; EOC, epithelial ovarian cancer; OS, overall survival; T, tumor.

Figure 5

ROC curve analysis for OS (A,B) and CSS (C,D) in the training and testing sets. The models display AUC values for 1-, 3-, and 5-year survival. AUC, area under the curve; CI, confident interval; CSS, cancer-specific survival; OS, overall survival; ROC, receiver operating characteristic.

Figure 6

Calibration curve analysis for OS (A,B) and CSS (C,D) in the training and testing sets. The models display AUC values for 1-, 3-, and 5-year survival. AUC, area under the curve; CSS, cancer-specific survival; OS, overall survival.

Figure 7

DCA of the OS-associated and CSS-associated nomograms. DCA curves of 1-, 3-, and 5-year OS in the training cohort (A-C) and validation cohort (D-F). DCA curves of 1-, 3-, and 5-year CSS in the training group (G-I) and validation group (J-L). CSS, cancer-specific survival; DCA, decision curve analysis; OS, overall survival.