Development and validation of a nomogram to predict overall survival of epithelial ovarian cancer patients: a population-based study (381)

Abstract

<b>Objectives:</b> To develop and validate a nomogram for the overall survival of epithelial ovarian cancer based on the National Union of Real- World Gynecological Oncology Research and Patient Management Platform (NUWA) in China. <b>Methods:</b> We enrolled patients with pathologically confirmed primary epithelial ovarian cancer in the NUWA platform between 2004 and 2018 as a training cohort to develop a nomogram. The cut-off for follow-up was December 2020. Patients with a history of other tumors or missing values of any variable were excluded. Variables were selected by univariate and multivariate Cox proportional hazard regression. Significant factors were applied to establish a nomogram to predict 3-year and 5-year survival, assessed by C-statistic and calibration curves. A testing cohort from SEER Database was applied to externally validate the nomogram by bootstrap resampling. Statistical analysis was conducted by SPSS version 26.0 (IL, USA) and R version 4.1.1 (www.r-project.org). <b>Results:</b> A total of 1262 primary epithelial ovarian cancer patients were filtered out from 5003 ovarian cancer patients in the NUWA platform. Total 10594 of 125682 patients in the SEER Database were screened out accordingly. Baseline age, histotype, histological grade, FIGO stage, longest diameter of the primary tumor, and residual disease after primary surgery were significant risk factors (<i>p</i><0.05). Compared with serous cancer, the hazard ratio (HR) of mucinous, endometrioid, clear cell, other epithelial cancer were 3.391 (95% CI: 2.201-5.223), 1.642 (95% CI: 1.045-2.580), 3.127 (95% CI: 2.134-4.582), and 2.071 (95% CI: 1.486-2.885), respectively. The HR of advanced stage and high grade of disease (FIGO stage III-IV vs I-II, High grade vs Low/moderate grade) were 1.921 (95% CI: 1.226-3.010) and 1.407 (95% CI: 1.048-1.889), respectively. Patients who received optimal (R1), suboptimal (R2) cytoreduction, and those who didn't receive primary debulking surgery had a higher risk of death compared to those without the residual disease (R0), with HRs of 5.530 (95% CI: 3.575-8.555), 7.757 (95% CI: 5.101-11.796) and 9.975 (95% CI: 5.831-17.063), respectively. For continuous variables, the HR of baseline age (years) was 1.024 (95% CI: 1.013-1.036), and the HR of the longest diameter of the primary tumor (cm) was 1.030 (95% CI: 1.007-1.054). The C-statistic values of the nomogram (Figure 1) in the training and testing cohorts were 0.771 (95% CI: 0.748-0.794) and 0.726 (95% CI: 0.718-0.734), respectively. The slopes of calibration curves of 3-year and 5-year survival in the training cohort were 1.165 (95% CI: 0.990-1.340) and 0.955 (95% CI: 0.920-0.990), respectively, whereas, in the testing cohort, these were 1.355 (95% CI: 1.108-1.602) and 1.374 (95% CI: 1.103-1.644), respectively. Fig. 1 <b>Conclusions:</b> The discrimination and goodness of fit of the nomogram indicated its predictive value for the survival of epithelial ovarian cancer in China. Furthermore, the nomogram also obtains certain values in external validation, which implies its feasibility in different populations.