Survival prognosis model for elderly women with epithelial ovarian cancer based on the SEER database.

Abstract

We aimed to analyze the risk factors of elderly women with epithelial ovarian cancer (EOC) using data on the SEER database, and to generate a nomogram model their 1-, 3-, and 5-year prognoses. The resulting nomogram model should be useful for clinical diagnoses and treatment. We collected clinical data of women older than 70 years with epithelial ovarian cancer (diagnosed on the basis of surgical pathology) from the SEER database including datasets between 2010 and 2019. We randomly grouped the data into two groups (7:3 ratio) using the R language software. We divided the independent prognostic factors obtained by univariate and multi-factor Cox regression analyses into training and validation sets, and we plotted the same independent prognostic factors in a nomogram model of overall survival (OS) at 1, 3, and 5 years. We used the C-index, calibration curve, and area under the curve to validate the nomograms. We further evaluated the model and its clinical applicability using decision curve analyses. We identified age, race, marital status, histological type, AJCC staging, differentiation degree, unilateral and bilateral tumor involvement, number of positive lymph nodes, chemotherapy, surgery, sequence of systemic treatment versus surgery, and time from diagnosis to treatment as independent prognostic factors for elderly women with EOC (P < 0.5). The C-indexes were 0.749 and 0.735 in the training and validation sets, respectively; the ROC curves showed that the AUC of each prognostic factor was greater than 0.7; and, the AUC values predicted by the line plot were similar in the training and validation sets. The decision curves suggest that this line plot model has a high clinical value for predicting overall survivals at 1, 3, and 5 years in elderly women with EOC. The nomogram model in this study can provide an accurate assessment of the overall survival of women older than 70 years with EOC at the time of the first treatment, and it provides a basis for individualized clinical treatment.