Prognostic Significance of Surgery and Radiotherapy in Elderly Patients with Localized Prostate Cancer: Establishing and Time-based External Validation a Nomogram from SEER-based Study

Objective: Prostate cancer (PC) is a significant disease affecting men's health worldwide. More than 60% of patients over 65 years old and more than 80% are diagnosed with localized PC. The current choice of treatment modalities for localized PC and whether overtreatment is controversial. Therefore, we wanted to construct a nomogram to predict the risk factors associated with cancer-specific survival (CSS) and overall survival (OS) in elderly patients with localized PC while assessing the survival differences in surgery and radiotherapy for elderly patients with localized PC.

Methods: Data of patients with localized PC over 65 years were obtained from the Surveillance, Epidemiology, and End Results (SEER) database. Univariate and multivariate Cox regression models were used to determine independent risk factors for CSS and OS. Nomograms predicting CSS and OS were built using multivariate Cox regression models. The consistency index (C-index), the area under the subject operating characteristic curve (AUC), and the calibration curve were used to test the accuracy and discrimination of the prediction model. Decision curve analysis (DCA) was used to test the potential clinical value of this model.

Results: A total of 90,434 patients over 65 years and diagnosed with localized PC from 2010 to 2018 were included in the study. All patients were randomly assigned to the training set (n = 63,328) and the validation set (n = 27,106). Univariate and multivariate Cox regression model analysis showed that age, race, marriage, T stage, surgical, radiotherapy, prostate-specific antigen (PSA), and Gleason score (GS) were independent risk factors for predicting CSS in elderly patients with localized PC. Age, race, marriage, surgery, radiotherapy, PSA, and GS were independent risk factors for predicting OS in elderly patients with localized PC. The c-index of the training and validation sets for the predicted CSS is 0.802(95%CI:0.788-0.816) and 0.798(95%CI:0.776-0.820, respectively). The c-index of the training and validation sets for predicting OS is 0.712(95%:0.704-0.720) and 0.724(95%:0.714-0.734). It shows that the nomograms have excellent discriminatory ability. The AUC and the calibration curves also show good accuracy and discriminability.

Conclusion: We have developed new nomograms to predict CSS and OS in elderly patients with localized PC. After internal validation and external temporal validation with reasonable accuracy, reliability and potential clinical value, the model can be used for clinically assisted decision-making.