Guiding adjuvant radiotherapy in stage III endometrial cancer: a prognostic model based on SEER.

Background: The effect of overall survival (OS) with adjuvant radiotherapy in stage III endometrial cancer (EC) remains controversial, and the adverse invents were unignorable. Methods: A total of 4,064 stage III EC patients who underwent adjuvant chemotherapy post-operatively were selected from Surveillance, Epidemiology, and End Results (SEER) Program. Independent risk factors were identified through Cox regression models. A nomogram was developed accordingly to predict OS. The concordance index (C-index), calibration, and Receiver Operating Characteristic (ROC) curves were applied to assess the model. Patients were divided into the low- and high-risk groups based on the optimal risk cutoff. Stratified analysis was conducted by radiation in both groups, and interactions between radiation and the risk groups were conducted to explore if any benefit less from adjuvant radiotherapy. Results: A total of five candidate factors were identified from the model showing good calibration and consistency discriminative power in the training (C-index: 0.73; 95% CI: 0.70-0.75), testing (C-index: 0.73; 95% CI: 0.69-0.77), and external validation cohorts (C-index: 0.88, 95% CI, 0.78-0.97). Patients were categorized into the low- and high-risk groups based on the optimal risk cutoff of 2.1048630. The women in the high-risk group experience significantly less (42% vs. 63% reduction) or none (0 vs. 63%) benefit (p-interaction = 0.049 vs. 0.016 in training and testing cohorts, respectively). Conclusion: A nomogram incorporating five variables was established to predict OS in stage III EC patients with adjuvant chemotherapy. The high-risk groups benefit less or none from adjuvant radiotherapy, which may serve as a useful reference for better guidance of radiotherapy in stage III EC patients.