Nomograms for predicting survival in early-onset esophageal cancer

Background: This study aimed to develop nomograms predicting the overall survival (OS) of patients younger than 50 years old with esophageal cancer.Methods: We selected patients included 2004-2015 in the Surveillance, Epidemiology, and End Results (SEER) database. Nomograms were constructed using significant variables from multivariable Cox analyses. The discrimination and calibration power of the models were evaluated using concordance indexes (C-indexes) and calibration curves. Decision curve analysis was used to assess the clinical net benefits of the nomograms.Results: Of 1,997 selected patients, 53.2% had advanced-stage tumor. Race, grade, T stage, N stage, and treatment were independent factors affecting OS in early-stage patients. The C-indexes of the corresponding nomogram were 0.710 (95% CI = 0.684-0.736) and 0.681 (95% CI = 0.640-0.722) in training and validation sets, respectively. Grade, marital status, and treatment were independent factors affecting OS in advanced-stage patients. The C-indexes of the corresponding nomogram were 0.677 (95% CI = 0.653-0.701) and 0.675 (95% CI = 0.638-0.712) in training and validation sets, respectively. Calibration curves demonstrated high consistency between predicted and actual survival.Conclusion: We constructed and verified nomograms that could accurately predict the survival rate of esophageal cancer in patients younger than 50 years old. This may help clinicians better understand prognostic factors.