Prognostic and predictive significance of tumor length in patients with esophageal squamous cell carcinoma undergoing radical resection

Background: The objective of this study was to investigate the prognostic and predictive significance of tumor length in patients with esophageal squamous cell carcinoma undergoing radical resection. Methods: Tumor length and other clinicopathological variables were retrospectively evaluated in 1435 patients with squamous cell carcinoma treated with radical resection between 2003 and 2010. Tumor length was analyzed as categorical and continuous variable. Associations with overall survival were assessed with Cox proportional hazards models. Model-based nomograms were constructed. Predictive accuracy was measured with C-index. Decision curve analysis was used to evaluate clinical usefulness of prediction models. Results: Both categorically and continuously coded tumor length were independent prognostic factors in multivariable analysis. Adding categorically and continuously coded tumor length to TNM staging model increased predictive accuracy by 0.2 and 0.4 % respectively. Decision curve analysis revealed that the models built by the addition of categorically or continuously coded tumor length did not perform better than TNM staging model. Conclusions: Tumor length is an independent prognostic factor in patients with esophageal squamous cell carcinoma treated with radical resection. It increases predictive accuracy of TNM staging system for overall survival in these patients. But it does not increase clinical usefulness of TNM staging system as a prediction model.