Objective To explore the influence of clinicopathological factors besides TNM stage, including preoperative tumor volume, length and maximum diameter, on survival prognosis of patients with thoracic esophageal squamous cell carcinoma (ESCC), and to evaluate the predictive survival rate of clinicopathological variables with statistical significance by nomogram. Methods A total of 296 patients with ESCC treated by radical resection at the Department of Thoracic Surgery of Affiliated Taixing People's Hospital of Yangzhou University from 2011 to 2014 were retrospectively analyzed. These patients were grouped for further analysis according to the optimal threshold of preoperative tumor volume, length and maximum diameter. Kaplan-Meier method was used to calculate survival rate and survival comparison was performed by log-rank test. The univariate and multivariate Cox models were used to analyze the relationships between clinical variables and survival prognosis. Finally, nomogram model was established by integrating statistically significant clinicopathological parameters, and the predictive value of this model was further verified by calibration curve, concordance index (C-index) and decision curve. Results The optimal thresholds of preoperative tumor volume were 32 cm 3 and 72 cm 3 by X-tile analysis, and among the patients whose tumor volume was <32 cm 3 (n=94), the 1-, 3- and 5-year survival rates were 100%, 84.0% and 68.1%; in the 32-72 cm 3 group (n=118), the 1-, 3- and 5-year survival rates were 98.3%, 42.4% and 24.6%; in the >72 cm 3 group (n=84), the 1-, 3- and 5-year survival rates were 94.1%, 25.0 and 7.1% (χ 2=86.639, P<0.001). The optimal cutoff values of tumor length were 3.0 cm and 5.0 cm, and among the patients with tumor length <3.0 cm (n=62), the 1-, 3-, and 5-year survival rates were 99.5%, 87.1% and 69.4%; in the 3.0-5.0 cm group (n=146), the 1-, 3-, and 5-year survival rates were 98.6%, 47.9% and 30.1%; in the >5.0 cm group (n=88), the 1-, 3-, and 5-year survival rates were 94.3%, 29.6%, 13.6%, respectively (χ 2=53.607, P<0.001). The thresholds of tumor maximum diameter were 2.5 cm and 3.5 cm, and among these, the 1-, 3- and 5-year survival rates were 99.5%, 84.3% and 74.5% in the maximum diameter <2.5 cm group (n=51); 98.3%, 57.0% and 36.4% in the 2.5-3.5 cm group (n=121); and 96.0%, 29.0% and 13.7% in the maximum diameter >3.5 cm group (n=124, χ 2=62.109, P<0.001). In univariate analysis, the following factors were significantly associated with overall survival (OS): tumor location, differentiation grade, T stage, N stage, TNM stage, adjuvant therapy, preoperative tumor volume, length and maximum diameter (all P<0.05). Furthermore, multivariate Cox regression analysis showed that differentiation grade (HR=0.514, 95%CI: 0.366-0.723, P=0.019), TNM stage (HR=1.757, 95%CI: 1.267-2.612, P=0.015), adjuvant therapy (HR=0.669, 95%CI: 0.503-0.889, P=0.006), preoperative tumor volume (set <32 cm 3 as the dummy variable, 32-72 cm 3: HR=3.689, 95%CI: 2.415-5.637, P<0.001; >72 cm 3: HR=5.720, 95%CI: 3.606-9.075, P<0.001) were independent risk factors for OS. Finally, the C-index of OS by nomogram incorporated the statistically significant clinicopathological parameters was predicted to be 0.722 (95%CI: 0.687-0.757), which was significantly higher than the 7th AJCC TNM stage, the C-index 0.633 (95%CI: 0.595-0.671). In addition, the calibration curve of nomogram model was highly consistent with actual observation for the five-year OS rate, and the decision curve analysis also showed that nomogram model had higher clinical application potentials than TNM staging model in predicting survival prognosis of thoracic ESCC after surgery. Conclusion The nomogram incorporated preoperative tumor volume is of great value in predicting survival prognosis of patients with thoracic ESCC.
