RESUMO
OBJECTIVE: To construct a predictive model for 3-year survival of patients after curative resection of colon cancer by nomogram. METHOD: A retrospective analysis was conducted to analyze the clinicopathologic data of 102 patients who underwent radical resection of colon cancer in Baoji Central Hospital from April 2015 to April 2017. The optimal cutoff values of preoperative CEA, CA125, and NLR for predicting overall survival were analyzed by receiver operating characteristic (ROC) curves. To observe the relationship between NLR, CEA and CA125 and clinicopathologic data, we performed multivariate Cox regression to analyze the independent factors affecting the prognosis of patients, and Kaplan-Meier test to identify the relationship between NLR, CEA and CA125 and patient survival. A nomogram prediction model was drawn for patients' 1-, 2-, and 3-year survival after radical resection of colon cancer, and the efficacy of the prediction model was evaluated. RESULTS: The area under the curve (AUC) of NLR, CEA and CA125 in predicting patient death was 0.784, 0.790 and 0.771, respectively. NLR was correlated with clinical stage, tumor diameter and differentiation (all P < 0.05); CEA was associated with clinical stage, tumor diameter, differentiation and lymph node metastasis (all P < 0.05); CA125 was only associated with tumor diameter in patients (P < 0.05). Differentiation, NLR, CEA and CA125 were independent risk factors affecting the prognosis of patients (all P < 0.05). The nomogram predicted a model C-index of 0.918 (95% CI 0.885-0.952), and the risk model score was found to have a high clinical value in the 3-year survival of preexisting patients. CONCLUSION: Preoperative NLR, CEA, CA125 and clinical stage are correlated with the prognosis of patients with colon cancer. The nomogram model constructed based on NLR, CEA, CA125 and clinical stage has good accuracy.
