High expression of serum carcinoembryonic antigen (CEA) associated with poor prognosis in early gastric cancer: a single-center retrospective study

Background: The level of carcinoembryonic antigen (CEA) in the diagnosis of early gastric cancer (EGC) seems obscure. This study is aimed to assess whether a high level of CEA is associated with the inferior prognosis for EGC. Methods: About 203 EGC patients, who received endoscopic mucosal resection or submucosal dissection or radical gastrectomy, were reviewed retrospectively. All the EGC cases were evaluated based on their clinicopathological features and surgical outcomes. The X-tile program was used to calculate the optimal cut-off points for the CEA using minimum P-value from log-rank ÷ 2 statistics. In the analysis of the overall cumulative probability of survival, the Kaplan-Meier method was employed. Their differences were evaluated through the log-rank test. The Cox multiple factors analysis was carried out using the logistic regression method. Results: The X-tile plots cut-off points for CEA were 15.54 ng/ml in EGC patients. They were classified as CEA-low and CEA-high groups. The percentage of the vessel carcinoma embolus was higher in CEA-high groups compared to their CEA-low counterparts (20% vs. 16.29%, P = 0.000). The five-year overall survival rate (OS) of the patients under CEA-high group was markedly inferior compared to the CEA-low group (68.95% vs. 99.42%, P < 0.05). The multivariate survival analysis showed that CEA (OR = 1.674), N stage (OR = 2.436) were significant prognostic factors for EGC (all P < 0.05). In the CEA-low subgroup, not signally risk factors were found (all P > 0.05), while N stage (OR = 2.632) was an independent risk factor in the CEA-high group by multivariate analysis (P < 0.05). Conclusion: The CEA, categorized by the cut-off points of 15.44 ng/ml could develop the best prognostic discriminatory ability and predictive accuracy for the EGC patients. It could be a reliable prognostic factor when combined with tumor node metastasis (TNM) evaluation system.