Background/aims: Preoperative CEA levels, depth of tumor penetration, and the number of positive lymph nodes were reported as independent factors prognostic of survival in colorectal cancer patients. This study was carried out in an effort to evaluate the prognostic significance of these three factors in patients with Dukes' C colorectal cancer in Taiwan.
Methodology: Between 1992 and 1994, a total of 112 patients with node-positive colorectal cancer were evaluated retrospectively at the Veteran General Hospital-Taipei. All patients underwent potentially curative surgery and received 5-fluorouracil based adjuvant chemotherapy. Reference to the Dukes' classification was according to the classical criteria described in 1932 for carcinoma of the rectum and adapted for use in colonic tumors. Data on the location of the tumor, depth of penetration, number of positive lymph nodes, degree of tumor differentiation, and preoperative CEA levels were analyzed to understand their association with survival. Blood samples for CEA measurement were taken a few days before operation. A multivariate analysis using the Cox's proportional hazards regression model was then performed to determine the most important independent predictors of survival among all the possible variables.
Results: Using univariate analysis the number of positive lymph nodes (P < 0.001), penetration of the bowel wall (P < 0.001), and preoperative CEA levels (P < 0.001) were found as significant prognostic factors, while the degree of tumor differentiation, location of the tumor, age and sex were not significant. Using multivariate Cox analysis the number of positive lymph nodes, penetration of the bowel wall, and preoperative CEA levels were still found as independent prognostic factors in node-positive colorectal cancer patients.
Conclusions: Data obtained from our study indicates that preoperative CEA levels, depth of tumor penetration, and the number of positive lymph nodes were independent prognostic factors in Dukes' C colorectal cancer patients. They could serve as appropriate modifications of the initial Dukes scheme in node-positive diseases.