PURPOSE:
Better prognostic predictors in colorectal cancer than the Dukes stage are necessary for individualized therapy and follow-up.
METHODS:
Survival among 212 patients operated on for colorectal cancer was examined regarding various clinical, histopathologic, cellular, and serologic tumor characteristics.
RESULTS:
Beside the Dukes stage, which was the most powerful variable, the erythrocyte sedimentation rate, leukocyte blood count, alkaline phosphatase, aspartate aminotransferase, six different serum tumor markers, number of small blood vessels, and age were found to be significantly associated with survival. The leukocyte blood count, alkaline phosphatase, and aspartate aminotransferase retained their significance in a multivariate model including tumor differentiation, local tumor stage, and age. Inclusion of tissue polypeptide antigen, the most powerful tumor marker in the multivariate model, showed that only the tumor stage, tissue polypeptide antigen, and age were statistically significantly correlated to survival. This was valid both for the group of patients considered as potentially curable and for those who potentially have been cured (Dukes Stages A-C).
CONCLUSIONS:
A great number of prognostic predictors failed to discard Dukes stage as the best one. One serum tumor marker, tissue polypeptide antigen, contains independent additional prognostic information.
