A prognostic score for colorectal cancer.

ABSTRACT

BACKGROUND/AIMS: Scoring systems are generally used for predicting prognosis in critically ill patients, but not frequently for predicting prognosis in cancer. Our aim was to develop a prognostic score for colorectal cancer. METHODOLOGY: Demographic, clinical, laboratory, radiologic, histopathologic and operative data of 112 patients who had resection for colorectal cancer, were analyzed for their effect on survival. Eight variables (invasion depth of tumor, nodal status, presence of metastasis, CEA level, differentiation, resectability, tumor location, presence of blood transfusion), which have the most significant effect on survival in univariate analysis and multivariate analysis, were chosen. Coefficients were calculated and the Prognostic Score for Colorectal Cancer (PSCRC) was designed. All patients were scored using this system and staged according to TNM, Dukes and Astler-Coller systems. RESULTS: The score range is between -49 and 80. Patients were grouped as Group 1 (score < -30), Group 2 (score -30 to -6), Group 3 (score -5 to 19), Group 4 (score > 19). Survival rates decreased and recurrence rates increased significantly with increasing scores. The associations of PSCRC and staging systems with survival and recurrence rates were determined by logistic and Cox regression analyses. PSCRC was proved to have the most significant association with survival and recurrence rates. Furthermore, PSCRC had better correlation with survival and recurrence rates when compared with staging systems according to Pearson correlation analysis. CONCLUSIONS: Inclusion of more variables in PSCRC seems to make it superior to the staging systems. PSCRC may be a practical and useful scoring system in clinical practice and is easily adapted to different patient populations.