Background/aims: The usefulness of preoperative CEA in CRC remains controversial as regards its biological function, and its use in the diagnosis, prognosis, and management and follow up of CRC patients. the aim of this study was to provide a critical and updated study for the value of CEA in CRC.
Methodology: From January 2000 to June 2005, a prospective randomized study involving 200 CRC patients for whom curative resection was performed, another 100 healthy persons as a control group was included. Basal CEA using chemilumescence technique and routine follow up were done.
Results: (1) The mean basal CEA in CRC patients (17.3 ng% +/- 1.67) was significantly higher than control (3.41 ng% +/- 1.1). (2) A significant linear association between basal CEA and Dukes' classes was evident with the mean basal CEA for Dukes' A, B, C were 7.8, 12.7, 25.8 respectively (expressed as ng%). (3) The validity of basal CEA in primary CRC diagnosis was highly positive (sensitivity 80%--PPV 86.95%--accuracy 73.66%), with hig her efficacy in advanced disease detection (sensitivity 93%--NPV 7%--accuracy 84.5%--odds ratio 30.3) and negative exclusion power for DFS prediction (specificity 13.84%). (4) The basal CEA was a discriminate factor in colorectal prognosis - B value (3.74). (5) Patients with CEA < or =5 ng% had better DFS (15%) and DFT (23.6 months) than those with CEA > 5 ng% as they had DFS (33.75%) and DFT (18.48 months). (6) Basal CEA above 15 ng% had a significant shift in the cumulative hazard of recurrence.
Conclusion: The CEA is a metastasis potentiator. The high serum CEA in CRC screening programs should be considered a marker of malignancy especially in patients with appropriate symptoms. The preop CEA in CRC patients identifies subsets with favorable, indolent and uneven biological behavior (< or =5 ng%, < or =15 ng%, > 15 ng% respectively). Moreover, the addition of preop CEA level to conventional staging forms a strong prognostic tool and supplies adopted practice guideline initiative for follow up and therapy in CRC.