Comparison of a CEA-EIA assay based on monoclonal antibody with a CEA-RIA assay with polyclonal antiserum

A monoclonal CEA-EIA assay is evaluated with respect to clinically pertinent data. Comparison is done with the conventional CEA-RIA assay (Roche). Good interlaboratory reproducibility was found, and the stability was very good over the one year evaluation period. The EIA assay could be performed in samples of serum and plasma with compatible results. The correlation between the EIA and RIA values was different in different diagnostic groups, with high correlation in colo-rectal cancer, and low in non-malignant diseases, in which the EIA assay had a lower frequency of CEA positive values. In colo-rectal cancer the RIA assay shows a 20% specificity improvement compared with the EIA assay. This was also reflected in better predictability for true positive and true negative cancer diagnosis in this group of patients as well as increased ability to discriminate between malignant and non-malignant diseases. In other groups of patients, like lung cancer and uterine cervical cancer, such improvement was not seen. The discrimination between malignant and non-malignant diseases was comparable to that of the RIA assay. In follow-up series the EIA and RIA assays detected recurrences and responses to treatment in a quite similar way. In most cases of recurrences from colo-rectal cancer, however, the EIA values increased faster and were a better indicator for recurrent disease than the RIA values.