Comparison of two CEA assays in primary and recurrent large bowel carcinoma with different DNA ploidy pattern.

Pre-operative CEA levels were measured in 100 patients with large bowel carcinomas with different DNA ploidy pattern and serial post-operative determinations performed in the 64 who had been operated for cure. The follow-up period was 3 1/2-8 yr. All CEA measurements were performed consecutively with a RIA (Roche), and subsequently repeated in one batch with an EIA (Roche) based on a monoclonal antibody. Both assays showed a similar number of 'false-negative' CEA levels pre-operatively--varying from 69% in aneuploid (AN) Dukes' A to 8% in AN Dukes' D tumours, and from 75% in near diploid (ND) Dukes' A to 40% in ND Dukes' D tumours. The sensitivity for detecting recurrence in patients with tumours of either ploidy pattern was slightly better with EIA than with RIA. A difference between the AN and ND group was shown somewhat better with RIA, the sensitivity in the AN group being 79% and the median lead time 7 months compared to 13% and 2 months in the ND group. The corresponding figures with EIA were 71% and 7 months for the AN group and 63% and 1 1/2 months for the ND group. However, all but one of the patients with ND DNA pattern who showed recurrence-associated CEA elevation with EIA also had an elevated level pre-operatively. We conclude that all patients operated for cure should be followed by regular CEA measurements post-operatively if they had an elevated CEA level prior to operation.(ABSTRACT TRUNCATED AT 250 WORDS)