DETECTION IN RECURRENT BREAST-CANCER BY C A 15-3 AND CEA - INFLUENCE OF LOCALIZATION AND EXTENT OF METASTASES

This retrospective study was designed in order to evaluate sensitivity , specifity, and accuracy of CA 15-3 and CEA serum levels as well as t heir combination in the follow-up of 2224 breast cancer patients (595 with proven metastases) depending on die localization and extent of me tastatic diseases. At an equivalent 97 % specificity, a higher sensiti vity of 60.2 % was determined for CA 15-3 in comparison with 47.9 % fo r CEA (p < 0.01). The combined determination of both marker levels inc reased the sensitivity by 11.4 %. The significantly higher sensitivity of CA 15-3 compared to CEA was only seen in patients with bone metast ases but not in patients with visceral or pulmonal metastases or in lo coregional recurrence. The sensitivity of CA 15-3 in bone metastases w as higher for extensive as well as for minimal metastatic spread. The sensitivity of CEA, CA 15-3, and their combination in 365 patients was significantly lower for the detection of a locoregional recurrence (1 8.2 %, 21.2 %, 31.8 %, respectively) compared with patients with isola ted distant metastases. With the exception of CEA levels in lung metas tases the levels of CEA and Ca 15-3 were significantly lower compared to other localizations of distant metastases. The sensitivity was 11.9 %, 28.6 %, and 33.3 % in 299 patients with minimal spread of metastas es into only one organ and correlated well with the extent of sensitiv ities (34.2 %, 39.5 %, 57.9 %) in patients with moderate metastatic sp read. Significantly highest sensitivities (54.8 %, 66.7 %, 78.5 %) wer e found for patients with extensive metastatic spread. This correlatio n between the sensitivity and the extent of the metastatic spread was also confirmed in 215 patients with metastases limited to the bone. In this group the lowest sensitivities (10.3 %, 34.5 %, 37.9 %) were fou nd in minimal metastatic disease. The correlation of the serum levels of CEA with the extent of metastatic spread was lower compared to CA 1 5-3. Therefore, CA 15-3 seems to be more useful for the determination of the metastatic spread. The low sensitivity of CA 15-3 and CEA for t he detection of locoregional recurrence and minimal metastatic spread (30-40 %) is a limiting factor for their use in follow-up. For the det ection or the rule-out of beginning distant metastasis, especially of bone metastases, these sensitivity levels seem to be too low.