[Diagnostic validity of CEA determination in metastasizing breast cancer]

The diagnostic validity of serial CEA determinations in metastatic breast cancer was investigated. First, the CEA values within 8 weeks after start of therapy were correlated with the response to therapy. Second, the CEA levels were used to predict progression after remission or stable disease. These investigations were performed in 150 patients with advanced breast cancer who had clinical follow-ups and serial CEA determinations every one to three months. CEA was not useful for monitoring response to therapy (sensitivity 63%, specifity 58%) or prediction of relapse (sensitivity 61%, specifity 82%) if CEA levels were correlated with clinical course in all patients. However, diagnostic validity of CEA was achieved if the patients were selected and appropriate definitions of significant changes in CEA used. Thus, 83% of the responders (sensitivity) could be identified by a significant decrease of CEA titers in patients with CEA levels of greater than or equal to 10 ng/ml. A decrease of more than 10% of pretreatment levels during the first 4-8 weeks after start of therapy proved to be the appropriate definition of a significant decrease of CEA titers. However, 32% of the nonresponders were misclassified as responders (unspecifity) using these criteria. The positive predictive value of a significant decrease of CEA for response to therapy was 72% (prevalence 45%), the negative predictive value 82% (prevalence 55%). Rising CEA titers specifically predicted progression of disease in patients with remission or stable disease. However, an appropriate sensitivity (86%) was achieved only in patients with baseline CEA levels of greater than or equal to 5 ng/ml.(ABSTRACT TRUNCATED AT 250 WORDS)