Performance of CEA and CA19-9 in identifying pleural effusions caused by specific malignancies.

ABSTRACT

OBJECTIVES: Tumor markers analysis has been proposed as a less invasive alternative for categorizing malignant and non-malignant pleural effusions. This study establishes diagnostic cutoffs for CEA and CA19-9 in pleural fluid to differentiate etiologies of effusions. DESIGN AND METHODS: Pleural effusions obtained from 198 patients (100 malignant, 98 non-malignant) were analyzed for CEA and CA19-9. ROC curve analysis was performed to determine analyte cutoffs, and cutoff performance was examined in various subsets of malignancies. RESULTS: CEA and CA19-9 concentrations were significantly higher in malignant effusions compared to non-malignant effusions, particularly in effusions caused by lung cancer and other cancers associated with elevated CEA and/or CA19-9 serum concentrations. Concentrations were not elevated in effusions caused by cancers not associated with serum tumor marker elevations. CONCLUSIONS: Measurement of CEA and CA19-9 in pleural fluid complements cytology and other classifying tests. Performance is specifically enhanced in effusions caused by malignancies known to secrete CEA or CA19-9, and their use should be tailored to patients suspected of having those malignancies. Routine analysis of these markers is therefore not recommended in all pleural effusions. Moreover, negative results should be correlated with serum levels to assist in the clinical interpretation.