Abstract Background Pleural effusions present a diagnostic challenge. Approximately 20% are associated with cancer and some 50% require invasive procedures to perform diagnosis. Determination of tumour markers may help to identify patients with malignant effusions. Two strategies are used to obtain high specificity in the differential diagnosis of malignant pleural effusions: a) high cut-off, and b) fluid/serum (F/S) ratio and low cut-off. The aim of this study is to compare these two strategies and to establish whether the identification of possible false positives using benign biomarkers – ADA, CRP and % of polymorphonuclear cells – improves diagnostic accuracy. Methods We studied 402 pleural effusions, 122 of them malignant. Benign biomarkers were determined in pleural fluid, and CEA, CA72-4, CA19-9 and CA15-3 in pleural fluid and serum. Results Establishing a cut-off value for each TM for a specificity of 100%, a joint sensitivity of 66.5% was obtained. With the F/S strategy and low cut-off points, sensitivity was 77% and specificity 98.2%, Subclassifying cases with negative benign biomarkers, both strategies achieved a specificity of 100%; sensitivity was 69.9% for single determination and 80.6% for F/S ratio. Conclusions The best interpretation of TM in the differential diagnosis of malignant pleural effusions is obtained using the F/S ratio in the group with negative benign biomarkers.
Find this article in full text from Springer Verlag. 
Full Text Finder