New diagnostic tools have emerged to assist the traditional diagnosis of malignant pleural effusion (MPE), such as high fluorescence cells (HFc) and tumor markers (TMs), determined by clinical laboratory automated pleural fluid workup. This study aimed to evaluate the diagnostic ability of the combination of HFc and TMs for diagnosing MPE. We recruited hospitalized patients with pleural effusion at Parc Taulí University Hospital. We collected and analyzed pleural fluid and serum samples in the clinical laboratory, and we sent a sample of pleural fluid to the Pathology Department for cytology workup. We determined the pleural fluid cell count by Sysmex XN-10 and assessed TMs (CEA, CA19.9, and CA15.3) using the ECLIA Cobas e801 Roche in both pleural fluid and serum samples. We established the final MPE diagnosis based on positive cytology and/or positive pleural biopsy. We classified patients based on these final diagnoses and conducted a comparison between variables, along with multivariate logistic regression. The study included 316 pleural effusions from 221 patients recruited. Multivariate logistic regression indicated the most significant predictor variables for MPE were CA15.3 in serum, CEA ratio, and HFc. We calculated two different models: one excluding HFc and one including it, with the latter displaying superior diagnostic ability (area under the curve 0.91). This model could identify 100 % of MPE cases with 30% specificity at low cut-offs, and higher values could help identify 60 % of MPE cases with 100 % specificity. Per our findings, this model has high diagnostic performance and could serve as a swift, automated, dependable, non-invasive tool for MPE detection.
Read full abstract