Diagnostic value of serous effusion cell morphology and count combined with adenosine deaminase test in tuberculous serous membrane inflammation

Objective To explore the diagnostic value of serous effusion cell morphology and count combined with adenosine deaminase (ADA) test for tuberculous serous membrane inflammation, and evaluate its application by receiver operating characteristic (ROC) curve analysis and Logistic regression. Methods A retrospective analysis was performed on 328 clinical cases (including 177 cases in the tuberculosis group and 151 cases in the non-tuberculosis group) from June 2018 to June 2019. All standard cytology and ADA tests were performed, Logistic regression model was established, and ROC curve was plotted to evaluate its predictive efficacy. Results Indicators that were eventually incorporated into the model included the number of nucleated cells (×106/L), lymphocyte percentage (%), the number of lymphocytes (×106/L), mesothelial cell percentage (%), mesothelial cell number (×106/L), ADA (mg/L), abnormal lymphocytes or lymphoid cells, and nuclear heterogeneous mesothelial cells. Logistic regression analysis showed that the area under the ROC curve of combined detection was 0.971 [95% CI (0.955-0.986)], and the sensitivity and specificity were 90.0% and 94.0%, respectively, all of which were higher than those of single detection. Conclusion Cell morphology combined with ADA detection can predict tuberculous serous membrane inflammation with high accuracy and clinical practicability. Key words: Tuberculous serous membrane inflammation; Cytology; Adenosine deaminase; Logistic models