Pleural Fluid and Bronchoalveolar Lavage Cytology in Patients With Lung Lesions and Pleural Effusion: A Prospective Paired-Cytology Study

BackgroundLung lesions with pleural effusion often have overlapping clinical and radiological features that make malignancy difficult to distinguish from tuberculosis and other benign disease.Pleural fluid cytology is a minimally invasive first-line test, but its sensitivity is limited, and bronchoalveolar lavage (BAL) cytology may detect abnormalities missed on pleural fluid alone.In this exploratory, single-centre, prospective paired-cytology study, the primary objective was to compare the agreement of pleural fluid, BAL, and combined (OR-rule) cytology with a composite clinical, radiological, and microbiological reference that included the index cytology; secondary objectives were inter-test concordance and clinico-pathological associations. MethodsThis prospective observational study was conducted in the Department of Pathology of a tertiary care hospital over 18 months after ethics committee approval.Seventy patients with suspected lung lesions and pleural effusion who underwent both pleural fluid aspiration and bronchoscopy with BAL were enrolled by consecutive sampling after informed consent.Smears were stained with Papanicolaou, May-Grünwald-Giemsa, and haematoxylin and eosin; cell blocks with immunocytochemistry were prepared where indicated.A specimen was cytology-positive only when malignant cells were identified; suspicious/atypical smears were indeterminate and counted as test-negative.Combined testing used an OR rule.The composite reference diagnosis incorporated the index cytology, so results are reported as agreement estimates with Wilson 95% CIs, Cohen's kappa, Fisher's exact test, and exact McNemar tests. ResultsMean age was 53.96 ± 15.54 years, and 43/70 (61.4%) were male.Malignant disease was diagnosed in 24 (34.3%) and non-malignant disease in 46 (65.7%); tuberculosis (18/70, 25.7%) and adenocarcinoma (12/70, 17.1%) were the commonest diagnoses.Malignancy was associated with older age, male sex, and smoking (all p < 0.05).Pleural fluid and BAL cytology each showed 70.8% sensitivity (17/24; 95% CI 50.8-85.1%),100% specificity (46/46; 95% CI 92.3-100%), and 90.0%agreement (63/70; 95% CI 80.8-95.1%).The combined strategy showed 87.5% sensitivity (21/24; 95% CI 69.0-95.7%)and 95.7% agreement (67/70; 95% CI 88.1-98.5%),with overlapping confidence intervals and unchanged specificity.Exact McNemar tests were not significant (pleural vs. BAL, 4 vs. 4 discordant cases, p = 1.00; combined vs. either test, 0 vs. 4, p = 0.125).Binary concordance between the two tests was 62/70 (88.6%; kappa 0.69, 95% CI 0.49-0.89);multi-category concordance was 56/70 (80.0%; kappa 0.58).Among the 24 malignant cases, 13 (54.2%)were positive on both tests, four (16.7%) on pleural fluid only, four (16.7%) on BAL only, and three (12.5%) on neither. ConclusionsPleural fluid and BAL cytology showed comparable standalone point estimates, and their combination showed higher point estimates, but differences were not statistically significant, and the 100% specificity is largely an artifact of the composite reference design.Because the reference diagnosis was not independent of the index tests, blinding was not documented, the sample was small, and itemised diagnoses differed by three patients from the disease-group classification (21 vs. 24 malignant), these findings are exploratory and hypothesis-generating.They await multicentre validation against an independent reference standard.