The present study aimed to compare the diagnostic accuracy between conventional smear (CS) and liquid-based preparation (LBP) in endoscopic ultrasonography-fine needle aspiration cytology (EUS-FNAC) of pancreatic lesions. Using 31 eligible studies, the diagnostic accuracy of cytologic examination in CS and LBP was evaluated through a conventional meta-analysis and diagnostic test accuracy review. Overall concordance rates were 82.8% (95% confidence interval [CI], 79.8-85.5%) and 94.0% (95% CI, 84.4-97.8%) in CS and LBP, respectively. CS with rapid on-site evaluation (ROSE) showed a higher concordance rate than CS without ROSE. In CS, the pooled sensitivity and specificity were 89.8% (95% CI, 85.2-93.1%) and 95.0% (95% CI, 90.0-97.6%), respectively. The diagnostic odds ratio (OR) and area under curve (AUC) of the summary receiver operating characteristic (SROC) curve were 90.32 (95% CI, 43.85-147.11) and 0.945, respectively. In LBP, the pooled sensitivity and specificity were 80.9% (95% CI, 69.7-88.7%) and 99.9% (95% CI, 1.5-100.0%), respectively. The diagnostic OR and AUC of the SROC curve were 57.21 (95% CI, 23.61-138.64) and 0.939, respectively. Higher concordance rates were found in CS with ROSE and LBP in EUS-FNAC of pancreatic lesions. Regardless of the cytologic preparation method, EUS-FNAC is a useful and accurate diagnostic tool for pancreatic lesions.
Keywords: conventional smear; diagnostic test accuracy review; endoscopic ultrasonography-fine needle aspiration cytology; liquid-based preparation; meta-analysis; pancreas.