Abstract
BACKGROUND/AIMS: Endoscopic ultrasound-guided fine needle aspiration (EUSFNA) and brushing cytology are used worldwide to diagnose pancreatic and biliary malignant tumors. Liquid-based cytology (LBC) has been developed and it is currently used to overcome the limitations of conventional smears (CS). In this study, the authors aimed to compare the diagnostic value of the CellPrepPlus (CP; Biodyne) LBC method with CS in samples obtained using EUS-FNA and brushing cytology. METHODS: This study prospectively enrolled 75 patients with pancreatic or biliary lesions from June 2012 to October 2013. For cytological analyses, including inadequate specimens, benign and atypical were further classified into benign, and suspicious and malignant were subcategorized as malignant. Sensitivity, specificity, accuracy, and positive predictive values (PPV) and negative predictive values (NPV) were evaluated. RESULTS: In the EUS-FNA based cytological analysis of pancreatic specimens, CP had a sensitivity of 60.7%; specificity, 100%; accuracy, 77.1%; PPV, 100%; and NPV, 64.5%. CS had a sensitivity of 85.7%; specificity, 100%; accuracy, 91.7%; PPV, 100%; and NPV, 83.3%. In the brushing cytology based analysis of biliary specimens, CP had sensitivity of 53.1%; specificity, 100%; accuracy, 54.5%; PPV, 100%; and NPV, 6.3%. CS had a sensitivity of 78.1%; specificity, 100%; accuracy, 78.8%; PPV, 100%; and NPV, 12.5%. CONCLUSION: Our study found that CP had a lower sensitivity because of low cellularity compared with CS. Therefore, CP (LBC) has a lower diagnostic accuracy for pancreatic EUS-FNA based and biliary brush cytology based analyses compared with CS.