Comparison of smear cytology and liquid-based cytology in EUS-guided FNA of pancreatic lesions: experience from a large tertiary center.

ABSTRACT

BACKGROUND AND AIMS: Studies comparing the diagnostic efficacy of liquid-based cytology (LBC) and smear cytology (SC) of pancreatic tissue sampling obtained via EUS-guided FNA (EUS-FNA) are still insufficient, mainly because results were controversial. We compared the diagnostic efficiency of LBC and SC of EUS-FNA of pancreatic lesions in one of the largest tertiary hospitals in China. METHODS: A retrospective database search (January 2015 to January 2019) was performed for patients who underwent EUS-FNA with both LBC and SC. Demographic, cytologic, and endosonographic data were collected from 819 patients; 514 cases met the inclusion criteria. Diagnostic accuracy, sensitivity, specificity, positive predictive value, and negative predictive value were compared. Rapid on-site evaluation was not available in all cases. RESULTS: Three hundred eighty-five cases (74.90%) had confirmed malignancy, and 40 cases (7.78%) confirmed benign neoplasm. Adequate tissue sampling rates showed no significant difference between the 2 groups. The sensitivity, accuracy, and negative predictive value (NPV) of LBC were higher than those of SC with statistical significance (71.4% vs 55.1%, 76.1% vs 61.6%, and 40.6% vs 27.7%, respectively). The sensitivity, accuracy, and NPV of combined SC and LBC were higher than those of LBC alone with statistical significance (83.9% vs 71.4%, 86.5% vs 76.1%, and 56.8% vs 40.6%, respectively). Multivariate analysis revealed that pancreatic neck/body/tail lesions (P = .003), solid lesions (P < .001), 22-gauge needle size (P < .001), and number of needle passage >3 (P = .041) were associated with higher diagnostic sensitivity in all participants using LBC, whereas number of needle passage >3 (P = .017) was associated with higher diagnostic sensitivity using SC. CONCLUSIONS: LBC was more accurate and sensitive than SC in EUS-FNA of pancreatic lesions with higher NPV when rapid on-site evaluation is unavailable. Pancreatic neck/body/tail lesions, solid lesions, 22-gauge needle, and more than 3 passes were associated with higher sensitivity when using LBC. Performing more than 3 passes is associated with higher sensitivity when using SC.