Molecular analysis of EUS-acquired pancreatic cyst fluid for KRAS and GNAS mutations for diagnosis of intraductal papillary mucinous neoplasia and mucinous cystic lesions: a systematic review and meta-analysis

Abstract

Although molecular analysis of pancreatic cyst fluid may aid pancreatic cyst classification, clinical practice remains highly variable. Therefore, we performed a systematic review and meta-analysis to evaluate the diagnostic performance of KRAS and GNAS mutations in EUS-acquired pancreatic cyst fluid for diagnosis of intraductal papillary mucinous neoplasms (IPMNs) and mucinous cystic lesions (MCLs). Individualized searches were developed in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) and Meta-Analysis of Observational Studies in Epidemiology (MOOSE) guidelines and meta-analysis analyzed according to the Cochrane Diagnostic Test Accuracy working group methodology. A bivariate model was used to compute the pooled sensitivity and specificity and to plot the summary receiver operating characteristics curve with summary point and corresponding 95% confidence interval (95% CI). Six studies (785 lesions) were included. For IPMNs and MCLs, KRAS + GNAS (combination) had significantly higher diagnostic accuracy than KRAS alone and GNAS alone (all P< .001). The pooled sensitivity, specificity, and diagnostic accuracy of KRAS + GNAS mutations for diagnosis of IPMNs were 94% (95% CI, 72-99; I2= 86.74%), 91% (95% CI, 72-98; I2= 89.83), and 97% (95% CI, 95-98), respectively, with each significantly higher compared with carcinoembryonic antigen (CEA) alone (all P< .001). For diagnosis of MCLs, KRAS + GNAS had a similar sensitivity and specificity compared with CEA alone; however, diagnostic accuracy was significantly improved (97% [95% CI, 95-98] vs 89% [95% CI, 86-91]; P< .001). Molecular analysis for KRAS + GNAS mutations in EUS-acquired pancreatic cyst fluid has high sensitivity and specificity with significantly improved diagnostic accuracy for diagnosis of IPMNs and MCLs when compared with CEA alone.