To evaluate the diagnostic accuracy of KRAS mutation in pancreatic cystic fluid and compare it with carcinoembryonic antigen and cytology, we identified studies with cyst fluid obtained by endoscopic ultrasound prior to surgery. We classified cysts as malignant, premalignant, and benign. A random-effects model was used for quantitative meta-analysis. Pooled sensitivities, specificities, and summary receiver operating characteristic curve analysis were conducted. We analyzed 16 studies, with 3429 patients, including 731 referred for surgery. Carcinoembryonic antigen was better for clinically significant cysts (premalignant and malignant) with sensitivity = 0.58 (95% confidence interval [CI], 0.53-0.65), specificity = 0.9 (95% CI, 0.76-0.97), and area under the curve (AUC) = 0.69. Cytology performed better in malignant cysts, with sensitivity = 0.37 (95% CI, 0.27-0.48), specificity = 0.96 (95% CI, 0.93-0.98), and AUC = 0.78. Isolated, KRAS mutation failed the diagnosis of malignant and significant cysts, with sensitivities = 0.43 (95% CI, 0.34-0.43) and 0.46 (95% CI, 0.42-0.51), specificities = 0.62 (95% CI, 0.56-0.68) and 0.97 (95% CI, 0.92-0.99), and AUCs = 0.56 and 0.53, respectively. Carcinoembryonic antigen and cytology are more accurate than KRAS. Additional studies are lacking to recommend KRAS as a single diagnostic test.