EUS elastography for diagnosis of solid pancreatic masses: a meta-analysis.

BACKGROUND The differential diagnosis of solid pancreatic masses has been a significant challenge up until now. EUS elastography is now used as a new technique to distinguish benign from malignant pancreatic masses. However, the sensitivity, specificity, and accuracy are still questionable. OBJECTIVE To evaluate the accuracy of EUS elastography for diagnosis of solid pancreatic masses. DESIGN Thirteen articles for EUS elastography diagnosing solid pancreatic masses were selected. The Mantel-Haenszel and DerSimonian Laird methods were used to analyze pooled results. SETTING PATIENTS This study involved 1044 patients. INTERVENTION EUS elastography. MAIN OUTCOME MEASUREMENTS The pooled sensitivity, specificity, likelihood ratios, diagnostic odds ratio, and summary receiver operating characteristic (sROC) curve. RESULTS The pooled sensitivity, specificity, and diagnostic odds ratio of EUS elastography distinguishing benign from malignant solid pancreatic masses were 0.95 (95% confidence interval [CI], 0.94-0.97), 0.67 (95% CI, 0.61-0.73), and 42.28 (95% CI, 26.90-66.46), respectively. The sROC area under the curve was 0.9046. The subgroup analysis based on excluding the outliers showed that the heterogeneity was eliminated, and the pooled sensitivity and specificity were 0.95 (95% CI, 0.93-0.97) and 0.7 (95% CI, 0.63-0.76), respectively. The sROC area under the curve was 0.8872. LIMITATIONS Varied diagnostic standards for EUS elastography were used in the enrolled studies. CONCLUSION EUS elastography is a reliable technique for the characterization of solid pancreatic masses and may be a useful complementary tool for EUS-guided FNA. However, a more accurate computer-aided diagnosis method for EUS elastography is in demand to reduce various biases and improve the accuracy of EUS elastography for diagnosis of solid pancreatic masses.