The Role of Liver and Spleen Elastography in the Screening of Esophageal Varices in Patients with Liver Cirrhosis: Do We Really Need Endoscopy?

Background: recently, the Baveno VI guidelines have suggested that esophagogastroduodenoscopy (EGD) to stage esophageal varices can be avoided in patients with advanced liver disease who have a liver stiffness (LS) <20 kPa and a platelet count >150x103/μL. Our study aims to analyze spleen stiffness (SS) as a non‐invasive method of diagnosis for clinically significant portal hypertension in order to avoid EGD in low‐risk patients for esophageal varices. We also want to compare the SS to other non‐invasive techniques and analyze their reproducibility and inter‐observer concordance. Patients and Methods: in this prospective study, we detected the SS and LS in 150 patients diagnosed with liver cirrhosis to be submitted for endoscopic screening for esophageal varices. In addition, we enrolled 70 healthy control individuals, who for other reasons, had undergone an endoscopic examination of the upper digestive tracts and who were negative for hepatic and lymphoproliferative disease. The discriminatory capacity for the presence of esophageal varices of the SS has been compared with that deriving from other non‐invasive procedures (LS, splenic diameter, splenic surface, platelet count, and combined scores deriving from these parameters). Optimal SS cut‐offs were sought to exclude the presence of varices. Particular emphasis was placed on the search for possible correlations of the with ultrasound parameters of portal hypertension and platelet count. Finally, we studied in a double‐blind fashion inter‐operator concordance with 50 measurements for LS, and 25 for SS. Results: cirrhotic patients have significantly higher SS and LS values than controls. The SS values were higher in cirrhotic patients with varices (n = 62) compared to patients without esophageal varices (n = 88) and to healthy controls (p <0.001). SS showed an AUROC of 0.93 (95% C.I., 0.89‐0.97), statistically different from the other predictors (p <0.001). The cut‐off, chosen according to Youden’s Index and equal to 38.55 kPa, showed sensitivity of 87%, specificity of 89%, NPV of 91%, and PPV of 84%. Instead, the cut‐off of 30.79 kPa has 100% sensitivity and 100% NPV. The cut‐off of 69.73 kPa demonstrated specificity and PPV of 100%. The SS demonstrated weak linear correlation with the splenic dimensions (bipolar diameter and surface measured at the organ’s hilum). Moreover, it has a linear correlation with the platelet count, which was greater than that present with LS (r = 0.5 vs r = 0.32). The test revealed an excellent intraclass correlation coefficient (ICC) equal to 0.96 for SS and 0.97 for LS. Conclusion: the results of this study show how the SS can play an important role in the daily clinical management of cirrhotic patients. The SS (alone or combined with other indicators) may play an important role as a non‐invasive screening test for predicting the risk of varices.