Role Of Different Non-invasive Methods For High Risk Varices In Patients With Cirrhosis

Objective: The purpose of this study is to explore the predictive value of different non-invasive methods in high risk varices(HRV)with cirrhosis.Methods: We recruited 105 patients with cirrhosis who attended the First Hospital of Lanzhou University from November 2019 to October 2020 and used two-dimensional shear wave elastography(2D-SWE)to measure spleen stiffness(SS)and liver stiffness(LS).These patients also received esophagogastroduodenoscopy(EGD)and laboratory examinations.HRV was defined as esophageal varices ≥ 5 mm or esophageal varices< 5 mm but advent of red spots or Child-Pugh class C.The risk factors affecting the occurrence of HRV were analyzed by binary logistic regression.Receiver operating characteristic(ROC)curves were established to predict the optimal threshold of LS、liver stiffness-spleen size-to-platelet(LSPS)based on2D-SWE and SS and their diagnostic efficacy for HRV.Results:1.Three cases failed to measure SS due to small spleen volume.Among 102 patients with cirrhosis,57 patients(55.9%)with compensated cirrhosis and 45patients(44.1%)with decompensated cirrhosis.EGD revealed 85 cases(83.3%)of gastroesophageal varices(GOV),of which 66 cases(64.7%)of HRV.2.Comparison between HRV and non-HRV groups: The mean platelets of the HRV group was significantly lower than that of the non-HRV group,and the spleen thickness,spleen long diameter and SS were significantly higher than those of the non-HRV group,the difference was statistically significant(P<0.001).In addition,the PT,INR,portal vein diameter,MELD score and LS of the HRV group were higher than those of the non-HRV group,and albumin was lower than that of the non-HRV group(P<0.05).3.Multivariate analysis of binary logistic regression showed that SS is an independent risk factor for HRV(P<0.001).Multivariate Analysis of Binary Logistic Regression4.Taking LS was 12.2 KPa as the optimal threshold to predict HRV,the area under the ROC curve(AUC)was 0.634,and the sensitivity 75.76%,specificity50.00%,positive predictive value 73.53%,negative predictive value 52.94% and accuracy 66.67%,respectively.5.Taking LSPS based on 2D-SWE was 2.68 as the optimal threshold to predict HRV,the AUC was 0.794,and the sensitivity 78.79%,specificity 69.44%,positive predictive value 82.54%,negative predictive value 64.10% and accuracy 75.49%,respectively.6.Taking SS was 37.6 KPa as the optimal threshold to predict HRV,the AUC was 0.922,and the sensitivity 90.91%,specificity 91.67%,positive predictive value95.24%,negative predictive value 84.62% and accuracy 91.18%,respectively.Conclusion:1.SS has high value in predicting HRV in patients with cirrhosis.The diagnostic efficacy of SS in predicting HRV with cirrhosis is higher than that of LS and LSPS based on 2D-SWE.2.The diagnostic efficacy of LSPS based on 2D-SWE in predicting HRV with cirrhosis is higher than that of LS.