The Value Of CT Radiomics In Predicting High Risk Esophageal Varices In Patients With Liver Cirrhosis

Objection Esophageal varices are a common complication in patients with cirrhosis and can develop into esophageal variceal bleeding,which is one of the leading causes of death in patients with cirrhosis.Many guidelines recommend the use of upper gastrointestinal endoscopy to detect the presence and assess the severity of esophageal varices,but endoscopy is an invasive procedure that some patients cannot tolerate,and endoscopy is not suitable for routine assessment of disease progression,frequent monitoring,and evaluation of therapeutic effect.Therefore,non-invasive detection methods with high accuracy and sensitivity are urgently needed in clinic,as a supplement or even replacement for endoscopy.This study aims to investigate whether the radiomics features of the spleen can predict high-risk esophageal varices in patients with cirrhosis,and construct a non-invasive prediction model for high-risk esophageal varices in patients with cirrhosis by combining the radiomics features,imaging indicators and serological indicators of the spleen,so as to supplement the endoscopic evaluation of the severity of esophageal varices in patients with cirrhosis.To provide reference for early prediction and intensive monitoring of high-risk esophageal varices in patients with cirrhosis.Methods A retrospective analysis was performed on 190 patients diagnosed with cirrhosis in the Second Affiliated Hospital of Anhui Medical University,,who were divided into high-risk group(142 cases)and low-risk group(48 cases)according to endoscopic examination results,and randomly divided into training set(n=133)and verification set(n=57)in a ratio of 7:3.Serological indicators of patients with cirrhosis after admission were collected,and portal phase abdominal enhanced CT images of the study population were collected.A cross-section of the splenic portal region was delineated along the organ margin as an area of interest,and its radiomics features were extracted.Feature correlation analysis and minimum absolute contraction and selection operator(LASSO)regression analysis were used to screen the features.Radiomics signature(RS)was established and Rad-score was calculated.Logistic regression analysis was used to construct a clinical model,and a nomogram was constructed by combining Rad-score,serological and imaging indexes.The predictive efficacy of the three models was evaluated using receiver operating characteristic curve(ROC)curves.The area under the curve(AUC)was compared using Delong test.Delong test results showed that the difference between the clinical model and the combined model was statistically significant,with Z value and P value being-2.53 and 0.011 in the training set and-1.98 and 0.047 in the validation set,respectively.The clinical practicability of the combined models was evaluated by decision curve analysis(DCA).Results A total of 1652 radiomics features were extracted from each ROI.The AUC of the constructed radiomics signature,clinical model and combination model in the training set were 0.94,0.92 and 0.96,respectively.The AUCs in the verification set were 0.79,0.91 and 0.94,respectively.In the optimal combination model,only Rad-score and left gastric vein diameter were risk factors for high-risk esophageal varices.The decision curve shows that the combination model has good clinical practicability.Conclusion in this study,the model constructed based on the radiomics,serological indexes and imaging indexes is helpful to predict patients with esophageal varices at high risk of cirrhosis.