Purpose: The aim of the present study was to propose and validate a standardized CT protocol for evaluating all the types of portosystemic collaterals (P-SC), including gastroesophageal varices and spontaneous portosystemic shunts (SPSS), and to evaluate the prognostic role of portal hypertension CT features for the prediction of the hepatic decompensation risk in cirrhotic patients.
Methods: A retrospective cohort study of 184 advanced chronic liver disease who underwent CT scan between January 2014 and December 2017. Patients with an interval > 6 months between the imaging, elastometric, endoscopic and biochemical evaluation were excluded, as well as patients with previous transjugular intrahepatic portosystemic shunt (TIPS), liver transplantation (LT) or terminal medical conditions. Data on liver disease history, co-morbidities, endoscopic and radiologic findings were collected. The incidence of hepatic decompensation and other events, such as portal vein thrombosis, HCC, TIPS placement, LT, death, and its cause, were also recorded. The procedure was performed at baseline and after the administration of contrast agent using a multiphasic technique and bolus tracking. Two senior radiologists working in different centres and a non-expert radiologist reviewed all CT examinations, to evaluate both intra-observer and inter-observer variability of the CT protocol and to obtain an external validation. The radiological variables were evaluated using both univariate and adjusted multivariate competing risk regression models.
Results: Both intra-observer and inter-observer agreement were excellent in detection and measurement of almost all types of P-SC. The presence of SPSS, a spleen diameter > 16 cm, a portal vein diameter > 17 mm and the presence of ascites resulted independent predictors of decompensation-free survival for cirrhotic patients and were incorporated in an easy-to-use score (AUROC = 0.799, p-value = 0.732) which can the risk of decompensation at 5 years, ranking it as low (11.3%), moderate (35.6%) or high (70.8%).
Conclusions: The CT protocol commonly performed during the HCC surveillance program for cirrhotic patients is valid for detecting all types of P-SC. The radiological score identified to predict the decompensation-free survival for cirrhotic patients could be an easy-to-use clinical tool.