Quantitative CT Predictors of Portal Venous Intervention in Uncontrolled Variceal Bleeding

OBJECTIVE. The purpose of this study was to quantify abdominal CT predictors of endoscopically refractory, uncontrolled variceal hemorrhage requiring portal venous intervention. MATERIALS AND METHODS. From 2009 to 2018, 64 patients with endoscopically refractory variceal hemorrhage requiring portal venous intervention (variceal hemorrhage group) and 67 patients without hemorrhage but with symptomatic, pressure gradient-proven portal hypertension (control group) underwent CT. CT scans were retrospectively reviewed for the following: varix size, variceal intraluminal protrusion, liver and spleen volumes, and portal vein diameter. RESULTS. Gastric variceal protrusion was found to be a strong CT parameter associated with refractory hemorrhage (mean depth, 0.75 mm in variceal hemorrhage group vs -2.91 mm in control group; p = 0.001). Gastric varix size was also associated with variceal hemorrhage (mean diameter, 8.03 vs 6.51 mm; p = 0.001). However, this trend was not observed in the sizes of the esophageal varices (mean diameter, 6.28 vs 6.43 mm; p = 0.370). Larger spleen volume (mean, 1312 vs 1152 cm³; p = 0.029) and liver volume (mean, 1514 vs 1143 cm³; p = 0.004) were also found to be predictors of variceal hemorrhage. Significant CT threshold findings included gastric variceal protrusion depth more than 0 mm (odds ratio [OR], 6.44), gastric varix size more than 6 mm (OR, 3.89), spleen volume more than 1000 cm³ (OR, 2.63), and liver volume more than 1000 cm³ (OR, 2.82). CONCLUSION. Quantitative imaging parameters on abdominal CT, such as intraluminal protrusion of gastric varices, gastric varix size, and larger spleen and liver volumes, were predictive of portal venous intervention, whereas esophageal varix size was not.