LAUSR

AIM To determine whether acoustic radiation force imaging (ARFI) of the liver/spleen could be used in patients with cirrhosis to predict the presence of gastroesophageal varices (GOVs). MATERIALS AND METHODS Fifty-eight patients with cirrhosis who were undergoing 6-monthly ultrasound examinations for hepatoma surveillance and who were due to have oesophagogastroduodenoscopy (OGD) within 6 months of their ultrasound were recruited. During routine ultrasound, the patient's liver and spleen were also assessed using ARFI. Other clinical parameters (platelet count, spleen size, and transient elastography measurements) were also collected. Logistic regression was used to determine which variables were significantly associated with presence or absence of varices univariably and multivariably RESULTS: Fourteen patients (24%) had GOVs. Patients with GOVs had higher ARFI measurements in the liver and spleen than patients without GOVs (liver: 2.39 versus 2.13, spleen: 2.89 versus 2.82), but these results were not statistically significant (odds ratio=1.75, 95% confidence interval [CI]=0.82, 3.91 and odds ratio=1.12, 95% CI=0.33, 3.97, respectively). The platelet/splenic ratio, in comparison, was associated with the presence or absence of GOVs in multivariate analysis (odds ratio=0.32, 95% CI=0.008, 0.91). CONCLUSION Although patients with GOVs had overall higher ARFI liver and spleen results, this was not statistically significant. As such, ARFI cannot yet replace OGD in predicting GOVs in this patient group.