Computed tomography during hepatic arteriography pattern may predict hepatocellular carcinoma recurrence following transarterial chemoembolization

Abstract

This study aimed to determine the role of morphological patterns seen on imaging in predicting hepatocellular carcinoma recurrence following transarterial chemoembolization therapy. Forty-seven patients from a single center who underwent transarterial chemoembolization to treat unresectable hepatocellular carcinomas between January 2011 and June 2012 were included in this study. We investigated whether the two pretreatment findings on computed tomography during hepatic arteriography (pattern 1, the single nodule pattern; pattern 2, at least one nodule showing the contiguous multinodular pattern) and other factors (age, sex, etiology, serum total bilirubin, serum albumin, prothrombin time, platelet count, serum level of protein induced by vitamin K absence/antagonist-II, serum α-fetoprotein, number of previous treatments for hepatocellular carcinoma, tumor number and maximum tumor size, presence of hypovascular lesions) could predict post-treatment recurrence. In a univariate analysis using Cox's proportional hazards model, serum total bilirubin, the serum level of protein induced by vitamin K absence/antagonist-II (≤100 vs ≥101 mAU/mL), tumor morphology (pattern 1 vs 2) and tumor number (≤3 vs ≥4) showed statistical significance (≤0.05). In a multivariate analysis of these factors, morphology and tumor number showed significance. According to Kaplan-Meier estimation, the cumulative disease-free survival rates were significantly lower in patients with four or more lesions than in those with three or less lesions and in patients showing pattern 2 than in those showing pattern 1. Patients with pattern 2 hepatocellular carcinoma and/or four or more lesions may have a relatively high recurrence rate after transarterial chemoembolization.