Evaluation of hepatocellular carcinoma after treatment with transcatheter arterial chemoembolization: comparison of Lipiodol-CT, power Doppler sonography, and dynamic MRI

BACKGROUND: The purpose of this study was to compare the ability of Lipiodol-computed tomography (CT), power Doppler (PD) sonography, and dynamic magnetic resonance imaging (MRI) in evaluating the therapeutic effect of transcatheter arterial chemoembolization (TACE) on hepatocellular carcinoma (HCC). METHODS: TACE was performed by injecting an emulsion consisting of Lipiodol and a chemotherapeutic drug, followed by gelatin sponge particles, into 54 patients with 84 HCC lesions. Five to 7 days later, Lipiodol-CT, PD sonography, and dynamic MRI were performed. Findings from the three modalities were correlated with relapse within 1 year after TACE. RESULTS: All lesions with blood flow on PD sonography or intratumoral enhancement on dynamic MRI relapsed regardless of the findings with Lipiodol-CT. None of the negatively enhanced lesions on dynamic MRI relapsed regardless of the Lipiodol-CT findings. However, the readers could not evaluate the contrast uptake in 14 lesions that were already hyperintense on the precontrast images. These cases were considered unsuitable for qualitative assessment and reduced the applicability of MRI to 83% of the examined lesions (70 of 84). Although PD sonography perfectly predicted relapse in superficial (0-5 cm from abdominal surface) lesions of the right hepatic lobe, blood flow in deep (>5 cm) or left lobe lesions was undetectable regardless of the occurrence of relapse. As a result, Lipiodol-CT displayed 76.0% sensitivity, 67.6% specificity, and 72.6% accuracy. The sensitivity, specificity, and accuracy of PD sonography were 34.0%, 100%, and 60.7%, respectively. In the 70 lesions in which evaluation was possible, dynamic MRI achieved 100% sensitivity, 100% specificity, and 100% accuracy. CONCLUSION: Of the three modalities, dynamic MRI was the best for evaluating the efficacy of TACE in the treatment of HCC. We also found that superficial lesions of the right lobe are good candidates for PD sonography. However, high signals on precontrast MR images, motion artifacts, and ultrasonic attenuation remain key limitations.