Optimal measurement modality and method for evaluation of responses to transarterial chemoembolization of hepatocellular carcinoma based on enhancement criteria.

PURPOSE: To determine the usefulness of enhancement by iodized oil deposits on computed tomography (CT) following transarterial chemoembolization for hepatocellular carcinoma (HCC), and to compare the reliability of such CT imaging with that of magnetic resonance (MR) imaging.
MATERIALS AND METHODS: Fifty-one patients for whom resected or explanted livers containing chemoembolized HCC lesions of at least 1 cm were available. Imaging responses were determined based on modified Response Evaluation Criteria In Solid Tumors (mRECIST) and European Association for the Study of the Liver (EASL) criteria for 59 target tumors on CT and MR scans before surgery. CT-based evaluation was performed per mRECIST and EASL criteria, considering iodized oil retention as indicating necrosis and, alternatively, as enhancing viable tissue ("mRECIST-Lipiodol" and "EASL-Lipiodol"). Pathologic necrosis was graded as 100%, 50%-99%, or less than 50%.
RESULTS: Goodman-Kruskal γ-values for radiologic-pathologic correlation were greater than 0.95 for mRECIST and EASL criteria on CT or MR imaging. However, mRECIST-Lipiodol and EASL-Lipiodol measurements showed weaker correlation with pathologic findings, with γ-values of 0.797 and 0.846, respectively. With respect to intermethod agreement, weighted γ-values for mRECIST by CT and MR, and for EASL criteria by CT and MR, both exceeded 0.80, whereas mRECIST-Lipiodol and EASL-Lipiodol showed only moderate levels of agreement with mRECIST/EASL criteria by CT or MR imaging, with γ-values of 0.522-0.631.
CONCLUSIONS: Response estimation based on measurement of iodized oil deposits as necrosis on CT when applying enhancement criteria after chemoembolization for HCC correlated well with actual pathologic class, and agreed with MR-based evaluation.