Spectral CT-Based Iodized Oil Quantification to Predict Tumor Response Following Chemoembolization of Hepatocellular Carcinoma

Purpose: To quantify iodized oil retention in tumors after transarterial chemoembolization using spectral computed tomography (CT) imaging in patients with hepatocellular carcinoma (HCC) and evaluate its performance in predicting 12-month tumor responses. Materials and Methods: From September 2017 to December 2018, 111 patients with HCC underwent initial conventional transarterial chemoembolization. Immediately after the procedure, unenhanced CT was performed using a spectral CT scanner, and the iodized oil densities in index tumors were measured. In tumor-level analyses, a threshold level of iodized oil density in the tumors was calculated using clustered receiver operating characteristic curve analyses to predict the 12-month tumor responses. In patient-level analyses, significant factors associated with a 12-month complete response, including the presence of tumors below the threshold value (ie, suspected residual tumors), were evaluated by logistic regression. Results: Forty-eight HCCs in 39 patients were included in the analyses. The lower 10th percentile of the iodine density was identified as the threshold for determining the 12-month nonviable responses. The area under the curve of the iodine density measurements in predicting the 12-month nonviable responses was 0.893 (95% confidence interval, 0.797–0.989). The threshold value of the iodine density of 10.68 mg/mL yielded a sensitivity of 82.76% and specificity of 94.74% (P <.001). In the patient-level analysis, the 12-month complete response was significantly associated with the presence of a suspected residual tumor, with an odds ratio of 72.0 (95% confidence interval, 7.273–712.770). Conclusions: Spectral CT imaging using quantitative analysis of the iodized oil retention in target HCCs can predict tumor responses after a conventional transarterial chemoembolization procedure.