Spectral CT Imaging-Based Quantification of Iodized Oil Retention following Chemoembolization: Phantom and Animal Studies

Purpose: To evaluate accuracy of iodine quantification using spectral CT and the potential of quantitative iodized oil analysis as an imaging biomarker of chemoembolization. Materials and Methods: A phantom of an artificial liver with 6 artificial tumors containing different amounts of iodized oil (0-8 vol %) was scanned by spectral CT, and iodized oil density (mg/mL) and Hounsfield unit (HU) values were measured. In addition, VX2 hepatoma was induced in 23 rabbits. After chemoembolization using iodized oil chemoemulsion, the rabbits were scanned by spectral CT. The accumulation of iodized oil in the tumor was quantified in terms of iodized oil density and HUs, and the performances in predicting a pathologic complete response (CR) were evaluated by receiver operating characteristic curve analyses. Results: The mean difference between true iodine densities and spectral image-based measurements was 0.5 mg/mL. Mean HU values were highly correlated with mean iodine density (r(2) = 1.000, P < .001). In the animal study, a pathologic CR was observed in 17 of 23 rabbits (73.9%). The range of area under the curve values of iodine and HU measurements was 0.863-0.882. A tumoral iodine density of 3.57 mg/mL, which corresponds to 0.7 vol% iodized oil in the tumor, predicted a pathologic CR with a sensitivity of 70.6% and a specificity of 100.0%. Conclusions: Spectral CT imaging has a potential to predict tumor responses after chemoembolization by quantitatively assessing iodized oil in targets.