This retrospective study examined the computed tomography (CT) criteria for judging the effectiveness of transcatheter arterial Lipiodol-chemoembolization (Lp-chemo-TAE) in 35 cases with hepatocellular carcinoma (HCC). Massive necrosis, defined as involving 97% or more of the HCC nodule, was observed in 15 cases after Lp-chemo-TAE, whereas nonmassive necrosis, defined as involving < or = 96% of the HCC nodule, was observed in the remaining 20 cases. In 12 of 15 cases (80%) with massive necrosis, uniform dense retention of Lipiodol (Lp) was observed throughout the HCC nodule on CT images 3-4 weeks after Lp-chemo-TAE as opposed to only one (5%) of 20 cases with nonmassive necrosis (p < 0.01). Eight of nine cases (89%) with massive necrosis had tumor attenuation values of 365 Hounsfield units (HU) or greater on CT images 3-4 weeks after embolization, as opposed to only four (27%) of 15 cases with nonmassive necrosis (p < 0.01). We conclude that the effectiveness of the Lp-chemo-TAE can be judged on CT from the degree and duration of Lp retention in the HCC nodule and the measurement of the attenuation value of the HCC nodule.