We evaluated the usefulness of CO2 US angiography in the detectability of and the effectiveness of TAE and/or PEIT for hepatocellular carcinoma (HCC). Twenty-three patients with HCC underwent CO2 angiography during the interventional procedure to treat HCC after examination of CT and conventional US. CO2 US angiography was observed on the US monitor by injecting CO2 microbubbles through a catheter placed in the hepatic artery. Contrast materials for CO2 US angiography were 3 ml of CO2 microbubbles prepared by vigorously mixing 3 ml of normal saline with 3 ml of 20% Intralipid, 3 ml of 20% albumin or 3 ml of the patient's own blood. In all patients, CO2 US angiography revealed equal or superior tumor detectability as compared with CT, conventional US and angiography. For demonstrating the inner structure of HCC, the image of CO2 microbubbles mixed with Intralipid was better than that of CO2 microbubbles mixed with albumin. In 9 of 23 patients, CO2 US angiography depicted nodules that had not been seen in the other images. TAE was performed in 21 patients with HCC who showed hypervascularity. In one patient in whom it was difficult to clearly depict the small lesion of HCC by conventional angiography and US, PEIT was successful under CO2 US angiography. The detectability of HCC was higher in CO2 US angiography than in CT, conventional US or angiography. The distribution of blood supply to HCC was observed easily by CO2 US angiography. In TAE of HCC, CO2 US angiography was useful to determine the dose of embolization materials without having to perform repeated angiography. It was possible to perform PEIT easily for non-detectable tumors without CO2 US angiography. CO2 US angiography was useful to evaluate the stage of HCC and to perform TAE and PEIT.