Effect of injection rate of contrast material on CT of hepatocellular carcinoma

The purpose of this study was to assess the effect of two injection rates of contrast material (3 mL/sec and 5 mL/sec) in hepatic arterial dominant phase MDCT for the detection of small (< 2 cm) hepatocellular carcinomas.

The injection rates 3 mL/sec and 5 mL/sec were used prospectively in imaging examinations of patients with the suspected diagnosis of hepatocellular carcinoma. Each group consisted of 30 patients by chance. The group that received injections at 3 mL/sec had 35 hepatocellular carcinoma lesions, and the 5 mL/sec group had 41 lesions. In all patients the dose and concentration of contrast material were 100 mL and 350 mg/mL iodine (total dose of iodine, 35 g). In each patient a mini-test-bolus technique was used with an additional 15 mL of contrast material to determine optimal scan delay after administration of contrast material. Receiver operating characteristic analysis was used to assess diagnostic performance with the two injection rates of contrast material.

There were no statistically significant differences between the two groups in regard to area under the curve and sensitivity. These values for the 3 mL/sec group were 0.97 and 28/35 (80%) and for the 5 mL/sec group were 0.96 and 36/41 (88%). However, the specificity and positive predictive values at 3 mL/sec (236/250 [95%] and 28/42 [67%]) were significantly higher than those at 5 mL/sec (227/265 [86%] and 36/73 [49%]) (p < 0.05). These results suggest there were more false-positive findings of contrast-enhanced lesions in cirrhotic livers on hepatic arterial dominant phase images obtained after injection of contrast material at 5 mL/sec than on images obtained after injection at 3 mL/sec.

In the detection of small hypervascular hepatocellular carcinoma in cirrhotic liver, the risk of false-positive findings of lesions on hepatic arterial dominant phase images is significantly greater with the higher injection rate (5 mL/sec) than with the medium rate (3 mL/sec).