Multidetector CT: diagnostic impact of slice thickness on detection of hypervascular hepatocellular carcinoma.

Abstract

The objective of our study was to evaluate the diagnostic impact of varying slice thickness on multidetector CT to optimize detection of hypervascular hepatocellular carcinoma. Forty-three patients with 87 hypervascular hepatocellular carcinomas (diameter: range, 3-80 mm; mean, 22 mm) and 19 patients with either chronic hepatitis or liver cirrhosis and without hepatocellular carcinoma who had undergone early arterial and late arterial phase imaging of the entire liver on multidetector CT were retrospectively enrolled in this study. The detector row configuration was 2.5 x 4 mm, the pitch was 6, and the scanning time was 10.5 sec for each phase. All patients received contrast medium (2 mL/kg of body weight) at a rate of 5 mL/sec; the mean scanning delay for the early arterial phase was 19.0 sec, and the mean delay for the late arterial phase was 34.5 sec. Eighty 2.5-mm-thick reconstruction images, forty 5-mm-thick reconstruction images, and twenty-six 7.5-mm-thick reconstruction images were obtained for each phase. Each image set was interpreted separately by three observers to detect hypervascular hepatocellular carcinoma by viewing images on a workstation monitor. Sensitivity, positive predictive value, and area under the receiver operating characteristic curve (A(z)) were calculated. We used retrospectively excellent follow-up and imaging or pathologic proof as the gold standard. The mean sensitivity and positive predictive value for hypervascular hepatocellular carcinoma were 76% and 69% on 2.5-mm images, 73% and 69% on 5-mm images, and 67% and 76% on 7.5-mm images, respectively. No significant difference in sensitivity among the images was detected, except by one observer who reported a significant difference in the sensitivity between 2.5- and 7.5-mm images (p < 0.05) and between 5- and 7.5-mm images (p < 0.05). The mean A(z) values were 0.79, 0.80, and 0.78 for 2.5-, 5-, and 7.5-mm images, respectively. No significant difference in A(z) values among the images obtained with different slice thicknesses was detected. For multidetector CT identification of hypervascular hepatocellular carcinoma, we found little or no advantage in reducing slice thickness to less than 5 mm.