Optimal techniques for intravenous administration of contrast material in computed tomographic (CT) examination of the liver remain controversial mainly because of inaccurate methods of data evaluation. This study tested the reproducibility of hepatic contrast material enhancement within the same patient and the accuracy of two methods of data analysis. Fourteen patients received identical CT examinations. Attenuation of liver and vessels was measured before and after administration of contrast material, and enhancement was calculated for each image. Data were analyzed by averaging enhancement values obtained over 5- and 10-second intervals and by using a new computer program that creates an enhancement-time curve and measures area under the curve (AUC) for any specific time interval. Data were compared between patients' first and second examinations. Mean liver enhancement was nearly identical at all time intervals with the AUC technique. Analysis of interpatient data showed marked variability of all parameters at all measured time intervals, suggesting that results from comparative studies within the same patient may be more accurate than those for studies of two different treatment groups. The AUC technique is preferable to other methods of data analysis for contrast enhancement evaluation.