PURPOSE: To evaluate the image quality of thin-section MSCT examinations of low-contrast objects such as liver lesions, and to find out whether a thin slice technique requires an increase in radiation dose. MATERIALS AND METHODS: MSCT examinations of the liver were performed in the portal venous phase on a Volume Zoom Scanner (Siemens, Erlangen) with a collimation of 4 x 2.5 mm, 120 kV, 125 mAs and table speed of 12.5 mm. Forty small hypodense liver lesions with a mean size of 3 mm (1 to 8 mm) were reconstructed using a slice thickness of 3, 5, 7, and 10 mm. All lesions were evaluated for each slice thickness by 3 independent radiologists using a 5-point scale (excellent to poor) for the following 7 criteria: contrast, demarcation and contour of the lesion, image noise and quality, demarcation of liver veins and liver contour. The mean values of the reviewers' scores were calculated. The correlation of the three observers in the evaluation of the criteria was determined using the Kendall's Tau coefficient. RESULTS: The 3-mm thin sections were excellent in the evaluation of lesion detection, lesion contrast and lesion contour with a mean score of 1.4 compared to 4.1 for 10-mm sections. Concerning the criteria image quality and liver contour, thin sections achieved the best results in our series. Image noise was pronounced in thin sections but did not affect negatively the image analysis. Interrater agreement was 0.53 for the criterium image noise. CONCLUSION: Thin sections in MSCT examinations of low contrast objects such as liver lesions do not require an increase in radiation dose because the increase in image noise is compensated by improved lesion contrast.
