Fat suppression has shown promise in improving the quality of T2-weighted spin-echo MR images of the upper part of the abdomen. The purpose of this study was to determine whether fat-suppressed images should be routinely used in lieu of conventional images. Accordingly, we prospectively compared the two techniques in a series of patients with both normal and abnormal findings in the upper part of the abdomen. Conventional and fat-suppressed T2-weighted spin-echo images (3000/80,160 [TR/TE]) were obtained in 45 consecutive patients referred for MR imaging of the upper part of the abdomen. Thirty-three patients had abnormal findings, and 22 of those 33 patients had histologic or follow-up confirmation of the diagnosis (14 with metastasis, one with hepatoma, four with hemangiomas, and three with cysts). Signal intensities (hepatic lesions, liver, spleen) and noise were measured to calculate signal-to-noise ratios and contrast-to-noise ratios. Qualitative comparison (liver, hepatic lesions, porta hepatis, spleen, pancreas, bowel, kidneys, adrenal glands, noise), evaluation of the number of hepatic lesions, and characterization of hepatic lesions were done by independent observers. Compared with conventional images, fat-suppressed images had higher signal-to-noise ratios (lesions, liver, spleen) and contrast-to-noise ratios (lesion-liver and spleen-liver) (p < .005). In qualitative comparison, three of three radiologists preferred fat-suppressed over conventional images for depiction of hepatic lesions and all upper abdominal organs except the liver, for which no clear preference was shown for either technique. Detection rates for hepatic lesions were similar with both types of images (observer 1: 112 lesions on fat-suppressed vs 118 on conventional images, observer 2: 142 vs 135), as was the characterization of hepatic lesions (91% accuracy on fat-suppressed images and 84% accuracy on conventional images, for 22 proved lesions and two observers). Fat-suppressed T2-weighted spin-echo MR images were better than non-fat-suppressed images for evaluation of the upper part of the abdomen. These results suggest that fat suppression should be routinely used in T2-weighted MR imaging of the upper part of the abdomen.