PURPOSE: To evaluate the usefulness of quantitative analysis of the degree of enhancement in dynamic MRimaging used to differentiate dysplastic nodule (DN) from small hepatocellular carcinoma (HCC), both of which showhigh signal intensity on T1-weighted images. MATERIALS AND METHODS: From 26 small HCCs and 71 DNs, all of whichshowed homogeneous high signal intensity on T1-weighted images among 42 patients with liver cirrhosis, weselected 16 small HCCs and 10 DNs of more than 1cm in diameter which were diagnosed by biopsy and follow-up imaging. Dynamic MR imaging of the entire liver was obtained using the breath-hold technique at postinjection 10sec. (phase 1), 35 sec. (phase 2), 60 sec. (phase 3), and 5 min. (delayed) after intravenous manual injection ofGd-DTPA (0.1 mmol/kg) at a velocity of 3-4 cc/sec. Nodule-to-liver contrast-to-noise ratios (CNR) during eachphase were calculated by measurement of the region of interest. RESULTS: On precontrast T1-weighted images, themean CNR of small HCCs was 2.873, and that of DNs was 3.854, there was thus no significant statistical difference(p>0.01). On postcontrast images, the CNR of small HCCs during each phase was 5.565, 3.790, 1.704, and 1.282, withpeak CNR phase 1 and a mostly decreasing trend thereafter. However, the CNR of DNs during each phase was 3.053,1.561, 0.919, and 1.038 ; there was thus showed no significant increase during phase 1 in comparison with the CNRsseen on precontrast images. During the precontrast stage and phase 1, the average difference in CNR was 2.691 forsmall HCCs and 0.801 for DNs the difference was thus significant (p<0.01). CONCLUSION: Quantitative analysis ofCNR, reflecting the degree of nodule-to-liver enhancement in dynamic MR imaging, was found to be useful for thedifferentiation of small HCCs from DNs, both of which show high signal intensity on T1-weighted images.