PURPOSE: To optimize MR imaging pulse sequences in the imaging of hepatic hemangioma and to evaluate on dynamic MR imaging the enhancing characteristics of the lesions. MATERIALS AND METHODS: Twenty patients with 35 hemangiomas were studied by using Turbo-spin-echo (TSE) sequence (T2-weighted, T2- and heavily T2-weighted breath-hold) and T1-weighted FLASH imaging acquired before, immediately on, and 1, 3 and 5 minutes after injection of a bolus of Gd-DTPA (0.1 mmol/kg). Phased-array multicoil was employed. Images were quantitatively analyzed for lesion-to-liver signal difference to noise ratios (SD/Ns), and lesion-to-liver signal ratios (H/Ls), and qualitatively analyzed for lesion conspicuity. The enhancing characteristics of the hemangiomas were described by measuring the change of signal intensity as a curve in T1-weighted FLASH dynamic imaging. RESULTS: ForT2-weighted images, breath-hold T2-weighted TSE had a slightly higher SD/N than other pulse sequences, but there was no statistical difference in three fast pulse sequences (p=0.211). For lesion conspicuity, heavily T2-weighted breath-hold TSE images was superior to T2-weighted breath-hold or non-breath-hold TSE (H/L, 5.75, 3.81, 2.87, respectively, p<0.05). T2-weighted breath-hold TSE imaging was more effective than T2-weighted TSE imaging in removing lesion blurring or lack of sharpness, and there was a 12-fold decrease in acquisition time (20 sec versus 245 sec). T1-weighted FLASH dynamic images of normal liver showed peak enhancement at less than 1 minute, and of hemangioma at more than 3 minutes; the degree of enhancement for hemangioma decreased after a 3 minute delay. CONCLUSION: T2-weighted breath-hold TSE imaging and Gd-DTPA enhanced FLASH dynamic imaging with 5 minutes delay are sufficient for imaging hepatic hemangiomas.