PURPOSE: To compare various breath-hold T2 weighted sequences in imaging normal pancreas with a phased-arraycoil. MATERIALS AND METHODS: Eighteen patients without pancreatic disease were studied with breath-hold turbo spinecho (TSE) (TR/TE/ETL, 3500/138/29), TSE with fat suppression (FS-TSE), half-Fourier acquisition single-shot turbospin echo (HASTE) (TE/ETL, 87/128) and HASTE with fat suppression (FS-HASTE) at 1.0T magnet strength and using aphased-array coil. Signal difference-to-noise ratio (SD/N) between the pancreas and peripancreatic fat wasmeasured ; the delineation of the pancreatic border and pancreatic duct, and the amount of artifact were evaluatedby two radiologists who reached a consensus. RESULTS: HASTE showed a higher SD/N than TSE or FS-HASTE (p < 0.01),TSE was superior to FS- TSE or HASTE in the delineation of pancreatic border(p < 0.001). HASTE was superior to TSEin the delineation of pancreatic duct(p < 0.001). TSE showed more artifacts than FS-TSE(p < 0.001) ; HASTE andFS-HASTE showed no artifact. CONCLUSION: TSE is better than HASTE for the delineation of pancreatic margin but HASTE shows less artifacts and a more conspicuous pancreatic duct. Fat suppression decreases artifacts but makes the pancreatic margin indistinct.