Sampling bandwidth and fat suppression: effects on long TR/TE MR imaging of the abdomen and pelvis at 1.5 T

In MR imaging, the sampling bandwidth is the rate at which the signal is digitized by the analog-to-digital converter. Reducing the sampling bandwidth can decrease noise in long TE images at the expense of increases in the artifacts of chemical-shift misregistration and motion. We compared 39 pairs of axial images with bandwidths of 32 kHz and 8-10 kHz. In 23 of these comparison studies (six female pelvis, seven male pelvis, 10 abdomen), all other variables were held constant, and in 16 comparisons (10 female pelvis, six male pelvis), signal from fat was suppressed in images with reduced bandwidth. Six patients with 11 liver lesions were included in those undergoing abdominal imaging. In three patients, fat was suppressed in images of the abdomen. The contrast-to-noise ratio was greater with reduced bandwidth for urine vs fat (24.5 vs 17.2; p less than .05) and central vs peripheral parts of the prostate gland (34.1 vs 22.7; p less than .02). In the abdomen, the contrast-to-noise ratio was increased between liver and right kidney (36.6 vs 25.2; p less than .01) and between liver and lesion (30.2 vs 18.2; p less than .005), but the motion-induced artifact was worse. An increase in chemical-shift misregistration did not affect the appearance of the internal structure of the uterus, prostate gland, or liver, but it made examination of the ovaries, seminal vesicles, and extrahepatic tissues difficult. The chemical-shift artifact in pelvic images could be eliminated by suppressing signal from fat with frequency-selective saturation pulses, but results were less satisfactory in the abdomen. When reduced bandwidth and fat suppression were combined, the contrast-to-noise ratio was improved for endometrium vs myometrium (22.9 vs 15.2; p less than .05) and central vs peripheral parts of the prostate gland (62.7 vs 28.5; p less than .02). Reduction of the sampling bandwidth is a promising technique for imaging the pelvis with small field of view and long TR/TE, but it currently appears less promising for images in the upper abdomen at 1.5 T. The increased chemical-shift artifact caused by reduced bandwidth can be eliminated by suppressing signal from fat.