MR imaging of the liver: comparison between single-shot echo-planar and half-Fourier rapid acquisition with relaxation enhancement sequences.

Abstract

To evaluate single-shot T2-weighted magnetic resonance sequences and their role in clinical practice in patients with hepatic lesions. Prospective comparison of echo-planar and half-Fourier rapid acquisition with relaxation enhancement (RARE) imaging was performed in 80 patients with focal hepatic lesions. Spin-echo (SE) single-shot echo-planar (echo times, 47 and 80 msec) and half-Fourier RARE (echo time, 59 msec) images were compared with turbo SE (repetition time msec/echo time msec = 3,200-7,600/90) images. Quantitative, qualitative, and receiver operating characteristic (ROC) analyses were performed. For liver, signal-to-noise ratios on half-Fourier RARE images were significantly higher than those on echo-planar images (P < .01). For cystic lesions, contrast on half-Fourier RARE and echo-planar images was slightly higher than that on turbo SE images. For solid lesions, contrast on echo-planar images was better than that on half-Fourier RARE or turbo SE images. Artifacts including ghosting, bowel motion, susceptibility difference, and chemical shift were negligible on half-Fourier RARE images in all patients, whereas susceptibility difference and chemical shift of various degrees were seen on all echo-planar images. On the basis of ROC analyses, tumor detection rates were significantly higher with half-Fourier RARE and turbo SE images than with echo-planar images (P < .01). Echo-planar images provide sufficient contrast to allow detection of both solid and cystic masses, but severe artifacts preclude routine use. Half-Fourier RARE images are free from artifacts (chemical shift and susceptibility) and diagnostic performance with them is similar to that with turbo SE images.