FOCAL HEPATIC LESION DETECTION - COMPARISON OF 4 T2-WEIGHTED MR-IMAGING PULSE SEQUENCES

PURPOSE: To evaluate T2-weighted magnetic resonance (MR) imaging with conventional spin-echo, breath-hold fast spin-echo, respiratory-trigge red fast spin-echo, and breath-hold multishot spin-echo echo-planar se quences for detection of focal hepatic lesions. MATERIALS AND METHODS: T2-weighted MR images obtained with the four sequences in 56 patients with 107 solid and 124 nonsolid lesions were retrospectively analyzed . Image review was conducted on a liver segment-by-segment basis; a to tal of 408 liver segments were reviewed separately and independently f or solid and nonsolid lesions by three radiologists. Diagnostic accura cy was evaluated with receiver operating characteristic analysis. RESU LTS: Lesions-to-liver contrast-to-noise ratio was highest with multish ot echo-planar images of both solid and nonsolid lesions. Diagnostic a ccuracy for solid lesions was statistically significantly better with conventional spin-echo images than with any other type of image (P<.00 01). Diagnostic accuracy for nonsolid lesions was statistically signif icantly better with respiratory-triggered fast spin-echo images than w ith any other type of image (P<.0001). Image quality was best with bre ath-hold fast spin-echo images. CONCLUSION: Conventional spin-echo MR imaging should not be replaced with breath-hold fast spin-echo or mult ishot spin-echo echo-planar imaging, despite the shorter acquisition t imes that are possible with the latter two sequences.