Non-breath-hold fast spin-echo versus breath-hold fast spin-echo and spoiled gradient-recalled echo MR imaging in the detection of hepatic tumors: correlation with surgical findings.

Abstract

Our objective was to compare the sensitivity of non-breath-hold T2-weighted fast spin-echo with and without fat suppression, breath-hold T2-weighted fast spin-echo without fat suppression, and spoiled gradient-recalled echo (GRE) MR imaging for detecting hepatic tumors using surgical findings as the standard of reference. Eighteen patients with 36 surgically proven hepatic tumors had non-breath-hold T2-weighted fast spin-echo (6000/117 [TR/effective TE; echo train length, 16) MR imaging with and without fat suppression, breath-hold T2-weighted fast spin-echo MR imaging (2700/105; echo train length, 20), and spoiled GRE images (10.1/1.9; flip angle, 30 degrees) obtained before and after injection of a gadolinium chelate. Images were analyzed separately by two independent readers, with disagreements resolved by consensus reading. Non-breath-hold T2-weighted fast spin-echo MR imaging with and without fat suppression depicted 22 (61%; 95% confidence interval [CI], 43-77%) and 20 (56%; 95% CI, 37-72%) of 36 hepatic tumors, respectively. Breath-hold T2-weighted fast spin-echo imaging allowed detection of 19 (53%; 95% CI, 35-69%) of 36 hepatic tumors. Unenhanced and gadolinium chelate-enhanced spoiled GRE images allowed depiction of 18 (50%; 95% CI, 33-67%) and 29 (81%; 95% CI, 63-91%) of 36 hepatic tumors, respectively. Gadolinium chelate-enhanced spoiled GRE images allowed depiction of significantly more hepatic tumors than any of the other pulse sequences. Gadolinium chelate-enhanced spoiled GRE imaging is more sensitive than T2-weighted MR imaging obtained with a breath-hold or a non-breath-hold technique.