T2-weighted MRI of the uterus: Fast spin echo vs. breath-hold fast spin echo

This study compared one routine T2-weighted fast spin echo (T2FSE) sequence with a breath-hold T2FSE (BH T2FSE) sequence of the female pelvis for image quality, uterine anatomy, lesion detection, and signal intensity measurements. Thirty-two consecutive women (mean age 41.7 years) were imaged at 1.5 T with one high-resolution routine T2FSE sequence and one BH T2FSE sequence in the sagittal plane as part of comprehensive pelvic magnetic resonance imaging. The different image sets were rated separately for imaging characteristics (overall image quality, uterine anatomy definition, lesion detection, and free fluid conspicuity) and then compared side by side. The image sets were also compared for artifacts (ghosting, blurring, pulsatility, and chemical shift misregistration). Signal-to-noise (S/N) and signal difference-to-noise (SD/N) ratios were calculated for the different uterine zones, uterine abnormalities, free fluid, rectus abdominis muscle, and bladder. Contrast-to-noise ratios (CNRs) were calculated for uterine abnormalities. Twenty-eight uterine abnormalities were detected in 20 patients and included leiomyomata (13 patients), adenomyosis (7 patients), benign endometrial polyps (6 patients), endometrial carcinoma (1 patient), and pregnancy (1 patient). BH T2FSE was superior or equivalent to T2FSE for overall image quality in 23/32 patients (71.8%), uterine anatomy definition in 19/32 patients (59.3%), and lesion detection in 13/20 patients (65%). BH T2FSE performed less well than T2FSE for free fluid conspicuity in 5/5 (100%) patients. BH T2FSE was equivalent to or less affected than T2FSE for ghosting artifact in 24/32 patients (75%) and blurring artifact in 29/32 patients (90.6%). Pulsatility and chemical shift artifacts were not problematic for either image set. S/N and SD/N were higher for all BH T2FSE determinations compared with T2FSE. For the endometrium, junctional zone, myometrium, and bladder, these differences were statistically significant. There were no statistically significant differences for CNR between the two image sets, although BH T2FSE values for leiomyomata, adenomyosis, and abnormal endometria were higher than those calculated for T2FSE. All pathology detected with T2FSE was detected on BH T2FSE despite the breath-hold sequence's inherently poorer spatial resolution compared with the non-breath-hold sequence. BH T2FSE may be able to replace T2FSE for some uterine applications with a substantial time savings. J. Magn. Reson. Imaging 1999;9:384–390. © 1999 Wiley-Liss, Inc.