Detection of focal hepatic lesions with 3-T MRI: comparison of two-dimensional and three-dimensional T2-weighted sequences

Abstract

To evaluate whether incorporation of a 3D turbo spin-echo sequence during T2-weighted MR imaging improves the detection of focal hepatic lesions by 3T MR imaging. Seventy-nine consecutive patients including 67 patients with 62 malignant and 71 benign lesions and 12 patients having no hepatic lesion underwent respiratory-triggered fat-suppressed axial T2-weighted turbo spin-echo imaging using two-dimensional (2D-TSE) and 3D (3D-TSE) sequences. Coronal multiplanar reformatted images (MPR-3D-TSE) were generated from 3D-TSE images. Breath-hold fat-suppressed 2D axial T2-weighted half-Fourier turbo spin-echo (HF-2D-TSE) images were combined for reading. Two independent radiologists reviewed three imaging sets, (1) 2D-TSE and HF-2D-TSE, (2) 3D-TSE and HF-2D-TSE, and (3) 3D-TSE, HF-2D-TSE and MPR-3D-TSE, for detection of malignant and benign lesions. Lesion-to-liver contrast ratio (CR) and the conspicuity of anatomical boundaries were assessed. For benign lesions, lesion-to-liver CRs with 3D-TSE (2.77 ± 1.91, p = .0002) and MPR-3D-TSE (2.47 ± 1.42, p = .012) were higher than with 2D-TSE (2.13 ± 1.80). Sensitivity for lesions of ≤ 10-mm was higher with 3D-TSE (86 %, p = .0039) and MPR-3D-TSE (84 %, p = .0078) than with 2D-TSE (72 %). However, the edge of left lateral lobe was less conspicuous with 3D-TSE (p < .0001) and MPR-3D-TSE (p = .0003) than with 2D-TSE because of susceptibility artifacts. Incorporation of 3D T2-weighted sequence may incrementally improve the detection of focal hepatic lesions.