Abstract P4-01-06: Evaluation of 3D T2-weighted Breast MRI

Abstract

Abstract Background: Although the dynamic contrast enhanced (DCE) sequence has long been considered the most important sequence to characterize benign and malignant breast processes with diagnostic breast MRI, the T2-weighted sequence can serve as a critical adjunctive sequence and improve the discriminating power of the interpreting radiologist. However, conventional T2-weighted acquisitions have inherent timing requirements that limit the resolution available in comparison to DCE images. Recently, 3D T2-weighted sequences have been developed that utilize modulated flip angle schedules and parallel imaging to provide higher resolution T2-weighted images in clinically feasible scan times (Hennig, J, et al., MRM 2003, Mugler, JP, et al., Radiology 2000, Busse, RF, et al., MRM 2006). While utilization of modulated flip angles allows for improved acquisition efficiency, contrast is potentially less purely T2-weighted. The goal of this study is to provide an initial assessment of a 3D T2-weighted acquisition (CUBE) (Busse, RF, et al., MRM 2006) in comparison to a conventional 2D Fast Spin Echo (FSE) in the breast Methods: Twenty-seven patients undergoing clinical MRI were scanned at 3T with both FSE and CUBE sequences as part of their exam. A total of 16 lesions were identified based on pathology (9 lesions) or on contrast-enhanced image assessment and/or stability from prior studies (7 lesions). Six lesions were malignant (all IDC) and 10 were benign (5 fibroadenomas, 4 cysts, 1 papilloma). Scan time (5 minutes) and in-plane resolution were equivalent between the CUBE and FSE images while slice thickness in CUBE (2 mm) was half that of FSE (4 mm). Lesion-to-fibroglandular tissue signal ratios (SL/SF) were calculated for both FSE and CUBE; SL is lesion signal and SF is fibroglandular tissue signal. The signal ratios were assessed with ordinary least squares linear regression. Along with lesion signal intensity with respect to surrounding tissue, depiction of lesion morphology on T2-weighted images can also contribute to differential diagnosis in breast MRI. A radiologist with breast MRI expertise evaluated the resolution difference between the two sequences based on the depiction of lesion morphology and the alignment of lesions between the T2-weighted and DCE images Results: SL/SF showed a correlation coefficient between the two methods of 0.93. Mean values of SL/SF for different methods and lesions were, FSE malignant: 1.01 ± 0.21, FSE benign: 2.05 ± 0.65, CUBE malignant 0.99 ± 0.25, CUBE benign 2.16 ± 0.62. Depiction of lesion morphology and signal intensity in small lesions improved in the CUBE images due to the reduced partial volume effect in the slice direction. Alignment of structures between T2-weighted and DCE images was facilitated with CUBE due to the higher through-plane resolution and the ability to reformat the images in orientations other than the acquisition plane. Conclusion: CUBE provides equivalent contrast and improved resolution in the breast in comparison to FSE. While T2-weighted images will continue to be an adjunct to DCE images, 3D T2-weighted sequences like CUBE have the potential to expand the contribution of T2-weighted images to the characterization of benign and malignant lesions in the breast. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P4-01-06.