Sequence design and evaluation of the reproducibility of water selective diffusion weighted imaging of the breast at 3T

摘要

Diffusion measurements derived from breast MRI can be adversely affected by unwanted signals from abundant fatty tissues if they are not suppressed adequately. To minimize this undesired contribution, we design and optimize a water selective DWI sequence, which relies on spectrally selective excitation on the water resonance, obviating the need for fat suppression. As this method is more complex than standard DWI methods, we also report a test-retest study to evaluate its reproducibility. In this study, a spectrally selective Gaussian pulse on water resonance is combined with a pair of slice selective adiabatic refocusing pulses for water only DWI. Field-map based shimming and manual determination of the center frequency are used for water selection. The selectivity of the excitation pulse is optimized by a spectrally selective spectroscopy sequence based on the same principles. A test-retest study of 10 volunteers in two separate visits is used to evaluate its reproducibility. Our results from all subjects show high quality diffusion weighted images of the breast without fat contamination. Mean ADC values for b = 0, 600 s/mm² and b = 50, 600 s/mm² all show good reproducibility as 95% confidence intervals (CI) of the apparent diffusion coefficient values are 4×10⁻⁵ mm²/s and 5×10⁻⁵ mm²/s and repeatability values are 1.09×10⁻⁴ and 1.31×10⁻⁴, respectively. In conclusion, water selective DWI is a feasible alternative to standard methods of DWI based on fat suppression. The added complexity of the method does not compromise the reproducibility of diffusion measurements in the breast.