High spectral and spatial resolution (HiSS) - ismrm

2 downloads 0 Views 22KB Size Report
Synopsis: High spectral and spatial resolution (HiSS) MR imaging has been used to improve the quality of anatomic and functional images. HiSS datasets ...
Images of the Fourier components of the water resonance in breast G. S. Karczmar1, M. Medved2, W. Du2, X. Fan2, M. Zamora2, P. MacEneaney2, Y. Du3, F. Kelcz4, G. Newstead2 1

University of Chicago, Chicago, United States, 2University of Chicago, Chicago, Illinois, United States, 3University of Colorado, Denver, Colorado, United States, 4 University of Wisconsin, Madison, Wisconsin, United States

Synopsis: High spectral and spatial resolution (HiSS) MR imaging has been used to improve the quality of anatomic and functional images. HiSS datasets provide detailed water and fat spectra associated with each voxel. In many voxels, distinct components of the water and fat lineshapes can be identified. Here we demonstrate that in breast, the different Fourier components of the water resonance and the fat resonance give rise to qualitatively different images. Introduction: High spectral and spatial resolution (HiSS) imaging in various forms has been used to improve the quality of anatomic and functional imaging (1-3). HiSS datasets from the breast provide a detailed water and fat spectrum associated with each small image voxel. These data demonstrate that water and fat resonances in small voxels are often inhomogeneously broadened (3,4). It is likely that the different components of water and fat signals come from distinct subvoxel environments and reflect local anatomy and physiology. Here, we demonstrate that in human breast the different Fourier components of the water and fat resonances from very small voxels give rise to images with significantly different features. Methods: HiSS data were acquired from women with suspicious breast lesions or normal breasts at 1.5 Tesla. Inplane spatial resolution was ~750 microns in 3 mm thick slices, and spectral resolution was ~2.6 Hz. Signals from water (or fat) in each image voxel were shifted so that the peak of the water (or fat) resonance was at the same resonance frequency in all voxels. Then images were calculated with intensity proportional to the Fourier components of the water (or fat) resonance at different offsets from the peak frequency. Results: FCIs (Fourier component images) often contained distinct anatomic features. For example, Figure 1 shows FCI’s acquired from a patient with infiltrating ductal carcinoma. The images are labeled with the corresponding frequency offset from the main peak. The FCI at -21 Hz shows a bright elongated area in the upper right quadrant of the lesion. A small bright area near the lesion center is clearly depicted in the 16 Hz image but not at the other frequencies. The texture at 0 Hz is relatively smooth with a few dark lines, while at 10 Hz and 16 Hz there is a more complicated pattern of bright and dark lines, perhaps arising from vasculature. Many features that are prominent in the FCI at 0 Hz have a narrow bandwidth and are not detectable in other FCI’s. Discussion: It is reasonable to suggest that different Fourier components of the water resonance produce different images because they arise from different -21 Hz -10 Hz 0 Hz 16 Hz 30 Hz components of the subvoxel environment. A. B. C. D. E. As a result each Fourier component Figure 1 represents a particular aspect of the anatomy and physiology within each voxel. Thus HiSS MRI may increase sensitivity to details of the subvoxel environment. This is a novel contrast mechanism and may yield information that cannot be duplicated with other methods of contrast enhancement. Acknowledments: Supported by the NCI and the Army Breast Cancer Research Program. The authors are grateful to Dr. Alan Koretsky and Dr. Charles Springer for helpful advice. References: 1) Kovar DA, et al., Acad Radiol 5:69-75, 1998 2) Sarkar S, et al., J Magn Reson Imaging 10: 1-7, 1999 3) Du W, et al., Radiology, 224: 577-85, 2002 4) Al-Hallaq HA, et al., NMR Biomed 15, 28-36,2002

Proc. Intl. Soc. Mag. Reson. Med. 11 (2003)

534