Keyhole Dixon Method for Faster, Perceptually Equivalent Fat Suppression

The acquisition time associated with the 2-point Dixon fat suppression technique was reduced by combining a keyhole in-phase ("Water+Fat") k-space data set with a full out-of-phase ("Water-Fat") k-space data set. A set of keyhole Dixon images was created by varying the number of lines in the keyhole data set. Off-resonance correction was incorporated into the image reconstruction process to improve the homogeneity of the fat suppression. A perceptual difference model (PDM) was validated with human observer experiments and used to compare the keyhole images to images from a full 2-pt Dixon acquisition. The PDM was used to determine the smallest keyhole width required to obtain perceptual equivalence to images obtained from the full 2-pt Dixon method. In experimental phantom studies, the keyhole Dixon image reconstructed from 96 of 192 "Water+Fat" k-space lines and 192 "Water-Fat" k-space lines was perceptually equivalent to the full (192+192) 2-pt Dixon images, resulting in a 25% reduction in scan time. Clinical images of a volunteer's knee, orbits, and abdomen created from the smallest, perceptually equivalent keyhole width resulted in a 27-38% reduction in total scan time. This method improves the temporal efficiency of the conventional 2-point Dixon technique and may prove especially useful for high-field systems where SAR limits will constrain RF-based fat suppression techniques.

Reconstructed keyhole Dixon images of a volunteer's knee with a) 40, b) 70, c) 100 and d) 192 line keyholes (2-point Dixon).