OPTIMIZATION OF MRI PROTOCOLS AND PULSE SEQUENCE PARAMETERS FOR EIGENIMAGE FILTERING

The eigenimage filter generates a composite image in which a desired f eature is segmented from interfering features. The signal-to-noise rat io (SNR) of the eigenimage equals its contrast-to-noise ratio (CNR) an d is directly proportional to the dissimilarity between the desired an d interfering features. Since image gray levels are analytical functio ns of magnetic resonance imaging (MRI) parameters, it is possible to m aximize this dissimilarity by optimizing these parameters. For optimiz ation, we consider four MRI pulse sequences: multiple spin-echo (MSE); spin-echo (SE); inversion recovery (IR); and gradient-echo (GE). We u se the mathematical expressions for MRI signals along with intrinsic t issue parameters to express the objective function (normalized SNR of the eigenimage) in terms of MRI parameters. The objective function alo ng with a set of diagnostic or instrumental constraints define a multi dimensional nonlinear constrained optimization problem, which we solve by the fixed point approach. The optimization technique is demonstrat ed through its application to phantom and brain images. We show that t he optimal pulse sequence parameters for a sequence of four MSE and on e IR images almost doubles the smallest normalized SNR of the brain ei genimages, as compared to the conventional brain protocol.