Removing unwanted background phase with a reference phantom for applications in susceptibility quantification

A method of removing the background phase with a reference phantom but without overcorrecting the induced phase from objects of interest is proposed. Several factors during the imaging procedure and post-processing are investigated for their accuracies.

A method using a reference phantom to remove eddy currents as well as using the least squares fit to quantify susceptibility and to remove the background phase is proposed. Phase induced from simulated spheroids was fitted and compared to their true magnetic moments, an important concept for the proposed method. A cylindrical phantom and its simulation, a phantom with straws filled with Gd-DTPA, and a simulated head model were used to study systematic errors due to some confounding factors. The feasibility for in vivo applications was demonstrated from an actual human head. Susceptibility and remaining phase after removing the background phase were measured in all cases.

Simulations show that magnetic moments of various spheroids and phantoms can be accurately quantified from images, regardless of the partial volume effect. All measured susceptibility values are within ±0.16 ppm of -9.4 ppm for agarose and 0.05 ppm of 1 ppm for Gd-DTPA. Most residual phase is within ±0.1 rad from the phantom center. Susceptibilities close to -9.4 ppm are also obtained for the simulated and actual head. Correspondingly, the remaining phase has a mean value less than two standard deviations.

The proposed method from phantom studies can reliably remove the background phase without overcorrections. The in vivo example demonstrates the feasibility of the method.