A Left Ventricular Motion Phantom for Cardiac Magnetic Resonance Imaging

The mammalian left ventricle (LV) has two distinct motion patterns: wall thickening and rotation. The purpose of this study was to design and build a low-cost, non-ferromagnetic LV motion phantom, for use with cardiac magnetic resonance imaging (MRI), that is able to produce physiologically realistic LV wall thickening and rotation. Cardiac MRI is continuously expanding its range of techniques with new pulse sequences, including new tissue tagging techniques which allow intra-myocardial deformation to be visualized. An essential step in the development of new cardiac MRI techniques is validating their performance in the presence of motion. MRI-compatible dynamic motion phantoms are of substantial benefit in the development of cardiac specific-magnetic resonance imaging techniques. These phantoms enable the investigation of motion effects images by mimicking the three dimensional motion of the heart. To date, no single study has succeeded in duplicating both LV motion patterns, in an MRI-compatible cardiac motion phantom. In addition, a phantom that is 100 MRI-compatible with low cost to build would be desirable to researchers. We have built two MRI-compatible phantoms, housed within a common enclosure and each filled with MRI-visible dielectric gel (as a surrogate to myocardium),which model the wall thickening and rotation motions of the left ventricle independently. The wall motion phantom is pneumatic, driven by a custom non-ferromagnetic pump which cyclically fills and empties a latex balloon within the phantom. The rotation phantom is manually driven by a plastic actuator which rotates the phantom through a specified angular rotation. Each phantom also generates a TTL pulse for triggering the MRI scanner. Although this circuitry contains ferromagnetic materials, it can be located outside the scanner bore. The wall thickening motion phantom has been tested using segmented cine, real time cine and grid tagged MRI acquisition sequences. Results were significant with 4 average variability and physiologically r