Improving the temporal resolution of 3D phase contrast MR angiography using keyhole technique at low tesla open-MRI system

Phase Contrast MR Angiography (PC MRA) takes longer scan time than other MRA technique. Especially, the 3D PC MRA images obtaining at low tesla MRI system takes more long scan time because the lower characteristic of hardware system than high tesla MRI system. The keyhole imaging was introduced that is a simple way to increase the temporal resolution of dynamic imaging studies. In this study, we applied the keyhole imaging technique to 3D PC MRA for improving the temporal resolution at low tesla MRI system. We also validated image quality and usefulness of combination of two techniques. All image data were obtained on a MagfinderII (Genpia, Korea) 0.32T MRI system. Using the 3D PC MRA pulse sequence, the vascular images for a human brain targeted on the Superior Sagittal Sinus (SSS) and Circle of Willis were obtained. The reference image was acquired fully sampled k-space data without velocity encoding gradients. Three flow sensitive images were obtained low frequency parts of k-space. The keyhole factor was used from 12.5% to 50% of the full k-space. We also used tukey window function to minimize the erroneous effects induced from frequency discontinuous. Moreover, the artifact power (AP) value was measured to validate the reconstructed image quality. Based on the our results, image artifacts that were considered as frequency discontinuous using the keyhole reconstruction were shown until use the 12.5% and 25% keyhole factors. Using above 50% keyhole factors, the images are similar to original image and the vessels in the brain are well observed. In conclusion, reconstructed images using keyhole technique combined 3D PC MRA data are enough to application at low tesla MRI system. Although our results are in the early stage of keyhole imaging and draw a conclusion from restrictive number of subjects, we considered that keyhole technique combined 3D PC MRA will give some advantages for improving the temporal resolution at low tesla MRI system.