Automated Assessment of Ghost Artifacts in MRI

Introduction: Flow artifacts in MR images can appear as image ghosts within and outside the body cavity. Technical improvements aimed at suppressing these image ghosts often rely on expert scoring (1,2) or on semi-automated methods demanding tissue segmentation and estimation of statistical properties of intensity distribution (3) to evaluate the efficacy of the methods. These approaches can be labor intensive, introduce observer bias, computationally demanding, and error-prone if tissue segmentation is used. Herein we propose two fully automated image-processing methods that rely on the statistical properties of background (noise) pixels to assess the presence of flow artifacts (appearing as image ghosts) without requiring tissue segmentation. The first method rapidly evaluates the presence of flow artifacts in a global fashion, while the second one provides a more detailed characterization of the artifacts. We evaluate the proposed methods in the setting of cardiac phase-resolved myocardial blood-oxygen-level-dependent (BOLD) MRI where different cine SSFP imaging strategies are proposed for overcoming flow artifacts. Finally we assess the utility of our automated approaches against expert scoring results.