ENTROPY BASED NON-INTERACTIVE NYQUIST GHOST CORRECTION ALGORITHM

点击翻译按钮获取中文摘要

This subproject is one of many research subprojects utilizing the resources provided by a Center grant funded by NIH/NCRR. The subproject and investigator (PI) may have received primary funding from another NIH source, and thus could be represented in other CRISP entries. The institution listed is for the Center, which is not necessarily the institution for the investigator. Introduction. Nyquist ghosting occurs in EPI due to misalignment between odd and even echoes. For this reason, clinical EPI scans are commonly performed with a reference scan incorporated into the prescan process from which phase information can be used to correct ghosting errors. For multi-shot, multi-coil and multi-slice acquisitions, this additional scan can often exceed a minute, which may be crucial in cases of acute stroke, for example. An interesting alternative to reference scans are post-processing methods that require no additional information. One such method (1) minimizes the intensity in a ROI selected in the ghost area by iteratively refining phase parameters. In general, however, this method requires user interaction. Another elegant method (2) uses the total image entropy as a cost function for iterative optimization and makes no assumptions about the shape or size of the object being imaged. However, the computation time to correct a 64-slice 128¿128 4-shot data set has been reported to be 2.5 min (2). For this method to be a dependable replacement for the reference scan in the clinic it may never fail, regardless of acquisition parameters chosen, anatomy of interest and how well the system is calibrated. Here we present a robust iterative entropy minimization implementation for ghost removal that we now use routinely as a replacement for reference scanning in clinical single- and multi-shot, GRAPPA-accelerated DWI and PWI EPI exams.