Task-Driven Trajectory Design for Endovascular Embolization
Embargo until
Date
2019-02-25
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Johns Hopkins University
Abstract
Computed Tomography (CT) is one of the most useful and widely applied imaging modalities, employed in both diagnostic and treatment planning purposes in the medical field. Circular and spiral acquisition trajectories are traditionally employed and work well in many cases. The advent of technologies such as robotic C-arms in interventional imaging allow for more complex data acquisitions, which enables potential improvements in image quality, increased field of view, and sampling. This capability has particular potential crucial in interventional cases where images may be compromised by complex anatomy or surgical tools. In this work, we present a paradigm that uses custom non-circular orbits and prior patient information along with segmentation and registration techniques to account for surgical tools and/or implants, to improve image quality. The framework leverages the anatomical model to optimize a parameterized source-detector trajectory for a variety of specific imaging tasks. We propose an overall workflow for orbit customization with investigations of the various workflow stages as well as the overall performance of the framework.
Description
Keywords
Imaging, Computed Tomography, Task-Driven, Optimization