Towards a Hall effect magnetic tracking device for MRI.

This paper presents the first prototype of a magnetic tracking device for Magnetic Resonance Imaging. The unique relationship between the space coordinates of a MRI scanner bore and the magnetic field gradients used in MRI allows building a localization system based on an accurate measurement of these gradients. These gradients are measured thanks to a 3D Hall device with a footprint of only 50µm(2), integrated with its specific conditioning circuit in a low cost, low voltage 0.35µm CMOS process. The first experimental results show that a sub-millimeter localization is possible. It opens the way to the development of MRI compatible magnetic tracking systems integrable in a surgical tool.

Evaluation of an image-based tracking workflow with Kalman filtering for automatic image plane alignment in interventional MRI.

Recently, a workflow for magnetic resonance (MR) image plane alignment based on tracking in real-time MR images was introduced. The workflow is based on a tracking device composed of 2 resonant micro-coils and a passive marker, and allows for tracking of the passive marker in clinical real-time images and automatic (re-)initialization using the microcoils. As the Kalman filter has proven its benefit as an estimator and predictor, it is well suited for use in tracking applications. In this paper, a Kalman filter is integrated in the previously developed workflow in order to predict position and orientation of the tracking device. Measurement noise covariances of the Kalman filter are dynamically changed in order to take into account that, according to the image plane orientation, only a subset of the 3D pose components is available. The improved tracking performance of the Kalman extended workflow could be quantified in simulation results. Also, a first experiment in the MRI scanner was performed but without quantitative results yet.

Evaluation of an image-based tracking workflow using a passive marker and resonant micro-coil fiducials for automatic image plane alignment in interventional MRI.

In this paper, an original workflow is presented for MR image plane alignment based on tracking in real-time MR images. A test device consisting of two resonant micro-coils and a passive marker is proposed for detection using image-based algorithms. Micro-coils allow for automated initialization of the object detection in dedicated low flip angle projection images; then the passive marker is tracked in clinical real-time MR images, with alternation between two oblique orthogonal image planes along the test device axis; in case the passive marker is lost in real-time images, the workflow is reinitialized. The proposed workflow was designed to minimize dedicated acquisition time to a single dedicated acquisition in the ideal case (no reinitialization required). First experiments have shown promising results for test-device tracking precision, with a mean position error of 0.79 mm and a mean orientation error of 0.24°.