Toolbox to Support Enhanced MR Imaging in Image-Guided Therapy
MR Image-guided therapy (IGT) applications often have requirements that are difficult to achieve using standard imaging approaches and tools. For example, achieving high temporal resolution may compromise volume coverage. To achieve imaging goals such as high temporal resolution while minimizing compromises elsewhere, imaging researchers in the NCIGT have developed enhanced imaging methods such as parallel imaging. They also have developed specialized tools such as real-time imaging controls that are important for MR image-guided therapy. To some extent, capabilities such as parallel imaging and real-time control have been implemented on commercial MR systems. However, such implementations are limited and rarely ideal for IGT. To ensure that IGT-compatible libraries exist that are implementable, imaging researchers in the NCIGT have built freely available libraries of functions, accessible on this page, to support enhanced imaging. The libraries contain functions to support or enable:
- real-time data acquisition, reconstruction and scanner control,
- reconstruction of accelerated imaging data including the principal parallel imaging methods and temporal sampling methods such as UNFOLD, and
- implementation of reduced field-of-view acquisition approaches by 2D RF selective excitation.
In addition to developing these libraries, researchers have applied them synergistically in a series of IGT applications such as thermal monitoring during focused ultrasound treatments and the tracking of devices during IGT procedures. While the dissemination of these libraries is part of the NCIGT's mission, the libraries are also meant to be used and refined within the IGT community to best address its needs going forward.
As part of an outreach effort, NCIGT imaging researchers presented in March 2008 on the Toolbox Project at the NCIGT's Clinical IGT conference held in Washington, DC.
Back to Research Projects.
Real-Time Imaging Library
As part of the real-time imaging library, researchers developed a flexible platform for real-time data acquisition, reconstruction and scanner control known as DARTS for Dynamically Adaptive Real-Time System. It supports parallel image reconstruction in Linux platforms running on multi-core computers, uses pipe structure to manage multiple processes with bulletin board to pass parameters between processes, has reconstruction modules for different sequences are easily plugged into system, and is tested on GE MR systems. DARTS can also be ported to other systems by replacing select modules.
The Reference Manual for the Real-Time Imaging Library is also available for download under Downloads with other NCIGT resources.
Fast Imaging Library
The Fast Imaging Library is a collection of MR image reconstruction functions, written in C and optimized for a multi-core real-time processing environment. This library currently provides a number of advanced MR imaging including:
- Temporal filtering methods such as UNFOLD
- Parallel Imaging Methods, including regularized SENSE and GRAPPA suitable for non-uniformly subsampled data,
- Reconstruction of Partial-Fourier data via homodyne synchronization
- EPI Ghost Correction and Regridding routines
- and a large number of utility functions to manipulate and visualize MRI data.
Further details can be found in he Fast Imaging Library Reference Manual. The complete package is available for download here and complements the other NCIGT resources available on the Downloads page.
2DRF Excitation Library
This library contains a set of functions to support implementation of 2DRF pulses in pulse sequences on GE MR platforms. Features of the library include:
- Macros and supporting function files for ease of implementation of 2DRF selective excitation.
- Incorporated into several product GE sequences such as 2dfast, fgre, fiesta, ssfse, epi.
- Tested for reduced FOV applications at 1.5T and 3T: FUS monitoring, needle tracking.
Involved Investigators
- Lawrence Panych: NCIGT Core PI
- Gary Zientara: Real-time imaging library
- Renxin Chu: Real-time imaging library
- Bruno Madore: Fast imaging library
- Scott Hoge: Fast imaging library
- Jing Yuan: Fast imaging and 2DRF excitation libraries
