Related Experiment Videos
Fast method for brain image segmentation: application to proton magnetic resonance spectroscopic imaging
David Bonekamp1, Alena Horská, Michael A Jacobs
1Division of Neuroradiology, Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins University School of Medicine, Baltimore, Maryland 21287, USA.
Magnetic Resonance in Medicine
|September 28, 2005
Summary
This study introduces a fast method for brain tissue segmentation using Fast Spin Echo (FSE) imaging and Eigenimage (EI) analysis. This rapid technique accurately segments gray matter, white matter, and cerebrospinal fluid for proton MRSI applications.
Area of Science:
- Neuroimaging
- Biomedical Engineering
- Medical Physics
Background:
- Accurate brain tissue segmentation (gray matter, white matter, cerebrospinal fluid) is crucial for interpreting proton magnetic resonance spectroscopic imaging (MRSI) data.
- Traditional segmentation relies on high-resolution T1-weighted MRI, which is time-consuming and limits MRSI analysis speed.
- Existing methods require high-resolution data, creating a bottleneck for rapid quantitative analysis.
Purpose of the Study:
- To develop and validate a rapid data acquisition and analysis procedure for brain tissue segmentation.
- To enable faster and more efficient analysis of proton MRSI data by improving segmentation speed.
- To provide a reliable alternative to time-consuming high-resolution MRI-based segmentation.
Main Methods:
- Utilized Fast Spin Echo (FSE) imaging with multiple contrasts and thick slices for rapid data acquisition.
- Employed linear Eigenimage (EI) filtering and normalization for tissue segmentation.
- Compared the novel FSE/EI method against standard high-resolution 3D T1-weighted MRI segmentation in five subjects.
Main Results:
- Achieved excellent correlation between FSE/EI segmentation and standard MRI techniques (GM: R2 = 0.893, WM: R2 = 0.892, ln(CSF): R2 = 0.831).
- Demonstrated excellent test-retest agreement (R2 > 0.926) for all tissue classes in a single individual.
- Applied FSE/EI segmentation to a proton MRSI dataset, yielding results consistent with previous studies.
Conclusions:
- Fast Spin Echo (FSE) imaging combined with Eigenimage (EI) analysis offers a rapid and reliable method for brain tissue segmentation.
- This technique is suitable for improving the efficiency of proton MRSI data analysis.
- The FSE/EI approach provides a valuable alternative for researchers needing faster segmentation without compromising accuracy.