Atlas-Free Brain Tissue Segmentation via MP2RAGE and Dixon MRI
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current MR neuro-imaging techniques face challenges in accurately differentiating brain tissues from non-brain tissues like the skull due to similar image intensities, leading to biased diagnostic outcomes and computational inefficiencies, especially when using pre-segmented digital atlases that may not account for individual anatomical variations or disease-induced changes.
Innovation Solution
An atlas-free magnetic resonance imaging method utilizing a Magnetization-Prepared 2 Rapid Gradient-Echo (MP2RAGE) sequence combined with the Dixon method for fat-water separation, followed by a multichannel image segmentation algorithm based on a Markov random field model, to enhance contrast and accurately segment brain tissues without relying on pre-computed digital brain atlases.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If pre-segmented digital atlases are used for brain tissue segmentation, then segmentation can be performed efficiently, but individual anatomical variations and disease-induced changes are not accounted for leading to biased diagnostic outcomes
Solution Approach 1:
The patent changes the imaging parameters by using MP2RAGE sequence with specific inversion times (TI1=700ms, TI2=2500ms) and echo times (TE1=2.44ms, TE2=6.06ms) to generate images with enhanced contrast between brain and non-brain tissues. This parameter optimization enables accurate segmentation without relying on population-based atlases, thus resolving the contradiction between segmentation efficiency and accuracy by achieving both through improved image quality.
2Loss of time
If standard T1-weighted MRI sequences are used, then imaging time is reduced, but contrast between brain tissues and non-brain tissues like skull is insufficient leading to poor skull-stripping
Solution Approach 1:
The patent employs periodic action by using two different inversion times (TI1=700ms, TI2=2500ms) within the MP2RAGE sequence to acquire two image volumes at different contrast weights. This periodic variation in inversion time allows optimization of tissue contrast while maintaining reasonable imaging time, resolving the contradiction between imaging speed and tissue differentiation capability.
Solution Approach 2:
The patent creates a composite imaging approach by combining two image volumes acquired at different inversion times and echo times. The combination of these multi-contrast images produces enhanced tissue differentiation, particularly for skull-stripping applications, while maintaining efficient imaging through the unified MP2RAGE sequence structure.
3Measurement precision
If MP2RAGE sequence with multiple inversion times is used, then tissue contrast is enhanced, but acquisition time increases
Solution Approach 1:
The patent merges multiple imaging objectives into a single unified MP2RAGE sequence that acquires two image volumes with different inversion times and echo times. By combining these acquisitions in one sequence rather than performing separate scans, the patent achieves enhanced tissue contrast for skull-stripping while minimizing total acquisition time, thus resolving the contradiction between image quality and imaging efficiency.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach provides improved skull-stripping outcomes with reduced T2* and PD weighting, enabling more accurate and efficient brain tissue segmentation, reducing computational bias and enhancing diagnostic reliability by providing a bias-free T1 contrast and distinct intensity ranges for fat and dura matter, thus facilitating better brain post-processing techniques.
Implementation Method 1
The time required for a substance to become magnetized after having been placed in a magnetic field or the time required for said substance to regain longitudinal magnetization following the radiofrequency pulse is usually called the longitudinal relaxation time T1
Implementation Method 2
a transverse relaxation time T2 (also called spin-spin relaxation) describes the interaction between neighbouring nuclei with identical precessional frequencies having different magnetic quantum states
Implementation Method 3
using a Dixon method combined with the MP2RAGE for acquiring a fat-water separated image of said part
Data Source
AI summary
An atlas-free magnetic resonance imaging method images at least one part of a brain. An MRI sequence configured for acquiring two image volumes of the part at different inversion times within a single acquisition is combined to a fat-water separation method for acquiring a fat-water separated image. For each echo time two image volumes are acquired, respectively a first image volume and a second image volume at the first echo time, and a first image volume and a second image volume at the second echo time, and combined to a uniform image. The acquired images are combined to form a final uniform image, a final fat-water separated image, and a final second image volume that are fed into a multichannel image segmentation algorithm using a Markov random field model for segmenting the part into multiple classes of cranial tissues, in order to obtain a segmented image of said part.