An optimized b-value sampling for the quantification of interstitial fluid using diffusion-weighted MRI, a genetic...

Area of Science:

• Neuroimaging
• Diffusion MRI Physics

Background:

• Multi-b-value diffusion-weighted MRI measures parenchymal diffusivity, microvascular perfusion, and interstitial fluid.
• Estimating these three components simultaneously is challenging due to SNR and b-value selection dependencies.
• Limited scanning time necessitates effective b-value selection for accurate interstitial fluid quantification.

Purpose of the Study:

• To evaluate an optimized b-value sampling strategy for interstitial fluid estimation in the brain.
• To determine the most effective b-values for improving the accuracy of interstitial fluid quantification.

Main Methods:

• A genetic algorithm was employed to determine an optimized b-value sampling scheme.
• The optimized scheme's performance was compared against linear, logarithmic, and prior sampling methods using Root Mean Square Error (RMSE).
• In vivo assessment utilized 7T MRI data with 101 b-values (0-1000 s/mm²), comparing fits with all b-values versus optimized subsets.

Main Results:

• The optimized b-value set for interstitial fluid estimation was identified as [0, 30, 90, 210, 280, 350, 580, 620, 660, 680, 720, 760, 980, 990, 1000] s/mm².
• Computer simulations showed the optimized sampling yielded lower RMSE across various Signal-to-Noise Ratio (SNR) levels.
• In vivo analysis confirmed that the optimized sampling scheme resulted in lower voxel-wise RMSE compared to other schemes.

Conclusions:

• A genetic algorithm-optimized b-value sampling scheme enhances the quantification of the interstitial fluid diffusion component.
• This optimized approach leads to a significant reduction in estimation errors (RMSE).