Automatic segmentation of rodent spinal cord diffusion MR images

Area of Science:

• Biomedical Imaging
• Neuroscience
• Medical Image Analysis

Background:

• Magnetic Resonance Imaging (MRI) is crucial for noninvasive spinal cord lesion analysis.
• Existing quantitative methods for spinal cord MRI segmentation are insufficient.
• Accurate segmentation is vital for understanding spinal cord injuries.

Purpose of the Study:

• To develop and validate a novel, quantitative MRI segmentation method for spinal cord tissues.
• To improve the accuracy of spinal cord lesion analysis using advanced algorithms.
• To enable precise segmentation of spinal cord, white matter, and hemorrhage.

Main Methods:

• A multistep, multidimensional approach using classification expectation maximization algorithm for MRI segmentation.
• Diffusion tensor imaging (DTI) to generate multi-weighted images of spinal cord slices.
• Joint estimation of maximum likelihood tissue classifications and edge detection for precise boundary identification.

Main Results:

• The proposed algorithm accurately segments the whole spinal cord, white matter, and hemorrhage.
• Segmentation accuracy is maintained even in the presence of significant spinal cord injury.
• The method's performance is comparable to expert manual segmentation.

Conclusions:

• The novel MRI segmentation technique provides accurate and quantitative analysis of spinal cord tissues.
• This method offers a robust tool for spinal cord lesion assessment in research and clinical settings.
• The algorithm demonstrates significant potential for improving the diagnosis and monitoring of spinal cord injuries.