Generation of Synthetic-Pseudo MR Images from Real CT Images

Area of Science:

• Medical Imaging
• Biophysics
• Computational Anatomy

Background:

• Magnetic Resonance Imaging (MRI) and Computed Tomography (CT) are crucial diagnostic modalities.
• Synthesizing MRI from CT could reduce scan times and costs.
• Accurate tissue characterization is key for medical image analysis.

Purpose of the Study:

• To develop a method for generating synthetic MRI (sMRI) from CT images.
• To validate the accuracy and clinical applicability of the generated sMRI.
• To assess the potential of sMRI in guiding further diagnostic imaging decisions.

Main Methods:

• CT image pixel values (mean and standard deviation) were mapped to tissue types.
• Tissue types were remapped to intrinsic MRI parameters (T1, T2, proton density).
• Synthetic weighted MR images were generated using a simulated spin-echo sequence.

Main Results:

• The approach successfully generated synthetic T1-, T2-, and proton density-weighted MR images from CT data.
• Quantitative analysis showed good agreement between synthetic and real images (e.g., phantom slope: 0.73-0.9, PRD: 8.7-18%).
• Human knee imaging demonstrated the clinical potential, with average slope of 0.89% and PRD of 18.8%.

Conclusions:

• Generating synthetic MRI from CT images is a valid and effective technique.
• The method provides accurate tissue parameter mapping and image synthesis.
• Synthetic MRI can aid physicians in deciding the necessity of acquiring real MRI scans.