MRI-to-CT Synthesis for Pediatric Cranial Evaluation
Published:Dec 29, 2025 23:09
•1 min read
•ArXiv
Analysis
This paper addresses a critical clinical need by developing a deep learning framework to synthesize CT scans from MRI data in pediatric patients. This is significant because it allows for the assessment of cranial development and suture ossification without the use of ionizing radiation, which is particularly important for children. The ability to segment cranial bones and sutures from the synthesized CTs further enhances the clinical utility of this approach. The high structural similarity and Dice coefficients reported suggest the method is effective and could potentially revolutionize how pediatric cranial conditions are evaluated.
Key Takeaways
- •Proposes a deep learning framework to synthesize CT scans from MRI data in pediatric patients.
- •Enables assessment of cranial development and suture ossification without ionizing radiation.
- •Achieves high structural similarity and Dice coefficients, indicating effective performance.
- •Allows for segmentation of cranial bones and sutures from synthesized CTs.
Reference
“sCTs achieved 99% structural similarity and a Frechet inception distance of 1.01 relative to real CTs. Skull segmentation attained an average Dice coefficient of 85% across seven cranial bones, and sutures achieved 80% Dice.”