Clinical Micro-CT Empowered by Interior Tomography, Robotic Scanning, and Deep Learning.

Area of Science:

• Medical Imaging
• Radiology
• Biomedical Engineering

Background:

• Current medical CT scanners lack the resolution for detailed analysis of structures like cochlear implants and temporal bones.
• Preclinical micro-CT systems provide high resolution but are not suitable for clinical use.

Purpose of the Study:

• To design and evaluate a clinical micro-CT (CMCT) system for ultrahigh-resolution, low-dose imaging.
• To enable detailed visualization of the temporal bone and cochlear implants for clinical applications.

Main Methods:

• Integration of spiral cone-beam CT, interior tomography, and deep learning techniques.
• Utilizing a micro-focus X-ray source, photon-counting detector (PCD), and robotic arms for localized tomography.
• A hybrid reconstruction algorithm leveraging prior information from a global scan for background compensation.

Main Results:

• The proposed CMCT system achieves accurate high-resolution local reconstruction.
• The reconstruction algorithm is robust to misalignments in image registration.
• Demonstrated feasibility of the system design for clinical applications.

Conclusions:

• The developed CMCT system offers a promising solution for high-resolution temporal bone imaging and cochlear implant visualization.
• The system design shows potential for various other clinical applications requiring detailed anatomical imaging.
• Deep learning integration is envisioned to further optimize imaging performance and dose efficiency.