Model-based perfusion reconstruction with time separation technique in cone-beam CT dynamic liver perfusion imaging.

Area of Science:

• Medical Imaging
• Interventional Radiology
• Oncology

Background:

• Cone-beam CT (CBCT) enables intraoperative evaluation of liver cancer embolization by assessing contrast agent flow.
• Challenges in CBCT include low temporal sampling and high noise levels, necessitating advanced reconstruction algorithms.

Purpose of the Study:

• To develop and evaluate a model-based perfusion reconstruction algorithm for CBCT to address its inherent limitations.
• To compare the performance of the proposed algorithm against static reconstruction methods.

Main Methods:

• Applied time separation technique (TST) using basis functions (analytical or prior knowledge-based) to model time attenuation curves.
• Simulated dynamic CBCT perfusion scans under varying noise levels to assess prior knowledge (PK) based TST.
• Compared TST with PK to TST with analytical basis functions and static reconstruction.

Main Results:

• A TST set of four basis functions effectively denoises data and preserves information, outperforming static reconstruction, especially at higher noise levels.
• TST with PK demonstrated superior performance compared to TST with analytical functions.
• Eight CBCT rotations were sufficient for comparable perfusion maps, reducing dose and reconstruction time.

Conclusions:

• CBCT perfusion reconstruction using TST shows significant potential, outperforming static methods and enabling dose reduction.
• Further validation with larger patient cohorts is required to confirm the clinical advantages of TST.