An Efficient Iterative Cerebral Perfusion CT Reconstruction via Low-Rank Tensor Decomposition With Spatial-Temporal...

Area of Science:

• Medical Imaging
• Radiology
• Computational Imaging

Background:

• Cerebrovascular diseases, particularly acute stroke, are a leading cause of long-term disability.
• Cerebral perfusion computed tomography (CPCT) offers rapid, high-resolution hemodynamic assessment for acute stroke.
• Standard CPCT protocols involve substantial radiation doses due to repeated scanning.

Purpose of the Study:

• To develop a low-dose CPCT image reconstruction method for acute stroke assessment.
• To improve the quality of CPCT images and the precision of hemodynamic maps.
• To reduce radiation exposure in CPCT while maintaining diagnostic utility.

Main Methods:

• A novel low-rank tensor decomposition with spatial-temporal total variation (LRTD-STTV) regularization was developed.
• The method utilizes the high similarity among sequentially acquired CPCT images.
• Tensor Tucker decomposition captures global spatial-temporal correlations, enhanced by spatial-temporal TV regularization for local structures.

Main Results:

• The LRTD-STTV model effectively reconstructs high-quality CPCT images from low-dose data.
• Hemodynamic maps derived from the reconstructed images show high precision and diagnostic accuracy.
• Experimental results on phantom and patient data demonstrate superior performance compared to existing algorithms.

Conclusions:

• The proposed low-dose CPCT reconstruction method significantly enhances image quality and diagnostic accuracy.
• This approach offers a viable solution for reducing radiation dose in acute stroke imaging.
• The LRTD-STTV model preserves essential spatial structures and underlying hemodynamic information for clinical decision-making.