Low-Dose Dynamic Cerebral Perfusion Computed Tomography Reconstruction via Kronecker-Basis-Representation Tensor...

Area of Science:

• Medical Imaging
• Radiology
• Computational Imaging

Background:

• Dynamic cerebral perfusion computed tomography (DCPCT) provides crucial hemodynamic information but involves high radiation doses.
• Concerns about radiation exposure necessitate the development of low-dose DCPCT techniques.
• Current methods struggle to maintain image quality at reduced radiation levels.

Purpose of the Study:

• To develop a novel algorithm for reconstructing high-quality DCPCT images from low-dose acquisitions.
• To improve the quality of perfusion maps derived from low-dose DCPCT data.
• To reduce patient radiation dose while preserving diagnostic information.

Main Methods:

• Proposed tensor-based robust principal component analysis (T-RPCA) models, including nonlocal patch-based T-RPCA (NL-T-RPCA).
• Utilized Kronecker-basis-representation tensor sparsity regularization to measure tensor low-rankness.
• Developed efficient algorithms using the alternating direction method of multipliers for model optimization.

Main Results:

• The proposed T-RPCA and NL-T-RPCA models significantly improved quantitative and visual quality of low-dose DCPCT images.
• Demonstrated superior performance compared to existing algorithms, especially at radiation doses as low as 20 mAs.
• Successfully reconstructed high-quality perfusion maps from low-dose data.

Conclusions:

• T-RPCA and NL-T-RPCA are effective for low-dose DCPCT image reconstruction and perfusion map enhancement.
• These methods offer a promising solution for reducing radiation dose in brain perfusion imaging.
• The developed algorithms provide substantial gains in image quality and diagnostic accuracy for low-dose DCPCT.