Low-dose CT reconstruction method based on prior information of normal-dose image

Area of Science:

• Medical Imaging
• Image Reconstruction
• Radiology

Background:

• Radiation risk from computed tomography (CT) is a significant concern, particularly for patients needing repeated scans.
• Low-dose CT protocols, such as sparse scanning, are employed to mitigate radiation exposure.
• Advanced reconstruction algorithms are essential for maintaining image quality with low-dose CT protocols.

Purpose of the Study:

• To develop a novel algorithm for sparse-view CT reconstruction that incorporates prior image information.
• To address the challenge of reconstructing high-quality CT images from limited projection data.

Main Methods:

• A prior information of normal-dose image (PI-NDI) method was developed for low-dose CT reconstruction.
• The PI-NDI method utilizes a transformed model for attenuation coefficients and prior information in the forward-projection process.
• Performance was evaluated using digital XCAT and head phantoms, measuring RMSE, PSNR, and MPAE.

Main Results:

• Visually, PI-NDI reconstructed images showed higher quality compared to existing methods using sparse-view data.
• Quantitative analysis revealed RMSE comparable to full-data methods (MLEM-TV, PWLS-TV, PWLS-PICCS).
• With very sparse data, PI-NDI achieved higher PSNR and lower MPAE than other algorithms.

Conclusions:

• A novel PI-NDI method for sparse-view CT image reconstruction was successfully developed.
• The PI-NDI method demonstrates superior performance compared to state-of-the-art techniques.
• This approach offers a promising solution for low-dose CT imaging in clinical practice.