An Iterative Least Squares Method for Proton CT Image Reconstruction

Area of Science:

• Medical Physics
• Image Reconstruction
• Computational Imaging

Background:

• Proton Computed Tomography (CT) imaging requires accurate algorithms to determine 3D proton stopping power distributions.
• Current methods may lack quantitative precision for clinical utility.

Purpose of the Study:

• To develop and present a robust least-squares iterative method for quantitative proton CT imaging.
• To enhance the accuracy and efficiency of reconstructing proton stopping power distributions.

Main Methods:

• Implemented a least-squares iterative algorithm with features for unique solution definition and uncertainty consideration.
• Incorporated individual and simultaneous optimization of step sizes for accelerated convergence.
• Utilized parallel processing with graphical processing units (GPUs) and defined stopping criteria.

Main Results:

• Demonstrated the ability to provide assurance of quantifiably close-to-optimal solutions for any imaged object.
• Showcased that step size optimization significantly reduces the number of iterations required for convergence.
• Validated the algorithm's performance using real-world proton imaging data.

Conclusions:

• The developed iterative method offers a more quantitative framework for proton CT imaging.
• The algorithm ensures image quality and computational efficiency, paving the way for clinical use.
• Optimization strategies lead to faster convergence and reliable image reconstruction.