A more accurate reconstruction system matrix for quantitative proton computed tomography.

Area of Science:

• Medical Imaging
• Computational Physics
• Radiological Sciences

Background:

• Quantitative proton CT (pCT) image reconstruction relies on an accurate system matrix.
• Previous methods using constant intersection chord lengths introduced systematic errors in proton stopping power calculations.
• Improving the system matrix accuracy is crucial for reliable pCT data.

Purpose of the Study:

• To develop a computationally efficient variable intersection chord length for enhanced system matrix accuracy in pCT.
• To reduce systematic errors in reconstructed proton stopping powers.
• To improve the overall accuracy and efficiency of quantitative pCT image reconstruction.

Main Methods:

• Developed an analytical expression for an angle-dependent effective mean chord length function, accounting for the discrete nature of pCT most likely path (MLP) reconstruction.
• Simulated a pCT dataset using GEANT4 with 200 MeV protons and a head phantom.
• Reconstructed phantom stopping powers using algebraic reconstruction technique with constant, exact, and effective mean chord length approaches.

Main Results:

• The effective mean chord length approach reduced mean reconstructed stopping power errors from approximately 10% to less than 0.5%.
• Reconstructions using the effective mean chord length were approximately 20% faster than those using the exact chord length.
• Both effective mean chord length and exact chord length methods significantly outperformed the constant chord length approach.

Conclusions:

• The effective mean chord length method significantly improves the accuracy of the system matrix for quantitative pCT.
• This approach offers a computationally efficient alternative to exact chord length calculations.
• The effective mean chord length method enables more accurate and efficient quantitative pCT reconstructions.