Shape-optimized Model-based Reconstruction Algorithm for Radiacoustic Imaging

Area of Science:

• Medical Imaging
• Computational Imaging
• Radiotherapy Physics

Background:

• Limited-view data acquisition in radiacoustic imaging leads to ill-posed inverse problems.
• Restricted detector placement causes characteristic artifacts in reconstructed images.

Purpose of the Study:

• To develop a shape-optimized model-based (SOMB) reconstruction framework.
• To address the limited-view problem and reduce artifacts in radiacoustic imaging.
• To improve accuracy for radiation therapy monitoring.

Main Methods:

• Reconstructing an approximate dose-region boundary using parametric level set-based shape optimization.
• Performing pointwise model-based dose reconstruction within the shape-constrained region.
• Reducing the number of unknowns by restricting reconstruction to the identified region.

Main Results:

• Demonstrated substantial artifact reduction and improved accuracy compared to standard model-based reconstructions.
• Achieved 3-25% improvement in correlation coefficients across various imaging settings.
• Showed more pronounced improvements under stronger limited-view conditions.

Conclusions:

• The SOMB framework effectively transforms severely ill-posed inverse problems into better-constrained ones.
• This approach offers a viable pathway for accurate radiation therapy monitoring.
• The method is validated through computational studies and experiments with clinical radiation sources.