Respiratory motion estimation from slowly rotating x-ray projections: theory and simulation

Area of Science:

• Medical Imaging
• Radiotherapy Physics
• Computational Anatomy

Background:

• Respiratory motion significantly impacts radiotherapy accuracy by causing tumor movement.
• Conventional X-ray CT imaging struggles to capture respiratory motion due to artifacts from inconsistent projection views.
• Accurate assessment of tumor motion is crucial for effective conformal radiotherapy.

Purpose of the Study:

• To develop and validate a method for estimating nonrigid, free-breathing respiratory motion parameters.
• To enable precise four-dimensional (4D) motion modeling for improved radiotherapy planning.
• To overcome limitations of conventional CT in visualizing dynamic thoracic structures.

Main Methods:

• Utilizing projection views acquired from a slowly rotating cone-beam CT scanner.
• Employing a nonrigid motion model to represent thorax deformation during breathing.
• Optimizing motion parameters by minimizing a regularized squared error cost function comparing simulated and measured projections.

Main Results:

• Simulation results demonstrated good agreement between estimated and true respiratory motion.
• The proposed method shows potential for accurate four-dimensional motion estimation.
• Successfully deforms a reference thorax volume based on estimated motion parameters.

Conclusions:

• The developed approach shows promise for accurate 4D respiratory motion estimation in radiotherapy.
• This technique can enhance the precision of conformal radiotherapy by accounting for tumor motion.
• Cone-beam CT combined with motion modeling offers a viable solution for dynamic imaging challenges.