Real-time spatiotemporal optimization during imaging.

Area of Science:

• Medical Imaging
• Radiation Oncology
• Computational Optimization

Background:

• High-quality medical imaging is crucial for effective patient care, particularly in precision treatments like radiation therapy.
• Patient motion during imaging compromises image quality, necessitating retrospective corrections or acceptance of reduced quality.
• Current methods often involve trade-offs between image quality, scan time, and radiation exposure.

Purpose of the Study:

• To formalize and implement a novel spatiotemporal optimization approach for real-time medical image acquisition.
• To evaluate the clinical efficacy of this approach in respiratory-correlated 4D cone-beam computed tomography (CBCT) for lung cancer radiation therapy.
• To assess the impact on image quality, scan time, and radiation dose compared to standard clinical practices.

Main Methods:

• Developed a general spatiotemporal optimization framework treating data acquisition as a real-time problem.
• Implemented this approach in a first-in-world clinical trial for lung cancer radiation therapy (NCT04070586).
• Applied the method to respiratory-correlated 4D CBCT, focusing on optimizing acquired data structure for reconstruction.

Main Results:

• Achieved maintenance or improvement in image quality compared to the clinical standard.
• Reduced scan time by 63% and radiation dose by 85%.
• Demonstrated enhanced clinical throughput and reduced risk of secondary tumors.

Conclusions:

• The spatiotemporal optimization approach offers significant benefits for medical imaging in radiation therapy.
• This method successfully improves efficiency and safety in lung cancer treatment.
• The generalizable framework holds potential for application to other patient motion types (e.g., cardiac) and imaging modalities (e.g., CT, MRI).