Data-driven head motion correction for PET using time-of-flight and positron emission particle tracking techniques

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Image Processing

Background:

• Patient head motion during Positron Emission Tomography (PET) scans is a significant challenge impacting diagnostic accuracy.
• Current methods like physical restraints or external tracking systems have limitations, and data-driven approaches often require manual intervention.
• Developing automated, data-driven solutions for motion correction is crucial for robust brain PET imaging.

Purpose of the Study:

• To introduce a fully automated, data-driven algorithm for detecting and correcting head motion in PET imaging.
• To enable retroactive processing of raw listmode data for motion correction.
• To improve the reliability and diagnostic quality of brain PET scans affected by patient movement.

Main Methods:

• Utilized a previously established time-of-flight (TOF) weighted positron emission particle tracking (PEPT) algorithm to identify motion-free frames.
• Employed weak radioactive point sources on glasses for estimating rigid transformations and registering static frames to a reference.
• Corrected raw event data by tracking point sources in listmode data for subsequent image reconstruction.

Main Results:

• The automated TOF-PEPT algorithm successfully detected and corrected head motion in five patient studies.
• Event-based correction produced images visually free of motion artifacts, comparable to a gold-standard frame-based image registration method.
• Quantitative analysis showed high Jaccard similarity indices (85-98%) between corrected and reference frames, indicating robust performance.

Conclusions:

• A fully automated, data-driven method for head motion detection and correction in PET raw listmode data has been successfully developed.
• The TOF-PEPT approach is easy to implement, offers high temporal resolution, and provides reliable post hoc correction of motion artifacts.
• This methodology enhances the assessment and correction of patient motion during PET imaging, improving overall data quality.