Docket #: S14-171

A method for tracking moving sources with PET

Stanford researchers at the Pratx Lab have developed a new trajectory reconstruction method for tracking moving sources labeled with positron-emitting radionuclides using PET. This method reconstructs the time-varying position of individual sources directly from raw list-mode PET, thereby bypassing conventional image reconstruction entirely.
Proof-of-concept experiments show that low-activity and fast-moving sources can be reliably tracked in vivo with 2 mm accuracy. This method can be applied for tracking of single cells in vivo, real-time tracking of a moving tumor during radiotherapy, and estimation of respiratory breathing in 4D-PET.

Figure

Figure description - Single-source tracking. List-mode PET data from a point source (1, 10, 100, 1000 Bq) moving at 5 mm/s in an S-like pattern was reconstructed using standard MLEM (top row) and the novel trajectory reconstruction method (bottom row). This result indicate that the invention can track a moving source better than can be achieved using conventional ML-EM.

Stage of Research

Proof-of-concept - Monte Carlo simulations (GATE) show that this algorithm can track a single cell inside a small-animal PET system with 2 mm accuracy provided that the activity of the cell (Bq) is greater than twice the square of its velocity (cm/s).

Applying method to tracking multiple sources in parallel.

Applications

In vivo cell tracking:

Cell-based therapies for cardiac and neural tissue regeneration and cancer immunotherapy
Preclinical research tool to study biological processes such as cancer metastasis
Small animal research for drug development

4D PET/CT imaging – estimation of respiratory breathing
Radiation Therapy - real time tracking of a moving tumor

Advantages

Real time
Accurate – can track moving sources with higher localization accuracy and up to higher velocities and lower activities
Novel concept - First proposed method to reconstruct the motion of a source directly from PET measurements, without forming an image
Can be extended to track multiple moving sources in parallel