Initial Evaluation of Direct 4D Parametric Reconstruction with Human PET Data.

Area of Science:

• Nuclear Medicine
• Medical Imaging
• Biophysics

Background:

• Direct EM methods enable kinetic parameter image generation from list mode PET data.
• The one-tissue compartment model (1T) describes voxel time-activity curves.
• Previous evaluations used simulations without motion or scatter effects.

Purpose of the Study:

• To propose and implement a practical direct 4D parametric reconstruction algorithm for human PET data.
• To evaluate the algorithm's performance using human subjects and a serotonin transporter tracer.
• To compare the direct 4D method with the conventional 2-step approach.

Main Methods:

• Developed a direct 4D parametric reconstruction algorithm extending previous frame-based physics correction.
• Implemented the algorithm for human data using the serotonin transporter tracer [(11)C]AFM.
• Compared results with a 2-step approach: frame-based reconstruction followed by voxel-wise parameter estimation.

Main Results:

• Encouraging initial results were obtained from 3 human subjects.
• Regional analysis showed similar K(1) and V(T) values between 2-step and 4D methods, with a consistent difference.
• Visual analysis indicated noise reduction in the 4D direct reconstruction.

Conclusions:

• Direct 4D parametric reconstruction is feasible with real human PET data.
• The 4D method shows potential for improved accuracy and precision over the 2-step frame method.
• Further research may validate the benefits of direct 4D parametric imaging in clinical settings.