Maximum-likelihood joint image reconstruction and motion estimation with misaligned attenuation in TOF-PET/CT

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Image Reconstruction

Background:

• Joint Activity Reconstruction/Motion Estimation (JRM) is crucial for accurate Positron Emission Tomography (PET) imaging.
• Previous JRM methods using conventional PET data exhibited slow convergence due to likelihood crosstalk.
• Misalignment between attenuation maps and PET data complicates accurate image reconstruction.

Purpose of the Study:

• To improve the convergence speed of JRM in PET imaging.
• To leverage Time-of-Flight (TOF) PET data for enhanced JRM performance.
• To validate the benefits of TOF-JRM over conventional JRM.

Main Methods:

• Utilized a probabilistic model assuming a shared motion field for activity and attenuation maps.
• Employed maximization of the penalized log-likelihood for JRM.
• Applied Time-of-Flight (TOF) PET data in the JRM framework.

Main Results:

• Demonstrated significantly improved convergence speed with TOF-PET data compared to conventional PET.
• Successfully reconstructed activity distributions even with misaligned attenuation maps.
• The TOF-PET approach mitigates the crosstalk issues inherent in conventional JRM.

Conclusions:

• Time-of-Flight PET data substantially accelerates the convergence of Joint Activity Reconstruction/Motion Estimation.
• TOF-JRM offers a more efficient and robust solution for PET image reconstruction.
• This advancement holds promise for improved diagnostic accuracy in PET imaging.