Joint PET-MR respiratory motion models for clinical PET motion correction

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Biophysics

Background:

• Respiration-induced patient motion causes artifacts and quantification errors in Positron Emission Tomography (PET) imaging.
• Integrated PET-MR scanners offer complementary data and high-resolution Magnetic Resonance (MR) imaging for motion monitoring and correction.

Purpose of the Study:

• To develop and validate a methodology for respiratory motion correction in PET data using a joint PET-MR motion model.
• To assess the impact of the motion correction technique on PET image quality and quantitative accuracy in clinical oncology datasets.

Main Methods:

• A joint PET-MR motion model was developed using 1-minute of simultaneously acquired PET and MR data per bed position.
• Dynamic 2D multi-slice MR images served as the dynamic imaging component, with PET data (low-resolution framing, principal component analysis) as the model surrogate.
• Various motion models (1D/2D linear, polynomial) were evaluated on 45 patient datasets; the methodology was applied to 5 clinical PET-MR oncology cases.

Main Results:

• The joint PET-MR motion model effectively captured respiratory variations (inter-cycle and intra-cycle).
• Qualitative assessment showed improved PET image quality and reduced artifacts; quantitative analysis revealed changes in Standardized Uptake Value (SUV) metrics in avid lesions.
• The model demonstrated the capability to predict respiratory motion, leading to significantly improved PET image quality compared to uncorrected data.

Conclusions:

• The proposed joint PET-MR motion modeling methodology successfully corrects for respiratory motion in PET data.
• This technique enhances PET image quality and quantitative accuracy with minimal additional scan time (1 minute per bed position) and no external hardware.
• The method is applicable to PET-MR oncology studies, improving diagnostic insights from PET-MR imaging.