MR-based motion correction for cardiac PET parametric imaging: a simulation study.

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Biophysics

Background:

• Dynamic PET imaging is susceptible to motion artifacts affecting kinetic parameter accuracy.
• Cardiac and respiratory motions introduce significant bias in quantitative PET measurements.
• Accurate kinetic parameter estimation is crucial for diagnosing and monitoring cardiac conditions.

Purpose of the Study:

• To evaluate the impact of MR-based motion correction on PET myocardial kinetic parameters using PET-MR simulations.
• To compare the performance of MR-based motion correction against conventional gating and no correction methods.
• To assess the reduction in bias and variance of kinetic parameters under realistic motion conditions.

Main Methods:

• Realistic PET-MR simulations were performed using the 4D XCAT phantom.
• Cardiac and respiratory motion patterns were incorporated into dynamic activity distributions.
• PET sinograms were reconstructed using static, gated, non-motion-corrected, and motion-corrected methods.
• Voxel-wise kinetic parameter (K1) estimation was conducted using unweighted least squares fitting.

Main Results:

• MR-based motion correction effectively reduced motion blurring in parametric images without increasing noise.
• Compared to non-motion-corrected methods, MR-based correction significantly reduced K1 bias (up to 29.9%).
• MR-based motion correction substantially decreased K1 standard deviation compared to gated methods (up to 95.2%).

Conclusions:

• MR-based motion correction in PET-MR imaging is highly effective in mitigating motion artifacts.
• This method improves the accuracy of myocardial kinetic parameter estimation by reducing bias and variance.
• The findings support the clinical utility of MR-based motion correction for quantitative dynamic PET studies.