Assessing Reliability of Myocardial Blood Flow After Motion Correction With Dynamic PET Using a Bayesian Framework

Area of Science:

• Nuclear Medicine
• Medical Imaging Physics
• Cardiovascular Imaging

Background:

• Dynamic positron emission tomography (PET) for myocardial blood flow (MBF) estimation is susceptible to errors, notably from patient motion.
• Patient motion causes frame misalignment and registration errors with CT attenuation maps, compromising MBF accuracy in clinical settings.
• Existing metrics for motion correction evaluation are often unsuitable for dynamic cardiac PET data.

Purpose of the Study:

• To present a novel method for quantitatively assessing motion correction quality in dynamic cardiac PET imaging.
• To introduce a Bayesian framework for estimating kinetic parameters and their uncertainties, specifically K1, as a surrogate for MBF.
• To evaluate the impact of motion correction on the reliability of MBF estimation.

Main Methods:

• Developed a Bayesian framework to model kinetic parameters as probability distributions, enabling uncertainty extraction.
• Calculated K1, a surrogate for MBF, and its associated uncertainty to assess motion correction effectiveness.
• Validated the framework using simulated data across various noise levels and applied it to 40 patient datasets with categorized motion magnitudes.

Main Results:

• The Bayesian framework demonstrated robust uncertainty estimation across different noise levels in simulated data.
• Application to patient data showed a significant reduction in K1 uncertainty, up to 60%, after manual motion correction.
• The degree of uncertainty reduction correlated with the magnitude of motion, validating the method's sensitivity.

Conclusions:

• The proposed Bayesian approach effectively quantifies motion correction quality in dynamic cardiac PET by measuring K1 uncertainty.
• Motion correction significantly improves the reliability of MBF estimation, as evidenced by reduced parameter uncertainty.
• This method offers a valuable tool for evaluating and ensuring the quality of dynamic cardiac PET imaging in clinical practice.