Change in positron emission tomography perfusion imaging quality with a data-driven motion correction algorithm
Yushui Han1, Ahmed Ibrahim Ahmed1, Charles Hayden2
1Houston Methodist Debakey Heart and Vascular Center, 6550 Fannin Street, Smith Tower-Suite 1801, Houston, TX, 77030, USA.
Summary
A new data-driven motion correction (DDMC) algorithm significantly improved the quality of cardiac PET images affected by motion. Machine measurements of motion correlated well with physician interpretations, enhancing diagnostic accuracy.
Area of Science:
- Nuclear Medicine
- Medical Imaging Analysis
Background:
- Cardiac motion is a significant challenge in Positron Emission Tomography (PET) imaging, often reducing image interpretability.
- Accurate assessment of myocardial perfusion is crucial for diagnosing and managing cardiac diseases.
Purpose of the Study:
- To evaluate the effectiveness of a prototype data-driven motion correction (DDMC) algorithm in improving cardiac PET image quality.
- To compare image quality, perfusion defect size, and severity between DDMC-corrected and non-corrected (NMC) images.
Main Methods:
- 40 consecutive patients with cardiac motion underwent PET imaging.
- Images were processed using both NMC and DDMC algorithms.
- Two blinded investigators assessed image quality using a 4-point scale; motion was quantified using Dwell Fraction.
Main Results:
- All DDMC-corrected images demonstrated improved quality and interpretability compared to NMC images.
- A significant correlation (Spearman's rho = 0.626, P < .001) was found between machine-based motion quantification and physician assessment.
- The study identified varying degrees of motion across the patient cohort, with 40% experiencing mild motion.
Conclusions:
- The DDMC algorithm effectively enhances the quality of cardiac PET images affected by motion.
- The strong correlation between automated motion quantification and expert interpretation validates the DDMC algorithm's performance.
