Development of an adjustable dynamic phantom for testing deviceless motion correction algorithm in PET

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Quantitative PET Imaging

Background:

• Respiratory motion degrades PET image quality, causing blurring and inaccurate quantification.
• Deviceless motion correction algorithms (MCA) offer a solution but require robust in vivo validation, which is challenging.
• Existing validation methods for MCA in PET are limited.

Purpose of the Study:

• To develop and validate an adjustable dynamic phantom with fillable spheres.
• To enable quantitative assessment of deviceless PET motion correction algorithms.
• To provide a ground-truth framework for MCA evaluation.

Main Methods:

• A microprocessor-controlled, two-axis dynamic phantom was engineered to simulate various respiratory motion patterns (craniocaudal, ventrodorsal, diagonal).
• Fillable 18F-FDG spheres of different volumes (4, 1, 0.25 mL) were imaged with and without MCA on a PET/CT scanner.
• Volumes, activities, and recovery coefficients (RCs) were extracted using two isocontour methods to assess MCA performance.

Main Results:

• Uncorrected scans showed significant volumetric overestimation, particularly for smaller spheres (up to 509%).
• MCA effectively reduced motion-induced distortions, improving accuracy for 1 mL spheres (e.g., from +166% to +82% for diagonal motion).
• Activity quantification was largely preserved with one isocontour method (IC0.1), while RCs showed partial improvement with MCA but did not reach static levels.

Conclusions:

• The developed dynamic phantom offers a reproducible method for evaluating deviceless PET motion correction.
• MCA successfully mitigates motion-induced blurring and volumetric errors, although performance varies with sphere size and motion direction.
• This system facilitates systematic benchmarking and optimization of PET motion correction technologies.