ORIGINALS
Optimal Frame Summation for Motion Correction in Sinusoidal Periodic Motion:
A Phantom Study
2026 Volume 76 Issue 2 Pages 137-145
Details
2026 Volume 76 Issue 2 Pages 137-145
Purpose: Patient motion in pediatric nuclear medicine often exhibits sinusoidal periodic patterns due to respiratory movements. Sinusoidal motion causes non-uniform velocity distribution, resulting in characteristic image distortion. This study investigated optimal frame summation conditions for inter-frame registration-based motion correction using ellipticity and center-of-mass stability as evaluation metrics.
Methods: A NEMA IEC body phantom (37 mm sphere) was moved in sinusoidal patterns at 25, 30, 40, and 60 beats per minute (BPM). Dynamic acquisition (1s/frame, 600 frames) was performed. Motion correction was applied with frame summation numbers of 2-30. Ellipticity (FWHM in motion direction / FWHM perpendicular) and center-of-mass shift were calculated. Optimal conditions were determined using a figure of merit (FOM) integrating both metrics.
Results: Sinusoidal motion increased ellipticity from 1.08 (static) to approximately 2.0. FOM-based optimal conditions were: 3-frame at 40 BPM (ellipticity 1.92 to 1.47, 23% improvement), 5-frame at 60 BPM (9%), 15-frame at 25 BPM (20%), and 10-frame at 30 BPM. Two-frame summation caused severe algorithm failure at 60 BPM (ellipticity 3.55, 16.9 pixels shift) and misregistration at 40 BPM (4.2 pixels shift).
Conclusion: Optimal frame summation conditions depend on motion speed. Two-frame summation is not recommended at any speed due to high misregistration risk. Evaluation of motion correction should include both ellipticity and position stability assessment.
