Scatter Correction with Combined Single-Scatter Simulation and Monte Carlo Simulation Scaling Improved the Visual...

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Radiological Physics

Background:

• 3D PET/CT imaging of the brain using ¹⁵O-gas inhalation is crucial for assessing cerebral blood flow and metabolism.
• High radioactivity in face masks used during imaging can cause cold artifacts, compromising quantitative accuracy.
• Conventional scatter correction methods, like single-scatter simulation with tail-fitting scaling (TFS-SSS), are susceptible to these artifacts.

Purpose of the Study:

• To evaluate the effectiveness of a novel scatter-correction method, Monte Carlo simulation-SSS (MCS-SSS), in addressing face mask-induced artifacts in 3D brain PET.
• To compare the performance of MCS-SSS against the conventional TFS-SSS method.

Main Methods:

• Phantom studies were conducted using a simulated face mask (plastic bottle) and brain (cylindric phantom) with varying levels of ¹⁸F-FDG radioactivity.
• Patient studies involved ¹⁵O-gas brain PET imaging in patients with cerebrovascular disease.
• Images were corrected using both TFS-SSS and MCS-SSS, with comparisons made against true activity concentrations and visual artifact assessment.

Main Results:

• TFS-SSS resulted in significant underestimation (up to 98%) of activity concentration in phantom studies, with noticeable cold artifacts near the simulated mask.
• MCS-SSS demonstrated an error within 5% in phantom studies, effectively minimizing cold artifacts, though minor high-activity artifacts were observed at very high radioactivity levels.
• In patient studies, TFS-SSS showed cold artifacts, while MCS-SSS produced artifact-free quantitative images of cerebral blood flow and metabolism.

Conclusions:

• MCS-SSS accurately corrects scatter in 3D ¹⁵O-gas brain PET imaging, effectively preventing cold artifacts caused by face masks.
• This advanced scatter-correction technique enhances the quantitative accuracy of PET imaging for brain studies.
• MCS-SSS is a valuable tool for reliable quantitative assessments in neuroimaging.