Modeling clustered activity increase in amyloid-beta positron emission tomographic images with statistical...

Area of Science:

• Neuroimaging
• Biophysics
• Medical Physics

Background:

• Amyloid-beta (Aβ) imaging using positron emission tomography (PET) detects Aβ plaques in the brain.
• Current analyses often rely on regional average tracer uptake, potentially missing detailed spatial information.
• Statistical image metrics, like the two-point correlation function (S2), offer advanced spatial pattern characterization.

Purpose of the Study:

• To adapt the S2 metric for analyzing human Aβ-PET brain imaging.
• To characterize longitudinal changes in Aβ plaque accumulation using 11C-Pittsburgh compound B (11C-PiB) scans.
• To evaluate a novel weighted two-point correlation function (wS2) for Aβ-PET analysis.

Main Methods:

• Developed a weighted two-point correlation function (wS2) for nonbinary, real-valued PET images.
• Calculated wS2 from serial 11C-PiB scans of cortical regions and whole-brain data.
• Compared the area under wS2 curves with standardized uptake value ratio (SUVR) and cerebrospinal fluid (CSF) Aβ measures.

Main Results:

• Longitudinal wS2 changes correlated with increased mean SUVR but exhibited lower variance.
• Whole-brain wS2 showed a stronger inverse correlation with CSF Aβ compared to mean SUVR.
• wS2 analysis was not confounded by region size or injected dose.

Conclusions:

• The wS2 metric detects subtle changes in Aβ accumulation.
• wS2 provides additional information on radiotracer binding characteristics beyond mean SUVR.
• This method enhances the understanding of Aβ plaque dynamics in the brain.