The main objective is to estimate the spatial resolution of brain PET images.

The spatial resolution of brain PET images is often estimated using Hoffman phantom data. Unfortunately, Hoffman phantom images may not always be readily available, and phantom-based approaches may yield sub-optimal results. We propose a new, computational approach that allows estimation of spatial resolution directly from the PET image itself.

We generalized the logarithmic intensity plots reported by Mizutani and colleagues (Mizutani, 2016) in order to perform spatial resolution estimation in both axial and in-plane directions. The FWHMs are estimated from multiple regression of the logarithm of the square norm of the image Fourier transform against the square distance from the origin in the Fourier domain. The proposed approach was applied to Amyloid and FDG PET images from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) study.

We obtained in-plane and axial FWHMs resolution estimators that vary between 3.5 mm and 8 mm for both [18F]florbetapir and [18F]FDG images. In both cases, we observed a low within-scan variability. We also obtained a strong cross-tracer consistency in FWHM resolution estimators, which is an expected result since the spatial resolution does not depend on the particular 18F-labeled tracer.

We obtained small (less than a voxel size) across-subject variability in groups of subjects sharing the same PET site and reconstruction parameters. Our novel approach not only eliminates the need for surrogate brain phantom data, but also provides a general framework that can be applied to a wide range of tracers and other image modalities, such as SPECT.

References

Mizutani et al., J. Microscopy, 261: 57–66, 2016.