Partial volume correction of PET image data using geometric transfer matrices based on uniform B-splines

Area of Science:

• Nuclear Medicine
• Medical Imaging
• Biophysics

Background:

• Traditional partial volume correction (PVC) in positron emission tomography (PET) relies on anatomical segmentation, limiting exploratory functional imaging.
• Existing methods struggle with noise and accuracy, particularly in low signal-to-noise ratio (SNR) scenarios common in PET.

Purpose of the Study:

• To introduce a novel, unbiased, voxel-wise partial volume correction (PVC) method for PET data.
• To evaluate the performance of the new method against established techniques using simulations and real-world PET datasets.

Main Methods:

• Developed the B-spline Geometric Transfer Matrix (bsGTM) method, combining B-spline basis functions with geometric transfer matrices.
• Evaluated bsGTM using Monte Carlo simulations, human PET data, and murine functional PET data.
• Compared bsGTM performance against iterative deconvolution, Gaussian convolution, and non-local means smoothing.

Main Results:

• bsGTM demonstrated comparable signal recovery to iterative deconvolution in simulations but with superior noise resilience.
• In murine PET data, bsGTM significantly enhanced sensitivity for detecting drug-induced binding potential changes.
• Human PET data showed bsGTM smoothing improved SNR with less degradation of binding potentials compared to other methods.

Conclusions:

• bsGTM provides superior performance for voxel-wise PVC compared to iterative deconvolution, especially in low SNR PET imaging.
• The method offers an anatomically unbiased approach to correct partial volume errors, valuable when segmentation is unavailable or unreliable.