Clustering-based linear least square fitting method for generation of parametric images in dynamic FDG PET studies

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Computational Biology

Background:

• Dynamic Positron Emission Tomography (PET) generates functional 3D images but is susceptible to noise.
• High noise levels in dynamic PET data compromise the quality and reliability of parametric images.
• Existing methods often struggle to balance image quality with computational efficiency.

Purpose of the Study:

• To introduce a novel fitting method, cLLS, for generating high-quality parametric images from noisy dynamic PET data.
• To improve the statistical reliability of parametric images derived from dynamic PET studies.
• To address the challenge of noise sensitivity in dynamic PET imaging.

Main Methods:

• A modified linear least square (LLS) fitting method named cLLS was developed.
• cLLS incorporates a clustering-based spatial constraint to classify dynamic PET data.
• K-means and hierarchical cluster analysis were combined for data classification.

Main Results:

• The cLLS method demonstrated improved parametric image quality compared to conventional LLS.
• High statistical reliability was achieved without a significant increase in computational burden.
• Effectiveness was validated through computer simulations and a human brain FDG PET study.

Conclusions:

• The cLLS method offers a robust solution for generating reliable parametric images from noisy dynamic PET data.
• This technique is particularly promising for dynamic FDG PET studies, enhancing diagnostic capabilities.
• cLLS represents a significant advancement in processing high-noise dynamic PET imaging.