A new iterative reconstruction technique for attenuation correction in high-resolution positron emission tomography

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Image Reconstruction

Background:

• Positron Emission Computed Tomography (PET) imaging is crucial for medical diagnostics.
• Accurate image reconstruction is vital for reliable PET analysis.
• Existing methods like filtered backprojection (FBP) have limitations in noise handling and resolution.

Purpose of the Study:

• To introduce and evaluate a novel iterative reconstruction technique (NIRT) for PET imaging.
• To assess NIRT's performance in nonuniform attenuation correction using transmission data.
• To compare NIRT against the standard FBP method for image quality and accuracy.

Main Methods:

• Developed a new iterative reconstruction technique (NIRT) based on general inverse problem theory.
• Derived a cost functional incorporating a noise term.
• Minimized the cost functional using a weighted-least-square maximum a posteriori conjugate gradient (CG) method.
• Acquired real PET data using a cylinder phantom and a Hoffman brain phantom with fluorine-18.

Main Results:

• NIRT demonstrated rapid convergence and produced high-quality reconstructed images.
• Images reconstructed with NIRT exhibited superior noise properties, with 1.75 times less rms% noise than FBP.
• NIRT achieved a lower distance metric (0.328) for the Hoffman brain phantom compared to FBP (0.526), indicating higher accuracy.
• Noise propagation was suppressed without significant loss of resolution.

Conclusions:

• The new iterative reconstruction technique (NIRT) effectively reduces noise in PET images.
• NIRT provides improved image accuracy and resolution compared to the standard FBP method.
• NIRT shows significant potential for enhancing diagnostic performance in PET imaging.