PET image denoising using a synergistic multiresolution analysis of structural (MRI/CT) and functional datasets.

Area of Science:

• Medical Imaging
• Image Processing
• Neuroscience

Background:

• Positron Emission Tomography (PET) provides functional imaging but suffers from lower resolution and signal-to-noise ratio compared to structural modalities like CT and MRI.
• Limitations in injected radioactivity, instrumentation, and inherent decay processes compromise PET image quality.
• Integrating high-resolution structural information can potentially enhance PET image quality.

Purpose of the Study:

• To develop and validate a multiresolution model for enhancing PET image quality using structural information from CT or MRI.
• To improve the signal-to-noise ratio of PET images while preserving spatial resolution.
• To demonstrate the method's applicability to other functional imaging modalities.

Main Methods:

• A multiresolution approach utilizing the wavelet transform (WT) was employed.
• Structural images (MRI/CT) were co-registered and downscaled to match PET resolution.
• A linear model integrated wavelet coefficients from both modalities, followed by Gaussian mixture modeling and k-means clustering for calibration.

Main Results:

• The method demonstrated effective noise reduction, achieving a 15% standard deviation decrease.
• Image resolution was successfully preserved during the enhancement process.
• Validation on simulated and clinical datasets confirmed the quantitative potential for both individual and group analyses.

Conclusions:

• The proposed methodology offers a robust and practical solution for enhancing PET image quality.
• It is computationally efficient and resilient to minor co-registration errors.
• The technique holds promise for improving the diagnostic value of functional imaging.