Evaluation of principal component analysis-based data-driven respiratory gating for positron emission tomography

Area of Science:

• Nuclear Medicine
• Medical Imaging
• Image Processing

Background:

• Respiratory motion significantly degrades PET/CT image quality and lesion uptake quantification.
• Respiratory gating is a key technique to mitigate motion artifacts in PET imaging.
• Current gating methods may have limitations in clinical settings.

Purpose of the Study:

• To evaluate a commercially developed data-driven gating (DDG) technique for clinical PET/CT.
• To assess the performance of a principal component analysis (PCA)-based DDG algorithm.

Main Methods:

• Applied a PCA-based DDG algorithm to phantom and patient FDG PET/CT data.
• Compared DDG-derived motion waveforms with an external infrared system (RPM).
• Reconstructed PET data using quiescent period gating (QPG) and analyzed image quality metrics.

Main Results:

• DDG demonstrated high correlation (r > 0.97) with the external RPM system for motion waveform extraction.
• QPG significantly improved image clarity and contrast compared to no motion compensation.
• Phantom recovery coefficients were comparable between DDG-QPG and RPM-QPG, and significantly higher than no motion compensation.

Conclusions:

• The PCA-based DDG algorithm reliably provides respiratory gating signals in phantom and patient PET/CT studies.
• The prototype commercial DDG algorithm shows potential for enabling reliable respiratory gating in routine clinical PET/CT.
• DDG offers a viable alternative for motion compensation in PET imaging.