An automatic analysis framework for FDOPA PET neuroimaging.

Area of Science:

• Neuroimaging
• Dopamine Neuroscience
• Medical Imaging Analysis

Background:

• FDOPA PET neuroimaging is crucial for assessing dopamine synthesis capacity.
• Existing analysis methods can be labor-intensive and vary in reproducibility.
• A standardized, automated approach is needed for large-scale FDOPA PET data analysis.

Purpose of the Study:

• To develop and validate a fully automated analytical framework for FDOPA PET neuroimaging data.
• To assess the framework's sensitivity to demographic and experimental variables.
• To ensure the replicability and reproducibility of FDOPA PET data quantification.

Main Methods:

• Re-engineered existing Matlab scripts into a Python-based automated analysis pipeline.
• Integrated the pipeline into the XNAT imaging platform.
• Utilized a repository of 892 FDOPA PET scans from 23 studies for validation.

Main Results:

• The automated pipeline demonstrated good reproducibility in striatal FDOPA uptake quantification (ICC values ranging from 0.71 to 0.88).
• Gender significantly influenced striatal dopamine synthesis capacity, with women showing higher capacity than men (F=10.7, p<0.001).
• The framework proved robust across a large dataset combining multiple studies.

Conclusions:

• The automated FDOPA PET analysis pipeline provides a standardized and robust method for quantifying dopamine synthesis capacity.
• The findings underscore the importance of considering gender in FDOPA PET studies.
• This validated framework facilitates comprehensive analysis and comparison across diverse neuroimaging studies.