Artificial Intelligence (AI)-assisted PET imaging is becoming a promising tool for Parkinson’s disease (PD) diagnosis. Our systematic review focused on the diagnostic accuracy of AI-assisted PET in detecting PD. We searched databases like Ovid MEDLINE, Ovid Embase, Web of Science, and IEEE Xplore for relevant studies developing AI algorithms in PET imaging, published by August 17, 2023. Binary diagnostic accuracy data for meta-analysis were extracted, yielding outcomes such as the area under the curve (AUC). Among 23 eligible studies, there was sufficient data for constructing contingency tables and calculating diagnostic accuracy.
This study underscores the acceptable diagnostic performance of PD using AI-assisted PET imaging across subgroups. To enhance future studies, we recommend implementing more rigorous reporting standards considering the unique challenges of AI research.
Source Nature