Detection of parkinson's disease with neuroimaging modalities using machine learning and artificial intelligence: a systematic review.
Authors
Affiliations (8)
Affiliations (8)
- Department of Radiology and Biomedical Imaging, School of Medicine, Yale University, 801 Howard Ave, PO Box 208048, New Haven, CT, USA. [email protected].
- Department of Radiology and Biomedical Imaging, School of Medicine, Yale University, 801 Howard Ave, PO Box 208048, New Haven, CT, USA.
- Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA.
- School of Medicine, Isfahan University of Medical Sciences, Isfahan, Iran.
- Cushing/Whitney Medical Library, School of Medicine, Yale University, New Haven, CT, USA.
- Department of Radiology, University of Pennsylvania, Philadelphia, PA, USA.
- Department of Psychiatry, School of Medicine, Yale University, New Haven, CT, USA.
- Department of Neurology, School of Medicine, Yale University, New Haven, CT, USA.
Abstract
The application of machine learning (ML) and artificial intelligence (AI) algorithms in medical imaging is an emerging area of interest, particularly in the context of clinical decision-making. Here, we report on the overall performance (i.e., sensitivity, specificity, and accuracy) of commonly used ML/AI techniques including convolutional neural networks (CNNs), support vector machines (SVMs), random forests, and ensemble approaches on the clinically relevant task of distinguishing between Parkinson's disease (PD) participants and matched healthy controls (HC). Our systematic review includes 130 studies from six different imaging modalities - dopamine transporter scans ([<sup>123</sup>I]Ioflupane single-photon emission computed tomography (SPECT)), positron emission tomography (PET) including [<sup>18</sup>F]FDG, [<sup>18</sup>F]DOPA, and [<sup>11</sup>C]raclopride, structural magnetic resonance imaging (MRI) (T1- and T2-weighted), functional MRI, and diffusion MRI. While some findings were in line with expectations for some modalities, such as the superior performance of dopamine SPECT and PET (> 90% sensitivity, specificity, and accuracy with methods like convolutional neural networks), others were more nuanced with the best-performing class of algorithms depending on the imaging modality, and sometimes, even the data source. Overall, we summarize the emerging trends across studies for each imaging technique and provide valuable recommendations for future lines of inquiry.