[en] The purpose of this study was to identify the specific metabolic brain pattern characteristic for Parkinson's disease (PD): Parkinson's disease-related pattern (PDRP), using network analysis of [18F]-fluorodeoxyglucose positron emission tomography (FDG-PET) brain images in a cohort of Slovenian PD patients. Twenty PD patients (age 70.1 ± 7.8 years, Movement Disorder Society Unified Parkinson's Disease Motor Rating Scale (MDS-UPDRS-III) 38.3 ± 12.2; disease duration 4.3 ± 4.1 years) and 20 age-matched normal controls (NCs) underwent FDG-PET brain imaging. An automatic voxel-based scaled subprofile model/principal component analysis (SSM/PCA) was applied to these scans for PDRP-Slovenia identification. The pattern was characterized by relative hypermetabolism in pallidum, putamen, thalamus, brain stem, and cerebellum associated with hypometabolism in sensorimotor cortex, posterior parietal, occipital, and frontal cortices. The expression of PDRP-Slovenia discriminated PD patients from NCs (p < 0.0001) and correlated positively with patients' clinical score (MDS-UPDRS-III, p = 0.03). Additionally, its topography agrees well with the original PDRP (p < 0.001) identified in American cohort of PD patients. We validated the PDRP-Slovenia expression on additional FDG-PET scans of 20 PD patients, 20 NCs, and 25 patients with atypical parkinsonism (AP). We confirmed that the expression of PDRP-Slovenia manifests good diagnostic accuracy with specificity and sensitivity of 85-90% at optimal pattern expression cutoff for discrimination of PD patients and NCs and is not expressed in AP. PDRP-Slovenia proves to be a robust and reproducible functional imaging biomarker independent of patient population. It accurately differentiates PD patients from NCs and AP and correlates well with the clinical measure of PD progression. (orig.)

[en] The purpose of this study is to develop and validate a 3D deep learning model that predicts the final clinical diagnosis of Alzheimer's disease (AD), dementia with Lewy bodies (DLB), mild cognitive impairment due to Alzheimer's disease (MCI-AD), and cognitively normal (CN) using fluorine 18 fluorodeoxyglucose PET (18F-FDG PET) and compare model's performance to that of multiple expert nuclear medicine physicians' readers. Retrospective 18F-FDG PET scans for AD, MCI-AD, and CN were collected from Alzheimer's disease neuroimaging initiative (556 patients from 2005 to 2020), and CN and DLB cases were from European DLB Consortium (201 patients from 2005 to 2018). The introduced 3D convolutional neural network was trained using 90% of the data and externally tested using 10% as well as comparison to human readers on the same independent test set. The model's performance was analyzed with sensitivity, specificity, precision, F1 score, receiver operating characteristic (ROC). The regional metabolic changes driving classification were visualized using uniform manifold approximation and projection (UMAP) and network attention. The proposed model achieved area under the ROC curve of 96.2% (95% confidence interval: 90.6-100) on predicting the final diagnosis of DLB in the independent test set, 96.4% (92.7-100) in AD, 71.4% (51.6-91.2) in MCI-AD, and 94.7% (90-99.5) in CN, which in ROC space outperformed human readers performance. The network attention depicted the posterior cingulate cortex is important for each neurodegenerative disease, and the UMAP visualization of the extracted features by the proposed model demonstrates the reality of development of the given disorders. Using only 18F-FDG PET of the brain, a 3D deep learning model could predict the final diagnosis of the most common neurodegenerative disorders which achieved a competitive performance compared to the human readers as well as their consensus.