AUTHOR=Li Xue-ning , Hao Da-peng , Qu Mei-jie , Zhang Meng , Ma An-bang , Pan Xu-dong , Ma Ai-jun 

TITLE=Development and Validation of a Plasma FAM19A5 and MRI-Based Radiomics Model for Prediction of Parkinson’s Disease and Parkinson’s Disease With Depression

JOURNAL=Frontiers in Neuroscience

VOLUME=Volume 15 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2021.795539

DOI=10.3389/fnins.2021.795539

ISSN=1662-453X

ABSTRACT=Background: In the clinical management of Parkinson’s disease (PD), prediction and early diagnosis of PD and Parkinson’s disease with depression (PDD) is essential. 
Objectives:Thus, a plasma FAM19A5 and MRI-based radiomics nomogram was developed to predict PD and PDD. 
Methods: 176 PD patients and 181 healthy controls (HC) were enrolled. The plasma FAM19A5 concentration of all participants was measured using ELISA technology. Among the enrolled subjects, MRI data from 164 individuals (82 in PD group, 82 in HC group) was collected. The bilateral amygdala, head of the caudate nucleus, putamen, and substantia nigra and red nucleus were manually labeled on the MR images. Radiomics features of labeled regions were extracted. Machine learning methods were applied to shrink the feature size and build a predictive radiomics signature. The radiomic signature, combined with the plasma FAM19A5 concentration and other risk factors was used to establish logistic regression models for prediction of PD and PDD. 
Results: The plasma FAM19A5 level (2.456±0.517) in the PD group was significantly higher than the HC group (2.231±0.457) (P<0.001); plasma FAM19A5 was also significantly higher in the PDD subgroup (2.577±0.408) compared to the non-depressive subgroup (2.406±0.549) (P= 0.045 < 0.05). The model of plasma FAM19A5 combined with the radiomics signature showed excellent predictive validity for PD and PDD with AUCs of 0.913 (95% CI: 0.861-0.955) and 0.937 (95% CI: 0.845-0.970), respectively. 
Conclusions: Nomograms incorporating a radiomics signature, plasma FAM19A5, and clinical risk factors may serve as potential tools for early prediction of PD and PDD in clinical.