Recent advances in using diffusion tensor imaging to study white matter alterations in Parkinson's disease: A mini review

PET, SPECT, and MRI imaging for evaluation of Parkinson's disease

This review assesses the primary neuroimaging techniques used to evaluate Parkinson's disease (PD) - a neurological condition characterized by gradual dopamine-producing nerve cell degeneration. The neuroimaging techniques explored include positron emission tomography (PET), single-photon emission computed tomography (SPECT), and magnetic resonance imaging (MRI). These modalities offer varying degrees of insights into PD pathophysiology, diagnostic accuracy, specificity by way of exclusion of other Parkinsonian syndromes, and monitoring of disease progression. Neuroimaging is thus crucial for diagnosing and managing PD, with integrated multimodal approaches and novel techniques further enhancing early detection and treatment evaluation.

Diffusion tensor metrics, motor and non-motor symptoms in de novo Parkinson's disease

INTRODUCTION

Parkinson's disease (PD) is a neurodegenerative disorder characterized by dopaminergic neurons' degeneration of the substantia nigra, presenting with motor and non-motor symptoms. We hypothesized that altered diffusion metrics are associated with clinical symptoms in de novo PD patients.

METHODS

Fractional Anisotropy (FA) and Mean (MD), Axial (AD), and Radial Diffusivity (RD) were assessed in 55 de novo PD patients (58.62 ± 9.85 years, 37 men) and 55 age-matched healthy controls (59.92 ± 11.25 years, 34 men). Diffusion-weighted images and clinical variables were collected from the Parkinson's Progression Markers Initiative study. Tract-based spatial statistics were used to identify white matter (WM) changes, and fiber tracts were localized using the JHU-WM tractography atlas. Motor and non-motor symptoms were evaluated in patients.

RESULTS

We observed higher FA values and lower RD values in patients than controls in various fiber tracts (p-TFCE < 0.05). No significant MD or AD difference was observed between groups. Diffusion metrics of several regions significantly correlated with non-motor (state and trait anxiety and daytime sleepiness) and axial motor symptoms in the de novo PD group. No correlations were observed between diffusion metrics and other clinical symptoms evaluated.

CONCLUSION

Our findings suggest microstructural changes in de novo PD fiber tracts; however, limited associations with clinical symptoms reveal the complexity of PD pathology. They may contribute to understanding the neurobiological changes underlying PD and have implications for developing targeted interventions. However, further longitudinal research with larger cohorts and consideration of confounding factors are necessary to elucidate the underlying mechanisms of these diffusion alterations in de novo PD.

Prediction of cognitive decline in Parkinson's disease based on MRI radiomics and clinical features: A multicenter study

OBJECTIVE

To develop and validate a multimodal combinatorial model based on whole-brain magnetic resonance imaging (MRI) radiomic features for predicting cognitive decline in patients with Parkinson's disease (PD).

METHODS

This study included a total of 222 PD patients with normal baseline cognition, of whom 68 had cognitive impairment during a 4-year follow-up period. All patients underwent MRI scans, and radiomic features were extracted from the whole-brain MRI images of the training set, and dimensionality reduction was performed to construct a radiomics model. Subsequently, Screening predictive factors for cognitive decline from clinical features and then combining those with a radiomics model to construct a multimodal combinatorial model for predicting cognitive decline in PD patients. Evaluate the performance of the comprehensive model using the receiver-operating characteristic curve, confusion matrix, F1 score, and survival curve. In addition, the quantitative characteristics of diffusion tensor imaging (DTI) from corpus callosum were selected from 52 PD patients to further validate the clinical efficacy of the model.

RESULTS

The multimodal combinatorial model has good classification performance, with areas under the curve of 0.842, 0.829, and 0.860 in the training, test, and validation sets, respectively. Significant differences were observed in the number of cognitive decline PD patients and corpus callosum-related DTI parameters between the low-risk and high-risk groups distinguished by the model (p < 0.05). The survival curve analysis showed a statistically significant difference in the progression time of mild cognitive impairment between the low-risk and the high-risk groups.

CONCLUSIONS

The building of a multimodal combinatorial model based on radiomic features from MRI can predict cognitive decline in PD patients, thus providing adaptive strategies for clinical practice.

Exploiting macro- and micro-structural brain changes for improved Parkinson's disease classification from MRI data

Parkinson's disease (PD) is the second most common neurodegenerative disease. Accurate PD diagnosis is crucial for effective treatment and prognosis but can be challenging, especially at early disease stages. This study aimed to develop and evaluate an explainable deep learning model for PD classification from multimodal neuroimaging data. The model was trained using one of the largest collections of T1-weighted and diffusion-tensor magnetic resonance imaging (MRI) datasets. A total of 1264 datasets from eight different studies were collected, including 611 PD patients and 653 healthy controls (HC). These datasets were pre-processed and non-linearly registered to the MNI PD25 atlas. Six imaging maps describing the macro- and micro-structural integrity of brain tissues complemented with age and sex parameters were used to train a convolutional neural network (CNN) to classify PD/HC subjects. Explainability of the model's decision-making was achieved using SmoothGrad saliency maps, highlighting important brain regions. The CNN was trained using a 75%/10%/15% train/validation/test split stratified by diagnosis, sex, age, and study, achieving a ROC-AUC of 0.89, accuracy of 80.8%, specificity of 82.4%, and sensitivity of 79.1% on the test set. Saliency maps revealed that diffusion tensor imaging data, especially fractional anisotropy, was more important for the classification than T1-weighted data, highlighting subcortical regions such as the brainstem, thalamus, amygdala, hippocampus, and cortical areas. The proposed model, trained on a large multimodal MRI database, can classify PD patients and HC subjects with high accuracy and clinically reasonable explanations, suggesting that micro-structural brain changes play an essential role in the disease course.