Advanced brain aging in multiple system atrophy compared to Parkinson's disease

Area of Science:

• Neuroimaging
• Neurodegenerative Diseases
• Biomarkers

Background:

• Multiple system atrophy (MSA) and Parkinson's disease (PD) are alpha-synucleinopathies with distinct clinical trajectories.
• Early differentiation between MSA and PD is crucial for effective treatment strategies.
• Neuroimaging-based brain age prediction offers a potential method to identify aberrant brain aging in neurodegenerative conditions.

Purpose of the Study:

• To investigate the utility of brain-predicted age difference (PAD) in differentiating between MSA and PD using deep learning on MRI data.
• To identify specific neuroanatomical features contributing to brain age differences in MSA and PD patients.

Main Methods:

• Recruited patients with MSA (N=23), PD (N=33), and healthy controls (HC, N=34).
• Applied a deep learning approach to estimate gray matter (GM-PAD) and white matter (WM-PAD) predicted age difference from T1-weighted and diffusion-weighted MRI scans.
• Utilized spatial normative models to quantify neuroanatomical impairments and their contribution to PAD.

Main Results:

• MSA patients exhibited significantly higher GM-PAD (9.33 years) and WM-PAD (9.27 years) compared to PD patients (GM-PAD: 0.75 years; WM-PAD: 1.90 years) and HC (GM-PAD: -1.47 years; WM-PAD: -0.74 years).
• No significant difference in PAD was found between PD and HC groups.
• Distinct image features contributed to PAD in MSA (orbitofrontal GM, central corpus callosum WM) versus PD (cuneus GM, uncinate fasciculus WM).

Conclusions:

• Deep learning-based brain age prediction can effectively differentiate MSA from PD.
• Significant differences in brain aging patterns, driven by distinct neuroanatomical contributions, exist between MSA and PD.
• These distinct imaging features hold potential as biomarkers for the differential diagnosis of MSA and PD.