Semiautomated Algorithm for the Diagnosis of Multiple System Atrophy With Predominant Parkinsonism

Area of Science:

• Neuroimaging
• Neurology
• Medical Diagnostics

Background:

• Putaminal iron deposition is a key differentiator between multiple system atrophy with predominant parkinsonism (MSA-p) and Parkinson's disease (PD).
• Previous differentiation methods relied on subjective visual inspection or complex manual quantitative techniques.
• Advanced imaging analysis is crucial for accurate and efficient diagnosis of neurodegenerative parkinsonian syndromes.

Purpose of the Study:

• To evaluate a novel semiautomated diagnostic algorithm for differentiating MSA-p from PD.
• To assess the algorithm's performance using 3-Tesla (3T) MR susceptibility-weighted imaging (SWI).
• To investigate the utility of putaminal iron distribution patterns in distinguishing these conditions.

Main Methods:

• Developed a two-step semiautomated algorithm to identify the putaminal margin and calculate phase-shift values reflecting iron concentration.
• Included 26 MSA-p patients, 68 PD patients, and 41 normal controls (NC).
• Optimized differentiation by analyzing combinations of phase-shift values, vertical pixels, and dominant sides after normalization.

Main Results:

• The algorithm successfully identified putaminal margins and revealed an anterior-to-posterior iron gradient in MSA-p.
• The optimized algorithm achieved an area under the receiver operating characteristic curve of 0.874 (80.8% sensitivity, 86.7% specificity) for MSA-p vs. PD.
• Algorithm performance improved in MSA-p patients with longer disease durations.

Conclusions:

• The developed semiautomated algorithm accurately detects putaminal margins and iron distribution.
• It demonstrates significant potential for differentiating MSA-p from PD.
• This method offers a more objective and efficient approach to diagnosing parkinsonian disorders based on iron deposition patterns.