Artificial intelligence-based rapid brain volumetry substantially improves differential diagnosis in dementia

Area of Science:

• Neurology
• Radiology
• Artificial Intelligence

Background:

• Accurate brain volumetry is crucial for diagnosing neurodegenerative diseases like Alzheimer's disease (AD) and frontotemporal dementia (FTD).
• Traditional volumetry methods are time-consuming and require specialized expertise.
• Deep learning-based artificial intelligence (AI) offers potential for rapid and automated analysis.

Purpose of the Study:

• To evaluate the clinical utility of a deep learning-based AI system for rapid brain volumetry.
• To assess the AI system's impact on diagnostic accuracy for AD and FTD.
• To determine performance gains across different levels of reader expertise.

Main Methods:

• Fifty-five patients (17 AD, 18 FTD, 20 controls) underwent MRI scans.
• Scans were analyzed twice by neuroradiologists, radiologists, and residents: once without AI and once with AI assistance.
• AI provided automatic lobe segmentation and age/sex-adjusted percentile comparisons.

Main Results:

• AI significantly improved diagnostic accuracy for AD (AUC: 0.800 vs. 0.926, p < 0.05).
• AI led to increased correct diagnoses (p < 0.01) and reduced errors (p < 0.03) for AD.
• Radiologists and residents showed notable performance gains with AI assistance (p < 0.05).
• AI improved FTD diagnosis accuracy for overall consensus, neuroradiologists, and radiologists (p < 0.05).

Conclusions:

• AI-assisted brain volumetry enhances diagnostic performance in differentiating AD and FTD.
• The AI tool significantly reduces processing time from hours to minutes.
• AI-supported volumetry has the potential to improve early diagnosis and patient management.