AI can identify nine types of dementia with one scan


2025-07-15 11:55:41 GMT+0800

 

Single-Scan AI Detects Nine Dementia Types with High Precision

A breakthrough AI diagnostic tool developed at Mayo Clinic can identify nine forms of dementia through a routine brain scan – significantly accelerating and improving detection accuracy for 55 million patients globally.

How StateViewer Works
The system analyzes fluorodeoxyglucose positron emission tomography (FDG-PET) scans to detect abnormal glucose metabolism patterns:

  1. Compares brain activity against dementia-specific databases

  2. Identifies disease signatures in distinct brain regions:

    • Alzheimer's: Memory processing areas

    • Lewy body dementia: Attention/motor regions

    • Frontotemporal dementia: Language/behavior zones

  3. Generates color-coded brain maps highlighting abnormalities

Performance Validation

  • Training data: 3,600+ scans (patients + healthy controls)

  • Accuracy: 88% overall disease identification

  • Speed: ~2× faster than standard methods

  • Precision: Up to 3× higher accuracy for subtype differentiation

Clinical Impact
Current dementia diagnosis requires:
☑️ Cognitive testing
☑️ Blood analysis
☑️ Multiple imaging studies
☑️ Specialist interpretation

StateViewer consolidates this complex workflow into one scan analysis, addressing critical challenges:

  • Alzheimer's is now the 5th leading cause of death globally

  • 10 million new dementia cases emerge annually

  • Overlapping symptoms frequently cause misdiagnosis

Future Development
Researchers plan to:
→ Enhance system performance across diverse populations
→ Expand clinical implementation
→ Investigate applications for early intervention

"By enabling precise, accessible diagnosis during routine scans," states the team, "StateViewer could transform treatment windows for millions before significant neurodegeneration occurs."



Please leave a message and We will get back to you in 12hrs.Thanks!