Idea
AI platform simulating clinical-grade PET from MRI to improve dementia diagnosis accuracy and accessibility.
Research Paper
Core Innovation
This paper introduces SiM2P, a 3D diffusion bridge-based framework that probabilistically maps MRI and patient data to simulate diagnostic-quality FDG-PET images. It significantly improves diagnostic accuracy and certainty over MRI alone and requires minimal site-specific data for deployment, enabling practical local use.
Why It Matters
FDG-PET is a gold standard for dementia diagnosis but is costly and less accessible than MRI. Simulating PET from MRI reduces costs and expands access, enabling earlier and more accurate dementia detection in diverse healthcare settings. This can transform diagnostic workflows and improve patient outcomes globally, especially in resource-limited environments.
Market Size (TAM)
$10–20B TAM for dementia diagnostic imaging; $2–5B SAM from hospitals and diagnostic centers. Driven by aging populations and demand for cost-effective neuroimaging.
Potential Customers & Pain Points
- Hospitals – High cost and limited access to FDG-PET imaging
- Diagnostic centers – Need for improved dementia differentiation
- Healthcare providers in low-resource settings – Lack of PET infrastructure
- Radiologists and neurologists – Desire for higher diagnostic certainty and agreement.
Business Model
Subscription-based SaaS platform with tiered pricing for hospitals and diagnostic centers; licensing for integration with MRI vendors; optional on-premise deployment for data-sensitive institutions.
Competitive Landscape
- GE Healthcare
- Siemens Healthineers
- Philips Healthcare
- Qure.ai
Implementation Challenges
- Regulatory approval for clinical use
- Integration into existing diagnostic workflows
- Clinician trust and adoption of simulated imaging
- Data privacy and site-specific adaptation challenges
Validation Strategy
- Conduct multi-center clinical trials comparing SiM2P simulated PET to standard FDG-PET
- Obtain regulatory clearance for diagnostic use
- Partner with hospitals for pilot deployments and workflow integration
- Collect real-world usage data to refine model and demonstrate cost savings
Research Paper Overview
Diffusion Bridge Networks Simulate Clinical-grade PET from MRI for Dementia Diagnostics
Summary
SiM2P uses 3D diffusion bridge models to simulate FDG-PET images from MRI and patient data, improving diagnostic accuracy for dementia. It enhances differentiation between Alzheimer's, frontotemporal dementia, and healthy controls, increasing diagnostic certainty and interrater agreement. The framework requires minimal local data and basic demographics, enabling cost-effective, accessible PET-quality imaging in resource-limited settings.