Real-time, acquisition parameter-free voxel-wise patient-specific Monte Carlo dose reconstruction in whole-body CT scanning using deep neural networks - Scorecard - MDSpire
Advertisement
Real-time, acquisition parameter-free voxel-wise patient-specific Monte Carlo dose reconstruction in whole-body CT scanning using deep neural networks
Clinical Scorecard: Patient-specific Monte Carlo dose reconstruction in whole-body CT imaging using deep neural networks without the need for real-time acquisition parameters
At a Glance
Category
Detail
Condition
Radiation dose estimation in whole-body CT imaging
Key Mechanisms
Monte Carlo simulations combined with deep neural networks to estimate patient-specific 3D dose maps without requiring real-time acquisition parameters
Target Population
Patients undergoing whole-body CT imaging, including PET/CT hybrid scans
Care Setting
Radiology and nuclear medicine departments performing CT and hybrid imaging
Key Highlights
CT is a high-dose imaging modality contributing significantly to patient ionizing radiation exposure.
Monte Carlo (MC) simulations using patient-specific computational models are the gold standard for organ dose estimation but are computationally intensive.
Deep learning algorithms can enable real-time, fully automated patient-specific MC-based dose map estimation without needing detailed acquisition parameters.
Guideline-Based Recommendations
Diagnosis
Use CT imaging judiciously to balance diagnostic benefits with radiation exposure risks.
Consider patient-specific dose estimation to optimize imaging protocols.
Management
Apply Monte Carlo simulations for accurate organ dose calculations when feasible.
Incorporate deep learning-based dose reconstruction methods to facilitate real-time dose estimation in clinical workflows.
Monitoring & Follow-up
Monitor cumulative radiation dose, especially in serial CT examinations such as patient follow-up.
Use patient-specific organ dose data to assess radiation risk and guide imaging frequency.
Risks
High radiation doses from CT can reach deterministic levels in serial examinations.
Tube current modulation affects dose distribution and must be accounted for in dose estimation.
Patient & Prescribing Data
63 patients undergoing whole-body PET/CT scans with activated tube current modulation
Deep neural networks trained on Monte Carlo simulations can estimate patient-specific dose maps accurately and rapidly, potentially improving radiation risk assessment without requiring detailed acquisition parameters.
Clinical Best Practices
Utilize patient-specific computational models for organ dose estimation to enhance radiation risk assessment.
Incorporate deep learning algorithms to overcome computational limitations of Monte Carlo simulations in clinical settings.
Validate dose estimation methods across different tube voltages (e.g., 90 and 120 kVp) to ensure generalizability.
Review and confirm automated body contour segmentation to ensure accuracy in dose mapping.
