Development of a new CBR-based platform for human contamination emergency situations

Area of Science:

• Medical Physics
• Radiological Protection
• Computational Modeling

Background:

• Dosimetry evaluation is critical following accidental human exposure in radiological emergencies.
• Current dosimetry often relies on generic 'Reference Man' models, limiting personalized accuracy.
• Personalized and realistic human phantoms are frequently unavailable for exposed subjects.

Purpose of the Study:

• To introduce the EquiVox platform for selecting the most adapted voxel phantom for individual dosimetry.
• To demonstrate the efficiency of EquiVox in optimizing in vivo lung monitoring.
• To present the development of artificial neural network tools for phantom adaptation.

Main Methods:

• Development of the EquiVox platform utilizing case-based reasoning principles.
• Retrieval of the most suitable existing phantom from a database to represent a victim.
• Application of the platform to optimize in vivo lung monitoring scenarios.
• Ongoing development of artificial neural networks for model adaptation.

Main Results:

• The EquiVox platform effectively retrieves the most appropriate phantom for dosimetry.
• Demonstrated efficiency in optimizing in vivo lung monitoring through appropriate phantom selection.
• Validation of the case-based reasoning approach for phantom selection.

Conclusions:

• The EquiVox platform enhances the accuracy of dosimetry evaluations in radiological emergencies.
• Personalized phantom selection via EquiVox improves treatment planning and accuracy.
• Artificial neural networks show promise for further refining phantom adaptation to individual victims.