Extravasation Modeling from a CT/PET Imaging for a Possible Digital Twin Approach

Abstract

Extravasation of radiopharmaceuticals during PET procedures can compromise both diagnostic accuracy and patient safety by altering tracer pharmacokinetics and delivering unintended local radiation doses. A clinically observed 18F-FDG extravasation case has been reconstructed through a superposition of co-recorded CT and PET DICOM scan images to localize and quantify the extravasated activity. Building on advances in personalized computational modeling, the Digital Twin paradigm could offer a powerful framework for patient-specific dosimetry. Anatomical segmentation has been performed using Synopsys® Simpleware™ Medical to generate an accurate 3D model of the injection site and surrounding tissues. This Digital Twin has been completed through the build-up of an Unstructured Mesh domain that can be studied with the MCNP6© Monte Carlo code for particle transport simulations, incorporating the effective half-life of 18F-FDG to reflect dynamic radiotracer clearance. The digital twin has been benchmarked through OLINDA/EXM® code and thanks to dose-rate measurement performed with the Thermo Fisher Scientific Inc.© RadEye SPRD detector directly on the oncological injected patient. By embodying a patient’s anatomy, physiology, and radiotracer kinetics in a digital twin, this methodology provides a robust tool for personalized dosimetry of extravasation events. The framework is readily extensible to other radiopharmaceuticals and therapeutic contexts, supporting safer and more effective nuclear medicine practices.