Research Proposal: Simulation Platform for Radionuclide Pharma Tracer Development

[devhliu]

Summary

This proposal outlines a technical development plan for a simulation platform to accelerate the design and verification of novel diagnostic and therapeutic radionuclide tracers. The system will simulate dynamic tracer uptake (small molecules, peptides, antibodies, etc.) in human organs and lesions, convert these models into PET/SPECT images for ultra-low-dose imaging validation, and compute dosimetry and therapeutic effectiveness.

Objectives

1. Simulate Dynamic Uptakes:
   • Model physiochemical and pharmacokinetic processes (PDPK, adsorption, distribution, extraction, etc.) for various tracer types.
   • Enable simulation in realistic organ and lesion geometries.
2. Imaging Simulation:
   • Convert simulated tracer uptake curves to PET or SPECT images.
   • Support ultra-low-dose imaging protocols for preclinical and clinical verification.
   • Techniques: organ segmentation, lesion inpainting, Monte Carlo-based imaging simulation.
3. Dosimetry and Effectiveness Calculation:
   • Use Monte Carlo methods to estimate dose distributions in segmented organs and lesions.
   • Analyze kinetics and accumulated Time Activity Curves (TACs) for efficacy prediction.

Technical Approach

• Organ and Lesion Segmentation: Deep learning or atlas-based segmentation for anatomical modeling.
• Lesion Inpainting: GAN or diffusion-based models to simulate diverse lesion appearances.
• Monte Carlo Simulation: For both imaging (PET/SPECT) and dosimetry (voxel-based dose calculation).
• Kinetics Modeling: PDPK modules for tracer distribution and TAC simulation.

Expected Impact

• Accelerate tracer development and preclinical imaging validation.
• Enable ultra-low-dose imaging studies to reduce patient exposure.
• Provide actionable dosimetry for therapeutic tracer evaluation.

References to Similar Techniques and Open-Source Projects (GitHub)

• EGSnrc: Monte Carlo simulation for dosimetry and imaging.
• OpenGATE: Voxel-based dosimetry, PET/SPECT simulation, memory leak reporting in voxelized phantoms.
• VICTRE_MCGPU: GPU-accelerated Monte Carlo X-ray imaging, used in breast phantom dosimetry.
• matRad: Monte Carlo and analytical methods for radiotherapy dose validation.
• MMC: Mesh-based Monte Carlo for photon transport.
• AutoPET II challenge: Multi-label segmentation for PET/CT lesion/organs using nnUNet (see paper).
• SAMUS: Segment Anything Model adapted for ultrasound and medical image segmentation.
• SHARM: Segmented anatomical head models for dosimetry/segmentation benchmarking.

Note: Search results are incomplete (37 results found, only 30 shown). For more, see GitHub issue search.

Next Steps

• Develop modular simulation pipeline for uptake modeling, imaging conversion, and dosimetry.
• Integrate and benchmark with existing open-source tools above.
• Establish protocols for ultra-low-dose PET/SPECT verification and therapy planning.