Development of a non-uniform rational B-Spline (NURBS) phantom for preclinical incorporation dosimetry with pigs

Introduction: Preclinical incorporation dosimetry using the MIRD scheme requires the availability of computer-modeled phantoms representing the body anatomy and geometry of the applied animal models. In addition to rodents such as mice and rats, pigs are often used as animal models. However, a corresponding phantom for incorporation dosimetry with pigs does not exist. In this work, a porcine NURBS phantom was developed and prepared for the implementation in appropriate software tools for incorporation dosimetry.

Methods: First, PET-CT and PET-MRI images of a 8 weeks old and 15 kg female pig were acquired. The image data was segmented using 3D Slicer and registered to a voxel model. Using the CAD tool Rhinoceros 3D the surface of the voxel model was 'polygonised' and then smooth NURBS surfaces were modeled for the different organs. Furthermore, to verify the validity of the developed phantom for dosimetry calculations, S-values for the radionuclide 18-F were simulated. For this purpose, the NURBS phantom were 'revoxelised' using a combination of the software tool Binvox and an in-house Matlab code. In addition, the software tool OEDIPE was used to create the input files for the MCNPX simulations and to calculate the S-values from the output files.

Results: The developed NURBS phantom of the pig consists of 19 different organs. All organs have anatomically realistic and smooth surfaces, which can be scaled to any body geometry by moving the NURBS control points. Furthermore, the S-values for the radionuclide 18-F were simulated for all combinations of source and target organs. By comparison of the relative organ mass with reference values from the literature, the plausibility of the segmented organ masses and simulated S-values were demonstrated.

Conclusions: The developed porcine NURBS phantom is suitable for the use in preclinical incorporations dosimetry studies. It can be used to calculate S-values of various radionuclides and body geometries. The use of kinetic data of the radionulides in healthy tissue and tumors, along with the phantom S-values, could contribute in the optimisation of dosimetry for radionuclide therapy.