Anatomical realism and dosimetric accuracy: The role of novel computational phantoms in internal dosimetry

Internal dosimetry involves the quantification of the radiation dose delivered to organs and tissues following the intake of radionuclides through inhalation, ingestion, or injection routes. Unlike external exposure, internal exposure involves the complex spatial and temporal distribution of radioactive material within the body, leading to the irradiation of both the source and target tissues. Computational models used in internal dosimetry help to establish dose conversion factors, which are based on the total number of radioactive transformations in the source organ, considering the biokinetics and S-values for a set of interested source and target organs. Estimation of these S-values requires input from radiation transport simulations performed using a realistic representation of human anatomy, known as computational phantoms. These phantoms provide a detailed anatomical framework for translating the radionuclide activity distribution in tissues to internal dose, thereby eliminating the need for ethically unreasonable experiments. As they are essential for the dosimetric model, the geometry, size, composition, and spatial relationship of organs defined by the computational phantom directly influence the internal dose estimates. Early internal dose calculations relied on stylised computational phantoms composed of simple geometric shapes. Although computationally efficient, these models provided limited anatomical realism and introduced significant uncertainties in organ dose estimates. The development of voxel phantoms derived from medical images of real individuals marked a major improvement, as they offered realistic organ shapes, volumes, and spatial relationships. The limitations of voxel phantoms, including the inadequate voxel resolution to represent thin target tissues and the presence of nonflexible surfaces, led to the development of more advanced mesh-type computational phantoms. They represent the human body using deformable polygon or tetrahedral meshes of varying sizes and offer intricate detailing, including the micrometre-thick structures of the skin, respiratory and alimentary tract walls, and the urinary bladder. The International Commission on Radiological Protection (ICRP) has introduced mesh-type reference computational phantoms (MRCPs) for adults and pediatric reference individuals in recent years, which can be directly implemented in Monte Carlo particle transport codes such as Geant4, PHITS, MCNP, or FLUKA. With the availability of MRCPs, internal dosimetry undergoes significant refinement in terms of dosimetric accuracy. Because MRCPs offer greater flexibility in terms of size/shape adjustments, many research groups have created population-specific reference phantoms tailored to the physical parameters, body postures, and anatomical details of interest for specific applications. Detailed steps in creating body size-dependent phantom libraries for adults[1] (108 adult male and 104 adult female phantoms representing Caucasian and Asian populations) and the pediatric population[2] (total of 637 phantoms representing infants, children, and adolescents) from MRCPs have been elaborated in the literature. These studies have observed that the organ dose estimation for external exposure was strongly influenced by body size, particularly for photon energies below 0.1 MeV for adults and 0.05 MeV for the pediatric population. This suggests that using a body size-dependent phantom library compared to reference-sized phantoms can substantially improve the accuracy of individual dose estimates. The phantom library can also be implemented to estimate specific absorbed fractions (SAFs), as briefly done in another study,[3] incorporating morphometric variability to achieve more accurate internal dose assessments. The established mesh-type phantom library can also be used for developing personal dosimetry, especially for patients undergoing diagnostic and therapeutic medical exposures. Another common application of the computational phantoms involves accurate calibration of instruments used for in vivo monitoring of occupational workers or patients. Monte Carlo codes are used to model these instruments based on technical specifications. These are used in conjunction with computational phantoms of suitable physical sizes, organ sizes, source distributions, or types of radionuclides to investigate the required calibration for practical applications. In such cases, the use of mesh-type phantom library can also lead to person-specific calibration for the monitoring of occupational workers or patients. Although MRCPs represent the highest level of detailing in terms of human anatomy among computational phantoms, the skeletal target tissues (hematopoietic stem cells and osteoprogenitor cells) are not represented in them due to their microscopic and intricate histology. Detailed bone models for various skeletal regions were developed based on the microcomputed tomography (microCT) images of bone regions established at the University of Florida. These detailed bone models were installed in the spongiosa regions of the MRCPs to estimate fluence-to-dose response functions.[4] Further estimation of absorbed dose in skeletal regions using MRCPs revealed significant changes, primarily due to the anatomical improvements of the skeletal system in the MRCPs compared to that in voxel phantoms. The microCT-detailed bone models enable the direct calculation of absorbed doses to the active marrow and bone endosteum, potentially improving the consistency and accuracy of skeletal dosimetry. The most recent development in the field of MRCPs pertains to the creation of reference phantoms for pregnant females. The first part of this research involves the development of fetal phantoms that represent the anatomical features of the developing fetus, such as anthropometric parameters, bone ossification, and contents formation time.[5] Estimated SAFs using these phantoms show the fetal age-dependent dose trends. In the second part of this study, mesh-type maternal phantoms were developed by installing the fetal phantoms inside the uterus of the maternal phantoms.[6] The fully developed maternal MRCPs for various fetal ages (8, 10, 15, 20, 25, 30, 35, and 38 weeks) will serve as the only ICRP reference phantoms of this kind to estimate reference dose coefficients for fetal members of the public. The dosimetric impact of maternal radionuclide intake on the fetus can be evaluated using a comprehensive dataset of SAFs to be derived using these maternal MRCPs in the near future. Computational phantoms form the backbone of modern internal dosimetry by providing the anatomical realism required for accurate estimation of internal dose following radionuclide intake. The evolution from stylised and voxel-based models to MRCPs has significantly reduced the uncertainties and enabled refined dose estimations for various exposure scenarios. The ongoing research using advanced computational phantoms highlights their crucial role in achieving reliable, ethical, and scientifically robust internal dosimetry for radiation protection purposes. Financial support and sponsorship Nil. Conflicts of interest There are no conflicts of interest.