Computational phantoms have become an important tool in radiological science since the 1960s. These models of the human body help experts assess radiation doses. As computing technology advances, these phantoms have also changed, evolving from the original simple geometric models to voxelized phantoms based on actual medical images, allowing for more accurate assessment of radiation dose. The development of the first generation of computational phantoms marked the beginning of radiation dose assessment methods, and their evolution has had a profound impact on the use and safety of radiation to this day.
The development of computational phantoms not only improves the accuracy of radiation doses, but also provides necessary support for the testing of medical image reconstruction algorithms.
Prior to the 1960s, assessment of radiation dose relied primarily on simple geometric models that simply represented each organ as a sphere with an "effective radius." Most of the calculations during this period were based on very crude mathematical formulas, which could not truly reflect the anatomy of the human body. As technology evolves, scientists have begun to develop more human-like samples, such as the Shepp-Logan phantom, a phantom that simulates a human head and is used to verify the effectiveness of image reconstruction algorithms.
First-generation computational phantoms were designed to improve organ dose assessment obtained from internal radioactive material, and they evolved gradually based on simple geometric models.
The MIRD phantom was developed by Fisher and Snyder at Oak Ridge National Laboratory in the 1960s. This phantom contains 22 internal organs and more than 100 subregions and is the first anthropomorphic phantom used for internal dose calculations. Subsequently, a series of derivative Phantoms based on MIRD were developed, such as the "Family" Phantom series developed by Cristy and Eckerman in the 1980s, and the German "ADAM and EVA". These models not only facilitate accurate dose calculations, but also promote the development of radiation therapy and medical imaging.
The birth of these phantoms makes it feasible to calculate doses for different age groups and genders, thereby helping the medical community better conduct dose assessments.
In the 1980s, with the further improvement of computing technology, scientists gradually abandoned the early simple models and turned to voxel phantoms based on CT and MRI images. These models can reconstruct the true form of humans through high-resolution digital images. As technology evolved, researchers discovered that they could convert these images into a voxel format, digitally reconstruct the human body, and perform a variety of dose calculations.
Currently, there are more than 38 voxel phantoms available for a variety of uses, which not only improves the accuracy of simulations but also facilitates future research.
Although voxel ghosting provides more accurate data, its development also faces many challenges. For example, the radiation dose absorbed by the patient during CT image acquisition is relatively high, and the processing time of MRI images is also relatively long. In addition, a complete body scan series is crucial to obtain useful data, but is often difficult to achieve in reality. This makes data management another pressing problem. Although new computers can store large amounts of data, the memory required to process these high-resolution images is also very large.
With the continuous advancement of computing technology and the increasing demand for radiation dose assessment, future computational phantoms will be more flexible and accurate. The emergence of new boundary representation phantoms (BREP) and statistical phantoms will open up new possibilities for individual-specific dose calculations. These phantoms are no longer limited to static models, but can deform based on the shape and posture of the body, allowing for more realistic dose assessment.
The development of the first generation of computational phantoms laid the foundation for the improvement of subsequent technologies. So how can we use these computational phantoms in the future to further improve the application and safety of radiological medicine?