Comparative analysis of the conversion coefficient for internal dosimetry using different phantoms

Abstract

There are many ways to estimate the absorbed dose for internal radionuclides. They are used to assess the stochastic risks associated with diagnostic applications of nuclear medicine as well as the deterministic risks to normal tissues associated with therapeutic nuclear medicine. The difference between these methods is the anatomical models used for the dosimetric calculation. The phantoms used by the International Commission on Radiological Protection (ICRP) Publication 89 are based on medical image data of the average population in Eastern Europe and North American. While the phantoms used in OLINDA/EXM software are based on the stylized anatomical models developed for the MIRD Committee of The Society of Nuclear Medicine in the 1960s. Its organs were mathematically based on descriptive and schematic materials from general anatomy references and were modeled after the prescriptions of the ‘Reference Man’ defined by ICRP-23. However, the difference between these phantoms and real patients can lead to a disparity of the real absorbed dose. This work aims to assess these two methods (OLINDA/EXM and ICRP-106) to an individualized dosimetric method. This individualized method consists of a voxels phantom constructed from computed tomography images from a patient with anatomical characteristics similar to those of ICRP-89’s reference man coupled with EGSnrc Monte Carlo code. The comparison of these methods was performed using the relative error. This is used to quantify the discrepancy between two sets of data. The relative errors were calculated between the conversion coefficients obtained by the three methods for skin, brain, lungs, liver, kidneys and cortical bone. The lowest relative error values were observed in the brain, liver, and lungs. This occurs because these organs have a higher uptake of the 18F-FDG and are recognized by the ICRP as source organs in this biokinetic model. Compared to OLINDA/EXM and to ICRP-106, the relative differences in these organs did not exceed 10%. In the kidneys, the significant relative error (>10%) can be justified due to the stage of elimination of the radiopharmaceutical at the time the image is taken, so that there is no way to perform comparative dosimetry in this organ. For the skin and bone organs, the significant relative error occurred due to non-representation of the whole human body by the patient-specific phantoms developed, unlike the phantoms used in dose estimation using OLINDA/EXM and ICRP-106. Thus, it is evident that dose estimation using reference models results in discrepancies in dose distribution, mainly in pediatric or obese patients in whom anatomical differences are greater.