Dosimetric analysis of (123)I, (125)I and (131)I in thyroid follicle models

BACKGROUND: Radioiodine is routinely used or proposed for diagnostic and therapeutic purposes: (123)I, (125)I and (131)I for diagnostics and (125)I and (131)I for therapy. When radioiodine-labelled pharmaceuticals are administered to the body, radioiodide might be released into the circulation and taken up by the thyroid gland, which may then be an organ at risk. The aim of this study was to compare dosimetric properties for (123)I, (125)I and (131)I in previously developed thyroid models for man, rat and mouse. METHODS: Dosimetric calculations were performed using the Monte Carlo code MCNPX 2.6.0 and nuclear decay data from ICRP 107. Only the non-radiative transitions in the decays were considered. The S value was determined for the cell nuclei in species-specific thyroid follicle models for mouse, rat and man for different spatial distributions of radioiodine. RESULTS: For the species-specific single follicle models with radioiodine homogeneously within the follicle lumen, the highest S value came from (131)I, with the largest contribution from the β particles. When radioiodine was homogeneously distributed within the follicle cells or the follicle cell nucleus, the highest contribution originated from (125)I, about two times higher than (123)I, with the largest contribution from the Auger electrons. The mean absorbed dose calculated for our human thyroid multiple follicle model, assuming homogenous distribution of for (123)I, (125)I, or (131)I within the follicle lumens and follicle cells, was 9%, 18% and 4% higher, respectively, compared with the mean absorbed dose according to Medical Internal Radiation Dose (MIRD) formalism and nuclear decay data. When radioiodine was homogeneously distributed in the follicle lumens, our calculations gave up to 90% lower mean absorbed dose for (125)I compared to MIRD (20% lower for (123)I, and 2% lower for (131)I). CONCLUSIONS: This study clearly demonstrates the importance of using more detailed dosimetric methods and models than MIRD formalism for radioiodine, especially (123)I and (125)I, in the thyroid. For radioiodine homogeneously distributed in the follicle lumens our calculations for the human multiple follicle models gave up to 90% lower mean absorbed dose compared with MIRD formalism.