Modeling radiation dose and effects from internal emitters in nuclear medicine: from the whole body to individual cells.

In nuclear medicine, proper application of radiation protection principles depends on balancing the potential risks of exposure to ionizing radiation against its possible benefits. Average doses to organs, in diagnostic or therapeutic applications, are not always representative of the doses received at the tissue or cellular level. Therefore, understanding of the relationship between the overall biological effect and absorbed dose delivered by the radiopharmaceutical may require study of doses at the organ, tissue, or cell level. In this paper, we review current models for radiation dose assessment, with consideration of the different models and assumptions employed for study at all levels of investigation.

Heterogeneous distribution of technetium-99m-labeled microspheres in rat lungs: microautoradiographic evidence and dosimetric consequences.

UNLABELLED: The heterogeneity of 99mTc-labeled microspheres distribution within rat lung was visualized and quantified using a microautoradiographic "track" method (MAR). METHODS: MAR was used to study the uptake of radioactivity by individual microspheres, thereby enabling calculation of the range of particle activity. MAR was also used to visualize in rat lung sections the intrapulmonary distribution of the microspheres within the lungs after intravenous administration. The mean doses delivered to the cells in close contact with the labeled microspheres were calculated taking only the 99mTc electron emissions into account. RESULTS: All the microspheres were labeled. Nevertheless, the spectrum of visible tracks varied by a factor of 10, inducing a variable activity per microsphere from < 36 Bq to 325 Bq (mean activity-94 Bq/microsphere). No correlation existed between the radioactivity uptake and the size of microspheres. A very heterogeneous tridimensional distribution of the microspheres within the lungs were demonstrated with interparticle distances ranging from 57-4400 microns. On the other hand, only 1 of 2000 rat lung capillaries was obstructed. Using the mean activity, calculated delivered doses were found to reach approximately 6 Gy for the closest endothelial cells and 2 Gy for epithelial cells. However, such high doses were delivered to only a few cells. CONCLUSION: The number of obstructed capillaries in human lungs is lower than in rat lungs; the distances between microspheres should be larger. Nevertheless, the individual doses absorbed by the pulmonary cells closest to the microspheres should be very important.

Dosimetry at the cellular level of Kupffer cells after technetium-99m-sulphur colloid injection.

The radiation dose to Kupffer cells was estimated at the cellular level after intravenous injection of 99mTc labeled sulphur colloids in rats. The results were then compared with those obtained using macroscopic dosimetry. From the microscopy appearance observed using a "track" microautoradiographic method (MAR), it was shown that only 0.2% of the Kupffer cells were actually involved in the pinocytosis of radioactive colloids. For each electronic emission from 99mTc (Auger and internal conversion), the fraction of the emitted energy actually absorbed within the Kupffer cell was calculated using the values provided by Berger. About 15% of the total energy emitted by electrons was absorbed in 0.2% of the Kupffer cells. If these results are extrapolated to humans, the dose absorbed by the labeled cells can be estimated to be between 0.5 and 0.9 Gy/MBq. This represents about 15,000 times the average electron dose to the liver as estimated from macrodosimetric methods. In cases such as this one where an important distribution heterogeneity is expected, dosimetric estimations at a cellular level may be particularly useful.

Ten highly sensitive thyrotropin assays compared by receiver-operating characteristic curves analysis: results of a prospective multicenter study.

We report a prospective multicenter study, undertaken to compare the efficacy of 10 highly sensitive thyrotropin assay kits for the diagnosis of hyperthyroidism. Performances of the kits were compared with a reference diagnosis based on clinical examination, pertinent biological tests, and determination by an independent laboratory of the concentrations in serum of free triiodothyronine and free thyroxin. No thyrotropin determination was used in establishing this reference diagnosis. Receiver-operating characteristic curves were obtained for results from 600 patients (217 hyperthyroid and 383 euthyroid) by each kit. Even though analyses were performed out of the working range of most kits, the clinical correlation was nevertheless excellent. The best results corresponded to a sensitivity of 97.5% associated with a specificity of 96.1% and were significantly better than those obtained with all other kits. Results of this comparison depended greatly on the heterogeneity of the "normal"/"abnormal" categories. When only diffuse hyperthyroidism was considered, sensitivity and specificity were improved for all kits, and there was no significant difference among the five best kits.