Monte-Carlo simulation of uncertainty in the estimation of 125I in the thyroid

Full text: At BARC a thin (76 mm dia x 2 mm thick) NaI (Tl) detector is used for the assessment of 125I in the thyroid of radiation workers engaged in preparation of radio-immunoassay kits. The detector is calibrated with a Remcal phantom with known amount of 125I activity filled in its thyroid cavity (volume 20 cc). Since 125I emits low energy photons ranging from 27 to 35.5 keV, its detection efficiency (DE) varies widely with 1) thickness of overlying tissue (OTT); 2) uncertainty in the neck to detector distance; 3) shape and 4) size of the thyroid. These factors introduce uncertainties in the estimate of 125I and therefore have to be accounted for while using DEs derived from experimental measurements with a physical phantom. Therefore a computer program based on Monte-Carlo photon transport techniques was developed to assess the uncertainties introduced in the activity estimates of the 125I in the thyroid due to the factors 1 to 3. For theoretical simulations, two thyroid models (volume: 20 cc) have been considered 1) Revised MIRD Head phantom (OTT=1.3 cm) and, 2) Ulanvosky model (OTT=0.6 cm). Response of the detector and the corresponding DEs have been calculated for a uniform distribution of the activity in the two models of the thyroid for neck to detector distances varying from 3 to 12 cm. The DEs computed for Ulanvosky model are found to be higher by about 30% than those for the MIRD model which is due to its much lower OTT values. The experimental DEs obtained with a Remcal phantom (OTT=1.3 cm) were found to be very similar to that of the MIRD model due to their similar OTTs. The Monte - Carlo simulation results obtained at detector distance of 3 cm for MIRD model indicated that an error in the measurement of detector distance of ± 0.5 cm would introduce a significant uncertainty in the DEs (∼ 11%). For a normal subject the thyroid OTT may vary from 0.6 to 2.0 cm, therefore the detector responses were determined for both the thyroid models for the above range of OTT simulated by shifting the thyroid location in the front-back direction w.r.t. the neck. The maximum uncertainty in DEs due to different OTT values (0.6 cm to 2.0 cm) than that of a calibrating phantom was found to be lying between +40% and -30% for a detector distance of 3 cm. The results obtained for the two thyroid models for same OTT (1.3 cm) and the same detector distance indicated that the effect of the thyroid shape on DE is small (< 8%) and is independent of the detector distance. In general, the uncertainties due to any of the factors at larger distances are found to be lower than those at smaller distances. This study has indicated that the uncertainties introduced due to differences in OTT are about 3 to 4 times higher than those caused by other factors and thus constitute a major source of uncertainty in the estimate of thyroidal burdens of 125I. These uncertainties should be taken into account along with other uncertainties caused by different sizes of the thyroid, placement errors of detector and calibration uncertainties to arrive at a reliable estimate of thyroidal burden of 125I in radiation workers. (author)