Jun 7 – 11, 2026
Prague, Czechia
Europe/Prague timezone

Uncertainty in Plutonium Dose Modeling: Lessons from USTUR Autopsy Data

Jun 8, 2026, 4:15 PM
15m
Auditorium 115

Auditorium 115

Břehová 7, Prague 1
Oral Presentation Dosimetry and radiation protection in industry Dosimetry and radiation protection in industry

Speaker

Dr Maia Avtandilashvili (United States Transuranium and Uranium Registries, Washington State University)

Description

The ICRP biokinetic and dosimetric models do not explicitly incorporate dosimetric uncertainty. Time‑dependent radionuclide activity retained in organs is directly related to radiation dose; hence, uncertainties in biokinetic model predictions of organ activities can be used as surrogates for uncertainties in dose. Bayesian methods, which provide full posterior dose distributions rather than point estimates, have been used in radiation epidemiology under the assumption that the true dose lies within the modeled posterior distribution. Thirty-five voluntary tissue donors to the United States Transuranium and Uranium Registries (USTUR), the former workers from Los Alamos, Rocky Flats, and Hanford with known internal plutonium deposition, were used to examine uncertainties in predicted organ activities. Available data included urine bioassay and post‑mortem measurements of 239Pu in the liver, skeleton, and respiratory tract. Measured organ activities ranged 3.5–920 Bq in the liver, 3.3–774 Bq in the skeleton, and 0.2–6550 Bq in the respiratory tract. Model uncertainty was evaluated by varying two parameters of the ICRP human respiratory tract model using Latin hypercube sampling: the rapidly dissolving fraction (fr) and the slow dissolution rate (ss). For each parameter set, intake was estimated by fitting urine data, and organ activities were predicted. The modeled distributions of organ activity predictions were then compared with measured post‑mortem values. The predicted activity distributions failed to cover the measured values in most cases. Specifically, modeled distributions did not include measured activities in 74% and 89% cases for liver and skeleton, respectively. For the respiratory tract, urine bioassay data were not informative; predicted distributions were extremely wide (geometric mean ratio of 2.5%–97.5% quantiles = 997), yet measured values still lay outside these bounds in 9 of 35 individuals (26%). This challenges the assumption that Bayesian posterior dose distributions always contain the true dose.

Authors

Dr Maia Avtandilashvili (United States Transuranium and Uranium Registries, Washington State University) Dr Martin Šefl (AMEDIS, spol. s r.o.) Dr Joey Zhou (United States Department of Energy) Dr Sergey Tolmachev (United States Transuranium and Uranium Registries, Washington State University)

Presentation materials

There are no materials yet.