Population-Based Pharmacokinetic Modeling for [153Sm]Sm-EDTMP Kidney Dosimetry in Mice Muhammad Syifa (a), Shang Peter Chin (a), Nur Rahmah Hidayati (b), Heru Prasetio (b). Deni Hardiansyah (a*).
a) Medical Physics and Biophysics, Physics Department, Faculty of Mathematics and Natural Sciences, Universitas Indonesia, Depok, Indonesia
b) Research Center for Safety, Meteorology and Nuclear Quality Technology, Research Organization for Nuclear Energy, National Research and Innovation Agency
*corr author: denihardiansyah[at]sci.ac.id
Abstract
Accurate estimation of time-integrated activity coefficients (TIACs) is essential in preclinical dosimetry because it determines the reliability of absorbed dose calculations in translational applications. Conventional biodistribution studies commonly use population-averaged data, which neglect inter-individual variability by assuming a single kinetic profile represents the entire population. This study evaluated the feasibility of population-based nonlinear mixed-effects modeling (NLMEM) for preclinical kidney dosimetry of [153Sm]Sm-EDTMP in mice. Additionally, each mouse is represented by an individual subject and all fitting was done simultaneously. Twenty-one BALB/c mice were intravenously administered 3.7-7.4 MBq of [153Sm]Sm-EDTMP and sacrificed at 0.5, 1, 3, 5, 24, 48, and 72 hours post-injection, with three mice per time point. Kidney radioactivity was measured using an automated gamma counter and expressed as percentage injected dose per gram of tissue (%ID/g), generating a sparse-sampling dataset with one biodistribution measurement per mouse. Afterwards, Renal biokinetic data were fitted using one- to three-exponential models. Population-based model selection (PBMS) using goodness-of-fit criteria and Akaike weights was performed under both population-averaging and NLMEM approaches. TIAC estimates from the best NLMEM model were used as reference values and compared with those from the conventional averaging approach using mean absolute percentage error (MAPE) and root mean square error (RMSE). PBMS identified the same bi-exponential model as the best-fitting function under both the NLMEM and population-averaging approaches. Using NLMEM, the bi-exponential model generated individual TIAC estimates ranging from 0.10-0.15 h, whereas the averaging approach produced a single TIAC estimate of 0.12 h. Comparison showed a MAPE of 13.3% and RMSE of 14.8%, indicating generally close agreement between methods. However, unlike the averaging approach, NLMEM estimated TIACs for each individual mouse and characterized biokinetic variability within the population, suggesting its potential for more personalized preclinical dosimetry.
Keywords: nonlinear mixed-effects modeling- preclinical study- [153Sm]Sm-EDTMP- population pharmacokinetic modeling
