Accuracy of Single Time-Point Renal Dosimetry for [177Lu]Lu-PSMA-617: An NLMEM-Based Virtual Patient Study of Typical and Deviating Kinetic Profiles Teuku Muhammad Zaki Yasykur Polem (a*), Heru Prasetio (b), Deni Hardiansyah (a)
a) Medical Physics and Biophysics, Physics Department, Faculty of Mathematics and Natural Sciences, Universitas Indonesia, Depok, Indonesia
*teuku.muhammad11[at]ui.ac.id
b) Research Center for Safety, Meteorology and Nuclear Quality Technology, Research Organization for Nuclear Energy, National Research and Innovation Agency, South Tangerang, Indonesia
Abstract
Accurate estimation of renal Time-Integrated Activities (TIAs) is crucial in [177Lu]Lu-PSMA-617 radioligand therapy to optimize treatment and minimize the risk of nephrotoxicity. While single time-point (STP) imaging reduces clinical burden compared to multi-time-point (MTP) protocols, the accuracy of STP in outlier patients whose kinetic parameters largely deviate from the population mean remains insufficiently characterized. A cohort of 80 virtual patients was simulated based on previously published population biokinetic parameters and distributions for renal [177Lu]Lu-PSMA-617 (Hardiansyah et al., 2024- PMID: 38423787). Simulations were performed at five time points: 1.8, 18.7, 42.6, 66.3, and 160.3 hours. Using this baseline virtual population, two sub-cohorts (n = 20 each) were generated: a typical cohort with biokinetic parameters within one standard deviation (SD) of the population mean, and an outlier cohort with biokinetic parameters exceeding two SDs from the population mean. Pharmacokinetic fitting was performed via Non-Linear Mixed Effects Modeling (NLMEM) using a sum-of-exponentials function established through population-based model selection. STP TIAs calculated at the 42.6-hour time point were evaluated against reference MTP TIAs using relative deviation, root-mean-square error (RMSE), and mean absolute percentage error (MAPE). In the typical cohort, the STP approach demonstrated high accuracy, with an RMSE of 8.43%, and a MAPE of 6.83%. In contrast, the outlier cohort showed substantial underestimation and increased error, with an RMSE of 23.48%, and a MAPE of 20.33%. Renal TIA calculation using an STP at 42.6 hours provides reliable estimates for patients with typical biokinetics. However, increased errors observed in the outlier cohort suggest increased variability of STP-based estimation in patients with biokinetic parameters deviating substantially from the population mean, and that MTP sampling may be required in such cases.
Keywords: [177Lu]Lu-PSMA-617, Single Time-Point, Time-Integrated Activity, Virtual Patients
