A Proof-of-Concept Study of Personalized Dosimetry for Targeted Radioligand Therapyusing Pre-treatment Diagnostic Dynamic PET/CT and Monte Carlo Simulation

Thanh-Tai Duong^1*Danny De Sarno¹HATIM FAKIR¹Glenn Bauman¹Martin Martinov²Rowan Thomson²Ting-Yim Lee¹

• ¹Western University, London, Canada
• ²Carleton University, Ottawa, Ontario, Canada

Theranostics combines diagnostic imaging (e.g., ¹⁸F-PSMA PET) with targeted radioligand therapy (TRT; e.g., ¹⁷⁷Lu-PSMA-617), but personalized dosimetry remains challenging due to complex dose calculations and the impractical need for multi-day post-treatment SPECT/CT imaging. We propose a proof-of-concept framework that uses pre-treatment PET/CT and Monte Carlo modeling to predict patient-specific TRT doses, eliminating the need for serial post-treatment imaging. Dynamic 18F-DCFPyL PET/CT scans (22 min) were obtained from six prostate cancer patients. Tissue time-integrated activity (TIA)—the total decays from accumulated radioligand—was estimated as the product of an extrapolated arterial time-activity curve (TAC) area and the Logan distribution volume (LDV) derived from graphical analysis. Voxel-wise LDV showed excellent linearity (mean R² = 0.99997 ± 0.00005). We generated patient-specific 177Lu-PSMA-617 TIA maps and, along with CT-derived tissue geometry, used the egs_mird Monte Carlo engine to compute absorbed dose in prostate tumors, normal prostate tissue, and bone marrow. Biological effective dose (BED) was estimated using an extended linear quadratic model accounting for dose rate, DNA repair, and repopulation effects. Substantial interpatient dose variability was observed using a uniform radioligand dose: 10.4 ± 4.9 Gy/GBq in tumors, 5.1 ± 0.7 Gy/GBq in normal prostate, and 1.0 ± 0.3 Gy/GBq in bone marrow. These differences were driven by variability in LDV (radioligand binding) and arterial TACs (influenced by tumor burden and clearance). Our framework enables pre-treatment, voxel-based TRT dose estimation using diagnostic PET/CT alone. This method addresses patient-specific differences in pharmacokinetics and biodistribution that impact therapeutic response and toxicity—critical for optimizing TRT and combining it with external beam radiation therapy (EBRT) or brachytherapy when appropriate. This LDV-based pre-treatment dosimetry approach represents a step toward precision radiotheranostics by allowing dose personalization without the logistical burden of post-treatment SPECT/CT. While promising, this method requires further validation in larger cohorts and comparison with post-treatment dosimetry to confirm its clinical utility