RT Journal Article SR Electronic T1 Towards Non-Tracer Kinetic Modeling: Our Past Returning Through the Future JF Journal of Nuclear Medicine JO J Nucl Med FD Society of Nuclear Medicine SP P1447 OP P1447 VO 64 IS supplement 1 A1 Rahmim, Arman A1 Abdollahi, Hamid A1 Fele Paranj, Ali A1 Lee, Kyung-Nam A1 Saboury, Babak A1 Uribe, Carlos YR 2023 UL http://jnm.snmjournals.org/content/64/supplement_1/P1447.abstract AB P1447 Introduction: Kinetic modeling and dosimetry appear to be entirely distinct fields. The former typically applies to dynamic PET images following injection of ‘tracer’ amounts of radioactivity (i.e. causing no change in physiology). The resulting time-activity-curves (TACs) (at region or voxel level) are commonly decay corrected, followed by fitting of a tracer kinetic model, to extract micro- or macro-parameters. By contrast, in dosimetry as applied to radiopharmaceutical therapies (RPTs), one quantifies absorbed doses from TACs as generated over days (e.g. from Lu-177 SPECT). The TACs are not decay corrected, and are typically fit using mono- or multi-exponential curves. Dosimetry can be performed retrospectively, or predicted from realistic models (e.g. personalized physiologically-based pharmacokinetic (PBPK) models). We argue that these two fields are highly related, and that non-tracer kinetic models (proposed in the distant past but infrequently deployed) need to enter common practice in the theranostics era.Methods: Pharmacokinetic models that capture and quantify radiopharmaceutical dynamics can provide significant value towards understanding phenomena involved in RPTs and optimization of therapies for individual patients. Radiopharmaceuticals used in RPTs have varying specific activities, resulting in a mix of radiolabeled (‘hot’) and non-radiolabeled (‘cold’) ligands that are coupled with one another due to: (1) competition for binding to ligand binding sites, and (2) radioactive decay of hot ligands into cold ligands (which we model; Figure 1). In tracer kinetic modeling, these phenomena are not considered. Hot ligands are not considered to saturate binding sites (they are tracer amounts). In some past works, assuming low specific activities, cold ligands have been modeled as to possibly saturate binding sites. This can result in time-varying rates of binding to receptors (thus no longer involving rate ‘constants’). We model these phenomena, given the therapeutic regimen, as hot and cold ligands can considerably occupy binding sites. Specifically, we considered 3 models (Figure 2): (i) non-tracer kinetic model (M1; integrating above-mentioned phenomena), (ii) a kinetic model M2 only accounting for cold ligands saturating binding sites, and (iii) a common tracer kinetic model M3 not accounting for hot or cold ligands able to saturate binding sites. We first simulated, using a PBPK model fine-tuned to 177Lu-PSMA patient data, varying amounts of hot and cold ligands, from typical values (e.g. Hot/Cold in nmol’s: 7.5nmol/90nmol), to higher specific activities and more aggressive therapies that would be involved in RPTs with fewer cycles (e.g. 30nmol), and varying receptor densities (e.g. 1, 10, 50 nmol/L) and sizes. We analyzed the resulting TACs from the 3 above-mentioned models. Results: Figure 3 shows few example TACs for the 3 models for different amounts of hot and cold ligands and receptor densities. Given numerous variations (many results generated and to be shown), we find that the M3 model (the most approximate tracer model), and to a lesser extent M2, can result in different TACs than the correct non-tracer M1 model. Receptor saturation plots (not shown) depict that non-tracer kinetic models can result in distinct behaviours than correct model, and that the more appropriate non-tracer kinetic model M1 (accounting for competitive contributions of hot and cold ligands) should be used.Conclusions: One can ask: should multi-time-point SPECT measurements of therapeutic agents that emit gamma rays (e.g. Lu-177) be fit using common tracer kinetic models? Our evidence suggests the contrary. TACs generated in RPTs can be misleadingly generated and/or fit by tracer kinetic models. Non-tracer kinetic models have been considered in our past, but it is in the theranostic era that they may find extensive usage. Our past appears to be returning through the theranostics future that is upon us, staring at us directly.