Abstract
2346
Introduction: Radiopharmaceutical therapies (RPTs) with alpha-emitting radioisotopes have shown promising results. However, cancers often manifest with large inter- or intra-patient variation in tumor phenotypes. The extent of necrotic and hypoxic regions or the size of the tumor can influence the radiopharmaceutical uptake patterns leading to different absorbed doses to cancer cells. To this end, we use longitudinal tumor growth simulations to compare 225Ac dose rates (DRs) to tumors of varying sizes and phenotypes in a prostate cancer model.
Methods: We used a stochastic, agent-based mathematical tumor growth model to generate 2D simulated cross-sections of tumors in vascularized normal tissue. The tumors contained normoxic, hypoxic, and necrotic cells, which transitioned dynamically between the states based on local glucose and oxygen availability. Realistic concentrations and diffusion rates of glucose and oxygen were used. Three distinct phenotypes of tumors, containing mostly necrotic (type 1), hypoxic (type 2), and normoxic (type 3) regions, were generated by varying the blood vessel density in healthy tissue and implicit angiogenesis rate. Two snapshots of longitudinal growth were analyzed per type (diameters of 5.5 mm and 10 mm, referred to as small and large). No activity was assigned to normal cells or blood vessels, and 96.5%, 3%, and 0.5% of the activity went to the normoxic, hypoxic and necrotic cells, respectively (based on the percentage of oxygen present in those cells). To determine the DR to the cell nuclei, two dose kernels (for activity in cytoplasm vs. in cell membrane) were created using GATE v9.0, assuming LnCaP cells. The kernel matrix sizes were 31 x 31 with 13.498 µm pixels matching the tumor images. To determine the ratio of PSMA uptake in the cytoplasm and membrane to scale the dose kernels, a standardized internalization assay was performed in which LNCaP cells were incubated with a 177Lu-labelled PSMA binding pharmaceutical as a surrogate to 225Ac, and the activity in each cell region was determined at multiple time points. The kernels were combined and convolved with the tumor images to yield DR maps per activity. For each tumor type, the mean DR per injected activity (DR/A) per tumor cell type was determined.
Results: The mean DR/A to the whole tumor was comparable between types with large tumors receiving lower DR/A than small tumors. Depending on the amount of hypoxia (types 1/2/3), hypoxic and normoxic regions received different mean DR/A. For type 1 tumors, the normoxic cells received higher DR/A (1.61x10-3 mGy/s/Bq) than hypoxic cells (1.34x10-4 mGy/s/Bq) and the mean DR/A decreased for larger tumors by factors of 2.2 and 2.4 for normoxic and hypoxic cells, respectively. A similar trend was observed for type 2 tumors. For type 3 tumors with little hypoxia, the hypoxic cells received a mean DR/A of 9.82x10-4 mGy/s/Bq and the normoxic cells 3.16x10-4 mGy/s/Bq for small tumors. However, for large type 3 tumors, mean DR/A of 3.53x10-5 mGy/s/Bq and 9.59x10-5 mGy/s/Bq were found for hypoxic and normoxic cells.
Conclusions: In general, normoxic cells received higher DR/A than hypoxic cells across phenotypes and tumor size; this is expected due to high vasculature and implicit angiogenesis rates in the corresponding regions. However, the results of our study indicate that at early stages of hypoxia (type 3, largely normoxic) a higher mean DR/A is delivered to hypoxic cells than in type 2 or 1 tumors. We observed higher whole tumor DR/A in smaller tumors; this is expected considering the short range (< 100 µm) of alpha particles emitted from 225Ac and low tumor mass. This work highlights the importance of radiopharmaceutical distribution within a tumor, which may differ amongst radiopharmaceuticals. Future work aims to test more tumor phenotypes and include isotopes such as 177Lu to see if these effects are altered when using a beta emitter with a much higher particle range, as well as determining the impact of free radicals.