Patlak modeling of pharmacokinetic tumor parameters from PET can detect chemotherapy response beyond that of conventional PET parameters in patient-derived endometrial cancer mouse models

1230 Objectives: In the clinic, positron emission tomography (PET)-imaging is conventionally acquired by static scanning, typically performed one hour post 18F-fluorodeoxyglucose (FDG)-injection. Further, the metabolic tumor measurements are routinely based on the semi-quantitative and variation-prone parameter; standardized uptake value (SUV) [1]. However, one of the main advantages of current PET technology is the ability to acquire dynamic metabolic data with absolute tumor quantifications. Pharmacokinetic parameters derived from dynamic imaging can potentially yield functional tumor information beyond that represented by SUV and allow better characterization of tumor heterogeneity and monitoring of therapeutic response [2,3]. In this work, we compare metabolic tumor parameters derived from static and dynamic FDG-PET following different therapy regimens in endometrial cancer mouse models. Methods: Two different endometrial cancer PDX models were generated by implanting patient-derived tumor tissue into the left uterine horns of immunocompromised mice. In the first model, mice were randomized in two groups, treatment (carboplatin 15mg/kg, twice weekly, n=5) or control (saline, n=5). In the second model, mice were randomized to receive paclitaxel (12 mg/kg, twice weekly, n=5), trastuzumab (10mg/kg, weekly, n=6) or control (saline, n=6). After 3 weeks of treatment and fasting for ≥12 hours, all mice underwent 1 hour dynamic PET following FDG injection. Static SUV parameters were calculated by analyzing the last 30 minutes of the scan, and tumors were segmented by applying a 2.5 SUV isocontour threshold. For the dynamic analyses, an image-derived input function from vena cava was used together with the segmented tumor VOIs derived from the static analyses to generate the net influx of tracer (Ki) using the Patlak model. Results: No significant differences between treatment groups and controls for the conventional static SUV tumor parameters: SUVmean, SUVmax or metabolic tumor volume (MTV) were found in either model. In contrast, treatment groups had lower tumor Ki compared to controls (p=0.052 for carboplatin, model 1 and p=0.041 for paclitaxel, model 2) (Figure 1) Conclusions: Low tumor Ki was observed in carboplatin- and paclitaxel treated mice whereas static PET parameters yielded no significant differences between treatment groups and controls. Ki derived from Patlak modeling of dynamic PET data represents a promising imaging marker supplementing that of conventional static SUV markers for monitoring treatment response in endometrial cancer PDX mouse models. References: 1. Boellaard, R. Standards for PET image acquisition and quantitative data analysis. J Nucl Med 2009, 50 Suppl 1, 11s-20s, doi:10.2967/jnumed.108.057182. 2. Veronese, M.; Rizzo, G.; Aboagye, E.O.; Bertoldo, A. Parametric imaging of (1)(8)F-fluoro-3-deoxy-3-L-fluorothymidine PET data to investigate tumour heterogeneity. Eur J Nucl Med Mol Imaging 2014, 41, 1781-1792, doi:10.1007/s00259-014-2757-z. 3. Kristian, A.; Holtedahl, J.E.; Torheim, T.; Futsaether, C.; Hernes, E.; Engebraaten, O.; Maelandsmo, G.M.; Malinen, E. Dynamic 2-Deoxy-2-[(18)F]Fluoro-D-Glucose Positron Emission Tomography for Chemotherapy Response Monitoring of Breast Cancer Xenografts. Mol Imaging Biol 2017,19, 271-279, doi:10.1007/s11307-016-0998-x.Figure legend: Box plot showing tumor Ki (min-1) derived from Patlak modeling after 1 hour dynamic FDG-PET scanning in the two models. A) No significant difference between the two groups, but a trend towards lower Ki in the carboplatin-group (p=0.052) B) Tumor Ki is significantly lower in paclitaxel-treated mice compared to control mice (p=0.041).