In-vivo PET characterization of lung cancer tumorgraft models

Aim: Lung cancer is the main cause of death for cancer in the world, with about 10% estimated survival at 5 years for advanced non-small-cell lung cancers (NSCLCs). Low therapies efficacy is mainly due to high tumoral heterogeneity. In recent studies, the presence of CD133 positive cancer stem cells (CSCs) has been observed. These cells can be intrinsically resistant to therapies and thus responsible for tumor relapse and metastasis. For these reasons, it is important to develop preclinical models that allow to test new therapies and that reproduce the heterogeneous aspect of human pathology. Aim of the study is to verify that tumorgraft models of lung cancer maintain the same metabolic and histological characteristics of the original tumor. Materials and methods: Tumorgraft models were obtained by direct subcutaneous implant of primary lung cancer fragments obtained from patients undergoing surgery after [18F]FDG-PET. Groups of animals (n=6 per group) with difference in patients' SUV and in CSC expression performed weekly [18F]FDG-PET studies for a month starting with a tumor dimension consistent with animal PET spatial resolution (approximately 2mm). For the quantification, a 3D ROIs (Region of Interest) analysis was performed and, after PET images calibration and correction for F-18 half-life, max radiotracer uptake values expressed as Standardized Uptake Value (SUVmax) were calculated and compared to correspondent patient's SUVmax. Results: Tumorgrafts recapitulate the phenotype of primary tumors concerning histological markers and the presence of CD133 positive cancer cells. In addition, we observed a tendency for higher tumor engraftment when tissue specimen originated from patients with SUV values higher than 8. Grafts originating from the same patients were reproducible and homogeneous for rate of growth and SUV values. In these grafts we observed also a good correlation between patients' and mice' SUVmax (r2=0.76, p<0.05). Finally graft SUV values did not correlate with the rate of tumor growth or the amount of CD133 positive cells. Conclusions: Lung cancers obtained from patient with a preoperative SUVmax higher than 8 showed a reproducible capability for in vivo engraftment maintaining the original metabolic and histological phenotype and represent an interesting preclinical model for the early assessment of therapy efficacy. However for a better in vivo characterization other biological properties such as cellular proliferation and regional hypoxia, that may better correlate with cancer stem cells content, will be further explored.