Glycolytic Plasticity of Metastatic Lung Cancer Captured by Noninvasive (18)F-FDG PET/CT and Serum (1)H-NMR Analysis: An Orthotopic Murine Model Study

We aim to establish a noninvasive diagnostic platform to capture early phenotypic transformation for metastasis using (18)F-FDG PET and (1)H-NMR-based serum metabolomics. Mice with implantation of NCI-H460 cells grew only primary lung tumors in the localized group and had both primary and metastatic lung tumors in the metastatic group. The serum metabolites were analyzed using (1)H-NMR at the time of PET/CT scan. The glycolysis status and cell proliferation were validated by Western blotting and staining. A receiver operating characteristic (ROC) curve analysis was performed to evaluate the diagnostic accuracy of SUV(mean) and serum metabolites in metastasis. In the metastatic mice, the SUV(mean) of metastatic tumors was significantly higher than that of primary lung tumors in PET images, which was supported by elevated glycolytic protein expression of HK2 and PKM2. The serum pyruvate level in the metastatic group was significantly lower than that in the localized group, corresponding to increased pyruvate-catalyzed enzyme and proliferation rates in metastatic tumors. In diagnosing localized or metastatic tumors, the areas under the ROC curves of SUV(mean) and pyruvate were 0.92 and 0.91, respectively, with p < 0.05. In conclusion, the combination of (18)F-FDG PET and (1)H-NMR-based serum metabolomics demonstrated the feasibility of a glycolytic platform for diagnosing metastatic lung cancers.