In this study, we utilized dynamic contrast-enhanced (DGE) MRI to explore glucose metabolism in tumor cells in real-time by employing extended duration times and in-depth analytical techniques. We also implemented self-organized mapping (SOM), a type of artificial neural network, to categorize the time progression of DGE MRI and produce a 2D image containing dynamic data. Our results indicate that different glucose consumption scenarios within tumor areas can distinguish alterations in the tumor microenvironment after glucose administration. Furthermore, our findings are consistent with mass imaging data, validating the reliability of our approach. This methodology holds promise for improving our comprehension of tumor metabolism and promoting the creation of tailored cancer treatments.