Cancer progression is commonly accompanied by an altered glucose metabolism. In general, spatially resolved imaging of glucose metabolism and its subtle alterations might provide valuable diagnostic information in vivo. A classical example is positron emission tomography that exploits this feature in obtaining preferential accumulation of fluorescent analog of glucose in tumors, thereby achieving an imaging contrast. We report a novel scaling analysis of glucose metabolism in mammary epithelial (NMuMG) cells by detrended fluctuation analysis of Cerulean (cyan fluorescent protein variant) fluorescence. Fluorescence fluctuations of Cerulean are reasoned to be indicative of dynamic pH changes associated with glucose metabolism. Normal parental cells and the spontaneously transformed (cancerous) NMuMG cells displayed robust scaling exponent that reflects nonrandom fluctuations in Cerulean fluorescence. Acute dependence of cancer cells on glycolysis as compared with normal cells is exploited to yield a statistically significant difference in scaling exponent, thereby providing discrimination between normal and cancer cells in vitro. By careful design of experiments in vivo, the proposed scaling approach might even have diagnostic potential for early detection of cancer lesions in small animal models.