PURPOSE: Variable uptake of the glucose analog 18fluorodeoxyglucose (FDG) has been noticed in positron emission tomography (PET) studies of breast cancer patients, with low uptake occurring especially in lobular cancer. At present, no satisfactory biologic explanation exists for this phenomenon. This study compared 18FDG uptake in vivo with biomarkers expected to be involved in the underlying biologic mechanisms. PATIENTS AND METHODS: Preoperative 18FDG-PET scans were performed in 55 patients. 18FDG activity was assessed visually by three observers using a four-point score. Tumor sections were stained by immunohistochemistry for glucose transporter-1 (Glut-1); Hexokinase (HK) I, II, and III; macrophages; hypoxia-inducible factor-1-alfa (HIF-1α); vascular endothelial growth factor (VEGF165); and microvessels. Mitotic activity index (MAI), amount of necrosis, number of lymphocytes, and tumor cells/volume were assessed. RESULTS: There were positive correlations between 18FDG uptake and Glut-1 expression (P < .001), MAI (P = .001), amount of necrosis (P = .010), number of tumor cells/volume (P = .009), expression of HK I (P = .019), number of lymphocytes (P = .032), and microvessel density (r = .373; P = .005). HIF-1α, VEGF165, HK II, HK III, and macrophages showed no univariate correlation with 18FDG. In logistic regression, however, HIF-1α and HK II added value to MAI and Glut-1. CONCLUSION: 18FDG uptake in breast cancer is a function of microvasculature for delivering nutrients, Glut-1 for transportation of 18FDG into the cell, HK for entering 18FDG into glycolysis, number of tumor cells/volume, proliferation rate (also reflected in necrosis), number of lymphocytes (not macrophages), and HIF-1α for upregulating Glut-1. Together, these features explain why breast cancers vary in 18FDG uptake and elucidate the low uptake in lobular breast cancer.