There is a growing need to develop suitable in vivo imaging strategies for sensitively detecting, characterizing, and monitoring therapeutic responses in malignant brain tumors and for complementing histologic analyses. Serial objective imaging readouts of key metabolic and physiologic processes may improve the diagnostic and therapeutic management of brain tumors by enabling the earlier identification of tumor progression, recurrence, and treatment failure. In patients with gliomas, tumor cell proliferation has been identified as an important surrogate marker of survival, as determined by staining of ex vivo biopsy specimens with proliferation markers, such as Ki67 (1-3). Such in vitro determinations, however, may fail to predict actual therapeutic responsiveness because of biopsy sampling errors, intratumoral variations in pertinent biologic properties, or nonuniform drug delivery (4). Cell proliferation imaging by PET may address the foregoing limitations, providing a serial 3-dimensional quantitative assessment of tumor growth or regression in vivo. The radiolabeled thymidine analog 3′-deoxy-3′-18F-fluorothymidine (18F-FLT) is well suited for this purpose and has been used to monitor tumor uptake in models of untreated (5-7) and treated (8-13) preclinical tumors in conjunction with radiation or antiproliferation therapies and, more recently, to characterize both static and kinetic changes in malignant brain tumor uptake in patients (14-20). After its intravenous administration, 18F-FLT is distributed extracellularly and is subsequently transported into the cytosol, where it is phosphorylated by thymidine kinase 1, an enzyme overexpressed during the DNA synthesis phase of the cell cycle (21). In its phosphorylated form, 18F-FLT is metabolically “trapped” in the cell, with essentially no incorporation into DNA. Its limited transport across the intact blood–brain barrier and the absence of proliferating cells in the normal adult brain result in low levels of 18F-FLT within the cranium. In contrast, 18F-FDG exhibits high physiologic uptake in the normal brain. Thus, tumor-to-normal brain uptake ratios and tumor specificity are greater for 18F-FLT than for 18F-FDG (18,22). Brain tumor 18F-FLT uptake measured at various times after injection has been shown to correlate with histopathologic proliferation markers, such as Ki67 or proliferative cell nuclear antigen, in preclinical models and patient studies (4,11,18). In addition, several clinical investigations have assessed uptake in brain tumors dynamically in conjunction with 2-tissue-compartment (4-rate-constant) models (14,15,19), with one such study correlating uptake ratios with rate constants in patients with gliomas (19). Kinetic modeling of dynamic data permits rate constants for the initial transport and subsequent metabolic (“trapping”) phases of 18F-FLT uptake to be independently evaluated, the latter rate constant being directly related to the number or rate of proliferating cells within the tumor volume. The ability to characterize these respective components of time–activity curves may be particularly important in distinguishing changes in tumor proliferation from alterations in 18F-FLT transport, because both factors contribute to overall radiotracer uptake. In preclinical models, there is a need to implement dynamic imaging and kinetic analysis methods to derive more accurate and specific estimates of functional parameters characterizing tumor metabolism. Unlike recent clinical investigations, only a few preclinical studies have used dynamic imaging for the purpose of determining the time after injection at which maximum tracer uptake occurs. Improving measurements of key biologic parameters that can serve as surrogate markers of tumor responses in clinically relevant mouse models is critical for enabling the future optimization of diagnostic and treatment planning protocols in patients. Along with correlative histologic or serum markers, such methods might enable earlier changes to be made to protocols prescribing conventional therapies or new single or combination pathway-specific inhibitors. Moreover, the application of kinetic analyses to preclinical models may permit relevant tumor biology to be elucidated under conditions that may not be feasible in patients with brain tumors. Such kinetic methods may additionally be applied to the in vivo evaluation and optimization of new diagnostic probes for tumor detection and characterization. In the present work, we investigated whether 18F-FLT can be used to detect and assess proliferative activity in genetically engineered mouse (GEM) high-grade glioma models that are thought to accurately recapitulate the human disease. We successfully developed and implemented a compartmental modeling approach that permits the discrimination of tracer transport and delivery from tumor proliferation. Compartmental model tracer rate constants were estimated by use of image-derived input and tissue time–activity data (i.e., left ventricular blood clearance and tumor uptake, respectively). To our knowledge, no studies reported to date have dynamically assessed uptake in preclinical brain tumor models with evaluation of the relevant rate constants. We sought to clarify the extent to which tracer transport and metabolism contribute to overall tumor uptake, and we correlated static uptake measurements (percentage injected dose per gram [%ID/g]) at 1 h (i.e., a 59- to 60-min time interval) with the metabolic rate constants.