It is becoming increasingly apparent that cell populations are heterogeneous in their functions, disease states and response to therapy. Tumor heterogeneity is one of the main factors contributing to acquire drug resistance. Substantial interest is now devoted to characterization methods that operate at the single-cell level, as opposed to bulk analyses that can only measure average properties over a given population. Fluorescence methods have long been used to measure molecular processes in single living cells. However, a vast number of small molecules remain invisible to fluorescence probing. These molecules lack inherent fluorescence and cannot be fluorescently labeled without altering their biochemical activity. Precise and sensitive quantitation of small molecules (e.g. in drug pharmacokinetics studies) remains the domain of radionuclide detection methods (scintillation counting, autoradiography, and positron emission tomography) since radiolabeling in most cases can preserve biochemical activity. The recent finding that radionuclide molecules too can be imaged at the cellular level in a microscope represents a radical departure from what was previously known. This new method, called radioluminescence microscopy, can measure the accumulation of a radionuclide molecule in single living cells. While the technique has been demonstrated for a variety of applications, the technology is still in its infancy. In this proposal, we are proposing several radical improvements that will allow us to measure radionuclide probe uptake in up to 1000 individual cells, in a single acquisition. Studies that investigate large cellnumbers are necessary for statistical reasons and due to the existence of rare cell subpopulations. Based on encouraging preliminary results, we are proposing a new instrument design called the radioluminescence microarray that can achieve this goal. This new design incorporates several radical improvements that will transform radioluminescence microscopy into a versatile tool for high-throughput studies with many potential applications. In Aim 1, we will develop a new device called the radioluminescence microarray. This device includes a micrometer-thin scintillator for high-resolution imaging, a microwell array for optimized cell placement, a fluidics platform for repeatable sample preparation, and an improved epifluorescence add-on for multi-modality imaging. In Aim 2, we will implement software for real-time display and automated analysis of radioluminescence microscopy images. Last, in Aim 3, we will validate the overall approach by investigating the interaction with single living cells of -fluorouracil (5-FU), a small-molecule non-fluorescent chemotherapeutic agent. Using the radioactive form of the drug ([18F] 5-FU), we will determine how 5FU distributes in heterogeneous cancer cell populations. Fluorescence microscopy will be used to assign cells to different subpopulations according to factors such as position in the cell cycle or stem-cell status. In summary, this project will develop new a new instrument with unmatched capabilities, which will be applied to deepening our understanding of drug resistance in heterogeneous cancer cell populations.