Non-melanoma skin cancer (NMSC) is the most common form of cancer in the US and its incidence is increasing. The current standard of care is visual inspection by physicians and/or dermatologists, followed by skin biopsy and pathologic confirmation. We have investigated the use of in vivo fluorescence imaging using fluorocoxib A as a molecular probe for early detection and assessment of skin tumors in mouse models of NMSC. Fluorocoxib A targets the cyclooxygenase-2 (COX-2) enzyme that is preferentially expressed by inflamed and tumor tissue, and therefore has potential to be an effective broadly active molecular biomarker for cancer detection. We tested the sensitivity of fluorocoxib A in a BCC allograft SCID hairless mouse model using a wide-field fluorescence imaging system. Subcutaneous allografts comprised of 1000 BCC cells were detectable above background. These BCC allograft mice were imaged over time and a linear correlation (R(2) = 0.8) between tumor volume and fluorocoxib A signal levels was observed. We also tested fluorocoxib A in a genetically engineered spontaneous BCC mouse model (Ptch1(+/-) K14-Cre-ER2 p53(fl/fl)), where sequential imaging of the same animals over time demonstrated that early, microscopic lesions (100 μm size) developed into visible macroscopic tumor masses over 11 to 17 days. Overall, for macroscopic tumors, the sensitivity was 88% and the specificity was 100%. For microscopic tumors, the sensitivity was 85% and specificity was 56%. These results demonstrate the potential of fluorocoxib A as an in vivo imaging agent for early detection, margin delineation and guided biopsies of NMSCs.