The development of novel cancer treatments is a challenging task, partly because results from model systems often fail to predict drug efficacy in humans, and also tumors are often inaccessible for biochemical analysis, preventing effective monitoring of drug activity in vivo. Utilizing a model system, we evaluated the use of drug‐induced DNA damage in surrogate tissues as indicators of drug efficacy. Samples of a commercially available melanoma skin model (Mattek MLNM‐FT‐A375) containing keratinocyte and fibroblast layers with melanoma nodules were subjected to various chemotherapeutic regimens for one, four, or eight days. At these times they were analyzed for DNA double‐stranded breaks (γH2AX foci) and apoptosis (TUNEL). A wide range of drug responses in both tumor and normal tissues were observed and cataloged. For the melanoma, the most common drug response was apoptosis. The basal keratinocyte layer, which was the most reliable indicator of drug response in the melanoma skin model, responded with γH2AX foci formation that was abrupt and transient. The relationships between tumor and surrogate tissue drug responses are complex, indicating that while surrogate tissue drug responses may be useful clinical tools, careful control of variables such as the timing of sampling may be important in interpreting the results.