Fourier transform infrared (FT-IR) spectroscopic imaging provides spatially resolved chemical information. Recent developments have shown that this chemical information can be used to determine tissue cell types. Our goal is then to use the spatial distribution of tissue cell types to perform accurate diagnosis of cancer. However, this step is limited by the spatial resolution provided by current imaging systems. In this paper, we demonstrate that these instruments can be designed to provide better spatial resolution for tissue chemistry. We present an optical model for the propagation of light through an FT-IR spectroscopic imaging system. Using this model, we minimally modify an existing FT-IR spectroscopic imaging system to obtain significantly higher resolution and image quality. We demonstrate that it is possible to identify previously obscured tissue types by performing histological classification based on bio-chemically derived spectral features (metrics).