An FTIR image of an 8 µm section of de-paraffinised bronchial biopsy that shows a histological transition from normal to severe dysplasia/squamous cell carcinoma (SCC) in situ was obtained in transmission by stitching together images of 256 x 256 µm recorded using a 96 x 96 element FPA detector. Each pixel spectrum was calculated from 128 co-added interferograms at 4 cm −1 resolution. In order to improve the signal to noise ratio, blocks of 4x4 adjacent pixels were subsequently averaged. Analyses of this spectral image, after conversion of the spectra to their second derivatives, show that the epithelium and the lamina propria tissue types can be distinguished using the area of troughs at either 1591, 1334, 1275 or 1215 cm −1 or, more effectively, by separation into two groups by hierarchical clustering (HCA) of the 1614-1465 region. Due to an insufficient signal to noise ratio, disease stages within the image could not be distinguished with this extent of pixel averaging. However, after separation of the cell types, disease stages within either the epithelium or the lamina propria could be distinguished if spectra were averaged from larger, manually selected areas of the tissue. Both cell types reveal spectral differences that follow a transition from normal to cancerous histology. For example, spectral changes that occurred in the epithelium over the transition from normal to carcinoma in situ could be seen in the 1200-1000 cm −1 region, particularly as a decrease in the second derivative troughs at 1074 and 1036 cm −1 , consistent with changes in some form of carbohydrate. Spectral differences that indicate a disease transition from normal to carcinoma in the lamina propria could be seen in the 1350-1175 cm −1 and 1125-1030 cm −1 regions. Thus demonstrating that a progression from healthy to severe dysplasia/squamous cell carcinoma (SCC) in situ can be seen using FTIR spectroscopic imaging and multivariate analysis.