Human leukocyte antigen (HLA) class I expression in melanoma is usually assessed using immunohistochemical staining. Here we report on the use of Fourier transform infrared (FTIR) hyperspectral imaging, a method widely used in two-dimensional analysis of chemical components, to study HLA class I expression in tissue. Two-dimensional cluster colour images derived from unsupervised hierarchical cluster analysis of FTIR hyperspectral data on melanoma sections were compared with consecutive sections that were immunohistochemically stained for class I expression. HLA-class-I-positive and -negative areas were differentiated by FTIR cluster images in all eight melanoma sections investigated without the need for antibody attachment. FTIR imaging enables the distinction of HLA-class-I-positive from class-I-negative areas in melanoma. This method is accurate, rapid and cost-effective.