Cytology by infrared micro-spectroscopy: Automatic distinction of cell types in urinary cytology

We report microscopically collected infrared spectra of cells found in human urine in an effort to develop automatic methods for bladder cancer screening. Unsupervised multivariate analysis of the observed spectral patterns reveals distinct spectral classes, which correlate very well with visual cytology. Therefore, we believe that spectral analysis of individual cells can aid cytology in rendering reliable diagnoses based on objective measurements and discriminant algorithms.