MULTIVARIATE CLASSIFICATION OF THE INFRARED-SPECTRA OF CELL AND TISSUE SAMPLES

Infrared microspectroscopy of biopsied canine lymph cells and tissue w as performed to investigate the possibility of using IR spectra couple d with multivariate classification methods to classify the samples as normal, hyperplastic, or neoplastic (malignant). IR spectra were obtai ned in transmission mode through BaF2 windows and in reflection mode f rom samples prepared on gold-coated microscope slides. Cytology and hi stopathology samples were prepared by a variety of methods to identify the optimal methods of sample preparation. Cytospinning procedures th at yielded a monolayer of cells on the BaF2 windows produced a limited set of IR transtensional spectra. These transmission spectra were con certed to absorbance and formed the basis for a classification rule th at yielded 100% correct classification in a cross-validated context. C lassifications of normal, hyperplastic, and neoplastic cell sample spe ctra were achieved by using both partial least-squares (PLS) and princ ipal component regression (PCR) classification methods. Linear discrim inant analysis applied to principal components obtained from the spect ral data yielded a small number of misclassifications. PLS weight load ing vectors yield valuable qualitative insight into the molecular chan ges that are responsible for the success of the infrared classificatio n. These successful classification results show promise for assisting pathologists in the diagnosis of cell types and offer future potential for in vivo LR detection of some types of cancer.