Automated detection of radiation-induced chromosome aberrations following fluorescence in situ hybridization.

The cytogenetic detection of rare chromosome aberrations induced by ionizing radiation requires the evaluation of large numbers of cells. In this report, an image analysis program based on thresholding of grey level histograms is described for a rapid automated detection of chromosome translocations in metaphase spreads from human lymphocytes following irradiation and chromosome in situ suppression (CISS) fluorescence in situ hybridization (FISH). To classify a metaphase spread as "normal", "aberrant", or "excluded", a minimum medium time of less than two seconds was required on a general purpose Personal Computer. Using appropriately stained specimen, for the false classification rate an upper limit of about 10% was estimated. The upper limit for the rate of "excluded" cells was estimated to be about 11% (99% confidence ranges). FISH-procedures allow to score chromosome aberrations also in the interphase nucleus. To apply simple threshold algorithms for segmentation of the FISH-stained nuclear areas, it is required that the grey level contrast is sufficiently high. The preliminary results presented here on the image analysis of human lymphocyte nuclei suggest the feasibility of such an approach.