Chromosome analysis by image processing in a computerized environment. Clinical applications.

Dealing with a routine regional cytogenetic activity, we have developed and adapted to clinical work a semi automatic karyotyping machine. Attempts for an accurate automated chromosome classification using a neural network have led to partial results. A specific adaptation to cancer cytogenetics is under development (determination of the modal number, translocations analysis with densitometric curves, automatic identification of markers). A specific program allows quantification of chromosome labelling with radioactive probes. Exchanges of digitized karyotypes are feasible with labs using automated karyotyping machines. A local network connects several karyotyping and metaphase finding stations. Guidelines for an international data bank concerning abnormal chromosome images have been elaborated. On the other hand the ISH techniques have been applied to the following topics: identification of human chromosome aberrations in amniotic and chorionic cells, chromosome studies of human gametes and embryos (including sex determination), identification of markers in cancer cells.