Automated image analysis of glioblastomas and other gliomas

Publisher Website
Google Scholar
Abstract
Tumours of the neuroglia, 172 in all (50 glioblastomas, 65 fibrillar and gemistocytic astrocytomas, 26 pilocytic astrocytomas and 31 oligodendrogliomas), were studied by automated microscopic picture analysis. Thirteen morphometric and densitometric parameters of the tumour cell nuclei as well as two mitotic parameters were determined on 4 μm thick Feulgen slide preparations. The correlation of the results with subjectively established grade of malignancy was examined. A close correlation was found between almost all morphometric-densitometric nuclear parameters and the morphologically established behaviour of the tumours under study. Based on these results both the method of automated microscopic picture analysis and the selected parameters have been assessed as efficient for the evaluation of the behaviour of glial tumours. The results are readily reproducible and can be obtained without requiring much time. An other advantage of this method is the use of histological preparations, which allows a comparison of the overall structure of the tumour as well as a pinpointed examination of those regions that are representative of the tumour. Subjective grading of glial tumours has been greatly objective by automation.