Image Analysis of Proliferating Cells in Tumors of the Human Nervous System

Abstract

Obtaining growth fractions from immunohistological preparations by the commonly used cell count calculation method is time consuming. For the first time, we investigated and compared the detection of proliferating cells in immunohistologically labeled tissue from tumors of the nervous system using the monoclonal antibody Ki-67 by a new computerized image analysis system and by cell count calculation. The two methods showed a high correlation (correlation index, 0.98) in 37 gliomas (2 pilocytic astrocytomas, 10 Grade II astrocytomas, 5 Grade III astrocytomas, 20 Grade IV astrocytomas and glioblastoma multiforme) and a heterogenous group of 10 additional tumors of the nervous system, including oligodendroglioma, pineoblastoma, primary central nervous system lymphoma, and neurofibroma. Advantages of and indications for image analysis are as follows: 1) time-saving evaluation of immunohistological preparations enables their use in neuro-oncological routine diagnostics and examination of larger cell populations, thus leading to more precisely reproducible results, especially in heterogenous tumors; 2) image analysis, calculating the area of cell nuclei rather than their number, avoids difficulties with fragmented or overlapping nuclei; 3) analysis of different antibodies (for example, Ki-67 and anti-proliferating cell nuclear antigen) may be performed with the same program; 4) investigation of a larger patient group may lead, in combination with the histopathological diagnosis and clinical parameters, to better adapted therapeutic concepts.