Quantitative data for intratumoral histologic heterogeneity were obtained by investigating ten small and ten large punched samples from 50 unembedded supratentonal gliomas. The 1000 samples were diagnosed according to the World Health Organization (WHO) classification and six histopathologic features associated with malignancy were evaluated (cellular density, nuclear pleomorphism, necroses, histologic architecture, vessels, and mitoses), each with defined gradations. The slides were read independently by two observers. The initially high interobserver variability (grade, 22.2%; type, 10.31, and tumor presence/ absence, 7.1%) was for the most part due to intermediate grades and types and was reduced to 1.7% after mutual review. Small samples showed lower mean grade than large samples and more often absence of tumor (7.6% versus 2.4%). Of all gliomas, 48% showed differently typed samples, 82% differently graded samples, and 62% benign and malignant grades. Intratumoral heterogeneity was higher for the necroses than for the other histopathologic features. Our results underscore the importance of extensive tissue sampling.
Cancer 64~442-447, 1989.
ISTOPATHOLOGIC TYPING AND GRADING of gliomas
H forms the basis of prognostic inferences and therapeutic decisions and is of paramount importance in evaluating the effectiveness of therapeutic strategies. However, some obstacles stand in the way of exact and reliable grading: different grading systems (e.g., threetiered Ringertz system' or four-tiered Kernohan2 and World Health Organization [WHO] systems3), different neuropathologists with nonuniform criteria of histologic malignancy, and different parts within the tumor to be graded. This intratumoral heterogeneity of type and grade in gliomas is common knowledge; quantitative data, however, are not available. Such data are all the more necessary because the limited volume of available surgical material, e.g., in stereotaxic biopsies, could yield nonrepresentative specimens. Therefore, we investigated the frequency and extent of intratumoral heterogeneity in gliomas by punching several samples from each tumor in a series of formalin-fixed gliomas and diagnosing each sample independently.