Prediction of oligodendroglial tumor subtype and grade using perfusion weighted magnetic resonance imaging

Abstract

Object: Treatment of patients with oligodendrogliomas relies on histopathological grade and characteristic cytogenetic deletions of 1p and 19q, shown to predict radio- and chemosensitivity and prolonged survival. Perfusion weighted magnetic resonance (MR) imaging allows for noninvasive determination of relative tumor blood volume (rTBV) and has been used to predict the grade of astrocytic neoplasms. The aim of this study was to use perfusion weighted MR imaging to predict tumor grade and cytogenetic profile in oligodendroglial neoplasms.

Methods: Thirty patients with oligodendroglial neoplasms who underwent preoperative perfusion MR imaging were retrospectively identified. Tumors were classified by histopathological grade and stratified into two cytogenetic groups: 1p or 1p and 19q loss of heterozygosity (LOH) (Group 1), and 19q LOH only on intact alleles (Group 2). Tumor blood volume was calculated in relation to contralateral white matter. Multivariate logistic regression analysis was used to develop predictive models of cytogenetic profile and tumor grade.

Results: In World Health Organization Grade II neoplasms, the rTBV was significantly greater (p < 0.05) in Group 1 (mean 2.44, range 0.96-3.28; seven patients) compared with Group 2 (mean 1.69, range 1.27-2.08; seven patients). In Grade III neoplasms, the differences between Group 1 (mean 3.38, range 1.59-6.26; four patients) and Group 2 (mean 2.83, range 1.81-3.76; 12 patients) were not significant. The rTBV was significantly greater (p < 0.05) in Grade III neoplasms (mean 2.97, range 1.59-6.26; 16 patients) compared with Grade II neoplasms (mean 2.07, range 0.96-3.28; 14 patients). The models integrating rTBV with cytogenetic profile and grade showed prediction accuracies of 68 and 73%, respectively.

Conclusions: Oligodendroglial classification models derived from advanced imaging will improve the accuracy of tumor grading, provide prognostic information, and have potential to influence treatment decisions.