Accuracy of magnetic resonance imaging diagnosis and grading of gliomas

Abstract

Background: Low- and high-grade gliomas differ in clinical presentation, natural history, treatment outcome, prognosis, survival pattern, histopathological, immunohistochemical and biomolecular profiles. Accurate pre-operative prediction of histopathological grade of gliomas remains challenging, and is critical for making optimum management plan and prognosticating the disease beforehand to determine the most cost-effective therapeutic choice with the best patient outcome. This prospective observational study on 54 patients aims to determine accuracy of pre-operative magnetic resonance imaging (MRI) in diagnosing and grading gliomas.Methods: Pre-operative grading of MRI-suspected gliomas was done by assigning scores of 0-2 to 9 criteria – midline crossing, perilesional edema, signal heterogeneity, intra-tumoral hemorrhage, tumor border definition, cystic/necrotic changes, mass effect, contrast enhancement and diffusion restriction. Total scores of 0-5, 6-9 and 10-18 were considered radiologically low, intermediate and high grades respectively and correlation with World Health Organization (WHO) grades I+II, III and IV respectively was determined.Results: MRI diagnosed 85.18% gliomas correctly. Pre-operative MR grading was 76-89% sensitive and 86-96% specific in predicting the histopathological grade of the gliomas. Signal heterogeneity and contrast enhancement had the highest whereas midline crossing and mass effect had the lowest correlation with histopathological grade.Conclusions: Pre-operative MRI is highly specific and somewhat less sensitive tool for grading gliomas pre-operatively. The diagnostic yield is higher for LGGs and GBMs, compared to anaplastic gliomas, probably due to their mixed or intermediate features.

Comparative study between dynamic susceptibility contrast magnetic resonance imaging and arterial spin labelling perfusion in differentiating low-grade from high-grade brain tumours

“…On conventional MRI, brain tumours display heterogeneous characteristics because the components within them have various degrees of cellular and nuclear pleomorphism, necrosis, mitotic activity, and vascular proliferation [2,3]. The accuracy of conventional MRI in grading brain tumours varies from 55% to 83.3% [4][5][6]. It is not sufficient to accurately characterize brain tumour only on the basis of conventional MR imaging.…”