MRI characterization and histopathological correlation of primary intra-axial brain glioma

OBJECTIVE: To determine the diagnostic accuracy of magnetic resonance imaging (MRI) in preoperative diagnosis and grading of intra-axial gliomas compared with histopathology. STUDY DESIGN: Prospective, comparative, study. PLACE AND DURATION OF STUDY: Radiology Department, Aga Khan University Hospital, Karachi from July 2004 to June 2006. METHODS: Fifty-three patients having different neurological symptoms referred to Radiology Department Aga Khan University Hospital for MRI examination were included in the study. These patients were provisionally diagnosed radiologically having intra-axial brain tumor or subsequently found to have pathologically proven primary intra-axial brain tumors. MRI scans were evaluated for location, consistency, hemorrhage, necrosis, margins, edema, MRI signals contrast enhancement and any additional features for staging the tumor. Preoperative diagnosis was compared with postoperative pathological diagnosis by using Pearson's Chi square test. Accuracy of magnetic resonance imaging in diagnosing and staging the brain tumors was determined. Sensitivity, specificity, accuracy, positive predictive value and negative predictive value of MRI in characterizing the lesion was also calculated. RESULTS: The study included 53 patients with age ranging from 1-year 10-months to 60 years (mean 32.7 years). Twenty-eight patients had supratentorial tumors (52%) and 25 had infratentorial tumors (47%). Twenty-eight patients had Astrocytoma (52%), 10 had Oligodendroglioma (19%), 6 had Medulloblastoma (11%), 5 had Lymphoma (9%), and 4 had Hemangioblastoma (7%). Preoperative MRI diagnosis was correct in 50 patients with accuracy of 94%. No significant difference was found between preoperative MRI grading and postoperative histopathological grade of intra-axial tumor (p-value>0.05). Sensitivity, specificity, positive predictive value and negative predictive value of MRI in detecting tumor necrosis were 93%, 77%, 80% and 90% while for detecting tumor hemorrhage were 57%, 93%, 57%, and 93% respectively. CONCLUSION: MRI is very accurate in preoperative diagnosis, staging and assessing the tumor characteristics of primary intra-axial brain tumors. It can be used reliably in our usual clinical practice.