CT and MRI findings in gliomatosis cerebri: a neuroradiologic and neuropathologic review of diffuse infiltrating brain neoplasms

The aim of the study was to develop objective criteria that might be helpfu l for the diagnosis of gliomatosis cerebri (GC) with the highest possible p robability based on both the neuroradiological and histopathological findin gs. Imaging findings in 14 patients with diffuse infiltrating brain neoplas ms were studied by two neuroradiologists. Computed tomography and MRI scans were compared with each other side by side. The extent and kind of disease were graded on a scale of 1-5. Interexamination agreement between the two methods was calculated using a kappa analysis. Neither of the neuroradiolog ists performed the examinations and both were blinded to the histopathologi cal findings, which were also available for all patients, based on biopsy a s well as follow-up CT and MRI studies. A neuroradiological-neuropathologic al correlation was performed. A score system helped to differentiate the fi ndings in three categories: 1 = suggestive of GC; 2 = GC cannot be excluded ; and 3 = others. Both CT and MRI were performed in 14 patients with clinic al signs and symptoms of an intracerebral tumor. All examinations had diagn ostic quality and showed the involvement of at least two brain lobes. Stere otactic biopsy was carried out in all patients. In 2 patients the neuropath ological diagnosis was suggestive of GC, in 1 patient glioblastoma, in 2 pa tients astrocytoma, and in 5 patients nonspecific astrogliotic proliferatio n. In the remaining 4 cases anaplastic tumor infiltration was diagnosed. Th e neuroradiological findings in 5 cases were suggestive of GC; in 6 cases a GC could not excluded; and in 3 patients only a slight probability of GC w as found. In 2 cases was the neuropathological and the neuroradiological di agnosis of GC concordant. Magnetic resonance imaging is significantly more sensitive than CT in the diagnosis of GC. However, even with multiple, MRI- guided stereotactic biopsies in correlation with intraoperative analysis of the sample by smear preparations by a neuropathologist the antemortem diag nosis of GC is still difficult. Discussion of neuropathological and neurora diological findings in each case in combination with a score system may hel p to resolve discrepancies.