P03.20 Comparison of histopathology profiles in surgically resected cerebral radiation necrosis with recurrent irradiated brain tumor

Abstract

INTRODUCTION: A conundrum exists in establishing the diagnosis of cerebral radiation necrosis (RTN), a dysregulated repair process following brain radiation therapy (RT), by tissue examination because specific histologic characteristics have not been defined and distinguished from those of recurrent brain tumor (RBT) after RT. The distinction of these two entities is critical in patient management. We compared the type and grade of histologic characteristics identified in surgically-derived tissue specimens obtained at imaging recurrence in brain tumor patients treated with RT and in whom the specimen contained predominantly or totally RTN versus predominantly or totally RBT. MATERIALS AND METHODS: Subjects in this retrospective IRB-approved study had diagnosis of RTN or RBT identified by a review of brain pathology reports at UHCMC from 2004-2013. RTN was defined by > 80% features of RTN; RBT was defined by >80% active tumor. All other cases (“mixed RTN and RBT”) were excluded. Clinical, imaging and treatment data were reviewed. Histologic profiling included vascular changes (telangiectasia, hyalinization, wall thickening and necrosis, thrombosis, endothelial hypertrophy, circumferential endothelial hyperplasia [CEN]), necrosis (coagulative [CN], pseudopalisading, zonal geographic [ZGN]), tumor features (perinuclear halos, mitosis, tumor satellitosis, epithelial differentiation), tissue reaction (granulation tissue, collagen scar, demyelination, “RT astrocytes” [RTA]), microcysts, dystrophic calcification [DC], blood products, and inflammatory infiltrates. Each characteristic was graded 0, <25%, 26-50%, 51-75%, 76-100% by a board-certified neuropathologist (MC). Profiles were compared with pretreatment tumor when available. RESULTS: 66 patients treated with EBRT and/or SRS were identified, 40 with RBT (glioma 25, metastasis 15) and 26 with RTN (metastasis 14, glioma 12). Using Fisher's exact and Chi-square tests we identified significant (p