A Retrospective Analysis Of Survival Following Treatment With Bevacizumab In Glioblastoma Patients With Radiographically-Confirmed Subependymal Or Leptomeningeal Spread Of Disease

Abstract OBJECTIVE We performed a retrospective analysis examining subependymal disease (SED) and leptomeningeal disease (LMD) in bevacizumab treated IDH-wildtype glioblastoma (GBM) patients to assess potential differences in survival. BACKGROUND GBMs are highly infiltrative and often invade the subependymal and leptomeningeal space making them challenging to treat. At recurrence, patients are commonly treated with bevacizumab, which leads to decreased tumor size and swelling. Patients have improved PFS, but there is no benefit in OS. This may reflect infiltrative disease progression--a hallmark of late-stage disease. We hypothesized patients with CSF spread manifested by SED and LMD may have worse outcomes after bevacizumab than patients with non-disseminated disease. METHODS We retrospectively reviewed charts of all IDH-wildtype GBM patients who received bevacizumab in 2016. MRI scans were reviewed by experienced neuro-radiologists to characterize patients into radiographically defined groups (“0” = no LMD/no SED; “1”= no SED/LMD contact; “2”= no SED/LMD dissemination, and “3’= LMD present). Survival from initiation of bevacizumab to death for each group was compared to group “0” (with no SED/LMD). Survival outcomes were analyzed via t-tests. RESULTS We evaluated 87 IDH-WT GBM patients. Radiographically, patients were grouped-- “0”= 16 (18%), “1”= 33 (38%), “2”= 20 (23%), and “3”= 18 (21%). There was no difference between group “0” and any radiographically defined subgroups. Median survival for group “0”= 187 days; group “1”= 219 days; group “2”= 217 days; group “3”= 169 days. CONCLUSION Contrary to our hypothesis, there was no significant difference in survival from the initiation of bevacizumab in patients with SED or LMD. This may be related to the fact that these patients also typically have a high burden of parenchymal disease. Further work is necessary to test this hypothesis in a larger cohort.