Dynamics of FLAIR Volume Changes in Glioblastoma and Prediction of Survival

Rachel Grossman, Nir Shimony, Dror Shir, Tal Gonen, Razi Sitt, Tali Jonas Kimchi, Carmit Ben Harosh, Zvi Ram

The extent of tumor resection (EOTR) calculated by enhanced T1 changes in glioblastomas has been previously reported to predict survival. However, fluid-attenuated inversion recovery (FLAIR) volume may better represent tumor burden. In this study, we report the first assessment of the dynamics of FLAIR volume changes over time as a predictive variable for post-resection overall survival (OS). Contemporary data from 103 consecutive patients with complete imaging and clinical data who underwent resection of newly diagnosed glioblastoma followed by the Stupp protocol between 2010 and 2013 were analyzed. Clinical, radiographic, and outcome parameters were retrieved for each patient, including magnetic resonance imaging (MRI)-based volumetric tumor analysis before, immediately after, and 3 months post-surgery. OS rate was 17.6 months. A significant incremental OS advantage was noted, with as little as 85 % T1-weighted gadolinium-enhanced (T1Gd)-EOTR measured on contrast-enhanced MRI. Pre- and immediate postoperative FLAIR-based EOTR was not predictive of OS; however, abnormal FLAIR volume measured 3 months post-surgery correlated significantly with outcome when FLAIR residual tumor volume (RTV) was <19.3 cm(3) and <46 % of baseline volume (p < 0.0001 for both). Age and isocitrate dehydrogenase (IDH)-1 mutation were predictive of OS (p < 0.0001, Cox proportional hazards). OS correlated with the immediate postoperative T1Gd-EOTR measured by enhanced T1 MRI, but not by FLAIR volume. Diminished abnormal FLAIR volume at 3 months post-surgery was associated with OS benefit when FLAIR-RTV was <19.3 cm(3) or <46 % of baseline. These threshold values provide a new radiological variable that can be used for prediction of OS in patients with glioblastoma immediately after completion of standard chemoradiation.