Monitoring Radiation Treatment Effects in Glioblastoma: FLAIR Volume as Significant Predictor of Survival

Glioblastoma is the most common adult central nervous system malignancy and carries a poor prognosis. Disease progression and recurrence after chemoradiotherapy are assessed via serial magnetic resonance imaging sequences. T2-weighted fluid-attenuated inversion recovery (FLAIR) signal is presumed to represent edema containing microscopic cancer infiltration. Here we assessed the prognostic impact of computerized volumetry of FLAIR signal in the peri-treatment setting for glioblastoma. We analyzed pre- and posttreatment FLAIR sequences of 40 patients treated at the Columbia University Medical Center between 2011 and 2014, excluding those without high-quality FLAIR imaging within 2 weeks before treatment and 60 to 180 days afterward. We manually contoured regions of FLAIR hyperintensity as per Radiation Therapy Oncology Group guidelines and calculated the volumes of nonenhancing tumor burden. At the time of this study, all but 1 patient had died. Pre- and posttreatment FLAIR volumes were assessed for correlation to overall and progression-free survival. Larger post-treatment FLAIR volumes from sequences taken between 60 and 180 days after conclusion of chemoradiotherapy were negatively correlated with overall survival (P = .048 on Pearson's correlation and P = .017 and P = .043 on univariable and multivariable Cox regression analyses, respectively) and progression-free survival (P = .002 on Pearson's correlation and P = < .001 and P = < .001 on univariable and multivariable Cox regression analyses). This study suggests that higher FLAIR volumes in the 2- to 6-month posttreatment window are associated with worsened survival.