Concurrent bevacizumab and temozolomide alter the patterns of failure in radiation treatment of glioblastoma multiforme

Background: We investigated the pattern of failure in glioblastoma multiforma (GBM) patients treated with concurrent radiation, bevacizumab (BEV), and temozolomide (TMZ). Previous studies demonstrated a predominantly in-field pattern of failure for GBM patients not treated with concurrent BEV. Methods: We reviewed the treatment of 23 patients with GBM who received 30 fractions of simultaneous integrated boost IMRT. PTV60 received 2 Gy daily to the tumor bed or residual tumor while PTV54 received 1.8 Gy daily to the surrounding edema. Concurrent TMZ (75 mg/m(boolean AND)2) daily and BEV (10 mg/kg every 2 weeks) were given during radiation therapy. One month after RT completion, adjuvant TMZ (150 mg/m(boolean AND)2 x 5 days) and BEV were delivered monthly until progression or 12 months total. Results: With a median follow-up of 12 months, the median disease-free and overall survival were not reached. Four patients discontinued therapy due to toxicity for the following reasons: bone marrow suppression (2), craniotomy wound infection (1), and pulmonary embolus (1). Five patients had grade 2 or 3 hypertension managed by oral medications. Of the 12 patients with tumor recurrence, 7 suffered distant failure with either subependymal (5/12; 41%) or deep white matter (2/12; 17%) spread detected on T2 FLAIR sequences. Five of 12 patients (41%) with a recurrence demonstrated evidence of GAD enhancement. The patterns of failure did not correlate with extent of resection or number of adjuvant cycles. Conclusions: Treatment of GBM patients with concurrent radiation, BEV, and TMZ was well tolerated in the current study. The majority of patients experienced an out-of-field pattern of failure with radiation, BEV, and TMZ which has not been previously reported. Further investigation is warranted to determine whether BEV alters the underlying tumor biology to improve survival. These data may indicate that the currently used clinical target volume thought to represent microscopic disease for radiation may not be appropriate in combination with TMZ and BEV.