Tailor-Made Setting of Radiation Doses Owing to the MGMT Gene Promoter Methylation for the Treatment of IDH Wild-Type Primary Glioblastoma

Abstract

Hypofractionated high-dose irradiation had shown excellent local control of glioblastomas (GBMs) but it also had high risk of radiation necrosis even with IMRT technique. Concurrent and adjuvant temozolomide (TMZ) also had prolonged the survival of patients with GBM, but its’ efficacy was depending upon methylation status of MGMT gene promoter. In this study, we evaluated the efficacy and safety of tailor-made setting of prescribed doses owing to the methylation status of MGMT in the treatment of GBM patients by hypofractionated high-dose IMRT with TMZ. Newly diagnosed GBMs without mutation of IDH were enrolled. Methylation status of MGMT was evaluated with methylation specific PCR. In all cases, residual enhanced lesions plus surgical cavity were defined as GTV, and 3-layered CTVs were contoured surrounding the GTV: CTV1 = GTV, CTV2 was 15mm surrounding GTV, and CTV3 was defined as high-intensity volume on FLAIR images. PTV1, PTV2, and PTV3 were defined by expanding CTVs with 5mm margin. Prescribed total doses for PTV2 and PTV3 were unified as 40 Gy and 32 Gy, but the dose for PTV1 was settled as 48 Gy for MGMT methylated (Met) but 68 Gy for unmethylated (UnMet) cases. These doses were delivered by 8 fractions in 10 days. Concurrent and adjuvant TMZ was administrated in all cases. The primary endpoint was overall survival (OS), and the secondary endpoints were progression-free survival (PFS), time to radiation necrosis (TRN), and independent survival (IS) which was defined as the time to KPS<70%. We used the Kaplan-Meier method to estimate survival distribution and a Cox proportional hazard model to calculate hazard ratios (HRs). From April 2008 through November 2015, 73 GBMs, 20 Met and 53 UnMet cases, were enrolled. There was no significant difference between Met and UnMet cases in patients’ age, sex, tumor volume, extent of resection, and KPS at registration. The median OS was 22.4 m. The OS of Met cases (median: 39.9 m) was significantly longer than that of UnMet cases (19.8 m, HR = 0.43, P = 0.007). The median local PFS was 38.5 m. Reduced-dose IMRT for Met cases still showed similar local PFS to high-dose IMRT for UnMet cases (HR = 0.73, P = 0.466). CSF dissemination-free survival was significantly longer in Met cases (not reached) than UnMet cases (14.1 m, HR = 0.36, P = 0.03). The TRN of Met cases (13.5m) was similar to that of UnMet cases (13.6 m, HR: 1.02, P = 0.945). The median IS was 14.6m. The IS of Met cases (21.3 m) was significantly longer than that of UnMet cases (11.7 m, HR = 0.49, P = 0.023). IMRT of 48 Gy/8 fractions for Met GBMs showed similar effect on tumor control and similar risk for radiation necrosis to 68 Gy/8 fractions for UnMet GBMs. With significant effect of TMZ on prevention of CSF dissemination, reduced dose IMRT showed favorable survival and functional outcome of patients after treatment. From these results, tailor-made setting of prescribed doses owing to the methylation status of MGMT thought to be desirable for the treatment of GBM patients.