A Phase 1/2 Trial Of 5 Fraction Stereotactic Radiosurgery With 5 Mm Margins With Concurrent And Adjuvant Temozolomide In Newly Diagnosed Supratentorial Glioblastoma Multiforme: Pattern Of Recurrence Analysis

We sought to determine the pattern of recurrence of tumors treated with 5 fractions of stereotactic radiosurgery (SRS) with 5 mm margins delivered with concurrent and adjuvant temozolomide (TMZ) in newly diagnosed glioblastoma (GBM). Thirty adult patients with newly diagnosed supra-tentorial GBM were enrolled on an institutional review board approved protocol of 5 consecutive days of SRS in escalating doses in a 3+3 design on 4 dose levels: 25 Gy, 30 Gy, 35 Gy, and 40 Gy targeting the GTV of the resection cavity/residual tumor with a 5 mm CTV margin and 0 mm PTV margin. MRI T2 edema signal was not purposely targeted. In-field recurrences were scored if the initial site of recurrence was contained within the 95% prescription isodose line (IDL), which corresponds to 5 mm margin on the cavity/residual or if contiguous with tumors within the 95% IDL. Recurrent tumor located outside of the 95% IDL, between 5 and 20 mm from the cavity (and otherwise covered with typical 2 cm margins for IMRT), were scored as marginal. Recurrences presenting more than 20 mm away from the cavity were considered distal. From 2010 to 2016, 26 patients had progressed and were eligible for pattern of recurrence analysis, with a median follow-up of 14 months (range 2-53). Seventeen out of 26 cases (65%) showed in-field recurrence within the 95% IDL, including one patient who also presented with a distal recurrence. Six patients (23%) showed purely distal recurrences, and 3 patients developed marginal failures (12%). Eight out of the 10 patients in the highest 40 Gy dose level recurred locally. Looking at marginal failures, 2 out of the 3 cases received doses above 30 Gy, including one who received 38 Gy to the failure region. The median time to progression was 3.5 months (range 0.5-17) for infield failures, 9 months (range 5-33) for marginal failures, and 17 months (range 6-22) for distal failures. Median OS was 12 months in patients with in-field recurrence vs. 21 months for distal recurrence (median OS not reached for marginal recurrence). On univariate analysis, pattern of failure was not associated with the dose level received (P = 0.422), with the MGMT methylation status of the tumor (P = 0.820), or with the extent of resection (P = 0.1051). Four out of the 6 patients who developed radionecrosis showed a distal pattern of recurrence compared to those who did not develop radionecrosis (P = 0.0245). Despite dose escalation up to 40 Gy in 5 fractions with concurrent TMZ, the dominant pattern of failure for GBM was in-field. Tumor Control Probability analyses are ongoing to determine the 2 Gy equivalent dose delivered to the 12% of tumors that progressed between 5 and 20 mm. Radionecrosis, but not total dose or MGMT status, correlated with in-field tumor control.