Multi-institutional study of reirradiation for recurrent high grade glioma.

2063 Background: Treatment options are limited for recurrent WHO Grade 3-4 glioma (HGG). Reirradiation (ReRT) is an option, but the ideal treatment regimen and the benefit of concurrent and adjuvant systemic therapy in the recurrent setting remains unclear. Methods: A retrospective review of patients with recurrent HGG treated with reRT was conducted at 12 institutions. Eligible patients were treated with two courses of fractionated RT (>3 fractions) for glioma with the second course being for HGG. To estimate overall and progression free survival times, the Kaplan-Meier method was used. To assess the relationship between survival times and study variables, Cox proportional hazard regression models were used to calculate hazard ratios and the corresponding 95% confidence interval along with the p-value. SAS (version 9.4, Cary, NC, USA) was used for all analyses. P values < 0.05 were assumed to be statistically significant. Results: 482 eligible patients were identified from 1997 to 2023. 235 (54%) had histologic confirmation of glioblastoma at reRT. The median age at reRT was 53.1 years and median KPS was 80. 196 patients (51%) were treated with reRT at their initial recurrence and 122 (30%) had an IDH mutation. The most common reRT dose and number of fractions were 35 Gy and 10 fractions. 192 patients (44%) and 95 (24%) received concurrent (conc) and adjuvant (adj) temozolomide (TMZ) respectively and 116 (27%) and 110 (28%) received conc and adj bevacizumab (BEV) respectively with reRT. Median OS and PFS were 9.8 and 5.3 months. OS (16.6 vs 7.8 months, HR 2.44 p < 0.01) and PFS (8.6 vs 4.6 months, HR 1.85 p < 0.01) were longer in patients with IDH mutations. Receipt of conc and adj TMZ with reRT was associated with improved OS (HR 0.67 p < 0.01 and HR 0.49 p < 0.01 respectively) and PFS (HR 0.66 p <0.01 and HR 0.47 p < 0.01 respectively). Receipt of concBEV with reRT was associated with worse OS (HR 1.66 p < 0.01) but not PFS (HR 1.15 p = 0.28). AdjBEV was not associated with OS or PFS. Concurrent and adjTMZ were associated with improved OS ( p = 0.04 and <0.01) and PFS ( p = 0.01 and <0.01) for IDHwt tumors. In IDHmt tumors, concTMZ was associated with improved OS but not PFS ( p = 0.03 and 0.11), while adjTMZ was associated with improved OS and PFS ( p = 0.05 and <0.01). Symptomatic adverse radiation effects and Grade 3 or greater neurologic toxicity were seen in 107 (25%) and 86 (20%) of patients respectively. Neither conc nor adjBEV nor TMZ were associated with symptomatic ARE. ConcBEV was associated with lower (13% vs 23%, p = 0.03) and concTMZ was associated with higher rates (25% vs 15%, p = 0.01) of Grade 3 or greater neurologic toxicity. Conclusions: The use of concurrent and adjuvant TMZ with reRT are associated with improved OS and PFS in recurrent HGG. OS and PFS are improved with conc and adjTMZ for IDHwt tumors, while there was no PFS benefit to concTMZ in IDHmt tumors. The rate of high grade neurologic toxicity was decreased with the use of concurrent BEV and increased with the use of concTMZ.