2233 - Factors Predicting Longer PFS and OS in Patients with Tile-Based Radiation Therapy for Recurrent GBM

Purpose/Objective(s): Tile-based radiation therapy (TBRT) is an emerging brachytherapy for brain neoplasms. In recurrent glioblastoma (GBM), for which there is no clearly superior standard treatment, TBRT may improve overall survival (OS) and progression-free survival (PFS). We hypothesize that concomitant cytotoxic chemotherapy enhances OS and PFS in patients receiving TBRT after reoperation for recurrent GBM.

Materials/Methods: Using data from a prospective multicenter observational basket study of patients receiving TBRT for brain neoplasms, we identified 137 patients with recurrent or progressive GBM who were previously treated with radiation therapy. We collected clinical characteristics and follow-up data, including patient age, sex, Karnofsky Performance Status (KPS), IDH genotype, MGMT methylation status, prior treatments, extent of resection (EOR), treatment volumes, concomitant medications, and PFS/OS status at follow-ups up to 36 months. We performed Kaplan-Meier, Cox regression, and Fine-Gray subdistribution hazards modeling to assess factors associated with OS and PFS. We included parameters with p < 0.2 in univariate analysis in multivariate Cox and Fine-Gray hazards modeling.

Results: Median age was 61 years (IQR 53-66), and 59.9% were male. Median preoperative KPS was 80 (IQR 70-90). 86.9% of patients had IDH wild-type tumors and 24.1% had MGMT promoter methylation. Median tumor volume was 20.7 cc (IQR 5.6-43.5). Gross total resection (GTR) was achieved in 64.2%, near-GTR in 21.2%, and subtotal resection in 10.9%. Patients received a median of 6 TBRT tiles (IQR 4-8), with a prescribed dose of 60 Gy at 5 mm depth. Median D90 was 54.2 Gy (IQR 45.2-66.3), with median V100% of 83.6% (IQR 66.8-93.3) and median V150% of 51.3% (IQR 39.0-61.1). 26.3% of patients received prior Tumor-Treating Field (TTF) therapy, and nearly all (97.8%) received concomitant chemotherapies. Reverse KM estimate of median follow-up time was 5.4 months. KM estimate of median PFS was 7.3 months, OS at 12 and 24 months was 66% and 51% respectively. Variables meeting criteria for multivariate modeling of PFS were prior TTF, GTV, D90, V150%, and concomitant cytotoxic chemotherapy. Variables meeting criteria for multivariate modeling of OS were EOR, age, prior TTF, GTV, V100%, and V150%, and concurrent bevacizumab. MGMT methylation, IDH genotype, concomitant bevacizumab, and cytotoxic chemotherapy were included in all multivariate models. On multivariate Cox modeling, concomitant cytotoxic chemotherapy was significantly associated with longer PFS, and greater EOR, younger age, and lower D90 were significantly associated with longer OS. On Fine-Gray modeling, none of these were significantly associated with progression with the competing risk of death.

Conclusion: Concurrent cytotoxic chemotherapy with TBRT is associated with enhanced PFS, but not OS in treatment of recurrent GBM.