Resection and surgically targeted radiation therapy for locally recurrent GBM.

Abstract

2054 Background: Recurrent GBM (rGBM) is a diffuse disease, and resection (R) alone does not provide durable local control (LC) or prolong overall survival (OS). Hypothesizing R plus immediate radiation (RT) may achieve durable LC and secondarily improve OS by permitting time for subsequent potentially effective but biologically slower treatments to have an impact, we prospectively evaluated R combined with a novel surgically targeted radiation therapy (STaRT) device utilizing Cs-131 embedded in bioresorbable collagen tiles. Methods: From 2/13-2/18 patients (pts) with locally recurrent GBM were treated on a prospective single arm trial (ClinicalTrials.gov, NCT#03088579) of maximum safe resection and immediate RT (GammaTile, GT Medical Technologies, Tempe AZ). Upon resection the at-risk areas of the surgical bed were lined with the GammaTile (GT) device, delivering 60-80 Gy at 5 mm. Follow up treatments were not specified but captured; no pt. underwent additional local therapy without progression, and no pt. was lost to follow up. We present study specified endpoints of local control (LC), overall survival (OS), and adverse events (AE), and a post hoc, hypothesis-generating analysis of outcomes by receipt of systemic (Sys) therapy. Results: 28 locally recurrent GBM were treated, 20 at first progression (range 1-3). Median age was 58 years (yrs.) (range 21-80), KPS 80 (60-100), female: male ratio 10:18 (36/64%). MGMT was methylated in 11%, unmethylated in 18%, and unknown in 71%. For all pts., median OS was 10.7 months (mo.) (range.1-42.3), and radiographic LC was 8.8 mo. (range.01-34.5). LC (defined as < 15 mm from surgical bed) was maintained in 50% of pts., and no first failure was local. 12 mo. OS was 75% for pts. < 50 yrs. vs. 43% for > 50 yrs. (HR.46, p =.009). MGMT, KPS, and sex were non-predictive. After R+GT, 17 pts. received > 1 cycle of systemic therapy (Sys), either as adjuvant or salvage, alone or in combination . Sys was bevacizumab (BEV) in 15 pts., temozolomide (TMZ) in 12, and lomustine (CCNU) in 8 (N > 17 as some pts. received > 1 Sys). Post hoc analysis disclosed a 15.1 mo. OS for pts. receiving > 1 cycle of Sys (Sys+, N = 17) vs. 6.5 mo. for no Sys (Sys-, N = 11) (hazard ratio (HR).38, p =.017)). LC was 11.4 mo. for Sys+ and 2.1 mo. for Sys- (HR.44; p =.16)). Median OS (mo.) for BEV+ vs. BEV- was 16.7/4.5 (HR.38, p =.017), for TMZ+ vs. TMZ- 17.5/6.7 (HR.40, p =.025) and for CCNU+ vs. CCNU- 17.5/7.9 (HR.61, p =.25), respectively. Three attributed AE occurred, 1 dehiscence requiring surgery and 2 radiation brain effects, medically treated. 4 unrelated deaths occurred < 60 days post-op, all in the Sys- cohort, impacting their opportunity for subsequent treatment. Conclusions: In this study local treatment alone was insufficient to achieve prolonged OS. Post hoc analysis suggests R+GT coupled with Sys may have potential to impact OS in rGBM patients. GT was FDA cleared in 2020 for use in newly diagnosed malignant and all recurrent intracranial neoplasms. Clinical trial information: NCT#03088579.