HOUT-21. CHARACTERISTICS OF SHORT-TERM SURVIVAL IN PATIENTS WITH GLIOBLASTOMA: A RETROSPECTIVE ANALYSIS

We sought to identify characteristics of glioblastoma (GBM) patients with short survival (< 10 months) in order to identify prognostic factors useful for guiding treatment management. This is an IRB-approved retrospective analysis of adult newly diagnosed GBM patients from 2008–2016 who survived < 10 months from diagnosis. We extracted demographics, tumor characteristics, and treatment details. We calculated survival from surgical diagnosis to date of death. The cohort includes 197 subjects (61% male) with a median age of 68 years (range 19–94). The majority (93%) are non-Hispanic white. The cohort has a median survival of 144 days (95% CI: 130–160). We focused on traditional prognostic indicators, including extent of surgical resection and KPS. A majority had biopsy only (n=92, 46.7%) rather than gross total (n=59, 29.9%) or subtotal (n=46, 23.4%) resection. Moreover, 160 out of 197 patients had a documented KPS with a majority being below 90 (KPS=70–80 (n=96); KPS < 70 (n=31)). Of 179 patients with data on RT course, 18% (n=32) received no RT or opted for hospice after diagnosis, 3% (n=6) received only RT, 54% (n=97) received RT+temozolomide (TMZ), and 24% (n=43) received RT+TMZ+bevacizumab. Of the 147 subjects receiving RT, 79% completed their RT course as prescribed. Most commonly, RT was prescribed as a 6- to 6-1/2-week course (85%), typically 59.4 Gy (45Gy primary, 14.4Gy boost) over 33 fractions or 60 Gy over 30 fractions. In contrast, 15% received a 3-week RT course, typically scheduled as 15 fractions of 2.667 Gy. We concluded that GBM patients with survival < 10 months were more likely to have biopsy only and a KPS < 90, notably associated with poorer prognosis. We continue to explore this dataset for further prognostic factors, particularly inability to complete planned RT course, and are comparing these traits to a larger cohort.