EPID-25. HOW MANY PATIENTS IN A REAL WORLD GLIOBLASTOMA POPULATION MEET ELIGIBILITY CRITERIA IN CLINICAL TRIALS?

Abstract

Abstract In population series of glioblastoma (GBM) the median survival is < 1 year. For patients eligible for combined oncological treatment (surgery, radiotherapy (RT) and temozolomide (TMZ)) the median survival raises to 15 months. These patients are younger, have higher performance status and less comorbidity compared to the total GBM population. The extent of selection bias of GBM patients enrolled in clinical trials is, however, unclear. Using the Cancer Registry of Norway and the Brain Tumor Database of Oslo University Hospital, all patients diagnosed with GBM between 2012 and 2016 within a defined geographical area that received all oncological treatment at Oslo University Hospital were identified. Their hospital charts were retrospectively investigated and all patients were evaluated for the possibility to be included in clinical trials according to standard criteria from recent published phase 3 studies. We defined three time points for possible enrollment in trials: before-, halfway through-, and after concomitant RT/TMZ. We identified 424 patients of whom 26 (6.1%) did not undergo surgery. The median survival in the population was 11.7 months. Of the patients that underwent surgery (biopsy or resection), the median survival progressively increased with additional oncological treatment; no RT/TMZ 1.1 months, only RT 7.9 months, RT/TMZ 13.9 months. A potential study patient group≥18 years undergoing surgery for a supratentorial GBM with ECOG ≤2, with normal bone marrow, hepatic and renal function and no severe comorbidity or oncological disease constituted 54% of the population before start of RT/TMZ. Patients adhering to RT/TMZ-treatment and without tumor progression represented 46% and 43% of the population halfway through- and after concomitant RT/TMZ, respectively. The survival of patients that could be included in clinical trials was significantly higher (p< 0.001) than patients not fulfilling standard criteria. Our data suggest a significant patient selection in clinical trials from a real-world GBM population.