RTHP-09. PROGRESSION AND PSEUDOPROGRESSION OF GLIOBLASTOMA MULTIFORME IN THE TEMOZOLOMIDE ERA

Standard of care for glioblastoma(GBM) includes temozolomide with radiotherapy. Pseudoprogression has been documented with increased frequency, complicating assessment of early imaging studies after chemoradiation (CRT). MATERIALS: Of 114 consecutive patients receiving CRT for newly diagnosed GBM at our institute between 1998 and 2015, 85 had pre and post-treatment charts and imaging suitable for retrospective evaluation using currently accepted RANO criteria. Patient status was classified at 3 months from baseline post-radiation imaging as, partial response/stable disease(PR/SD), pseudoprogression(PsP), and true progression(TP). At the three-month time point there were 37(43.5%) with PR/SD, 16(18.8%) PsP, and 23(27.0%) TP. Nine(10.7%) patients were not evaluable due to early initiation of bevacizumab. Age and sex were equally distributed between groups. Trends toward increased rates of gross total resection in PR/SD, and periventricular location for TP, were seen. Rates of MGMT methylation were 50%(PR/SD), 55.6%(PsP), and 16.7%(TP). Early imaging one month after the post-radiation baseline demonstrated increased contrast enhancement in all PsP cases (none had stabilized or improved), 93.7% had increased tumor perfusion (relative cerebral blood volume), and 50% had increased FLAIR signal. No differences were noted in mean tumor volumes, or location of pseudoprogression in reference to radiation isodose levels. Most importantly, for PR/SD, PsP, and TP, median PFS was 9.80, 9.07, and 4.16 months (p ˂ 0.0001) and median OS was 27.0, 20.7, and 14.0 months(p = 0.0042), respectively. Contrary to published reports, our data do not demonstrate an improvement in survival outcomes for patients with pseudoprogression compared to stable disease or response on first post-radiation imaging. Our data may be limited by patient numbers but importantly, there were no early imaging characteristics, including perfusion imaging, that differentiated pseudoprogression from true progression. These results underscore the need for continued investigations to develop non-invasive techniques to predict pseudoprogression and accurately predict overall survival.