HOUT-14. TIME TO PROGRESSION AND VOLUME AT PROGRESSION DEPEND ON DEFINITION OF PROGRESSION IN GLIOBLASTOMA

Time to progression (TTP) is an established response measure, and at progression patients can be included in trials. Standards for imaging interpretation of progression are lacking. We determined the differences in time to progression and in tumor volumes at progression between three methods to assess progression. From a consecutive cohort of 97 patients with glioblastoma in 2012 or 2013, 63 had MRI follow-up after initial treatment to evaluate progression. TTPclinical was determined by multidisciplinary evaluation in clinical practice; TTPRANO by the RANO criteria for trial inclusion; and TTPconsensus by multidisciplinary consensus review (neuroradiologist, neurosurgeon, 2 radiation oncologists) looking back on all MRI information (on average 3.9 follow-up MRIs per patient, range 1-11) with knowledge of further progression and death, postulated as gold standard. MRIs were co-registered to facilitate the consensus review and the maximum perpendicular tumor diameters and volumes were based on enhancing tumor segmentations. No patient was without progression with consensus review, one with clinical practice evaluation and 22 with RANO criteria. The median TTPconsensus was 36 weeks, TTPclinical 40 weeks and TTPRANO 57 weeks. The median overall survival was 64 weeks. The median progression volume was 8.8 mL with consensus review, 17 mL with clinical practice evaluation, and 38 mL with RANO. TTP and volume at progression vary considerably depending on definition of progression. Different purposes may require different progression criteria. Early detection with small volumes may be useful to evaluate progression locations in relation to initial treatment, but can only be determined after the course of disease. RANO criteria may be useful for reproducible clinical trial inclusion, but at the price of later detection with larger volumes. Progression according to RANO criteria is considerably later than clinical practice evaluation in this cohort, potentially introducing lead time bias in trials.