Abstract ONC201 is the first monotherapy beyond radiation to improve outcomes in H3K27M diffuse midline glioma (DMG). However, individual response to ONC201 is variable, and no radiographic biomarkers predict response. Diffusion imaging and ADC reflect tissue cellularity, necrosis and potentially response to treatment. We hypothesize ADC at early treatment timepoints may correlate with overall ONC201 response. Imaging and clinical data were abstracted from chart review of patients treated with ONC201 at University of Michigan. Two neuroradiologists performed centralized review for RAPNO measurements and mean ADC. ROIs were measured using consistent anatomic locations, excluding areas of cystic degeneration or necrosis. Histogram analysis was performed to better capture tumor heterogeneity and mitigate interoperator variability. N = 81 were identified for review. Preliminary analysis was performed on 37 patients with upfront ONC201 therapy, including eighteen on ONC014 (NCT03416530), twelve completing extended access protocol (NCT03134131), and seven from the Xcures database. Primary tumor location included brainstem (n = 22) and thalamic (n = 13). Baseline characteristics of age, ADC and RAPNO size at diagnosis or cycle 3 follow-up size were not statistically different. Stratifying by ADC, median overall survival (OS) for pre-progression post radiation patients was 344 days when ADC decreased and 537 days when ADC increased (log-rank p-value 0.0051) with progression-free survival (PFS) 93 days and 190 days respectively (log-rank p-value 0.0151). In preliminary parametric analysis, increased fraction of tumor with decreased ADC from baseline to cycle 3 may correlate with decreased PFS (spearman r= -0.5250, p-value 0.0471). In H3K27M-DMG patients treated with ONC201, increased ADC after two cycles of ONC201 was predictive of longer OS and PFS. Ongoing work will validate these findings in an independent external cohort, clarify in a cohort of long-term survivors whether this is an ONC201 specific effect and further integrate parametric assessments, tumor enhancement and MRI perfusion, and non-imaging biomarkers (cf-tDNA and metabolomics).
Lau et al. (Fri,) studied this question.