To identify risk factors for the development of tumor-related epilepsy (TRE) and investigate its natural history in adults with new diffuse gliomas (DG).

TRE is common in patients with DG and causes significant morbidity. Future clinical trials for TRE will depend on population enrichment strategies.

Adults with newly diagnosed DG were enrolled in a 1-year prospective cohort study. Data collected included: personal/family seizure history, neuro-imaging, surgical outcome, and tumor histology including isocitrate dehydrogenase mutation, excitatory amino acid transporters 1 and 2, and system Xc antiporter. Logistic regressions were used to examine associations between clinical variables and risk of seizure at diagnosis. At study end, subjects were identified as having TRE (controlled or refractory) or not. Refractory was defined as failure to control seizures despite appropriate therapy. Cox proportional hazards models were used to examine the associations between the potential risk factors and time to first or second seizure.

64 subjects were recruited between 2/2018-2/2020. 34/64 experienced seizures at diagnosis. Temporal lobe location was associated with seizures at diagnosis, but not significantly (OR 2.6, p = 0.08). 12/29 seizure-naïve subjects developed TRE. Subjects with multifocal (HR 3.7, p =0.03) or diffuse tumor distribution (HR 20.2, p = 0.02) were at higher risk of TRE than those with unifocal tumors. Thalamic involvement was associated with higher risk of developing TRE (HR 9.6, p = 0.01). Seizure recurrence was more common with involvement of insula (HR 7.4, p = 0.02), thalamus (HR 3.2, p = 0.04) or both hemispheres (HR 7.1, p=0.02). Proportion of unmethylated MGMT was higher in rTRE, but this was not significant (p=0.08). Other histologic markers were not associated with TRE risk.