Patients aged ≥12 years with grade 2 mIDH1/2 oligodendroglioma or astrocytoma, no prior glioma treatment other than surgery, and no uncontrolled seizures, were randomized 1:1 to vorasidenib 40 mg or placebo daily. On-treatment seizure numbers were analyzed in patients with ≥1 seizure during baseline or on-treatment periods using a negative binomial regression model. Potential association between seizure activity and tumor volume was assessed using mixed-effect model with repeated measurements. Analyses were exploratory; P-values were not prespecified and should be interpreted with caution.

This analysis included 168 patients treated with vorasidenib (oligodendroglioma n=88; astrocytoma n=80) and 163 treated with placebo (oligodendroglioma n=84; astrocytoma n=79). Patients treated with vorasidenib versus placebo had lower on-treatment rates of seizures; model-estimated rate per person-year: 18.2 (95% confidence interval [CI] 8.4–39.5) vs 51.2 (95% CI 22.9–114.8); rate ratio: 0.36 (95% CI 0.14–0.89), P=0.0263. There was a highly positive association between tumor volume (log scale) and seizure number (estimate of coefficient: 0.7, standard error [SE]: 0.26, P=0.007). Sensitivity analyses per tumor volume suggest that vorasidenib was negatively associated with both number and severity of seizures (estimates of coefficient [SE], -4.81 [1.41], P=0.0008; -0.55 [0.25], P=0.0289, respectively).

Vorasidenib was associated with lower seizure rate and severity than placebo in patients with mIDH1/2 glioma. Smaller tumor volume was associated with lower seizure rate. Our analyses suggest a potential mechanism of seizure control with vorasidenib through tumor size reduction.