Medical Student Columbia University Irving Medical Center, Department of Neurological Surgery
Introduction: Convection-enhanced delivery (CED) of MTX110 (aqueous panobinostat) shows promise in treating diffuse midline glioma (DMG). Repeated CED infusions may improve efficacy, but optimizing safe delivery to the pons requires monitoring infusion-induced volumetric changes.
Methods: Seven patients with biopsy-confirmed DMG underwent two 9.6 mL CED infusions of MTX110 over 48 hours with a 3-5 day rest interval. MRIs were acquired pre-infusion, 48 hours post-infusion and at 2-month follow-up. Tumor extent was defined as the T2-hyperintense region, and the drug distribution volume (Vd) was assessed via co-infused gadolinium. Images were nonlinearly registered to track changes in the brainstem morphology across infusions. The pons volume and bicaudate index were measured at each time point.
Results: For the first and second infusion, mean (SD) tumor coverage was 69% (15%) and 58% (22%) (P = 0.09), mean pons coverage was 53% (16%) and 44% (20%) (P = 0.11), and Vd was 20 cm3 and 17 cm7 (P = 0.23). Nonlinear registration revealed dorsal pons expansion toward the fourth ventricle and cerebellum, with pons volume increasing by 17% (6%) across the first infusion and 12% (5%) across the second (P = 0.02). During the rest period, the pons volume decreased by 6% (5%) but remained 10% above baseline (P < 0.001). The bicaudate index increased by 14% (7%) and 32% (10%) across infusions. No patients developed hydrocephalus, and all morphological changes except in one patient (due to tumor recurrence) returned to baseline at a 2-month follow-up.
Conclusion : Repeated CED infusions safely achieved excellent tumor and pons coverage. Importantly, coverage is underestimated as post-infusion MRIs were acquired 48 hours after infusion and gadolinium inherently underestimates drug distribution. Reversible morphological changes in the brainstem were observed, demonstrating pontine elasticity that was well-tolerated. These non-invasive imaging techniques can be incorporated into future CED clinical trials to monitor safety and optimize drug distribution.