Association between enlarged perivascular space (ePVS) density and stereo‐electroencephalography (SEEG) biomarkers of epileptogenicity

Abstract Objective We aim to explore the association between enlarged perivascular space (ePVS) density and stereo‐electroencephalography (SEEG) biomarkers of epileptogenicity. Methods We retrospectively analyzed consecutive SEEG patients from an Australian site. We automatically segmented ePVSs from 3T pre‐SEEG T1‐weighted magnetic resonance imaging (MRI) scans and calculated ePVS (1) hemispheric, (2) sub‐lobar, and (3) contact‐level density. We defined the epileptogenic zone (EZ) SEEG contacts as those identified as the primary EZ and then selected for radiofrequency thermocoagulation (RF‐TC). We classified contacts generating the top 10% of interictal epileptogenicity biomarkers (spikes, fast ripples, and cross‐rates of high‐frequency oscillations HFO*spikes). We assessed the relationship at each level for the whole cohort and its subgroups: MRI‐negative‐only, different voxel sizes (.9 mm 3 /1.0 mm 3 ), contact locations (mesial temporal/neocortical), and seizure‐free patients. Results From 53 RF‐TC patients, ePVS density was not associated with the EZ at the hemispheric ( p = .995), sub‐lobar ( p = .090), or contact ( p = .999) level in the whole cohort. In the MRI‐negative‐only subgroup and 1.0 mm 3 isotropic subgroup, ePVS density was inversely associated with EZ (odds ratio OR .76, 95% confidence interval CI: .61–.93, p = .009 and OR .83, 95% CI: .69–.99, p = .036, respectively). There was no association between ePVSs and the interictal epileptogenic biomarkers at all levels in the whole cohort. At the subgroup level, mesial temporal ePVS density was inversely associated with the top 10% of fast ripples (OR .09, 95% CI: .04–.24, p < .001) and cross‐rates of HFO*spikes (OR .22, 95%CI: .11–.44, p < .001), whereas in the neocortex, ePVSs were positively associated with fast ripples (OR 1.10, 95% CI: 1.01–1.18, p = .018), which may be driven by the interictal biomarker organization rather than epileptogenicity. Significance ePVS density may not be a consistent biomarker of SEEG‐defined epileptogenicity. The relationship between reduced ePVS and epileptogenicity in MRI‐negative cases warrants further study, as this may shed light on underlying pathobiology and EZ biomarkers. Different imaging techniques may be able to capture the relationship between the glymphatic system disruption and epileptogenicity.