Abstracts

Rationale: Febrile status epilepticus (FSE) is an important risk factor for developing temporal lobe epilepsy (TLE). Although only 30-40% of children who experience FSE will go on to develop TLE (Mathern, et al. [1995] Brain. 118, 105?"118), there is currently no way to predict which patients will develop epilepsy. Recently, we discovered a magnetic resonance imaging (MRI) signal that is predictive of TLE in a rat model of FSE (Choy, et al. [2014] J Neurosci. 34, 8672?"84). Specifically, a reduction of T2- relaxation times in the basolateral amygdala (BLA) 2-4 hours after FSE distinguished which rats became epileptic months later. However, clinically it would be difficult to enroll and scan infants within that time period. Thus, we investigated if our observed signal changes persist until 6 hours after FSE. We also compared MRIs at 18 and 48 hours to evaluate if the evolution of the T2 signal can provide more information about the evolving epileptogenic process. Methods: Experimental FSE was induced in 10-11 day old Sprague Dawley rats as previously described (Dub鮠et al. [2006] Brain 129, 911?"22). Briefly, half of the pups were exposed to a stream of warm air to induce behavioral seizures. Elevated core temperatures and seizures were maintained for 60 minutes. In the first cohort, rats were imaged using T2-weighted MRI scans on a Bruker Avance 11.7 T MR scanner at 2, 18, and 48 hrs after the end of FSE. Bipolar electrodes were implanted bilaterally in the hippocampus one month after FSE and were monitored with video-electroencephalography (EEG) for up to 10 months. EEGs were manually analyzed for seizures. In a second cohort, animals were imaged 2 and 6 hrs following FSE. In both cohorts, regions of interest were manually drawn on T2 maps. Manual analysis was done without knowledge of experimental group. Animals were labeled with decreased T2 signal relaxation if the value was =2 standard deviations below the mean of normothermic controls. Receiver Operator Characteristic (ROC) curves were use to determine diagnostic strength. Results: In the long-term cohort, 6/19 (32%) of FSE rats had spontaneous seizures. ROC curve indicated that the area under the curve (AUC) at 2 hr for the BLA was 0.91 0.08 (p= 0.005). Measuring the change in T2 signal in the BLA between the 2 and 48 hr revealed an increased AUC to 0.99 0.02 (p = 0.001). The T2 signal change between 2 and 48 hr in the medial amygdala predicted epileptogenesis perfectly, with an AUC of 1.00 0, (p = 0.001). In the second cohort, there was no statistical difference (p =0.35) and excellent congruence (r=0.73) between values obtained at 2 and 6 hrs. Conclusions: These findings are significant because they support the clinical use of T2-weighted MRI to predict epilepsy following FSE. The predictive T2 decrease in the BLA found at 2 hrs is still apparent at 6 hrs, allowing for easier clinical translation to infants. The difference the T2 relaxation values between 2 and 48 hrs emerged as another predictor of epileptogenesis. These changes may reflect biological processes occurring in the days following FSE and we are currently investigating how they may relate to epileptogenesis. Funding: NIH Grant NS35439, NS78279