Prevalence, distribution, and clinical associations of suspected postictal changes on brain magnetic resonance imaging in epileptic dogs.

Abstract

To determine the prevalence of presumed postictal changes (PC) on brain MRI in epileptic dogs, describe their distribution, and recognize possible correlations with different epilepsy features. 540 client-owned dogs with epilepsy and a complete medical record that underwent brain MRI at 4 veterinary referral hospitals between 2016 and 2019. Data were collected regarding signalment, seizure type, seizure severity, time between last seizure and MRI, and etiological classification of epilepsy. Postictal changes were considered when solitary or multiple intraparenchymal hyperintense lesions were observed on T2-weighted and fluid-attenuated inversion recovery images and were hypointense or isointense on T1-weighted sequences, which were not confined to a vascular territory and showed no to mild mass effect and no to mild contrast enhancement. Sixty-seven dogs (12.4%) showed MRI features consistent with PC. The most common brain sites affected were the piriform lobe, hippocampus, temporal neocortex, and cingulate gyrus. Dogs having suffered cluster seizures or status epilepticus were associated with a higher probability of occurrence of PC, compared to dogs with self-limiting seizures (OR 2.39; 95% confidence interval, 1.33 to 4.30). Suspected PC were detected both in dogs with idiopathic epilepsy and in those with structural epilepsy. Dogs with unknown-origin epilepsy were more likely to have presumed PC than were dogs with structural (OR 0.15; 95% confidence interval, 0.06 to 0.33) or idiopathic epilepsy (OR 0.42; 95% confidence interval, 0.20 to 0.87). Time between last seizure and MRI was significantly shorter in dogs with PC. MRI lesions consistent with PC were common in epileptic dogs, and the brain distribution of these lesions varied. Occurrence of cluster seizures or status epilepticus, diagnosis of unknown origin epilepsy, and lower time from last seizure to MRI are predictors of suspected PC.