Computed tomography perfusion as a diagnostic tool for seizures after ischemic stroke

Cerebral cortical ischemia is a risk factor for post-stroke seizures. However, the optimal imaging method is unclear. We investigated CT perfusion (CTP) in detecting cortical ischemia and its correlation with post-stroke seizures compared with non-contrast CT (NCCT). We included patients with acute ischemic stroke admitted to the Royal Melbourne Hospital between 2009 and 2014. Post-stroke seizure information was collected. Cortical involvement was determined on acute NCCT and CTP (T_max, cerebral blood volume [CBV], and cerebral blood flow [CBF]). The association between cortical involvement detected by different imaging modalities and post-stroke seizures was examined. Three-hundred fifty-two patients were included for analysis. Fifty-nine percent were male, and median age was 73 years (inter-quartile range 61-82). Follow-up was available for 96 %; median follow-up duration was 377 days (inter-quartile range 91-1018 days). Thirteen patients had post-stroke seizures (3.9 %). Cortical involvement was significantly associated with post-stroke seizures across all modalities. CBV had the highest hazard ratio (11.3, 95 % confidence interval (CI) 1.1-41.2), followed by NCCT (5.3, 95 % CI 1.5-18.0) and CBF (4.2, 95 % CI 1.1-15.2). Sensitivity was highest for T_max (100 %), followed by CBV and CBF (both 76.9 %) and NCCT (63.6 %). Specificity was highest for CBV (77.8 %), then NCCT (75.6 %), CBF (54.0 %), and T_max (29.1 %). Receiver-operating characteristic area under the curve was significantly different between imaging modalities (p < 0.001), CBV 0.77, NCCT 0.70, CBF 0.65, and T_max 0.65. CTP may improve sensitivity and specificity of cortical involvement for post-stroke seizures compared to NCCT. (orig.)