Neurological Disorders

Background: Patients who present with acute ischaemic stroke require both accurate and timely diagnosis in order to facilitate prompt treatment. Up to 30% of acute stroke presentations can actually consist of stroke mimics such as seizures or migraine. Recent advances in Artificial Intelligence (AI) have enabled automatic analysis and quantification of both CT angiography and CT perfusion studies. AI analytic programs can assist clinicians in both detecting ischaemic penumbra, as well as recognising atypical perfusion patterns in stroke mimics.

Methodology/Case presentation: A 68-year-old lady with a background of cardiovascular risk factors presented to Emergency as a stroke activation. She exhibited right sided weakness, aphasia and neglect. Urgent computed tomography imaging showed neither evidence of established infarct or haemorrhage, nor any large vessel occlusion. Collateral history revealed a generalized tonic-clonic movements preceding her right-sided weakness, and clinical seizure was diagnosed. Anti-epileptic medications were commenced, and her clinical deficits rapidly improved. The next morning, our patient’s right sided neglect and aphasia recurred.

Findings: Repeat CT angiography showed reduced opacification of cortical branches of the left middle cerebral artery. CT perfusion imaging showed areas of mild hypoaemia in the left frontoparietal lobes, quantified by commercially available artificial intelligence processing software as slightly reduced perfusion time of 4 seconds involving only 18 mL of territory. These findings indicated a final diagnosis of post-ictal paresis. Post-ictal Todd’s paresis leads to reduced cerebral
perfusion via metabolic exhaustion and the action of inhibitory interneurons.

Conclusion: This case study demonstrates the utility of CT perfusion studies and AI-analysis in differentiating ischaemic strokes from stroke mimics stroke mimics, facilitating accurate and timely diagnosis in order to streamline management.

Limitations: This case report’s unitary sample size allows for only case-restricted conclusions, and does not allow its findings to be generalized to a greater population. Larger studies could correlate its findings on a greater sample size and allow for further analysis and application to patient populations.