Role of cardiac computed tomography in hyperacute stroke assessment

Background: Incorporating cardiac CT with hyperacute stroke imaging may increase the yield for cardioembolic sources. It is not clarified whether stroke severity influences on rates of intracardiac thrombus. We aimed to investigate a National Institutes of Health Stroke Scale (NIHSS) threshold below which acute cardiac CT was unnecessary. Methods: Consecutive patients with suspected stroke who underwent multimodal brain imaging and concurrent non-gated cardiac CT with delayed timing were prospectively recruited from 1st December 2020 to 30th November 2021. We performed receiver operating characteristics analysis of the NIHSS and intracardiac thrombus on hyperacute cardiac CT. Results: A total of 314 patients were assessed (median age 69 years, 61% male). Final diagnoses were ischemic stroke (n=205; 132 etiology-confirmed stroke, independent of cardiac CT and 73 cryptogenic), transient ischemic attack (TIA) (n=21) and stroke-mimic syndromes (n=88). The total yield of cardiac CT was 8 intracardiac thrombus and 1 dissection. Cardiac CT identified an intracardiac thrombus in 6 (4.5%) with etiology-confirmed stroke, 2 (2.7%) with cryptogenic stroke, and none in patients with TIA or stroke-mimic. All of those with intracardiac thrombus had NIHSS ≥4 and this was the threshold below which hyperacute cardiac CT was not justified (sensitivity 100%, specificity 38%, positive predictive value 4.0%, negative predictive value 100%). Conclusions: A cutoff NIHSS ≥4 may be useful to stratify patients for cardiac CT in the hyperacute stroke setting to optimize its diagnostic yield and reduce additional radiation exposure.