A stroke severity monitoring system based on quantitative modified multiscale entropy

Stroke is the leading cause of death and disability worldwide. This requires significant resources in health-care costs. In addition to physical examination and brain imaging, medical staff need a more quantitative and continuous method to reveal and monitor the severity of patients. This paper proposed a novel stroke severity monitoring system based on a nonlinear method-quantitative modified multiscale entropy (qmMSE). National Institutes of Health Stroke Scale (NIHSS) is adopted as reference of the severity of stroke patients. In the intensive care unit (ICU) admitted acute stroke patients, our proposed qmMSE feature has significant (p-value equals to 0.0026) difference between high NIHSS groups and low NIHSS groups. QmMSE not only highly correlates to NIHSS, but also consists with other clinical parameters, such as stroke volume and Glasgow Coma Scale (GCS). Beside, our proposed method has better significance in patient with ischemic stroke in cortical region. The p-value reaches 0.0008.