Modern Prediction Methods for Recurrent Ischemic Stroke. An Analytical Review

Significance. Ischemic stroke is a common disease with a high mortality and disability rate. The prevalence of recurrent acute cerebrovascular accidents, combined with the limited capabilities of their prediction methods, significantly reduces the effectiveness of personalized risk management for recurrent stroke. Purpose. To assess the capabilities of methods available in world practice for predicting the risk of recurrent ischemic stroke Material and Methods. Selecting relevant studies from the PubMed database for the period of 2015-2025 in accordance with the inclusion criteria. The screening stage was carried out via the CADIMA service of the 2.2.4.2 version. The review was written in accordance with the PRISMA 2020 criteria. Results. The data from the studies focused on recurrent ischemic stroke risk prediction was analyzed. The possibilities of using prediction methods based on laboratory, instrumental, genetic factors, as well as prediction methods based on machine learning technologies were studied. Machine learning methods had higher metric assessment rates, compared to classical statistics methods. Conclusion. The results show that the effectiveness of machine learning methods is high in regards to recurrent stroke prediction.