High-level cytokine release syndrome risk prediction method and system

Abstract

The invention belongs to the technical field of medical care informatics, and particularly relates to a high-level cytokine release syndrome risk prediction method and system. Clinical observation data of cytokine release syndrome of a novel coronavirus patient is utilized, and a machine learning model is used for predicting the risk of the cytokine release syndrome; according to the present invention, after the characteristics significantly related to the occurrence of the high-level cytokine release syndrome are screened out, the oxygen saturation, the D-dimer, the diastolic pressure and the international standardization ratio of the CAR-T cell treatment patient are adopted as the prediction evaluation indexes, the high-level cytokine release syndrome risk prediction is performed through the XGBoost model, the sensitivity is high, the specificity is high, and the good prediction effect is provided.

Images