Establishment And Verification Of Hematoma Expansion Prediction Score Of Intracerebral Hemorrhage In Qinghai Area

Objective: Intracerebral hemorrhage(ICH)was the most fatal and incurable type of stroke.Studies indicated that clinical data and non-contrast computed tomography(CT)image markers could predict hematoma expansion(HE)in early stage.The purpose of this study was to establish a nomogram model to predict HE after ICH based on the results of above factors,and validated it in an internal cohort.Methods: From January 1,2012 to May 22,2018,all patients with ICH were collected in Qinghai Provincial People’s Hospital.This was a single center retrospective study.Time from onset to baseline CT scan was less than 24 hours and re-examination CT scan was completed within 24 hours after baseline CT in all patients.CT imaging markers included hematoma shape,hematoma heterogeneity,island sign,satellite sign,blend sign,black hole sign and swirl sign.The clinical data and imaging data of these patients were analyzed by multivariate logistic regression analysis,and the independent predictors of HE were obtained.The nomogram model was built based on these independent risk factors.The weight of these factors was assigned reasonably.The next internal validation was carried out.Result: A total of 406 patients with ICH and the incidence of HE was 30.9%.The incidence of HE in basal ganglia was 36.4% and non-basal ganglia was 17.2%.Island sign,diabetes,black hole sign,basal ganglia hemorrhage,swirl sign,time from onset to baseline CT and blend sign were independent risk factors for HE(all p < 0.05).Island sign was the most risk factor for predicting HE,while swirl sign was the lowest risk factor.A nomogram model was established based on these independent predictive factors which were significantly associated with HE.As cut-off was 50%,the point of one patient was higher than 254 points,which meant that the patient might have HE.The accuracy,sensitivity and specificity of the model were high.The internal validation results further confirmed the accuracy of the prediction ability of this model.Conclusion: This model had high accuracy of predicting HE in Qinghai Area.Because of the non-invasive,rapid and low-cost,the model was easy to promote and had wide application prospects in clinical practice in Qinghai Area.