Combination Of Noncontrast Computed Tomography Hypodensities And Ultraearly Hematoma Growth For Outcome Prediction After Spontaneous Intracerebral Hemorrhage

BackgroundSpontaneous intracerebral hemorrhage(sICH)is the second most common type of stroke after ischemic stroke,although it only accounts for 6.9-19.6%of all stroke types,it has the highest morbidity and mortality.Accurate evaluation of the prognosis of sICH patients is important for the choice of treatment options,noncontrast computed tomography hypodensities(HD)and ultraearly hematoma growth(uHG)are reliable markers for outcome prediction in patients with sICH,but there is not a joint analysis of the two yet.ObjectiveThe present study aimed to assess whether the combination of these 2 markers could improve the prognostic value for sICH.Methods1.We recruited 242 patients with sICH who had been admitted within 6 hours from the onset of symptom.2.HD was defined as any hypodense region inside the hematoma,regardless of its morphology and size,but not be connected with the surrounding brain parenchyma.HD was assessed by 2 independent blinded readers.3.uHG was calculated as baseline ICH volume/onset-to-imaging time.Receiver operating characteristic(ROC)curve was used to determine the optimal cutoff point of uHG for outcome prediction,which was 6.16 mL/h.According to the cutoff point,uHG was dichotomized as uHG(H):uHG is greater than or equal to 6.16mL/h,and uHG(L):uHG is less than 6.16mL/h.We divided the study population into 4 groups:uHG(L)HD(-)(uHG<6.16 mL/h and HD negative),uHG(L)HD(+)(uHG<6.16 mL/h and HD positive),uHG(H)HD(-)(uHG≥6.16 mL/h and HD negative),and uHG(H)HD(+)(uHG≥6.16 mL/h and HD positive).4.The outcome at 90 days was evaluated by the modified Rankin Scale(mRS)score and was dichotomized as good(mRS score 0-3)and poor(mRS score 4-6).5.The association between the combined indicators and unfavorable outcome was investigated using multivariable logistic regression models,The predictive value of HD and uHG in predicting 90 day prognosis of sICH was compared.6.Three variables,GCS score,HD and uHQ were included to draw the nomogram with R software,and verify the discrimination and calibration of the nomogram internally.According to the weight assigned to each index in nomogram,a new short-term prognosis prediction score model of sICH was constructed——GHU score,and its predictive value is compared with the widely used ICH Score in clinic.Results1.Compared with patients with good prognosis,patients with poor outcomes were more likely to have HD and higher uHG in univariate analysis.2.In multivariate logistic regression analysis,compared with uHG(L)HD(-),the risk of unfavorable outcomes was increased in uHG(H)HD(-)(odds ratio[OR]=2.957,P=0.044)and uHG(L)HD(+)(OR=1.924;P=0.232),while that in the uHG(H)HD(+)group was the highest(OR=5.710;P<0.001).3.The proportions of unfavorable prognoses were 32.6%in uHG(L)HD(-)48.3%in uHG(L)HD(+),2.2%in uHG(H)HD(-),and 87.5%in uHG(H)HD(+)(P<0.001).4.We compared the predictive value of both HD and uHG to predict the 90-day prognosis of sICH alone and in different combinations.It is found that when uHG(H)and HD(+)coexist,the specificity and positive predictive value of the prognosis are the highest,while when the presence of any one or both of HD(+)and uHG(H),the sensitivity and negative predictive value of the prognosis are the highest.5.Our nomogram has good discrimination and calibration after internal verification,and the accuracy of GHU score(AUC=0.861)for short-term prognosis of sICH is higher than that of ICH Score(AUC=0.788).(P=0.008)ConclusionThe combination of uHG and HD improves the stratification of unfavorable prognoses in patients with sICH.