Predictive value of computed tomography texture analysis for early basal ganglia hemorrhage expansion
LI Cheng
LI Xiaojun
LI Xin
WEI Lijie
GUO Kai
Abstract：Objective To investigate the discriminatory value of texture analysis of CT for early basal ganglia hemorrhage en-largement.Methods A total of 104 hypertensive basal ganglia hemorrhage patients were selected from the basal ganglia hemor-rhage region using Mazda texture analysis software.Misclassified rate(MCR)was calculated and ROC(Region of curve)curve analysis was performed for texture parameters to judge texture parameters with diagnostic efficacy.Results The best texture pa-rameters were selected by three different dimensionality reduction methods including Fisher's coefficient,classification error probability combined average correlation coefficient(POE+ACC),and mutual information(MI).The MCR calculated by non-linear discriminant analysis(NDA)was 0%.Among the texture parameters,S(4,-4)Difvarnc,S(3,-3)Difvarnc,S(2,-2)Difvarnc,GrVariance,S(1,-1)SumAverg,S(5,0)Difvarnc,Perc.90%,S(0,1)Difvarnc had better diagnostic efficacy for tex-ture parameters.This indicated that the texture parameters of gray level histograms and gray level co-occurrence matrices had greater discriminatory significance for basal ganglia hemorrhage.Conclusion CT texture analysis has discriminatory value for early basal ganglia hemorrhage enlargement.The lowest MCR is calculated by the analytical method of NDA,in which the dis-criminative value of texture parameters under gray level histograms and gray level co-occurrence matrices is greater.
Keywords：Basal ganglia hemorrhageTesture analysisRadiomicsPrediction modelTomographyX-ray computed
Publication Date：2025-07-30
Online Publishing Date：2025-08-26(First online date of this platform, not the publication date of the document)
Pages：5( 5-9 )