Prediction of Basal Ganglia Hematoma Expansion Based on Radiomics and Clinical Characteristics: A Novel Multivariate Predictive Nomogram

Figure 3

Constructed nomogram for predicting HE. (a) The nomogram is derived from multivariate analysis. (b) The AUCs of m-CTD in the training cohort and validation cohort were 0.776 and 0.730, respectively. (c, f) The AUCs for the training cohort and validation cohort were 0.874 and 0.883, respectively. (d, g) Calibration curves for the training and validation cohorts. (e, h) DAC curve analyses for both cohorts show the nomogram’s optimal predictive efficacy at risk thresholds approximating 0.4. AUC = area under the curve; DAC = decision analysis curve; ROC = receiver operating characteristic; m-CTD = mean CT densities; BHV = baseline hemorrhage volume; BS = blend sign; BHS = basal ganglia hemorrhage; SBP = systolic blood pressure; Hcy = homocysteine.