Fig. 5: Weighted combinations of plasma ratio of metabolites and micro-RNAs (miRNAs) improve the performance of the proteins-based cavernous angioma with symptomatic hemorrhage (CASH) biomarker. | Communications Medicine

Fig. 5: Weighted combinations of plasma ratio of metabolites and micro-RNAs (miRNAs) improve the performance of the proteins-based cavernous angioma with symptomatic hemorrhage (CASH) biomarker.

From: Plasma metabolites with mechanistic and clinical links to the neurovascular disease cavernous angioma

Fig. 5

The canonical values estimated with the optimal integrative weighted model were (a) higher (unpaired Student’s t test, p < 0.0001) in the CASH (n = 20 biologically independent patient samples) compared with cavernous angioma without symptomatic hemorrhage (non-CASH) patients (n = 20 biologically independent patient samples) included in the independent propensity matched validation cohort. b The receiver operating characteristic (ROC) analyses showed that this integrative model differentiated CASH patients from non-CASH patients with 80% sensitivity and 95% specificity (area under the curve [AUC] = 90.25%). c An alternative optimal integrative weighted model developed using a parallel 10-fold cross-validation machine learning approach showed higher canonical values higher (unpaired Student’s t test, p < 0.0001) in CASH compared with non-CASH patients. d The ROC analysis showed that the alternative optimal integrative model differentiated CASH from non-CASH with 85% sensitivity and 80% specificity (AUC = 90.75%). CI, confidence interval; ***p < 0.001. Error bars represent standard error of the mean.

Back to article page