3.3 ROC curve analysis
We next constructed ROC curves using RDW, LVMI, and a combination of
LVMI and RDW to predict HCM patient outcomes (Fig. 1). The
cut-off value of RDW for the all -cause mortality and MACE was 0.13, and
the AUC was 0.753 and 0.746 respectively. For LVMI, the AUC for
predicting all-cause mortality and MACE incidence was 0.855 and 0.838
respectively. When RDW and LVMI were evaluated together, the AUC value
was 0.890 (95% CI 0.849-0.923, P <0.01) in predicting
the occurrence of all-cause mortality, and was 0.885 (95% CI
0.844-0.919, P <0.01) in predicting the occurrence of
MACE. As such, the combined evaluation of RDW and LVMI was superior to
either of these variables alone when predicting the risk of all-cause
mortality and MACE in patients with HCM. (Table 3).