Skip to main content
. 2021 Dec 15;11(12):2366. doi: 10.3390/diagnostics11122366

Figure 3.

Figure 3

ROC analysis of ETA receptor expression score for detecting AMR status. The most appropriate cut-off for the test of the chosen binary classifier is between 2 and 3 of the summarized ETA receptor expression score at the highest true positive rate, together with the lowest false positive rate, which amount to 72% and 80%, respectively (see Figure 2). Based on the calculated AUC, it can be established that the accuracy of the classification of positive AMR using the level of summarized ETA receptor expression score reached 80% (since the lower limit of the AUC 95% confidence interval is greater than 50%, this result should be considered statistically significant, i.e., p > 0.05). Then, it allows to state that 80% of patients would be correctly diagnosed taking only the level of the summarized ETA receptor expression score into account.