Skip to main content
. 2023 May 29;14:1126217. doi: 10.3389/fimmu.2023.1126217

Figure 4.

Figure 4

Diagnostic performance of the biomarker panel in validation cohort I and results compared to those for hemoglobin. Validation cohort I (n = 343) was randomly divided into a training set (n = 241) and a testing set (n = 102). The biomarker panel included CAT, LTF, MMP9, RBP4, and SERPINA3. (A) Receiver operating characteristic (ROC) curve analysis of the biomarker panel and hemoglobin in the training set. The area under the curve (AUC) was compared using the DeLong test. (B) Calibration curve analysis of the biomarker panel and hemoglobin in the training set. The closer the curve is to the reference line (black), the more accurate the prediction. (C) Decision curve analysis (DCA) for the biomarker panel and hemoglobin in the training set. The larger the area under the curve is, the greater the clinical benefit that can be expected. (D) ROC curves, (E) calibration curve, (F) DCA curve for the biomarker panel and hemoglobin in the testing set.