Fig. 5

Related biomarkers of lung pathological recovery in severe COVID-19 patients in the MSC group.A Nomogram of five protein combinations to predict the percentage decline rate of consolidation ≤  − 51.99% after MSC therapy (28-day good outcome). To use the nomogram, the protein expression level of an individual value is located on each variable axis, and a line is drawn upward to determine the number of points received for each variable protein expression level. The sum of these numbers is located on the total points axis. A line is drawn downward to the 28-day good outcome probability axes to determine the probability of the percentage decline rate of consolidation ≤  − 51.99% after MSC therapy. B Calibration plot and confusion matrix to assess the accuracy of the model in 58 patients. C The receiver operating characteristic (ROC) curve and confusion matrix to distinguish responders from non-responders outcomes in the training set. The ROC curves were created by plotting the sensitivity (i.e., true-positive rate) against 1 − specificity (i.e., false-positive rate). The line in each plot represents the area under the curve (AUC). D ROC curve and confusion matrix to distinguish responders from non-responders outcomes after MSC therapy in the validation set-1