Fig. 3

Validation and risk stratification of BPD prognostic model. A-B Risk score distribution plot for children with BPD (Figure A uses continuous variables for risk assessment, while Figure B converts continuous variables into categorical variables based on the optimal intercept value for risk assessment). Patients are arranged in ascending order of risk scores from left to right (a). Survival status of each patient (b). The y-axis represents overall survival time. Color code: blue represents alive cases, and red represents deceased cases. Heatmap of expression levels for selected variables (c). C (a-c)  Calibration curves of the 1-, 5-, and 7-year overall survival for children with BPD. The light blue line represents the ideal reference line, where the predicted probabilities match the observed survival probability. The red dots, calculated through bootstrapping (sample: 1000), represent the performance of the prognostic model. The closer the solid red line is to the light blue line, the more accurate the model's predicted survival probability. D (a-c)  Decision curve analysis for the BPD prognostic model. The y-axis represents net profit, which is calculated by summing up the gains (true positives) and subtracting the harms (false positives). The x-axis represents threshold probability.  E ROC curve of the 1-, 5-, and 7-year survival prediction for BPD prognostic model. F Kaplan–Meier overall survival curves for the low-risk, medium-risk and high-risk BPD patients stratified by the BPD prognostic model. Abbreviations: A, adrenaline; ALT, alanine transaminase; AUC, area under the curve; BPD, bronchopulmonary dysplasia; NA, noradrenaline; OS, overall survival; PH, pulmonary hypertension; RS, respiratory support; ROC, receiver operating characteristic curve