Fig. 4: Associations between AI-predicted aggressiveness and survival outcomes.
From: Artificial intelligence links CT images to pathologic features and survival outcomes of renal masses
A Kaplan–Meier analysis of disease-specific survival (DSS) in the internal test set. B Kaplan–Meier analysis of recurrence-free survival (RFS) in the internal test set. C Kaplan–Meier analysis of overall survival (OS) in the internal test set. D Kaplan–Meier analysis of DSS in the external test set. E Kaplan–Meier analysis of RFS in the external test set. F Kaplan–Meier analysis of OS in the external test set. G Forest plot of AI-aggressiveness score cox regression analyses performed in different subgroups in the external test set. The centers for the error bars represent the hazard ratios. H C-indexes of AI-aggressiveness score, radiomics model, the nephrometry nomogram, TNM stage, and ISUP grade in predicting DSS, RFS, and OS in the internal and external test set. The error bands in the survival curves in Fig. 4A–F represent the 95% confidential intervals. P values were two-sided and not adjusted for multiple comparisons. Source data for survival analyses are provided as a Source Data file.
