Fig. 3: Pathway-based super signature for pre-treatment samples.

a–c The heatmap of ssGSEA values of pre-treatment responders (R) and nonresponders (NR) in the Gide et al. cohort (a), Lee et al. cohort (b), and MGH cohort (c). Nonresponders are presented on the left side, and responders are presented on the right side. FDR-corrected two-sided Welch t-test was conducted to compare ssGSEA values between R and NR samples, only FDR < 0.05 showed here. d–f The boxplot of pre-treatment samples’ PASS-PRE signature scores in the Gide et al. cohort (d), Lee et al. cohort (e), and MGH cohort (f).The pre-treatment samples were separated into responders (R) and none-responders (NR) with number of samples showed in each cohorts. The P-values were computed via a one-sided rank-sum test. Boxplot center lines indicate medians, box edges represent the interquartile range, whiskers extend to the minimum and maximum, and the outliers are plotted individually using the ‘*’ symbol. g ROC and AUC of PASS-PRE on pre-treatment samples from Gide et al., Lee et al. and MGH cohorts. h ROC and AUC on the all pre-treatment test samples, which combined with Gide et al., Lee et al. and MGH cohorts. i–k The Kaplan–Meier analysis of progression-free survival (PFS) of pre-treatment samples in the combined test pre-treatment samples (i), Gide et al. cohort (j), and MGH cohort (k). The two-sided log-rank test compared high and low subgroups based on the mean of pre-treatment samples odd ratio as cutoff. Hazard ratio (HR) was calculated and shown with confidence interval (CI). Source data are provided as a Source Data file.