Fig. 1

Graphical flowchart"

Fig. 2

PCA of neoadjuvant and adjuvant cohorts"

Fig. 3

Immune deficiency in trastuzumab-resistant patients A, B: Volcano plot of RNA-seq gene expression changes in the trastuzumab-resistant neoadjuvant and adjuvant cohorts. C, D: KEGG pathway analysis in the trastuzumab-resistant neoadjuvant and adjuvant cohorts. E, F: Top 10 candidate GO terms and pathways in the functional annotation of GSEA of the trastuzumab-resistant neoadjuvant and adjuvant cohorts. G: Representative pathways of GSEA in the trastuzumab-resistant group (P-adjust<0.05) of neoadjuvant and adjuvant cohorts.."

Fig. 4

TIME analysis of neoadjuvant and adjuvant cohorts Significant differences between the two subgroups were assessed using the Wilcoxon test (NS: Not significant; *: P<0.05; ***: P<0.001)."

Fig. 5

Trastuzumab-response related hub genes A: Correlation coefficient heatmap to demonstrate the immune characteristics of the hub genes. B: Kaplan-Meier analysis of the DDFS curves for the four genes (of the 12 hub genes) with significant prognostic value (log-rank test, P<0.05). C: Univariate COX regression analysis of the 12 hub genes. D: RTFQ-PCR validation of the four prognostic genes in breast cancer cell lines. E. Expression analysis of the four prognostic genes in TCGA cohort. Significant differences between the two subgroups were assessed using the Wilcoxon test (NS: Not significant; *: P<0.05; **: P<0.01)."

Fig. 6

TRGPI construction and evaluation Multivariate COX regression analysis of clinicopathological factors and TRGPI in the training set (A) and test set (B). C: Kaplan-Meier survival analysis of the TRGPI subgroups in the training set (left) and test set (right) (log-rank test, P<0.05). D: ROC curve analysis of the prognostic value of TRGPI for DDFS at 18 months and 36 months in the training set (left) and the test set (right). E: The TRGPI score curve showing the distribution of patients under trastuzumab therapy in the test set. Significant differences between the two subgroups were assessed using the Wilcoxon test (*: P<0.05; ***: P<0.001)."

Fig. 7

Genomic features in TRGPI subgroups A: Waterfall plots of 30 highly variant mutant genes demonstrated the mutation landscape in the TRGPI-low group (left) and TRGPI-high group (right). B: Distribution of the top 10 somatic mutations VAF in the TRGPI-low group (upper) and TRGPI-high group (lower). C: The mutation co-occurrence and exclusion analyses in TRGPI-low group (upper left) and TRGPI-high group (lower right). NS: No significance. D: Differences in TMB among different TRGPI subgroups. Wilcoxon test was used to compare the statistical difference (NS: Not significant)."

Fig. 8

TIME characteristics in different TRGPI subgroups of the training set A: The proportions of TIME cells in different TRGPI subgroups. Significant differences between the two subgroups were assessed using the Wilcoxon test. B: The TRGPI grouping and proportions of TIME cells for 91 patients. Age, ER status, PR status, metastasis, DDFS, and TRGPI were used as patient annotations. NS: Not significant; *: P<0.05; ***: P<0.001."

Fig. 9

TIME characteristics of TRGPI subgroups in neoadjuvant and adjuvant cohorts A: The percentage of immune cells; B:The expression levels of PD-1, CTLA-4, IL-2, and IL-21 from different TRGPI subgroups were compared in the neoadjuvant and adjuvant cohorts. Significant differences between the two subgroups were assessed using the Wilcoxon test. C: Immune subtype analysis of TRGPI subgroups in the neoadjuvant and adjuvant cohorts. NS: Not significant; *: P<0.05; **: P<0.01; ***: P<0.001; ****:P<0.000 1.."

Fig. 10

The TRGPI-low subgroups benefit more from ICI A: MHC-I scores; B: TIDE, MSI, T cell exclusion and dysfunction scores between the two TRGPI subgroups were compared using the Wilcoxon test. NS: Not significant. *: P<0.05; ***: P<0.001; ****: P<0.000 1."