Artificial Intelligence Meets Whole Slide Images: Deep Learning Model Shapes an Immune-Hot Tumor and Guides Precision Therapy in Bladder Cancer

Figure 5

AI score shapes an inflamed TME in BLCA. (a) Differences in the expression of 122 immunomodulators (chemokines, receptors, MHC, and immunostimulators) between AI clusters in BLCA. (b) Differences in the various steps of the cancer immunity cycle between AI clusters. (c) Correlation between AI score and the infiltration levels of five types of TIICs (CD8+ T cells, NK cells, macrophages, Th1 cells, and dendritic cells), which were calculated using six independent algorithms. (d, e) The differences in infiltration levels of TIICs between AI clusters in the TIMER and MCP-counter algorithms. The asterisks indicate a statistically significant p value calculated using the Mann–Whitney or -test (, , and ). (f) The differences in the expression of effector genes of several anticancer TIICs (including CD8+ T cells, NK cells, macrophages, Th1 cells, and dendritic cells) between AI subtypes.

(a)

(b)

(c)

(d)

(e)

(f)