Fig. 6

Detailed characterization of monocytes/macrophages in HPSCC. a UMAP plot of monocyte/macrophage cells colored by cell type. b Frequency (left) and proportion (right) of four major mononuclear/macrophage cell types in tumor and normal tissue samples. c Heatmap showing signature DEGs between mononuclear/macrophage cell types. d Bubble heatmap showing marker genes across mononuclear/macrophage cell types. Dot size indicates fraction of expressing cells, colored according to expression normalized to z-score. e Dot plot of representative M1, M2, angiogenic, and phagocytic signatures in monocyte/macrophage clusters [Z-score normalized log₂ (count + 1)]. f Differential pathways enriched in C1QC⁺ and SPP1⁺ TAMs according to GSVA. Two-sided unpaired limma-moderated t-test. g Absolute infiltration proportion of SPP1⁺ TAMs compared between normal (n = 43) and tumor (n = 43) tissues in the TCGA-HNSC cohort. h Kaplan–Meier curve of OS in the TCGA-HNSC cohort stratified by optimal cut-off point for SPP1 expression and SPP1⁺ TAM infiltration. i Pseudotime trajectory analysis of mononuclear/macrophage cells. Each dot represents one cell, colored according to its cluster label. Inlet plot showed each cell with a pseudotime score from dark blue (early state) to light blue (terminal). Jitter plot showing expression changes in macrophage differentiation-associated genes over pseudotime. j Correlation of mCAF signature with SPP1⁺ TAMs based on TCGA-HNSC data. Each dot represents a patient (Pearson’s correlations). k Kaplan–Meier OS analyses of four subgroups in the TCGA-HNSC cohort, stratified by infiltration of both mCAFs and SPP1⁺ TAMs. l Dot plot of predicted ligand–receptor interactions between mCAFs and SPP1⁺ TAMs in tumor samples. m Representative images showing mIHC staining of panCK, FAP, and SPP1 in HPSCC tumor samples, in individual and merged channels. Scale bar = 20 μm. Significance in (h) and (k) was determined with two-sided log-rank tests