Correction: Multimodal sequencing identifies synergistic mechanisms driving resistance to neoadjuvant nivolumab treatment in hepatocellular carcinoma Authors: Fanhong Zeng, Jiaxin Guo, Zheng Zeng, Vanilla Xin Zhang, Yu-Man Tsui, Po-Man Li, Abdullah Husain, Qingyang Zhang, Joyce Man-Fong Lee, Tan-To Cheung, Daniel Wai-Hung Ho, Irene Oi-Lin Ng Published: 2026 Correction: Mol Cancer 25, 190 (2026) https://doi.org/10.1186/s12943-026-02682-x Following publication of the original article [1], the authors reported that during their review, they found that the supplementary figures and legends are missing from the online version, and that several figure legends, one cell-line designation in Table 3, and four abbreviation entries require correction. The requested changes are listed below. Figure legends The published wording for the following figure panels should be corrected with the corrected wording shown in the third column.PanelPublished textCorrected textFig. 3FBox plot shows the MAPK pathway score statistical comparison between the R and NR groups in the G1 area. R: responders; NR: Non-responders.Violin/box plots compare MAPK pathway scores (left) and lipogenesis scores (right) between the NR and R groups in the G1 area.Fig. 3GBox plot shows the lipogenesis and MAPK pathway score statistical comparison between the R and NR groups among intra-spot and para-spot. R: responders; NR: Non-responders.Correlation between lipogenesis and MAPK pathway scores in the G1 area.Fig. 7KCorrelation between expression of RAMP1 (lipo hotspot) and targets (lipo hotspot) in bulk-seq.Correlation between expression of RAMP1 (expressed in lipo hotspot) and targets (expressed in lipo hotspot) in bulk-seq.Fig. 8ASpatial representation shows the expression of ligand (ADM, M2 hotspot), ligand (RAMP1, lipo hotspot) and target (EBP, lipo hotspot).Spatial representation shows the expression of ligand (ADM, M2 hotspot), receptor (RAMP1, lipo hotspot), and target (EBP, lipo hotspot). PanelPublished textCorrected textFig. 3FBox plot shows the MAPK pathway score statistical comparison between the R and NR groups in the G1 area. R: responders; NR: Non-responders.Violin/box plots compare MAPK pathway scores (left) and lipogenesis scores (right) between the NR and R groups in the G1 area.Fig. 3GBox plot shows the lipogenesis and MAPK pathway score statistical comparison between the R and NR groups among intra-spot and para-spot. R: responders; NR: Non-responders.Correlation between lipogenesis and MAPK pathway scores in the G1 area.Fig. 7KCorrelation between expression of RAMP1 (lipo hotspot) and targets (lipo hotspot) in bulk-seq.Correlation between expression of RAMP1 (expressed in lipo hotspot) and targets (expressed in lipo hotspot) in bulk-seq.Fig. 8ASpatial representation shows the expression of ligand (ADM, M2 hotspot), ligand (RAMP1, lipo hotspot) and target (EBP, lipo hotspot).Spatial representation shows the expression of ligand (ADM, M2 hotspot), receptor (RAMP1, lipo hotspot), and target (EBP, lipo hotspot). Table 3 cell-line designation In Table 3, the cell-line designation MHCC97-H should be corrected to MHCC97-L. Abbreviations The published definitions should be replaced with the corrected definitions shown in the second column.Published textCorrected textPD-1: Cell death-1 receptorPD-1: Programmed cell death 1PD-L1: Or its ligandPD-L1: Programmed cell death 1 ligand 1(Prolif-TAMs: Proliferating TAMsProlif-TAMs: Proliferating TAMsReg-TAM: Immune regulatory TAMsReg-TAMs: Immune regulatory TAMs Published textCorrected textPD-1: Cell death-1 receptorPD-1: Programmed cell death 1PD-L1: Or its ligandPD-L1: Programmed cell death 1 ligand 1(Prolif-TAMs: Proliferating TAMsProlif-TAMs: Proliferating TAMsReg-TAM: Immune regulatory TAMsReg-TAMs: Immune regulatory TAMs Supplementary Information Supplementary Material 1: Fig. S1. BANKSY clustering results and NMF factor distribution. A. The spatial map shows the distribution of the cluster of ST-seq samples. B. Heatmap of immune function analysis of NIS region. Nav: Navie T cell; CTL: Cytotoxic T cell; TEX: Exhausted T cell; M1_like: M1_like TAM; M2_like: M2_like TAM. C. The box plot compares the lipogenesis score in the tumor niche between the R and NR groups in ST-seq. R: responders; NR: Non-responders. D. The line plot shows the process of choosing the best number of factors for non-negative matrix factorization (NMF). E. The spots with the highest activity (Top 1%) for each factor will be selected. The stacked bar chart shows the proportion of each factor in the top 1% activity spots in each sample. R: responders; NR: Non-responders. Fig. S2. Cellular deconvolution and NMF composition of the tumor niche. A. The spatial map shows the characterization of ST-seq data by cell-type deconvolution. CD4: CD4 T cell; CD8: CD8 T cell; TAM: tumor-associated macrophage; Endo: Endothelial cells; Fib: Fibroblast. B. The lollipop chart shows the differences in the proportion across groups of the top 1% spots for each NMF factor. NMF: non-negative matrix factorization. C. The heatmap shows expression profiles of the top marker genes for selected NMF factor. NMF: non-negative matrix factorization. D. The spatial map shows distribution of tumor niche (upper panel) and the distance of each spot from the border of the tumor niche (bottom panel), where positive values ​​represent the radial distance outward from the tumor niche and negative values ​​represent the radial distance toward the center of the tumor niche. E. Dot plot showing pathway enrichment comparison of stromal spots across groups in tumor niche. F. Line plot showing lipogenesis and immune score changes from tumor core to outside (NR vs R). CTL: Cytotoxic score; TEX: Exhausted score; M1: M1_like score; M2: M2_like score; Pro: pro-inflammatory; Anti: Anti-inflammatory. G. Line plot showing immune activation function scores changes from tumor core to outside (NR vs R). CTL: Cytotoxic score; TEX: Exhausted score; M1: M1_like score; M2: M2_like score; Pro: pro-inflammatory; Anti: Anti-inflammatory. Fig. S3. Spatial distance and cellular spatial colocalization analysis. A. The network diagram shows intra-spot and para-spot cellular colocalization in the G1 area.CD4: CD4 T cell; CD8: CD8 T cell; TAM: tumor-associated macrophage; Endo: Endothelia; Fib: Fibroblast. B. Spatial distribution of lipo hotspots across samples. C. The line plot shows the spatial distribution differences between lipo hotspots and non-lipo hotspots across groups in the G1 area. D. Bar plot showing pathway enrichment of cell neighborhood of lipo hotspots across groups. R: responders; NR: Non-responders. Fig. S4. Spatial distribution of hotspots across samples (in-house cohort). A. Spatial distribution of M2 hotspot across samples. B. Spatial distribution of TEX hotspot across samples. C. Pie chart shows the proportion of M2 (upper panel) and TEX (bottom panel) hotspots between the R and NR groups, respectively. R: responders; NR: Non-responders. Fig. S5. Spatial distribution of lipo hotspots (A), M2 hotspots (B) and TEX hotspots (C) across samples (validation cohort). Fig. S6. Patients in the NR group exhibited an immunosuppressive microenvironment(T cells). A. The heatmap shows the landscape of the copy number variation in scRNA-seq samples. B. The heatmap shows the enriched lipid-related pathway in the NR group. NR: Non-responders. C. Bar plot showing pathway enrichment comparison of HCC between NR and R groups. R: responders; NR: Non-responders. D. Pseudotime analysis shows pseudo-time progression in tumor cells. E. The t-SNE plot shows the T cell subtype distribution. F. The box plot compares the T cell function between the R and NR groups in scRNA-seq. R: responders; NR: Non-responders; * p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. G. The dot plot shows the immune checkpoint expression of T cell between the R and NR groups in scRNA-seq. R: responders; NR: Non-responders; * p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. H. The line plot shows the correlation of T cell between pseudotime and exhaustion score in NR group. I. The line plot shows the correlation between pseudotime and immune checkpoint in T cell. J. Bar plot showing pathway enrichment comparison of Treg between NR and R groups. R: responders; NR: Non-responders. K. The line plot shows Treg's GSEA results in the NR v.s. R. GSEA: Gene set enrichment analysis. Fig. S7. Patients in the NR group exhibited an immunosuppressive microenvironment (TAMs). A. The dot plot shows myeloid subtype marker expression in different clusters. B. Bar plot showing pathway enrichment comparison of TAM between NR and R groups. R: responders; NR: Non-responders. C. The line plot shows the GSEA results of TAM in the NR v.s. R. GSEA: Gene set enrichment analysis. D. Pseudotime's directional diagram (left panel) and correlation between pseudotime and LA-TAM score (right panel). E. The network diagram shows the strength of cellular communication among different groups. F. The network shows the top 50 HCC ligand-receptor pairs with the most remarkable differences in communication strength between groups (NR vs R). R: responders; NR: Non-responders. G. Spatial representation shows the co-occurrence of ligand-receptor (LR, ADM-RAMP1, M2 hotspot – lipo hotspot), ligand-target (LT, ADM-EBP, M2 hotspot – lipo hotspot) and receptor-target (RT, RAMP1-EBP, lipo hotspot). Fig. S8. Results of the inflammatory cytokine array in supernatant between the RAMP1 knockdown cells and control cells in direct co-culture experiments. (THP1 cells: Figure S8A; CD8T cells: Figure S8B). Supplementary Material 1: Fig. S1. BANKSY clustering results and NMF factor distribution. A. The spatial map shows the distribution of the cluster of ST-seq samples. B. Heatmap of immune function analysis of NIS region. Nav: Navie T cell; CTL: Cytotoxic T cell; TEX: Exhausted T cell; M1_like: M1_like TAM; M2_like: M2_like TAM. C. The box plot compares the lipogenesis score in the tumor niche between the R and NR groups in ST-seq. R: responders; NR: Non-responders. D. The line plot shows the process of choosing the best number of factors for non-negative matrix factorization (NMF). E. The spots with the highest activity (Top 1%) for each factor will be selected. The stacked bar chart shows the proportion of each factor in the top 1% activity spots in each sample. R: responders; NR: Non-responders. Fig. S2. Cellular deconvolution and NMF composition of the tumor niche. A. The spatial map shows the characterization of ST-seq data by cell-type deconvolution. CD4: CD4 T cell; CD8: CD8 T cell; TAM: tumor-associated macrophage; Endo: Endothelial cells; Fib: Fibroblast. B. The lollipop chart shows the differences in the proportion across groups of the top 1% spots for each NMF factor. NMF: non-negative matrix factorization. C. The heatmap shows expression profiles of the top marker genes for selected NMF factor. NMF: non-negative matrix factorization. D. The spatial map shows distribution of tumor niche (upper panel) and the distance of each spot from the border of the tumor niche (bottom panel), where positive values ​​represent the radial distance outward from the tumor niche and negative values ​​represent the radial distance toward the center of the tumor niche. E. Dot plot showing pathway enrichment comparison of stromal spots across groups in tumor niche. F. Line plot showing lipogenesis and immune score changes from tumor core to outside (NR vs R). CTL: Cytotoxic score; TEX: Exhausted score; M1: M1_like score; M2: M2_like score; Pro: pro-inflammatory; Anti: Anti-inflammatory. G. Line plot showing immune activation function scores changes from tumor core to outside (NR vs R). CTL: Cytotoxic score; TEX: Exhausted score; M1: M1_like score; M2: M2_like score; Pro: pro-inflammatory; Anti: Anti-inflammatory. Fig. S3. Spatial distance and cellular spatial colocalization analysis. A. The network diagram shows intra-spot and para-spot cellular colocalization in the G1 area.CD4: CD4 T cell; CD8: CD8 T cell; TAM: tumor-associated macrophage; Endo: Endothelia; Fib: Fibroblast. B. Spatial distribution of lipo hotspots across samples. C. The line plot shows the spatial distribution differences between lipo hotspots and non-lipo hotspots across groups in the G1 area. D. Bar plot showing pathway enrichment of cell neighborhood of lipo hotspots across groups. R: responders; NR: Non-responders. Fig. S4. Spatial distribution of hotspots across samples (in-house cohort). A. Spatial distribution of M2 hotspot across samples. B. Spatial distribution of TEX hotspot across samples. C. Pie chart shows the proportion of M2 (upper panel) and TEX (bottom panel) hotspots between the R and NR groups, respectively. R: responders; NR: Non-responders. Fig. S5. Spatial distribution of lipo hotspots (A), M2 hotspots (B) and TEX hotspots (C) across samples (validation cohort). Fig. S6. Patients in the NR group exhibited an immunosuppressive microenvironment(T cells). A. The heatmap shows the landscape of the copy number variation in scRNA-seq samples. B. The heatmap shows the enriched lipid-related pathway in the NR group. NR: Non-responders. C. Bar plot showing pathway enrichment comparison of HCC between NR and R groups. R: responders; NR: Non-responders. D. Pseudotime analysis shows pseudo-time progression in tumor cells. E. The t-SNE plot shows the T cell subtype distribution. F. The box plot compares the T cell function between the R and NR groups in scRNA-seq. R: responders; NR: Non-responders; * p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. G. The dot plot shows the immune checkpoint expression of T cell between the R and NR groups in scRNA-seq. R: responders; NR: Non-responders; * p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. H. The line plot shows the correlation of T cell between pseudotime and exhaustion score in NR group. I. The line plot shows the correlation between pseudotime and immune checkpoint in T cell. J. Bar plot showing pathway enrichment comparison of Treg between NR and R groups. R: responders; NR: Non-responders. K. The line plot shows Treg's GSEA results in the NR v.s. R. GSEA: Gene set enrichment analysis. Fig. S7. Patients in the NR group exhibited an immunosuppressive microenvironment (TAMs). A. The dot plot shows myeloid subtype marker expression in different clusters. B. Bar plot showing pathway enrichment comparison of TAM between NR and R groups. R: responders; NR: Non-responders. C. The line plot shows the GSEA results of TAM in the NR v.s. R. GSEA: Gene set enrichment analysis. D. Pseudotime's directional diagram (left panel) and correlation between pseudotime and LA-TAM score (right panel). E. The network diagram shows the strength of cellular communication among different groups. F. The network shows the top 50 HCC ligand-receptor pairs with the most remarkable differences in communication strength between groups (NR vs R). R: responders; NR: Non-responders. G. Spatial representation shows the co-occurrence of ligand-receptor (LR, ADM-RAMP1, M2 hotspot – lipo hotspot), ligand-target (LT, ADM-EBP, M2 hotspot – lipo hotspot) and receptor-target (RT, RAMP1-EBP, lipo hotspot). Fig. S8. Results of the inflammatory cytokine array in supernatant between the RAMP1 knockdown cells and control cells in direct co-culture experiments. (THP1 cells: Figure S8A; CD8T cells: Figure S8B).