Skip to main content
NCBI home page
JHEP Reports logoLink to JHEP Reports
. 2023 Apr 15;5(8):100762. doi: 10.1016/j.jhepr.2023.100762

Spatial immunophenotypes predict clinical outcome in intrahepatic cholangiocarcinoma

Chunbin Zhu 1,†, Jiaqiang Ma 1,2,†, Kai Zhu 1,†, Lei Yu 1, Bohao Zheng 3, Dongning Rao 1,2, Shu Zhang 1, Liangqing Dong 1, Qiang Gao 4,5,6,7,∗, Xiaoming Zhang 2,∗, Diyang Xie 8,∗
PMCID: PMC10285646  PMID: 37360908

Abstract

Background & Aims

Intrahepatic cholangiocarcinoma (iCCA) is a severe malignant tumour that shows only modest responses to immunotherapy. We aimed to identify the spatial immunophenotypes of iCCA and delineate potential immune escape mechanisms.

Method

Multiplex immunohistochemistry (mIHC) was performed to quantitatively evaluate the distribution of 16 immune cell subsets in intratumour, invasive margin and peritumour areas in a cohort of 192 treatment-naïve patients with iCCA. Multiregion unsupervised clustering was used to determine three spatial immunophenotypes, and multiomics analyses were carried out to explore functional differences.

Results: iCCA displayed a region-specific distribution of immune cell subsets with abundant CD15+ neutrophil infiltration in intratumour areas. Three spatial immunophenotypes encompassing inflamed (35%), excluded (35%) and ignored (30%) phenotypes were identified. The inflamed phenotype showed characteristics of abundant immune cell infiltration in intratumour areas, increased PD-L1 expression and relatively favourable overall survival. The excluded phenotype with a moderate prognosis was characterized by immune cell infiltration restricted to the invasive margin or peritumour areas and upregulation of activated hepatic stellate cells, extracellular matrix and Notch signalling pathways. The ignored phenotype, with scarce immune cell infiltration across all subregions, was associated with MAPK signalling pathway elevation and a poor prognosis. The excluded and ignored phenotypes, constituting non-inflamed phenotypes, shared features of an increased angiogenesis score, TGF-β and Wnt-β catenin pathway upregulation and were enriched for BAP1 mutations and FGFR2 fusions.

Conclusion

We identified three spatial immunophenotypes with different overall prognoses in iCCA. Tailored therapies based on the distinct immune evasion mechanisms of the spatial immunophenotypes are needed.

Impact and implications

The contribution of immune cell infiltration in the invasive margin and peritumour areas has been proved. We explored the multiregional immune contexture of 192 patients to identify three spatial immunophenotypes in intrahepatic cholangiocarcinoma (iCCA). By integrating genomic and transcriptomic data, phenotype-specific biological behaviours and potential immune escape mechanisms were analysed. Our findings provide a rationale to develop personalized therapies for iCCA.

Keywords: Multiplex immunohistochemistry, Spatial immunophenotype, Immune evasion, Tumour microenvironment, Intrahepatic cholangiocarcinoma

Graphical abstract

graphic file with name ga1.jpg

Open in a new tab

Highlights

  • •

    Spatial immunophenotypes in iCCA can stratify patients into three subgroups with distinct clinical outcomes.

  • •

    Immune escape mechanisms of different spatial immunophenotypes in iCCA are explored.

  • •

    Association between fibrosis grades and immune cell infiltrates in iCCA are identified.

  • •

    PD-L1 expression in iCCA displays a region-specific pattern.

Introduction

Intrahepatic cholangiocarcinoma (iCCA) is the second most common primary liver cancer. Its incidence has been increasing globally for decades.1 Established risk factors for iCCA include primary sclerosing cholangitis, hepatitis virus infection and intrahepatic cholelithiasis.2 Resection remains a major radical treatment. However, iCCA is still associated with a poor prognosis because of high recurrence rates (due to aggressive biological behaviour) and limited systemic treatment options.3

A proportion of iCCAs are infiltrated with immune cells overexpressing coinhibitory receptors, suggestive of the potential efficacy of immune checkpoint inhibitors for patients with iCCAs.4 Nevertheless, current clinical trials have reported a modest overall response rate ranging from 3% to 22% for programmed death 1 (PD-1) inhibitor monotherapy in patients with unresectable cholangiocarcinoma,5 partly owing to substantial inter- and intrapatient heterogeneity of the immune atlas. Therefore, detailed characterization of the immune milieu in iCCA is needed to effectively select patients that might respond to immunotherapy.

Several molecular classifications of iCCA have been proposed,[6], [7], [8], [9], [10] including two tumour microenvironment (TME)-based classifications.9,10 Notably, these classifications were determined by the immune characteristics of intratumour areas. The contribution of immune cell infiltration in the invasive margin (IM) and peritumour areas has been overlooked. According to previous studies, high densities of macrophages in peritumoral liver tissue of hepatocellular carcinoma and increased infiltration of CD3+ lymphocytes in IM areas of colon cancer were found to be associated with poor survival.11,12 The immune score based on the densities of CD3+ T cells and cytotoxic CD8+ T cells in the intratumour and IM areas could optimize the estimation of tumour recurrence and overall prognosis.13

Herein, we explored the immune contexture using multiplex immunohistochemistry (mIHC) across matched distinct subregions, including peritumour, IM and intratumour areas, and identified three spatial immunophenotypes in a cohort of 192 patients with iCCA. By integrating genomic and transcriptomic data, we also analysed phenotype-specific biological behaviours and explored potential immune escape mechanisms, providing clues for tailored therapies for patients with iCCA.

Materials and methods

Clinical samples and TMA construction

A total of 192 patients with iCCA who underwent surgical resection without prior anti-cancer treatment at Zhongshan Hospital of Fudan University from 2012 to 2017 were retrospectively reviewed and enrolled in this study. Detailed clinicopathologic data are shown in Table S1. Formalin-fixed paraffin-embedded tissue samples were obtained to perform serial sections and construct a tissue microarray (TMA) as described previously.14 Tissue cores of 1 mm diameter were collected from peritumour, IM and intratumour areas based on the review of H&E staining. Informed consent was obtained from each patient and the study was approved by the Research Ethics Committee of Zhongshan Hospital.

Multiplex immunohistochemistry and quantitative analysis

mIHC staining was performed according to the manufacturer’s instruction as described previously.15 The primary antibodies are shown in Table S2 and detailed markers used to define the 16 immune cell subsets are shown in Table S3. All pixel intensity information was filed into image cytometry data analysis software, FCS Express 7 v7.10.0007 (De Novo Software) (Fig. S1A,B). Details on mIHC, multispectral imaging and quantitative analysis are provided in the supplementary materials and methods. Considering wide variation in absolute densities of immune cells, comparisons of immune cell distribution were made via ln(densities (cells/mm2) +1).

Multiomics data analysis

The transcriptomic and genomic data of 44 patients with iCCA in our study were derived from the previous CPTAC iCCA cohort.8 Detailed clinicopathologic data of 44 patients with iCCA are shown in Table S4. Molecular Signatures Database (MSigDB, c2.cp.kegg.v7.1.symbols.gmt) was used for enrichment analysis. Single-sample gene set-enrichment analysis was performed to identify pathway alterations among different subgroups. Immune score and microenvironment score were estimated using the xCELL algorithm. The TIDE (tumour immune dysfunction and exclusion) score16 (http://tide.dfci.harvard.edu) was calculated to predict the efficacy of immune checkpoint inhibitors.

Comparison of the spatial immunophenotypes with previously reported classifications of iCCA

Unsupervised clustering was performed as previously defined.[6], [7], [8], [9] The comparisons between our spatial immunophenotypes and four reported classifications, including Andersen’s CCA signatures,6 Sia’s classes,7 Dong’s proteomic subgrouping8 and Job’s immune subtypes,9 were analysed by Fishers’ exact test.

Statistical analyses

Statistical analyses were performed using R software (4.1.2) and SPSS 26 software (IBM, Armonk, NY, USA). Comparisons were performed using Wilcoxon rank-sum test, Wilcoxon matched-pair signed-ranks test, Kruskal-Wallis test or Fishers’ exact test as appropriate. The median values of immune cell densities were set as the cut-off points when their impacts on prognosis were analysed. Clinicopathological factors associated with overall survival (OS) and recurrence-free survival (RFS) were identified using univariate Cox proportional hazards regression models. Significant factors in univariate analyses were further subjected to multivariate Cox regression analyses. Survival curves were plotted via the Kaplan-Meier method and compared using the log-rank test. The Spearman correlation coefficient was applied to assess linear relationships between continuous variables. The pheatmap package was used for hierarchical cluster analyses (distance metric: euclidean distance, linkage method: ward.D2 linkage, scale: row) and heatmap generation. The driving immune cells were analysed in the glmnet package and receiver-operating characteristic (ROC) curves were generated using the pROC package. The clusterProfiler R package and GSVA package were used for KEGG pathway enrichment and single-sample gene set-enrichment analyses, respectively. The number of nearest neighbours was calculated on the assigned coordinates of each cell and carried out through the spatstat package. A two-sided p <0.05 was considered statistically significant.

Results

The distribution and prognostic significance of immune cell subsets

To comprehensively characterize the immune landscape of iCCA, we conducted an optimized sequential mIHC workflow encompassing six, six, five and three distinct markers in every four serial formalin-fixed paraffin-embedded tissue sections to identify 16 immune cell subsets in intratumour, IM and peritumour areas. Representative mIHC staining images are shown in Fig. 1A and Fig. S1C. Interpatient heterogeneity and spatial clustering of immune cells were identified (Fig. 1B). Consistent with previous studies,9,17 intratumour areas displayed scarce infiltration of most immune cell subsets except for CD15+ neutrophils (median, 249.25 vs. 134.75 vs. 152.00/mm2, p <0.001) compared with IM and peritumour areas, respectively (Fig. 1C). CD3+ T cells (median, 382.25 vs. 228.00 vs. 14.50/mm2, p <0.001), CD20+ CD79a+ B cells (median, 183.50 vs. 125.25 vs. 5.00/mm2, p <0.001), CD20- CD79a+ plasma cells (median, 2.25 vs. 1.00 vs. 1.50/mm2, p = 0.016), CD123+ plasmacytoid dendritic cells (pDCs: median, 198.25 vs. 20.00 vs. 18.50/mm2, p <0.001) and CD208+ myeloid dendritic cells (mDCs: median, 70.25 vs. 51.00 vs. 6.00/mm2, p <0.001) were all predominant in IM areas compared with peritumour and intratumour areas, respectively (Fig. 1C). On the other hand, peritumour areas presented abundant Tryptase+ mast cells (median, 195.00 vs. 57.50 vs. 38.25/mm2, p <0.001), CD68+ macrophages (median, 590.75 vs. 372.50 vs. 114.00/mm2, p <0.001) and their subsets CD68+ CD206- macrophages (median, 515.50 vs. 335.75 vs. 104.25/mm2, p <0.001) and CD68+ CD206+ macrophages (median, 36.25 vs. 23.75 vs. 6.25/mm2, p <0.001) compared with IM and intratumour areas, respectively (Fig. 1D). Notably, Tryptase+ mast cells, CD68+ macrophages and their subsets showed a downwards trend in infiltration from peritumour to intratumour areas (Fig. S1D). The densities of the remaining immune subsets were comparable between the IM and peritumour areas (Fig. S1E). Collectively, the current data demonstrate a region-specific distribution of immune cell subsets in iCCAs.

Fig. 1.

Fig. 1

Open in a new tab

Identification of immune subsets by multiplex immunohistochemistry in iCCAs.

(A) Representative images of serial FFPE sections stained with four biomarker panels in intratumour areas. Scale bars, 100 μm. (B) Heatmap showing immune cell infiltration in peritumour, IM and intratumour areas. (C) Violin plots showing immune cells enriched in intratumour (left upper) and IM areas (right upper and lower) (Wilcoxon matched-pair signed-ranks test). (D) Violin plots showing immune cells enriched in peritumour areas (Wilcoxon matched-pair signed-ranks test). FFPE, formalin-fixed paraffin-embedded; iCCA, intrahepatic cholangiocarcinoma; IM, invasive margin.

The correlations between immune cell distribution and clinicopathological factors were investigated. As shown in Fig. 2A,B, patients with grade 3 and 4 fibrosis showed elevated infiltration of some immune cell types in intratumour areas, including CD8+ T cells, CD20+ CD79a+ B cells, CD68+ CD206+ macrophages and CD208+ mDCs. Additionally, iCCAs with grade 3 fibrosis exhibited the most abundant infiltration of lymphoid cells, especially CD8+ T cells and CD20+ CD79a+ B cells. Besides, severer fibrosis was also associated with increased CD68+ CD206+ macrophage and CD208+ mDC infiltration in peritumour areas (Fig. S2A,B). There were no significant differences in peritumoral immune cell infiltration among patients with other underlying liver conditions, including cirrhosis, hepatolithiasis and steatosis, or with a relatively healthy liver (Fig. S2C). Additionally, increased infiltration of CD4+ T cells, CD8+ T cells, CD208+ mDCs and CD123+ pDCs in intratumour areas was associated with poor differentiation, consistent with their roles in hepatocellular carcinoma (Fig. S2D).18 Moreover, patients with early TNM stages exhibited significantly higher densities of CD4+ T cells and the aryl hydrocarbon receptor (AHR)+ T helper 22 (Th22) cell subset in intratumour and peritumour areas, suggesting a negative relationship between CD4+ T cells and tumour progression.19

Fig. 2.

Fig. 2

Open in a new tab

Clinical significance of immune cell subsets.

(A-B) Heatmaps (A) and the corresponding violin plots (B) showing relative densities of immune cells (against the lowest value of each line) stratified by fibrosis grades in intratumour areas. The immune cell subsets with significant differences are indicated in bold (Kruskal-Wallis test). (C) Forest plots showing prognostic significances of various immune subsets in intratumour areas. (D) Heatmap showing correlation between immune cell subsets in three areas using Spearman’s correlation coefficients. ∗p <0.05; ∗∗p <0.01; ∗∗∗p <0.001.

The impact of immune cell infiltration in different regions on OS was analysed (Fig. 2C and Fig. S2E,F). Univariate analysis showed that the abundances of CD3- CD56+ natural killer (NK) cells, CD3+ CD56+ NKT cells and CD20+ CD79a+ B cells in intratumour areas; CD208+ mDCs in IM areas; and CD20+ CD79a+ B cells and CD3- CD56+ NK cells in peritumour areas were correlated with prolonged OS (Fig. 2D). On multivariate analysis, CD20+ CD79a+ B cell infiltration in intratumour areas served as an independent indicator of prognosis (Table S5-6). The infiltration of most immune cell types was positively correlated in the same areas (Fig. 2D).

Definition of spatial immunophenotypes in iCCA

Based on unsupervised cluster analysis of immune cell infiltration in intratumour, IM and peritumour areas, each area was divided into immune-high and immune-low subgroups. Three spatial immunophenotypes of iCCA were identified (Fig. 3A). Tumours (35%) with immune-high features in intratumour subregions were classified as inflamed phenotype, whereas remnant tumours (65%) with immune-low characteristics in intratumour areas were defined as non-inflamed. Among non-inflamed tumours, 30% were termed the ignored phenotype due to immune-low features across all three areas, whereas the other 35% with immune cell infiltration restricted to the IM or peritumour areas were identified as the excluded phenotype. Representative immunostaining features of the three spatial immunophenotypes are shown in Fig. 3B and Fig. S3A. These classifications were generally consistent with the recently proposed spatial distribution of tertiary lymphoid structures in iCCA and other cancer types.20,21 When the region information was not considered, two immune subtypes were identified: high- and low-infiltration groups (Fig. S3B). The majority of the inflamed and excluded phenotypes belonged to the high-infiltration group, while most ignored phenotypes overlapped with the low-infiltration group (Fig. S3C). Based on time-dependent ROC curves, the three spatial immunosubtype classifications displayed a more favourable performance and discrimination for predicting RFS and OS than the two immunosubtype classifications without regard to regional distribution (Fig. S3D,E).

Fig. 3.

Fig. 3

Open in a new tab

Definition of three spatial immunophenotypes.

(A) Defining strategies of three spatial immunophenotypes. (B) Representative multiplex immunostaining of three spatial immunophenotypes with panel 1 markers. Scale bar, 100 μm. (C) Kaplan-Meier curves of OS and RFS stratified by the spatial immunophenotypes. (D) ROC curves showing the driving immune cells of three spatial immunophenotypes. (E) Comparisons of proportion of patients harbouring the indicated clinical covariates among the spatial immunophenotypes (Fishers’ exact test). The mean levels of HBV load are shown (Kruskal-Wallis test). HBV, hepatitis B virus; OS, overall survival; RFS, recurrence-free survival; ROC, receiver operating characteristic.

Patients with the inflamed phenotype had significantly more favourable OS (median OS, >60 vs. 32.3 vs. 14.9 months, p = 0.001) and RFS (median RFS, >50 vs. 21.3 vs. 11.8 months, p <0.001) than those with the excluded or ignored phenotype, respectively (Fig. 3C). Multivariate analysis confirmed the spatial immunophenotype as an independent indicator for OS (hazard ratio 0.506; 95% CI 0.301-0.850; p = 0.010) and RFS (hazard ratio 0.241; 95% CI 0.124-0.466; p <0.001) (Table S7-8). We then identified that the combination of AHR+ Th22 cells in intratumour, IM and peritumour areas could predict the inflamed (AUROC 0.916; 95% CI 0.868-0.965), excluded (AUROC 0.832; 95% CI 0.774-0.890) and ignored (AUROC 0.941; 95% CI 0.910-0.972) phenotypes with good discriminability (Fig. 3D). Therefore, AHR+ Th22 cells were the driving cells of the spatial immunophenotypes, and the IHC evaluation of AHR+ Th22 cells in the three areas could serve as a surrogate marker for classification.

For each spatial immunophenotype, the distribution of the immune cell subset was investigated (Fig. S3F). For the inflamed phenotype, apart from CD15+ neutrophils (median, 395.75 vs. 254.00 vs. 165.50/mm2, p <0.001), Foxp3+ regulatory T cells (median, 88.00 vs. 78.75 vs. 56.75/mm2, p = 0.042) and CD20- CD79a+ plasma cells (median, 12.75 vs. 4.00 vs. 1.00/mm2, p <0.001) were also more abundant in intratumour areas than in IM and peritumour areas. CD15+ neutrophils and CD20- CD79a+ plasma cells showed an uptrend in infiltration from peritumour to intratumour areas in the inflamed phenotype (Fig. S3G). However, there were no significant differences in CD20- CD79a+ plasma cell infiltration among the three areas in the excluded and ignored phenotypes. CD15+ neutrophils in the ignored phenotype displayed a trend contrary to that of the inflamed phenotype.

The associations between the spatial immunophenotypes and clinicopathological features are displayed in Fig. 3E and Table S9. Patients with the inflamed phenotype had a higher proportion of HBV infection (p = 4.8e-02) and a higher viral load (p = 1.3e-02) than those with the excluded and ignored phenotypes. The inflamed phenotype also had a higher ratio of single tumours (p = 1.3e-02). In addition, consistent with the association between fibrosis and increased immune cell infiltration, the inflamed and the excluded phenotypes were associated with higher fibrosis grades (p = 2.0e-03). Taken together, these findings support spatial immunophenotypes as prognostic indicators.

Spatial immunophenotypes and programmed death ligand 1 expression

Programmed death ligand 1 (PD-L1) overexpression has been established as an indicator for PD-1 inhibitor treatment in lung cancer,22 but its significance in iCCAs remains inconclusive. Since PD-L1+ cells within a 20 μm radius of CD8+ T cells were proposed to exert a direct inhibitory effect on CD8+ T cells,23 overall PD-L1 expression and its expression within the functional activity zone were evaluated in iCCAs (Fig. 4A).

Fig. 4.

Fig. 4

Open in a new tab

Correlations between spatial immunophenotypes and programmed death ligand 1 expression.

(A) Representative multiplex immunostaining (left) and spatial location rayplot (right). Black lines connected PD-L1+ cells within 20 μm of CD8+ T cells. Scale bar, 100 μm. (B and C) Boxplots showing PD-L1 expression at the overall level (B) and within a 20 μm radius of CD8+T cells (C) (Wilcoxon matched-pair signed-ranks test). (D-G) Violin plots showing relative densities of PD-L1+ CK19+ tumour cells (D), PD-L1+ CD15+ neutrophils (E), PD-L1+ CD206+ (F) and PD-L1+ CD206- macrophages (G) in intratumour areas stratified by spatial immunophenotypes at the overall level and within a 20 μm radius of CD8+ T cells (Kruskal-Wallis test).

Previous studies indicated that PD-L1 is mainly expressed on tumour cells24 and myeloid cells25 in cholangiocarcinoma. We found a region-specific pattern of PD-L1 expression in iCCAs. PD-L1 was mainly expressed on CK19+ tumour cells and CD15+ neutrophils in intratumour and IM areas, whereas its expression in peritumour areas was predominantly on CD15+ neutrophils and CD206+ macrophages (Fig. 4B and Fig. S4A). PD-L1 expression within a 20 μm radius of CD8+ T cells had comparable distribution patterns (Fig. 4C and Fig. S4B). Notably, although the intratumour areas had a significantly higher density of PD-L1+ CD15+ neutrophils at the overall level (Fig. S4A), the IM and peritumour areas exhibited more PD-L1+ CD15+ neutrophil infiltration within 20 μm around CD8+ T cells (Fig. S4B). Together, our results suggest that the antitumour activity of CD8+ T cells in intratumour areas was mainly inhibited by PD-L1+ CK19+ tumour cells, whereas the dominant suppressors in IM and peritumour areas were PD-L1+ CD15+ neutrophils.

We then analysed PD-L1 expression in distinct spatial immunophenotypes. Compared with the ignored phenotype, the inflamed and excluded phenotypes presented more abundant PD-L1+ CK19+ tumour cells in intratumour and IM areas at the overall level and within 20 μm of CD8+ T cells (Fig. 4D and Fig. S4C). The inflamed phenotype exhibited abundant PD-L1+ CD15+ neutrophils across the three subregions at the overall level and within 20 μm of CD8+ T cells (Fig. 4E and Fig. S4D). For PD-L1 expression on CD68+ macrophages, the inflamed phenotype displayed higher densities of PD-L1+ CD206+ and PD-L1+ CD206- macrophages in intratumour areas than the non-inflamed phenotypes (Fig. 4F-G). The ignored phenotype showed less infiltration in peritumour areas at the overall level and within 20 μm of CD8+ T cells (Fig. S4E-F). The ignored phenotype also presented significantly less abundant functional PD-L1+ CD206+ and PD-L1+ CD206- macrophages in IM areas, although there was no significant difference at the overall level. These results indicate that a more direct interaction between CD8+ T cells and PD-L1+ cells, especially PD-L1+ CD15+ neutrophils, occur in the inflamed phenotype.

Immune escape mechanisms of spatial immunophenotypes

To explore the biological processes underlying the distinct spatial immunophenotypes, we performed transcriptomic and genomic analyses based on 44 iCCA samples present in our cohort and the previous CPTAC iCCA cohort8 (Fig. 5). First, we compared our spatial immunophenotypes with four existing classifications[6], [7], [8], [9] based on the transcriptomic data. In our cohort, the inflamed phenotype was largely concordant with Sia’s inflammation (upregulated immune response–related pathways) and Job’s I2 (enriched innate and adaptive immune cell infiltration) phenotypes, and was associated with a favourable prognosis, whereas the non-inflamed phenotypes co-clustered with Andersen’s C2 (upregulated angiogenesis-associated genes) and Dong’s S2 (abundant fibroblasts and endothelial cell infiltration) subtypes, and were associated with a poor prognosis.

Fig. 5.

Fig. 5

Open in a new tab

Heatmap of main molecular features and immune evasion mechanisms of the spatial immunophenotypes.

∗p values shown are calculated by Wilcoxon rank-sum test for continuous variables or Fisher’s exact test for categorical variables. aComparisons between the inflamed phenotype and the rest of the cohort. bComparisons between the excluded phenotype and the rest of the cohort. cComparisons between the ignored phenotype and the rest of the cohort. TIDE, tumour immune dysfunction and exclusion.

Next, we studied the immune and metabolic differences among the spatial immunophenotypes. The inflamed phenotype displayed an ambivalent immune profile enriched with both immunostimulatory and immunosuppressive features. While overexpression of immune activity-related genes and elevation of immune-related scores revealed an immunostimulatory feature, upregulation of immune checkpoint genes, such as PDCD1 (PD-1) (p = 2.1e-03) and cytotoxic T lymphocyte antigen 4 (CTLA4) (p = 5.0e-07), indicated immunosuppressive characteristics. Correspondingly, the inflamed phenotype had lower TIDE scores than the non-inflamed phenotypes (p = 5.4e-03), suggesting that patients with the inflamed phenotype might benefit from anti-PD-1 or anti-CTLA4 therapies.16 On the other hand, the non-inflamed phenotypes exhibited upregulation of the immune checkpoint genes VTCN1 (p = 1.5e-02) and CD276 (p = 5.6e-03), in line with the previous finding that the coexistence of CD276 and VTCN1 indicates immune-cold features with low CD8+ T-cell infiltration.26 The inflamed phenotypes displayed a hypermetabolic pattern in lipid and amino acid metabolism, which are necessary for the activation of effector immune cells,27 whereas the non-inflamed phenotypes showed increased aerobic glycolysis, which instigates immunosuppressive networks.28 Among the non-inflamed phenotypes, the excluded phenotype showed enhanced gene expression of RASAL2 (p = 7.4e-04), a molecule that has been reported to be associated with the Hippo signalling pathway and tumour progression.29 The ignored phenotype was characterized by the upregulation of DHX32 (p = 4.0e-03), the high expression of which promotes angiogenesis through the β-catenin pathway.30

Subsequently, we evaluated the immune evasion modalities of the spatial phenotypes. The inflamed phenotypes exhibited IL2, IL4, IL12 and IL21 upregulation, all of which can stimulate T-cell expansion.[31], [32], [33] In addition, the inflamed phenotype also showed elevation of the PPAR signalling pathway, which is essential for lipid metabolic homeostasis,34 and elevated NF-kB signalling pathway activity, which is related to the stimulation of innate immunity. The non-inflamed phenotypes shared an increased angiogenesis score, which is associated with restriction of immune cells entering into tumours;35 upregulation of the Wnt-β catenin pathway, which is correlated with the absence of T-cell infiltration;36 and upregulation of the TGF-β pathway, which is involved in promoting naïve T-cell differentiation towards regulatory T cells and dampening the antigen-presenting function of dendritic cells.37 Among the non-inflamed phenotypes, the excluded phenotype presented significant enrichment in activated hepatic stellate cells, extracellular matrix and Notch signalling pathways, all of which are notably correlated with liver fibrogenesis.38 The ignored phenotype showed upregulation of the MAPK signalling pathway, which is involved in the regulation of both angiogenesis and vascular permeability.35

Next, we investigated the profiles of somatic alterations among distinct spatial immunophenotypes. Consistent with a previous study,10 KRAS mutations were remarkably enriched in the inflamed phenotype (29% vs. 4%, p = 4.2e-02), whereas BAP1 mutations (30% vs. 0%, p = 9.4e-03) and FGFR2 fusions (43% vs. 5%, p = 4.4e-03) were more frequent in the non-inflamed phenotypes. However, in our study, TP53 mutations predominated in the inflamed phenotype (38% vs. 9%, p = 3.1e-02), which differed from the results reported in another study.10 In terms of chromosomal aberrations, the non-inflamed phenotypes exhibited significantly increased gains in 1q32.1 (65% vs. 20%, p = 9.5e-03) and 1q43 (50% vs. 15%, p = 4.1e-02), both of which harbour angiogenesis-associated genes, including IL19,39 IL20,40 and AKT.41 Moreover, the non-inflamed phenotypes presented elevated losses in 8p22 (45% vs. 10%, p = 3.1e-02) and 3p22.2 (35% vs. 0%, p = 8.3e-03), which harbour chemotaxis-associated genes, including MTUS142 and CX3CR1,43 respectively.

Discussion

The modality of immune cell infiltration differed considerably among tumour types and histological subtypes. Our study has confirmed that iCCA as a whole exhibits mild immune cell infiltration in intratumour areas, while hepatocellular carcinoma shows a more active immune reaction.44 Notably, neutrophils were more abundant in intratumour areas than in IM and peritumour areas in our study. A recent study established abundant neutrophil infiltration as a marker of poor prognosis in iCCA.45 Although CD15+ neutrophil infiltration was not significantly related to OS in our mIHC analysis, the interaction between tumour-associated neutrophils and macrophages has been demonstrated to promote iCCA progression by activating STAT3.46

Among the three immunophenotypes identified in our study, the inflamed phenotype showed an abundance of cytotoxic adaptive and innate immune cells and a favourable prognosis, whereas the ignored phenotype was devoid of immune cell infiltration and was associated with a poor prognosis. The excluded phenotype was characterized by relatively high infiltration in IM or peritumour areas and a moderate prognosis. The inflamed phenotype was associated with a high proportion of HBV infection in our study. A recent study has revealed that immunotherapy-based treatment is associated with a superior survival benefit in patients with virus-related hepatocellular carcinoma.47 Likewise, immunotherapy may also render more survival benefit in patients with HBV-associated iCCA, and further subgroup analyses are needed in the future. Severe fibrosis was related to increased CD206+ macrophage infiltration in both intratumour and peritumour subregions, consistent with the interaction between cancer-associated fibroblasts and myeloid-derived suppressor cells.48 Therefore, patients with iCCA and fibrosis may benefit from agents targeting cancer-associated fibroblasts, such as navitoclax and 1D11.49

According to the immune evasion mechanisms, patients with the inflamed phenotype are more likely to be the target population for traditional immunotherapy, such as PD-1/PD-L1 or CTLA4 inhibitors, while patients with non-inflamed phenotypes may benefit from novel immunotherapies targeting CD276 (B7–H3) or VTCN1. Our mIHC analyses demonstrated that the distribution of PD-L1 expression was consistent with the distribution of myeloid immune cells, in line with a previous study.25 As the inflamed immunophenotype displayed the highest infiltration of neutrophils and macrophages, PD-L1 expression was most abundant in the inflamed phenotype. Currently, B7–H3 inhibitor (NCT03406949) and B7–H3 CAR (chimeric antigen receptor) T-cell (NCT03198052) therapies are being tested in clinical trials.

The excluded phenotype displayed high levels of fibrogenesis-related pathways, which might be associated with abundant fibrous stroma and the reduced ability of immune cells to migrate.50 Although the potential of Notch-directed therapy to address fibrosis has been demonstrated,38 approaches that enable liver-targeted Notch antagonism without major side effects remain to be developed. The efficacy of agents targeting angiogenesis, such as ramucirumab for biliary tract cancer, has been investigated with promising results,51 and whether more survival benefit can be achieved for the ignored phenotype remains to be determined in future stratification studies.

Actionable somatic alterations have been identified in nearly 40% of iCCA cases, and several effective targeted therapies have been approved for clinical use in recent years. FGFR2 fusions or other rearrangements occur in 10-15% of iCCAs.52 In our study, FGFR2 fusions predominated in non-inflamed phenotypes, consistent with recent findings that FGFR2 genetic alterations are correlated with low mutational burden and reduced immune infiltration.53 The FGFR1-3 inhibitor pemigatinib has been approved as a second-line therapy for iCCAs refractory to chemotherapy.54 Whether pemigatinib can alter “immune-cold” tumours and render synergetic effects in combination with immunotherapies is under investigation in other cancer types (NCT05004974). On the other hand, the inflamed phenotype showed enrichment in KRAS mutations, suggesting potential efficacy of BI-3406, an agent with the ability to increase sensitivity to PD-1 therapy in patients with KRAS-mutant lung cancer.55

Our study has several limitations. The transcriptomic and genomic analyses were based on a relatively small number of patients, which might lead to bias in delineating immune escape mechanisms. In addition, quantitative assays of immune cell infiltration were mainly based on digital image analysis. Although strict manual verification was carried out, the calculation of the densities of immune cells might harbour a certain degree of error. Finally, the immune escape mechanisms and specific treatment of spatial phenotypes warrant further functional analysis and clinical validation.

In conclusion, we characterized three spatial immunophenotypes of iCCAs and identified distinct immune evasion mechanisms, providing a rationale for potential tailored therapies for iCCAs.

Financial support

This study was supported by the National Natural Science Foundation of China (Grants. 82130077 and 81961128025); the Research Projects from the Science and Technology Commission of Shanghai Municipality (Grants 21JC1410100, 21JC1401200, 20JC1418900); the Natural Science Funds of Shanghai (Grants 21ZR1413800).

Authors’ contributions

Conception and design: CZ, JM, KZ, QG, XZ, DX. Development of methodology: CZ, JM, LD. Acquisition of data (acquired and managed patients, provided facilities, etc.): CZ, BZ, DR, SZ, LY. Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): CZ, JM. Writing, review, and/or revision of the manuscript: CZ, JM, KZ, QG, XZ, DX. Administrative, technical, or material support: DR, QG, XZ. Study supervision: QG, XZ, DX. All authors read and approved the final manuscript.

Data availability statement

The datasets used and analysed during the current study are available from the corresponding author on reasonable request.

Conflicts of interest

The authors declare no conflicts of interest that pertain to this work.

Please refer to the accompanying ICMJE disclosure forms for further details.

Footnotes

Author names in bold designate shared co-first authorship

Supplementary data to this article can be found online at https://doi.org/10.1016/j.jhepr.2023.100762.

Contributor Information

Qiang Gao, Email: gaoqiang@fudan.edu.cn.

Xiaoming Zhang, Email: xmzhang@ips.ac.cn.

Diyang Xie, Email: xie.diyang@zs-hospital.sh.cn.

Ethics approval and consent to participate

The ethics approval of this study was granted by the institutional review board committee at Zhongshan Hospital of Fudan University (Shanghai, China).

Supplementary data

The following are the supplementary data to this article:

Multimedia component 1
mmc1.pdf (2.5MB, pdf)
Multimedia component 2
mmc2.docx (49.5KB, docx)
Multimedia component 3
mmc3.pdf (500.7KB, pdf)
Multimedia component 4
mmc4.xlsx (75.7KB, xlsx)
Multimedia component 5
mmc5.pdf (5.8MB, pdf)

References

  • 1.Bridgewater J., Galle P.R., Khan S.A., Llovet J.M., Park J.W., Patel T., et al. Guidelines for the diagnosis and management of intrahepatic cholangiocarcinoma. J Hepatol. 2014;60:1268–1289. doi: 10.1016/j.jhep.2014.01.021. [DOI] [PubMed] [Google Scholar]
  • 2.Zhang H., Yang T., Wu M., Shen F. Intrahepatic cholangiocarcinoma: epidemiology, risk factors, diagnosis and surgical management. Cancer Lett. 2016;379:198–205. doi: 10.1016/j.canlet.2015.09.008. [DOI] [PubMed] [Google Scholar]
  • 3.Rahnemai-Azar A.A., Weisbrod A., Dillhoff M., Schmidt C., Pawlik T.M. Intrahepatic cholangiocarcinoma: molecular markers for diagnosis and prognosis. Surg Oncol. 2017;26:125–137. doi: 10.1016/j.suronc.2016.12.009. [DOI] [PubMed] [Google Scholar]
  • 4.Zhou G., Sprengers D., Mancham S., Erkens R., Boor P.P.C., van Beek A.A., et al. Reduction of immunosuppressive tumor microenvironment in cholangiocarcinoma by ex vivo targeting immune checkpoint molecules. J Hepatol. 2019;71:753–762. doi: 10.1016/j.jhep.2019.05.026. [DOI] [PubMed] [Google Scholar]
  • 5.Kelley R.K., Bridgewater J., Gores G.J., Zhu A.X. Systemic therapies for intrahepatic cholangiocarcinoma. J Hepatol. 2020;72:353–363. doi: 10.1016/j.jhep.2019.10.009. [DOI] [PubMed] [Google Scholar]
  • 6.Andersen J.B., Spee B., Blechacz B.R., Avital I., Komuta M., Barbour A., et al. Genomic and genetic characterization of cholangiocarcinoma identifies therapeutic targets for tyrosine kinase inhibitors. Gastroenterology. 2012;142:1021–1031 e1015. doi: 10.1053/j.gastro.2011.12.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Sia D., Hoshida Y., Villanueva A., Roayaie S., Ferrer J., Tabak B., et al. Integrative molecular analysis of intrahepatic cholangiocarcinoma reveals 2 classes that have different outcomes. Gastroenterology. 2013;144:829–840. doi: 10.1053/j.gastro.2013.01.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Dong L., Lu D., Chen R., Lin Y., Zhu H., Zhang Z., et al. Proteogenomic characterization identifies clinically relevant subgroups of intrahepatic cholangiocarcinoma. Cancer Cell. 2022;40:70–87.e15. doi: 10.1016/j.ccell.2021.12.006. [DOI] [PubMed] [Google Scholar]
  • 9.Job S., Rapoud D., Dos Santos A., Gonzalez P., Desterke C., Pascal G., et al. Identification of four immune subtypes characterized by distinct composition and functions of tumor microenvironment in intrahepatic cholangiocarcinoma. Hepatology. 2020;72:965–981. doi: 10.1002/hep.31092. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Martin-Serrano M.A., Kepecs B., Torres-Martin M., Bramel E.R., Haber P.K., Merritt E., et al. Novel microenvironment-based classification of intrahepatic cholangiocarcinoma with therapeutic implications. Gut. 2023;72:736–748. doi: 10.1136/gutjnl-2021-326514. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Zhu X.D., Zhang J.B., Zhuang P.Y., Zhu H.G., Zhang W., Xiong Y.Q., et al. High expression of macrophage colony-stimulating factor in peritumoral liver tissue is associated with poor survival after curative resection of hepatocellular carcinoma. J Clin Oncol. 2008;26:2707–2716. doi: 10.1200/JCO.2007.15.6521. [DOI] [PubMed] [Google Scholar]
  • 12.Laghi L., Bianchi P., Miranda E., Balladore E., Pacetti V., Grizzi F., et al. CD3+ cells at the invasive margin of deeply invading (pT3–T4) colorectal cancer and risk of post-surgical metastasis: a longitudinal study. Lancet Oncol. 2009;10:877–884. doi: 10.1016/S1470-2045(09)70186-X. [DOI] [PubMed] [Google Scholar]
  • 13.Pagès F., Mlecnik B., Marliot F., Bindea G., Ou F.-S., Bifulco C., et al. International validation of the consensus Immunoscore for the classification of colon cancer: a prognostic and accuracy study. Lancet. 2018;391:2128–2139. doi: 10.1016/S0140-6736(18)30789-X. [DOI] [PubMed] [Google Scholar]
  • 14.Gao Q., Qiu S.J., Fan J., Zhou J., Wang X.Y., Xiao Y.S., et al. Intratumoral balance of regulatory and cytotoxic T cells is associated with prognosis of hepatocellular carcinoma after resection. J Clin Oncol. 2007;25:2586–2593. doi: 10.1200/JCO.2006.09.4565. [DOI] [PubMed] [Google Scholar]
  • 15.Stack E.C., Wang C., Roman K.A., Hoyt C.C. Multiplexed immunohistochemistry, imaging, and quantitation: a review, with an assessment of Tyramide signal amplification, multispectral imaging and multiplex analysis. Methods. 2014;70:46–58. doi: 10.1016/j.ymeth.2014.08.016. [DOI] [PubMed] [Google Scholar]
  • 16.Jiang P., Gu S., Pan D., Fu J., Sahu A., Hu X., et al. Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. Nat Med. 2018;24:1550–1558. doi: 10.1038/s41591-018-0136-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Mao Z.-Y., Zhu G.-Q., Xiong M., Ren L., Bai L. Prognostic value of neutrophil distribution in cholangiocarcinoma. World J Gastroenterol. 2015;21:4961–4968. doi: 10.3748/wjg.v21.i16.4961. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Kurebayashi Y., Ojima H., Tsujikawa H., Kubota N., Maehara J., Abe Y., et al. Landscape of immune microenvironment in hepatocellular carcinoma and its additional impact on histological and molecular classification. Hepatology. 2018;68:1025–1041. doi: 10.1002/hep.29904. [DOI] [PubMed] [Google Scholar]
  • 19.Ge P., Wang W., Li L., Zhang G., Gao Z., Tang Z., et al. Profiles of immune cell infiltration and immune-related genes in the tumor microenvironment of colorectal cancer. Biomed Pharmacother = Biomedecine Pharmacotherapie. 2019;118 doi: 10.1016/j.biopha.2019.109228. [DOI] [PubMed] [Google Scholar]
  • 20.Ding G.-Y., Ma J.-Q., Yun J.-P., Chen X., Ling Y., Zhang S., et al. Distribution and density of tertiary lymphoid structures predict clinical outcome in intrahepatic cholangiocarcinoma. J Hepatol. 2022;76:608–618. doi: 10.1016/j.jhep.2021.10.030. [DOI] [PubMed] [Google Scholar]
  • 21.Fridman W.H., Meylan M., Petitprez F., Sun C.-M., Italiano A., Sautès-Fridman C. B cells and tertiary lymphoid structures as determinants of tumour immune contexture and clinical outcome. Nat Rev Clin Oncol. 2022;19:441–457. doi: 10.1038/s41571-022-00619-z. [DOI] [PubMed] [Google Scholar]
  • 22.Reck M., Remon J., Hellmann M.D. First-line immunotherapy for non-small-cell lung cancer. J Clini Oncol Official J Am Soc Clin Oncol. 2022;40:586–597. doi: 10.1200/JCO.21.01497. [DOI] [PubMed] [Google Scholar]
  • 23.Carstens J.L., Correa de Sampaio P., Yang D., Barua S., Wang H., Rao A., et al. Spatial computation of intratumoral T cells correlates with survival of patients with pancreatic cancer. Nat Commun. 2017;8 doi: 10.1038/ncomms15095. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Kitano Y., Yamashita Y.-I., Nakao Y., Itoyama R., Yusa T., Umezaki N., et al. Clinical significance of PD-L1 expression in both cancer and stroma cells of cholangiocarcinoma patients. Ann Surg Oncol. 2020;27:599–607. doi: 10.1245/s10434-019-07701-4. [DOI] [PubMed] [Google Scholar]
  • 25.Xia T., Li K., Niu N., Shao Y., Ding D., Thomas D.L., et al. Immune cell atlas of cholangiocarcinomas reveals distinct tumor microenvironments and associated prognoses. J Hematol Oncol. 2022;15:37. doi: 10.1186/s13045-022-01253-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Si S., Wang L., Cao H., Xu Y., Zhan Q. Co-deficiency of B7-H3 and B7-H4 identifies high CD8 + T cell infiltration and better prognosis in pancreatic cancer. BMC Cancer. 2022;22:211. doi: 10.1186/s12885-022-09294-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Leone R.D., Powell J.D. Metabolism of immune cells in cancer. Nat Rev Cancer. 2020;20:516–531. doi: 10.1038/s41568-020-0273-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Ganapathy-Kanniappan S. Linking tumor glycolysis and immune evasion in cancer: emerging concepts and therapeutic opportunities. Biochimica Et Biophysica Acta. Rev Cancer. 2017;1868:212–220. doi: 10.1016/j.bbcan.2017.04.002. [DOI] [PubMed] [Google Scholar]
  • 29.Pan Y., Tong J.H.M., Lung R.W.M., Kang W., Kwan J.S.H., Chak W.P., et al. RASAL2 promotes tumor progression through LATS2/YAP1 axis of hippo signaling pathway in colorectal cancer. Mol Cancer. 2018;17:102. doi: 10.1186/s12943-018-0853-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Lin H., Fang Z., Su Y., Li P., Wang J., Liao H., et al. DHX32 promotes angiogenesis in colorectal cancer through augmenting β-catenin signaling to induce expression of VEGFA. EBioMedicine. 2017;18:62–72. doi: 10.1016/j.ebiom.2017.03.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Spolski R., Li P., Leonard W.J. Biology and regulation of IL-2: from molecular mechanisms to human therapy. Nat Rev Immunol. 2018;18:648–659. doi: 10.1038/s41577-018-0046-y. [DOI] [PubMed] [Google Scholar]
  • 32.Tepper R.I. The eosinophil-mediated antitumor activity of interleukin-4. J Allerg Clini Immunol. 1994;94:1225–1231. doi: 10.1016/0091-6749(94)90336-0. [DOI] [PubMed] [Google Scholar]
  • 33.Yan J., Smyth M.J., Teng M.W.L. Interleukin (IL)-12 and IL-23 and their conflicting roles in cancer. Cold Spring Harbor Perspect Biol. 2018;10 doi: 10.1101/cshperspect.a028530. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Jonker J.W., Suh J.M., Atkins A.R., Ahmadian M., Li P., Whyte J., et al. A PPARγ-FGF1 axis is required for adaptive adipose remodelling and metabolic homeostasis. Nature. 2012;485:391–394. doi: 10.1038/nature10998. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Claesson-Welsh L., Welsh M. VEGFA and tumour angiogenesis. J Intern Med. 2013;273:114–127. doi: 10.1111/joim.12019. [DOI] [PubMed] [Google Scholar]
  • 36.Grasso C.S., Giannakis M., Wells D.K., Hamada T., Mu X.J., Quist M., et al. Genetic mechanisms of immune evasion in colorectal cancer. Cancer Discov. 2018;8:730–749. doi: 10.1158/2159-8290.CD-17-1327. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Batlle E., Massagué J. Transforming growth factor-β signaling in immunity and cancer. Immunity. 2019;50:924–940. doi: 10.1016/j.immuni.2019.03.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Zhu C., Kim K., Wang X., Bartolome A., Salomao M., Dongiovanni P., et al. Hepatocyte Notch activation induces liver fibrosis in nonalcoholic steatohepatitis. Sci Translat Med. 2018;10 doi: 10.1126/scitranslmed.aat0344. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Kako F., Gabunia K., Ray M., Kelemen S.E., England R.N., Kako B., et al. Interleukin-19 induces angiogenesis in the absence of hypoxia by direct and indirect immune mechanisms. Am J Physiol. 2016;310:C931–C941. doi: 10.1152/ajpcell.00006.2016. Cell Physiology. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Hsieh M.Y., Chen W.Y., Jiang M.J., Cheng B.C., Huang T.Y., Chang M.S. Interleukin-20 promotes angiogenesis in a direct and indirect manner. Genes Immun. 2006;7:234–242. doi: 10.1038/sj.gene.6364291. [DOI] [PubMed] [Google Scholar]
  • 41.Li Y., Cai B., Shen L., Dong Y., Lu Q., Sun S., et al. MiRNA-29b suppresses tumor growth through simultaneously inhibiting angiogenesis and tumorigenesis by targeting Akt3. Cancer Lett. 2017;397:111–119. doi: 10.1016/j.canlet.2017.03.032. [DOI] [PubMed] [Google Scholar]
  • 42.Zhao T., Ding X., Chang B., Zhou X., Wang A. MTUS1/ATIP3a down-regulation is associated with enhanced migration, invasion and poor prognosis in salivary adenoid cystic carcinoma. BMC Cancer. 2015;15:203. doi: 10.1186/s12885-015-1209-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Schmall A., Al-Tamari H.M., Herold S., Kampschulte M., Weigert A., Wietelmann A., et al. Macrophage and cancer cell cross-talk via CCR2 and CX3CR1 is a fundamental mechanism driving lung cancer. Am J Respir Crit Care Med. 2015;191:437–447. doi: 10.1164/rccm.201406-1137OC. [DOI] [PubMed] [Google Scholar]
  • 44.Kurebayashi Y., Ojima H., Tsujikawa H., Kubota N., Maehara J., Abe Y., et al. Landscape of immune microenvironment in hepatocellular carcinoma and its additional impact on histological and molecular classification. Hepatology (Baltimore, Md.) 2018;68:1025–1041. doi: 10.1002/hep.29904. [DOI] [PubMed] [Google Scholar]
  • 45.Lin J., Dai Y., Sang C., Song G., Xiang B., Zhang M., et al. Multimodule characterization of immune subgroups in intrahepatic cholangiocarcinoma reveals distinct therapeutic vulnerabilities. J Immunother Cancer. 2022:10. doi: 10.1136/jitc-2022-004892. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Zhou Z., Wang P., Sun R., Li J., Hu Z., Xin H., et al. Tumor-associated neutrophils and macrophages interaction contributes to intrahepatic cholangiocarcinoma progression by activating STAT3. J Immunother Cancer. 2021;9 doi: 10.1136/jitc-2020-001946. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Cheng A.-L., Qin S., Ikeda M., Galle P.R., Ducreux M., Kim T.-Y., et al. Updated efficacy and safety data from IMbrave150: atezolizumab plus bevacizumab vs. sorafenib for unresectable hepatocellular carcinoma. J Hepatol. 2022;76:862–873. doi: 10.1016/j.jhep.2021.11.030. [DOI] [PubMed] [Google Scholar]
  • 48.Fabris L., Sato K., Alpini G., Strazzabosco M. The tumor microenvironment in cholangiocarcinoma progression. Hepatology. 2021;73(Suppl 1):75–85. doi: 10.1002/hep.31410. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Affo S., Yu L.X., Schwabe R.F. The role of cancer-associated fibroblasts and fibrosis in liver cancer. Annu Rev Pathol. 2017;12:153–186. doi: 10.1146/annurev-pathol-052016-100322. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Chen D.S., Mellman I. Elements of cancer immunity and the cancer-immune set point. Nature. 2017;541:321–330. doi: 10.1038/nature21349. [DOI] [PubMed] [Google Scholar]
  • 51.Lee S., Shroff R.T., Makawita S., Xiao L., Danner De Armas A., Bhosale P., et al. Phase II study of ramucirumab in advanced biliary tract cancer previously treated by gemcitabine-based chemotherapy. Clini Cancer Res Official J Am Assoc Cancer Res. 2022;28:2229–2236. doi: 10.1158/1078-0432.CCR-21-3548. [DOI] [PubMed] [Google Scholar]
  • 52.Goyal L., Kongpetch S., Crolley V.E., Bridgewater J. Targeting FGFR inhibition in cholangiocarcinoma. Cancer Treat Rev. 2021;95 doi: 10.1016/j.ctrv.2021.102170. [DOI] [PubMed] [Google Scholar]
  • 53.Lin Y., Peng L., Dong L., Liu D., Ma J., Lin J., et al. Geospatial immune heterogeneity reflects the diverse tumor-immune interactions in intrahepatic cholangiocarcinoma. Cancer Discov. 2022;12:2350–2371. doi: 10.1158/2159-8290.CD-21-1640. [DOI] [PubMed] [Google Scholar]
  • 54.Abou-Alfa G.K., Sahai V., Hollebecque A., Vaccaro G., Melisi D., Al-Rajabi R., et al. Pemigatinib for previously treated, locally advanced or metastatic cholangiocarcinoma: a multicentre, open-label, phase 2 study. Lancet Oncol. 2020;21:671–684. doi: 10.1016/S1470-2045(20)30109-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Koga T., Suda K., Fujino T., Ohara S., Hamada A., Nishino M., et al. KRAS secondary mutations that confer acquired resistance to KRAS G12C inhibitors, sotorasib and adagrasib, and overcoming strategies: insights from in vitro experiments. J Thorac Oncol: Official Publ Int Assoc Study Lung Cancer. 2021;16:1321–1332. doi: 10.1016/j.jtho.2021.04.015. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Multimedia component 1
mmc1.pdf (2.5MB, pdf)
Multimedia component 2
mmc2.docx (49.5KB, docx)
Multimedia component 3
mmc3.pdf (500.7KB, pdf)
Multimedia component 4
mmc4.xlsx (75.7KB, xlsx)
Multimedia component 5
mmc5.pdf (5.8MB, pdf)

Data Availability Statement

The datasets used and analysed during the current study are available from the corresponding author on reasonable request.

Articles from JHEP Reports are provided here courtesy of Elsevier

RESOURCES