Hepatocellular carcinoma molecular subtype classification and subtype-specific treatment for hepatocellular carcinoma

Abstract

The invention provides a method for hepatocellular carcinoma (HCC) subtype classification and treatment. The method can include identifying one or more features associated with a hepatic epithelial cell lineage for the hepatic epithelial cell lineage. The one or more features can be specified as a representation of a molecular subtype associated with hepatocellular carcinoma (HCC), such as, for example, a bile duct-like subtype, a hepatocyte-like subtype, or a progenitor cell-like subtype. If these characteristics are detected within a tumor sample of a patient, the patient can be determined as exhibiting the molecular subtype. In addition, the treatment of the patient can be determined based on the molecular subtype exhibited by the patient. For example, treatment of the patient can include additional therapies, such as GPC3 / CD3 bispecific antibodies, to overcome subtype-specific resistance to combinatorial immunotherapies associated with the progenitor cell-like subtypes. Related systems and computer program products are also provided.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims priority to and the benefit of U.S. Provisional Patent Application No. 63 / 337,007 filed on April 29, 2022 and U.S. Provisional Patent Application No. 63 / 373,696 filed on August 26, 2022, the entire contents of both provisional patent applications are incorporated herein by reference for all purposes. Technical Field

[0003] The subject matter described herein relates generally to digital pathology, and more specifically to hepatocellular carcinoma (HCC) molecular subtype classification and subtype-specific treatment for HCC. Background Art

[0004] Hepatocellular carcinoma (HCC) is a common malignant tumor with a high mortality rate. The combination therapy of atezolizumab (anti-PD-L1) and bevacizumab (anti-VEGF) has shown strong anti-tumor activity in clinical trials and has become a new standard of care as a first-line treatment for HCC. However, nearly one-third of patients still experience disease progression, highlighting the great need to understand the tumor heterogeneity of HCC and the response or resistance mechanisms to guide innovative treatment strategies. Summary of the invention

[0005] Systems, methods, and articles (including computer program products) for hepatocellular carcinoma (HCC) subtype classification and treatment are provided. In some exemplary embodiments, a system comprising at least one processor and at least one memory is provided. The at least one memory may include program code that provides operations when executed by at least one processor. These operations may include: for a hepatic epithelial cell lineage, identifying one or more features associated with the hepatic epithelial cell lineage; designating the one or more features as representative of a molecular subtype associated with hepatocellular carcinoma (HCC); receiving a tumor sample from a patient; and determining that the patient exhibits a molecular subtype associated with hepatocellular carcinoma based on the one or more features detected in the tumor sample from the patient.

[0006] In another aspect, a method for subtype classification and treatment of hepatocellular carcinoma (HCC) is provided. The method may include: for a hepatic epithelial cell lineage, identifying one or more features associated with the hepatic epithelial cell lineage; designating the one or more features as representative of a molecular subtype associated with hepatocellular carcinoma (HCC); receiving a tumor sample from a patient; and determining that the patient exhibits a molecular subtype associated with hepatocellular carcinoma based on the one or more features detected in the tumor sample from the patient.

[0007] In another aspect, a computer program product is provided, comprising a non-transitory computer-readable medium storing instructions. The instructions may cause operations that may be performed by at least one data processor. The operations may include: for a hepatic epithelial cell lineage, identifying one or more features associated with the hepatic epithelial cell lineage; designating the one or more features as representative of a molecular subtype associated with hepatocellular carcinoma (HCC); receiving a tumor sample from a patient; and determining that the patient exhibits a molecular subtype associated with hepatocellular carcinoma based on the one or more features detected in the tumor sample from the patient.

[0008] In some variations of the methods, systems, and non-transitory computer-readable media, one or more of the following features may optionally be included in any feasible combination.

[0009] In some variations, the one or more characteristics may include genetic characteristics.

[0010] In some variations, a plurality of molecular subtypes associated with hepatocellular carcinoma (HCC) may be identified based at least on transcriptomic data associated with a plurality of hepatocellular carcinoma (HCC) tissue samples.

[0011] In some variations, multiple subtypes may be identified by applying cluster analysis to the transcriptome data to identify the number of subpopulations present within the transcriptome data.

[0012] In some variations, cluster analysis may be applied to identify one or more subpopulations associated with the greatest co-phenotypic correlation values.

[0013] In some variations, the cluster analysis may include non-negative matrix factorization (NMF).

[0014] In some variations, the cluster analysis may include one or more of connectivity-based clustering, centroid-based clustering, distribution-based clustering, density-based clustering, subspace-based clustering, group-based clustering, and graph-based clustering.

[0015] In some variations, the multiple subtypes may be identified and / or validated by applying a classifier to the transcriptome data.

[0016] In some variations, the classifier may include a random forest classifier.

[0017] In some variations, the one or more characteristics may include at least one of a tumor cell intrinsic characteristic and a tumor microenvironment characteristic.

[0018] In some variations, the one or more features may include immunohistochemistry of cytochrome P450, expression levels of cytochrome P450, Hippo signaling pathway, and / or expression levels of YES-associated protein (YAP).

[0019] In some variations, the one or more features may include the amount of fibroblast activation protein in the stroma, the vascular density, the density of cluster of differentiation 8 (CD8) in the tumor epithelium, the number of MHCI+ tumor cells, the density of cluster of differentiation 8 (CD8) in the tumor epithelium, the density of PDL1+, the density of activated T cells, and / or the density of exhausted T cells.

[0020] In some variations, a molecular subtype associated with hepatocellular carcinoma may include a cholangioid subtype. The hepatic epithelial cell lineage may include a cholangiocyte.

[0021] In some variations, a molecular subtype associated with hepatocellular carcinoma may comprise a hepatocyte-like subtype.The liver epithelial cell lineage may comprise a hepatocyte.

[0022] In some variations, a molecular subtype associated with hepatocellular carcinoma may include a progenitor-like subtype.The hepatic epithelial cell lineage may include a bipotent progenitor cell.

[0023] In some variations, treatment for hepatocellular carcinoma (HCC) may be determined based at least on the molecular subtype of the patient.

[0024] In some variations, treatment for hepatocellular carcinoma (HCC) may include combination immunotherapy based at least on whether the patient has the cholangiocarcinoma subtype or the hepatocellular subtype.

[0025] In some variations, treatment for hepatocellular carcinoma (HCC) may include atezolizumab (anti-PD-L1) plus bevacizumab (anti-VEGF) combination therapy based at least on whether the patient has the cholangiocarcinoid subtype or the hepatocellular subtype.

[0026] In some variations, treatment for hepatocellular carcinoma (HCC) may include one or more additional therapies based at least on the patient having a progenitor-like subtype to overcome subtype-specific resistance to combination immunotherapy associated with the progenitor-like subtype.

[0027] In some variations, treatment for hepatocellular carcinoma (HCC) may include a GPC3 / CD3 bispecific antibody in addition to combination immunotherapy based at least on the patient having a progenitor-like subtype.

[0028] In some variations, the one or more features may include cancerous epithelial tissue, necrotic tissue, and / or normal tissue present in the image of the tumor sample.

[0029] In some variations, the one or more features may comprise a growth pattern present in an image of the tumor sample.

[0030] In some variations, the one or more features include one or more of cancerous epithelial cells, fibroblasts, endothelial cells, and normal cells present in the image of the tumor sample.

[0031] In some variations, the one or more features may include one or more of hepatocellular carcinoma (HCC) hepatocellular carcinoma epithelial cells, hepatocellular carcinoma (HCC) carcinoma epithelial cells with Mallory hyaline bodies or globules, and hepatocellular carcinoma (HCC) hepatoblastoid carcinoma epithelial cells.

[0032] In some variations, a patient's response to treatment and / or pathological remission may be determined based at least on the patient's molecular subtype.

[0033] Specific implementations of the current subject matter may include, but are not limited to, methods consistent with the description provided herein and articles including tangibly embodied machine-readable media that are operable to cause one or more machines (e.g., computers, etc.) to cause operations that implement one or more of the features described. Similarly, a computer system that may include one or more processors and one or more memories coupled to the one or more processors is also described. A memory that may include a non-transitory computer-readable or machine-readable storage medium may include, encode, store, etc., one or more programs that cause one or more processors to perform one or more of the operations described herein. A computer-implemented method consistent with one or more implementations of the current subject matter may be implemented by one or more data processors present in a single computing system or multiple computing systems. Such multiple computing systems may be connected and may exchange data and / or commands or other instructions, etc., via one or more connections, including, for example, via a network (e.g., the Internet, a wireless wide area network, a local area network, a wide area network, a wired network, etc.) via a direct connection between one or more computing systems in the multiple computing systems, etc.

[0034] The details of one or more variations of the subject matter described herein are set forth in the accompanying drawings and the following description. Other features and advantages of the subject matter described herein will become apparent with reference to the description and drawings, as well as the claims. Although certain features of the subject matter disclosed herein are described for illustrative purposes related to hepatocellular carcinoma (HCC), it should be readily understood that such features are not intended to be limiting. The claims following this disclosure are intended to define the scope of the subject matter protected. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The accompanying drawings, which are incorporated into and constitute a part of this specification, illustrate certain aspects of the subject matter disclosed herein and, together with the description, help explain some principles associated with the disclosed embodiments.

[0036] Figure 1 depicts a system diagram illustrating an example of a digital pathology system according to some exemplary embodiments;

[0037] Figure 2 depicts a schematic diagram showing liver epithelial cell lineages, tumor cell intrinsic characteristics, tumor microenvironment (TME) features, and immune landscapes of different hepatocellular carcinoma (HCC) subtypes according to some exemplary embodiments;

[0038] Figure 3 Depicted are the relationships between molecular subtypes and tumor intrinsic heterogeneity and tumor microenvironment heterogeneity in advanced hepatocellular carcinoma (HCC) according to some exemplary embodiments. Figure 3 Results shown are from 99 hepatocellular carcinoma samples in the training dataset. Figure 3 (a), Experimental design of the study; Figure 3 (b), phenotypic correlations of clusters 1 to 5 from NMF analysis with the top 3,000 most differentially expressed genes; Figure 3 (c), Unsupervised hierarchical clustering with the top 3,000 most differentially expressed genes. Box plots show the expression of known hepatic epithelial cell lineage markers in NMF subtypes. Significant p-values ​​of unpaired two-tailed t-tests between groups are shown; Figure 3 (d), Signature gene set enrichment (GSEA) analysis showing unique pathway enrichment for each subtype relative to the other two subtypes; Figure 3 (e), Box plots showing the expression of selected features of pathways representing each subtype. The significant p-values ​​of unpaired two-tailed t-tests between groups are shown;

[0039] Figure 4 Depicted are cell lineage-associated molecular subtype signatures reproduced in hepatocellular carcinoma and validated in situ according to some exemplary embodiments. Figure 4 (a), Pie chart comparing the incidence (% of total) of molecular subtypes in the training dataset GO30140 A arm and the IMbrave 150 biomarker population (BEP); Figure 4 (b), expression profiles of the 288 core gene set in 90 GO30140 A-arm baseline samples ranked by subtype classified by the developed random forest classifier and superimposed with etiology, PD-L1 IHC status, and TP53 and CTNNB1 alterations; Figure 4 (c), expression profiles of the 288 core gene sets sorted by subtype in 177 IMBrave 150 baseline samples and superimposed with etiology, PD-L1 IHC status, and TP53 and CTNNB1 alterations;

[0040] Figure 5Depicted are the relationships between molecular subtypes, Response Evaluation Criteria in Solid Tumors (RECIST) responses, and the benefit of atezolizumab plus bevacizumab over sorafenib, according to some exemplary embodiments. Figure 5 (a) , Waterfall plot showing the maximum percentage change in the longest dimension (SLD) from the size of the baseline target lesion sorted by subtype and magnitude of change; Figure 5 (b), KM curves showing PFS of the three subtypes in the GO30140 A arm; Figure 5 (c), PFS (top row) and OS KM curves within subtypes stratified by treatment Atezo + Bev vs. sorafenib. Hazard ratios and p-values ​​are for treatment arm terms from the Cox proportional hazards multivariate model, which included age, sex, MVI or EHS, AFP, and ECOG performance status terms;

[0041] Figure 5 The data in show that molecular subtypes of HCC may be characterized not only by different tumor-intrinsic characteristics and tumor microenvironment features, but they may also have clinical relevance and be associated with different outcomes of immunotherapy, for example, the cholangioid subtype is associated with the greatest benefit of combination therapy, while the progenitor-like subtype with high expression of GPC3 may appear to benefit less. In some embodiments, biological insights into the heterogeneity of HCC and strategies to target subtype-specific weaknesses to overcome resistance to combination immunotherapy are provided.

[0042] Figure 6 It is depicted that, according to some exemplary embodiments, the anti-tumor activity of ERY974 in progenitor-like tumors is associated with increased T cell infiltration and cytolytic activity. Figure 6 (a), GPC3 mRNA levels in bile duct (cholangiocyte-like) cells and progenitor (progenitor-like) cells. Statistical significance was determined by two-tailed unpaired t-test; Figure 6 (b), Cell surface expression of GPC3 determined by flow cytometry using anti-GPC3 antibody. The black line indicates the shift in the absence of antibody; the green line indicates the shift in the presence of isotype control; the red line indicates the shift in the presence of anti-GPC3 antibody; Figure 6 (c), Antibody binding capacity (ABC) of GPC3 in SK-HEP1 (cholangiocyte-like), huh-1 (progenitor-like), and huh-7 (progenitor-like). Figure 6 (d), T cell-dependent cytotoxicity assay (TDCC); Figure 6(e), Antitumor efficacy of ERY974 in huh-1 / huNOG (left) and huh-7 / T cell injection (right) models (n=5). The dose of ERY974 was 5 mg / kg in the huh-1 / huNOG model and 1 mg / kg in the huh-7 / T cell injection model. The red arrow indicates the time of ERY974 administration. TGI values ​​are shown in each figure. Data are shown as mean ± standard deviation (SD) (n=5). p value is derived from two-tailed unpaired t-test. ns, not significant (Wilcoxon test); Figure 6 (f), study design for PD analysis; Figure 6 (g), Tumor infiltrating lymphocyte analysis of huh-1 tumors by flow cytometry on day 3 after ERY974 administration (n=8 per group). The significant p-values ​​of the unpaired two-tailed t-test between vehicle and ERY97 treated groups are shown; Figure 6 (h), Histopathological analysis of huh-1 tumors treated with ERY974 (day 3). Representative HE staining (left), CD3 IHC (top row, middle and right), and CD8 IHC (bottom row, middle and right) are shown. The scale bar in the left and middle column images is 100 μm, and the scale bar in the right column images is 50 μm. The rectangular inset in the middle column image represents the corresponding field enlarged in the right column image; Figure 6 (i), Heat map analysis of T cell marker genes and T cell activation marker genes in huh-1 tumor RNA treated with ERY974 on day 3 (n=8);

[0043] Figure 7 depicts Gene Set Enrichment Analysis (GSEA) and xCell analysis of molecular subtypes of hepatocellular carcinoma according to some exemplary embodiments;

[0044] Figure 8 depicts molecular subtypes of hepatocellular carcinoma reproduced in the GO30140 A-arm study and the IMBrave150 study according to some exemplary embodiments;

[0045] Fig. 9 depicts molecular subtypes and pathway enrichment present in The Cancer Genome Atlas (TCGA) hepatocellular carcinoma (HCC) dataset according to some exemplary embodiments;

[0046] Fig.10 depicts the expression of known hepatocellular carcinoma cell lineage marker genes in a training set according to some exemplary embodiments;

[0047] Fig.11 Depicts a higher number of regulatory T cells (Tregs) in the progenitor-like subtype according to some exemplary embodiments reg ) and effector T cells (T eff) ratio;

[0048] Fig.12 Depicted are fluorescence activated cell sorting (FACS) gating strategies for characterizing T cell infiltration and activity in xenograft tumors according to some exemplary embodiments;

[0049] Fig.13 depicts a flow chart showing an example of a process for hepatocellular carcinoma (HCC) subtype classification and treatment according to some exemplary embodiments;

[0050] Fig.14 Depicted is a block diagram illustrating an example of a computing system in accordance with some example embodiments.

[0051] When applicable, like reference numerals refer to like structures, features, or elements. DETAILED DESCRIPTION

[0052] Hepatocellular carcinoma (HCC) is a highly heterogeneous disease with complex etiologies and diverse molecular and cellular dysfunctions. Several molecular classifications of hepatocellular carcinoma have been determined based on gene expression signatures, genetic / epigenetic landscapes, and metabolic networks. Interestingly, the heterogeneity of these subtypes is characterized not only by diverse molecular signatures (including oncogenic pathways and immune cell infiltration patterns), but also by morphology and cell differentiation stages. However, the clinical relevance of these hepatocellular carcinoma subtypes, especially in the context of immunotherapy, has not been fully characterized. In fact, the treatment options for hepatocellular carcinoma (HCC) remain a disease with a poor prognosis and limited treatment options. Although combination immunotherapy, such as atezolizumab (anti-PD-L1) plus bevacizumab (anti-VEGF) combination therapy, has recently become a new standard of care for patients with unresectable hepatocellular carcinoma, not all patients benefit from these treatments. Therefore, obtaining biological insights into the heterogeneity of hepatocellular carcinoma remains crucial for understanding the response and resistance mechanisms to combination immunotherapy and identifying effective new therapeutic targets.

[0053] In some exemplary embodiments, pathological remission of treatment (e.g., major pathological remission (MPR), etc.), response to treatment, and / or one or more treatments suitable for hepatocellular carcinoma (HCC) can be determined based on the subtype of hepatocellular carcinoma present in the patient. In particular, three different molecular subtypes of hepatocellular carcinoma were identified in three separate hepatocellular carcinoma cohorts (including GO30140 Phase 1b and IMbrave150 Phase 3 trials) by transcriptomics, genomics, and in situ analysis. The three subtypes of bile duct-like, progenitor-like, and hepatocyte-like are identified and validated based on their association with different hepatic epithelial cell lineages. In addition, each subtype is associated with different tumor cell-intrinsic characteristics, tumor microenvironment (TME) characteristics, and immune landscapes. For example, the progenitor-like subtype is associated with bipotent progenitor cells, the bile duct-like subtype is associated with bile duct cells, and the hepatocyte-like subtype is associated with hepatocytes. Each subtype may exhibit different tumor cell-intrinsic characteristics, such as immunohistochemistry of cytochrome P450, expression level of cytochrome P450, Hippo signaling pathway, expression level of YES-associated protein (YAP), etc. In addition, each subtype may exhibit different tumor microenvironment (TME) characteristics, including, for example, the amount of fibroblast activation protein in the stroma, vascular density, density of cluster of differentiation 8 (CD8) in tumor epithelium, the amount of MHCI+ tumor cells, density of cluster of differentiation 8 (CD8) in tumor epithelium, density of PDL1+, density of activated T cells, density of exhausted T cells, etc.

[0054] In some exemplary embodiments, the above-mentioned biological insights into the heterogeneity of hepatocellular carcinoma can be used to formulate subtype-specific treatment strategies, including those strategies that target subtype-specific weaknesses to overcome resistance to combination immunotherapy such as atezolizumab (anti-PD-L1) plus bevacizumab (anti-VEGF) combination therapy. That is, the molecular subtype exhibited by a hepatocellular carcinoma (HCC) patient may indicate the patient's pathological response (e.g., major pathological response (MPR), etc.) or response to one or more treatments for hepatocellular carcinoma. For example, patients exhibiting a cholangioid subtype benefit the most from combination immunotherapy, while patients exhibiting a progenitor-like subtype benefit less. Therefore, in some cases, treatment of hepatocellular carcinoma (HCC) patients exhibiting a progenitor-like subtype may include a GPC3 / CD3 bispecific antibody in addition to the combination immunotherapy in order to overcome subtype-specific resistance to the combination immunotherapy associated with the progenitor-like subtype.

[0055] Figure 1 A system diagram illustrating an example of a digital pathology system 100 is depicted in accordance with some exemplary embodiments. Figure 1 , the digital pathology system 100 may include a digital pathology platform 110, an imaging system 120, and a client device 130. Figure 1As shown, the digital pathology platform 110, the imaging system 120, and the client device 130 can be communicatively connected via a network 140. The network 140 can be a wired network and / or a wireless network, including, for example, a local area network (LAN), a virtual local area network (VLAN), a wide area network (WAN), a public land mobile network (PLMN), the Internet, etc. The imaging system 120 may include one or more imaging devices (including, for example, a microscope, a digital camera, a whole slide scanner, an automated microscope, etc.). The client device 130 can be a processor-based device, including, for example, a workstation, a desktop computer, a laptop computer, a smart phone, a tablet computer, a wearable device, etc.

[0056] Reference again Figure 1 , the digital pathology platform 110 may include an analysis engine 115 configured to determine a molecular subtype of hepatocellular carcinoma exhibited by the patient based on a hepatocellular carcinoma (HCC) tumor sample of the patient. In addition, the analysis engine 115 may determine a hepatocellular carcinoma (HCC) treatment for the patient based at least on the molecular subtype of hepatocellular carcinoma exhibited by the patient. Figure 2 As shown, hepatocellular carcinoma is a heterogeneous disease in which different molecular subtypes are closely associated with liver epithelial cell lineages and unique tumor cell intrinsic characteristics, tumor microenvironment (TME) characteristics, and immune landscapes. Therefore, the analysis engine 115 can determine, based at least on the patient's hepatocellular carcinoma tumor sample, that the patient exhibits a progenitor-like subtype associated with a bipotent progenitor cell, a cholangiocyte-like subtype associated with a cholangiocyte, or a hepatocyte-like subtype associated with a hepatocyte. In the case where it is determined that the patient exhibits a cholangiocyte-like subtype or a hepatocyte-like subtype, the analysis engine 115 can determine a treatment including a combination immunotherapy (e.g., atezolizumab (anti-PD-L1) plus bevacizumab (anti-VEGF) combination therapy, etc.). Alternatively, if the analysis engine 115 determines that the patient exhibits a progenitor-like subtype, the analysis engine 115 can determine a treatment including a GPC3 / CD3 bispecific antibody in addition to the combination immunotherapy to overcome subtype-specific resistance to the combination immunotherapy.

[0057] In some exemplary embodiments, different hepatocellular carcinoma (HCC) subtypes, such as progenitor cell-like, bile duct-like and hepatocellular-like molecular subtypes, can be identified based on transcriptome data. For example, the analysis engine 115 can identify different hepatocellular carcinoma (HCC) subtypes by applying cluster analysis (such as non-negative matrix decomposition (NMF), connectivity-based clustering, centroid-based clustering, distribution-based clustering, density-based clustering, subspace-based clustering, group-based clustering, and graph-based clustering, etc.) to transcriptome data associated with various hepatocellular carcinoma tumor samples. Cluster analysis can be applied to generate the number of clusters associated with the maximum co-phenotype correlation value. In some cases, in addition to or in place of cluster analysis, the analysis engine 115 can identify different hepatocellular carcinoma (HCC) subtypes by applying a classifier (e.g., an integrated learning model such as a random forest classifier, etc.) to transcriptome data associated with various hepatocellular carcinoma tumor samples. For example, in some exemplary embodiments, while the analysis engine 115 may apply a first technique (e.g., cluster analysis) to identify different hepatocellular carcinoma (HCC) subtypes, the results associated with the first technique may be verified by the analysis engine 115 applying a second technique (e.g., a classifier).

[0058] In some exemplary embodiments, different hepatocellular carcinoma (HCC) molecular subtypes identified based on transcriptome data can also be identified based on histological features present in an image of a hepatocellular carcinoma (HCC) sample. In some cases, the image of the hepatocellular carcinoma (HCC) sample may be a whole slice image (WSI). In some cases, the image of the hepatocellular carcinoma (HCC) sample may have been processed, for example, by a stain such as hematoxylin and eosin (H&E) stain, etc. Therefore, the analysis engine 115 can identify a variety of features (including, for example, different types of tissues, cells, growth patterns, etc.) within an image (e.g., a whole slice image (WSI), etc.) of a hepatocellular carcinoma (HCC) sample.

[0059] In some exemplary embodiments, the analysis engine 115 may identify one or more regions corresponding to different types of tissue within an image of a hepatocellular carcinoma (HCC) sample. For example, the analysis engine 115 may distinguish and annotate regions corresponding to cancerous epithelial tissue, necrotic tissue, normal tissue, etc. Alternatively and / or in addition, the analysis engine 115 may distinguish and annotate various regions, such as bile ducts, benign lobules, etc. In some cases, the analysis engine 115 further identifies the growth pattern of hepatocellular carcinoma (HCC) present within the image of the hepatocellular carcinoma (HCC) sample, including, for example, trabecular, trabecular or sinusoidal, solid, pseudoacinar growth patterns, etc.

[0060] In some exemplary embodiments, the analysis engine 115 can identify one or more regions corresponding to different types of cells in an image of a hepatocellular carcinoma (HCC) sample. For example, the analysis engine 115 can distinguish and mark regions corresponding to cancer epithelial cells, fibroblasts, endothelial cells, and normal cells (e.g., non-cancerous cells). Alternatively and / or in addition, the analysis engine 115 can distinguish and mark regions corresponding to lymphocytes, plasma cells, and macrophages. In some cases, the analysis engine 115 can distinguish and mark regions corresponding to cancer epithelial cells, fibroblasts, endothelial cells, lymphocytes, plasma cells, macrophages, and normal cells (e.g., non-cancerous cells). In addition, in some cases, the analysis engine 115 can distinguish and mark different types of cancer epithelial cells, including, for example, hepatocellular carcinoma (HCC) hepatocellular cells, hepatocellular carcinoma (HCC) with Mallory hyaline bodies or globules, hepatoblastoids (e.g., smaller cells showing a nuclear-cytoplasmic ratio), etc.

[0061] In some exemplary embodiments, to identify the correlation between features present in an image of a hepatocellular carcinoma (HCC) sample and various hepatocellular carcinoma (HCC) molecular subtypes, the analysis engine 115 may subject these features to one or more clustering analyses, fractal dimension measurements, and the like. For example, the presence of certain features within an image of a hepatocellular carcinoma (HCC) sample, such as the presence of certain types of tissues, growth patterns, and cells, may indicate that the hepatocellular carcinoma (HCC) sample is positive for one of a cholangioid subtype, a hepatocellular subtype, and a progenitor-like subtype of hepatocellular carcinoma (HCC). In some cases, the molecular subtypes present in an image of a hepatocellular carcinoma (HCC) sample may be identified in whole or in part by applying a feature discrimination model, an end-to-end model (e.g., collecting a full-slice image as input), a spatial feature model (e.g., collecting a graph neural network (GNN) with cell and tissue heat maps as input), and the like.

[0062] To further illustrate, Figure 3 (a) depicts an example in which the analysis engine 115 applies non-negative matrix factorization (NMF) to transcriptome data of various hepatocellular carcinoma tumor samples to determine the optimal number of subpopulations present therein. For example, using non-negative matrix factorization (NMF), the analysis engine 115 can iteratively select the most robust clustering pattern for the transcriptome data, including the optimal number of clusters (e.g., within a configurable range, such as 2 to 8) such that the intra-cluster correlation of each resulting cluster is maximized. Figure 3 In the example shown in (a), 99 stage IVA hepatocellular carcinoma resection tissue samples were used as the training set, and the GO30140 Ph1b A group and IMbrave150 biomarker population (BEP) were used as the validation set. Figure 3(b) shows that cluster analysis (e.g., non-negative matrix factorization (NMF)) can be applied to a portion of the training set that exhibits the most variable genes. Figure 3 As shown in (b), the three clusters associated with the highest co-phenotypic correlation value (0.98) may be the optimal number of subpopulations. Figure 3 (c) shows that these clusters ( Figure 3 The non-negative matrix factorization (NMF) subtypes in the genus Hoshida highly overlap with the S1 to S3 subtypes identified by metadata consensus cluster analysis. In addition, Figure 3 (c) shows that when superimposed with metabolic subtypes iHCC1-3 previously identified by transcriptome and metabolic functional network analysis of The Cancer Genome Atlas (TCGA) hepatocellular carcinoma set, iHCC3 superimposed with the first and second subtypes NMF1 and NMF2 identified by non-negative matrix factorization, while the third non-negative matrix factorization subtype NMF3 was enriched in iHCC1 and iHCC2 subtypes with similar metabolic programs that differed from iHCC3.

[0063] Each non-negative matrix factorization (NMF) subtype identified by non-negative matrix factorization (NMF) may be associated with unique tumor-intrinsic characteristics and tumor microenvironment (TME) characteristics. These relationships can be identified by examining the association between molecular subtypes and liver epithelial lineages. For example, in the liver lobule, bipotent progenitor cells in the Hering duct can produce cholangiocytes lining the bile duct near the portal vein, or mature hepatocytes of the central vein under the guidance of molecular cues such as Notch signaling. Therefore, the liver is associated with three epithelial lineages (cholangiocytes, bipotent progenitor cells, and hepatocytes), as shown in the expression patterns of the corresponding cell lineage characteristics. It is worth noting that, as Figure 3 (c) shows that the expression of specific cell lineage features possessed by each of the subtypes identified by non-negative matrix decomposition (NMF) is significantly higher than that of other subtypes. Therefore, based on the main cell lineage features expressed by each non-negative matrix decomposition (NMF) subtype, the corresponding clusters are designated as bile duct-like subtypes, progenitor cell-like subtypes, and hepatocyte-like subtypes.

[0064] Figure 3 (d) and Figure 7 To more comprehensively discover the signaling pathways enriched in each subtype, we performed gene set enrichment analysis (GSEA). Figure 7Figure 3. Cell type enrichment analysis (e.g., using xCell) of RNA-seq data from the training set to discover signaling pathways enriched in each subtype. The progenitor-like subtype displayed a robust cell cycle transcriptional program (G2M, E2F and MYC targets, mitotic spindle signature) but had relatively mild inflammation compared to the other two subtypes, with low interferon-α and interferon-γ immune responses ( Figure 3 (d)) and fewer conventional dendritic cells (aDCs) and stromal cells ( Figure 7 In contrast, the cholangioid subtype is characterized by a high inflammatory response, such as TNF-α and interferon-γ, as well as epithelial-mesenchymal transition (EMT), a pathway driven by TGFb ( Figure 3 (d)). Deconvolution analysis including cell type enrichment analysis (e.g., using xCell) further confirmed cholangiocyte-like enrichment for both innate and adaptive immune cell types, immune scores, and cholangiocyte-like signatures ( Figure 7 Finally, the hepatocyte-like subtype was characterized by enrichment of signatures related to hepatocyte function, including bile acid and xenobiotic metabolic pathways ( Figure 3 (c)), and cell type enrichment for hepatocytes ( Figure 7 The hepatocyte-like subtype is also less proliferative, consistent with its overlap with the nonproliferative S3 subtype of Hoshida.

[0065] In some exemplary embodiments, tumor-intrinsic features and tumor microenvironment (TME) features of cell lineage-associated subtypes can be further analyzed based on expression patterns between subtypes, wherein gene expression features represent pathways identified by gene set enrichment analysis (GSEA) analysis, oncogenic signaling pathways involved in hepatocellular carcinogenesis and prognosis, markers associated with clinical outcomes of atezolizumab plus bevacizumab, and liver metabolic pathways with gene expression features. Table 1 below lists examples of gene expression features.

[0066] Table 1

[0067]

[0068]

[0069]

[0070] In addition to confirming the inflammatory immune microenvironment of the cholangioid subtype with high expression of CD274 (PD-L1 mRNA), T effector, antigen presentation, and NK cell signatures, we also determined that the cholangioid subtype exhibited high YAP / TAZ activation and fibroblast TGFb response signature (F-TBRS) ( Figure 3(e)). Interestingly, the progenitor-like subtype had re-expressed oncofetal genes, including GPC3 and AFP, as well as many imprinted genes such as DLK1, IGF2, and ZIM2 (Table 1), which indicated a lower degree of differentiation. The high expression of MYCN was consistent with the high cell cycle activity identified as progenitor-like by GSEA ( Figure 3 (d)). The hepatocyte-like subtype is characterized by significantly higher expression of pathways and signatures representing key biosynthetic and metabolic functions of well-differentiated hepatocytes, such as cytochrome 450s (CYPs) associated with xenobiotic detoxification and cellular metabolism. This subtype also has lower expression of cell cycle genes compared to the cholangioid and progenitor-like subtypes ( Figure 3 (e)).

[0071] In some exemplary embodiments, the MSigDB ( http: / / www.gsea- msigdb.org / gsea / msigdb / collections.jsp ) were used to analyze subtype-specific metabolic pathways based on the expression patterns of the signatures. Similar to metabolic subtype iHCC3, cholangioid and progenitor-like subtypes had more influx in glycolysis and fatty acid biosynthesis (FAB), but less in TCA cycle and fatty acid oxidation (FAO) compared to hepatocyte-like, which had similar metabolic dependencies as iHCC1-2 ( Figure 3 (e)).

[0072] Reference now Figure 4 (a) Genetic features associated with hepatocellular carcinoma subtypes were evaluated by FoundationOne gene panel analysis of 99 tumor tissues in the training set. No significant differences were observed between subtypes in terms of tumor mutation burden (TMB) or genomic instability, such as loss of heterozygosity (LOH) levels. On the other hand, the mutational landscape of the training set confirmed common disease-associated somatic mutations in hepatocellular carcinoma, including TP53 (58%), TERT promoter (39%), and CTNNB1 mutations (17%). Importantly, CTNNB1 mutations were primarily confined to the hepatocellular subtype (12 / 14, 86%, P value associated with the subtype was 0.0026); whereas TP53 alterations were more prevalent in cholangioid and progenitor cell-like subtypes (P = 0.022) and were mutually exclusive with CTNNB1 alterations (Fisher's exact test, P = 0.019) ( Figure 4 (a)). In conclusion, subtype-specific genetic signatures further highlight the importance of tumor cell-intrinsic features in shaping the distinct oncogenic mechanisms of HCC subtypes.

[0073] One or more in situ assays that validate subtype-specific molecular signatures are performed to further distinguish the effects of tumor cell intrinsic characteristics and tumor microenvironment (TME) characteristics in shaping HCC subtypes. Figure 4(b) shows that genes downstream of YES-associated protein (YAP) activation are highly expressed in the cholangioid subtype. YES-associated protein (YAP) activation begins with the translocation of unphosphorylated YES-associated protein (YAP) to the nucleus, and then YES-associated protein (YAP) forms a transcription factor complex with TAZ and TEAD to promote cell proliferation, dedifferentiation, and transdifferentiation. Therefore, YAP activation can be assessed by performing nuclear YES-associated protein (YAP) immunohistochemistry (IHC) analysis of 94 FFPE sections in the training set for assessing YES-associated protein (YAP) activation. Representative images and box plots summarizing the H-score of YAP staining in nuclear tumor cells are shown in Figure 2. Figure 4 (b). Figure 4 As shown in (b), the H scores of cholangioid tumors and progenitor cell-like tumors were significantly higher than those of hepatocellular tumors (P = 0.025 for cholangioid tumors vs. hepatocellular tumors and P = 0.010 for progenitor cell-like tumors vs. hepatocellular tumors, respectively).

[0074] Next, the enrichment of CYP expression in the hepatocellular-like subtype was confirmed by immunohistochemistry (IHC) assays detecting CYP2C8, CYP2C9, CYP2C18, and CYP2C19. The H-score was used to measure cytoplasmic CYP expression in tumor cells from the 90 tissue sections in the training set. Consistent with the RNA-seq, CYP expression was highest in the hepatocellular-like HCC ( Figure 4 (b)), the average tumor cell H score was 163 (0-290) ( Figure 4 (b)). Almost all hepatocellular carcinomas of the hepatocellular carcinoma type (e.g., 44 of 45 (98%)) were expressed CYP (H score > 0). In contrast, only 64% (18 / 28) of cholangioid and 82% (14 / 17) of progenitor-like subtype tumor cells had detectable CYP ( Figure 4 (b)). The CYP H score of hepatocyte-like cells was significantly higher than that of other subtypes (p values ​​for cholangiocyte-like cells vs. hepatocyte-like cells and for progenitor-like cells vs. hepatocyte-like cells were both < 0.0001) ( Figure 4 (b)).

[0075] Two multiplex immunofluorescence (IF) panels were developed to investigate the tumor-stroma-immune background (panel 1) and T cell function (panel 2) characteristics of cholangioid, progenitor-like, and hepatocyte-like subtypes in 64 baseline biopsies in the GO30140 A group. A digital pathology scoring algorithm was developed to measure the density of cells with certain phenotypes (single or multiple markers) in the epithelial tumor bed, stromal area, or total tumor area for each panel. A cocktail of antibodies against hepatocyte paraffin 1 (HepPar1) and arginase 1 (Arg1) was incorporated into both panels to label hepatocellular carcinoma tumor cells. Figure 4(c) shows representative images of combination 1 staining of all 5 markers in different molecular subtypes, and numerical scores summarizing key tumor-stroma-immune background features are shown. Figure 4 (c) also shows representative images and numerical scores of key features of combination 2 in molecular subtypes. Figure 4 (c) HepPar1 / Arg1 in tumor regions of the hepatocellular-like subtype + Cell density (per mm 2 HepPar1 / Arg1 in tumor regions + The progenitor-like tumor cells have an intermediate average density of HepPar1 / Arg1 + cell( Figure 5 (e) to (h)) and variable expression of HepPar1 / Arg1. In addition, in our dataset, the expression of two known cholangiocyte-related genes, SOX9 and KRT19, was highest in the cholangioid subtype ( Fig.10 Interestingly, even in hepatocellular carcinoma, adjacent normal liver tissue (when available) showed stronger HepPar1 / Arg1 staining than tumor cells ( Figure 5 (e) and (g)).

[0076] A combination of multiplex immunohistochemistry (IHC)1 was designed to investigate CD8 T cell infiltration, antigen presentation, TGFβ-stimulated reactive stroma, and angiogenesis in the TME, measured by antibodies against CD8, MHC-I, FAP, and CD31, respectively. eff ) and transcriptome activation of TGFβ signaling ( Figure 3 (b) In this subtype, CD8 density within the tumor area and %FAP+ area within the stromal area were also significantly higher (not significant). Combination 2 contained antibodies against PD1, GZMB, CD3, and PDL1 to investigate T cell functional status ( Figure 5 (g) and (h)). Figure 4 (B) As shown, PDL1+ cells within the tumor bed (tumor epithelium), CD3 T cells, and activated (CD3+GZMB+) and exhausted T cells (CD3+PD1+) within the tumor area were highest in the cholangioid subtype.

[0077] In some exemplary embodiments, the molecular subtypes of hepatocellular carcinoma (HCC) identified by the above-mentioned cluster analysis (e.g., non-negative matrix factorization (NMF), etc.) can be validated in an independent cohort by applying a classifier (e.g., a random forest classifier) ​​trained on a gene set of RNA sequence data from a training set. The classifier predicted the same three hepatocellular carcinoma subtypes from tumor samples of the GO30140 Phase 1b Group A (n=90) and the IMbrave150 Phase 3 hepatocellular carcinoma clinical trial (n=178). In addition, the incidence of each subtype in the training cohort and the two test cohorts was also substantially consistent (one-way ANOVA P value>0.9999) ( Figure 8 (a) to (c)), confirming the robustness of these subtypes. Consistent with the high immune presence, the majority of cholangioid samples (94% in GO30140 A group and 79% in IMbrave150) contained immune cells (IC) or tumor cells (TC) with PD-L1 (≥1%) expression. In addition, the frequencies of small nucleotide variants (SNVs) and small insertions and deletions (Indels) in CTNNB1 and TP53 genes (hcc biomarkers manuscript) were significantly higher in the three datasets and in the TCGA hepatocellular carcinoma dataset. 9 Comparison of the distribution of known demographic prognostic factors between the subtypes in the datasets showed that the proportion of well-differentiated tumors was highest in hepatocellular carcinomas. The results of this comparison are shown in Tables 2 to 4 below.

[0078] Table 2 below depicts the demographic characteristics of the molecular subtypes in the training set.

[0079] Table 2

[0080]

[0081] Table 3 below depicts the demographic characteristics and known prognostic factors for hepatocellular carcinoma of the molecular subtypes in the Ph1b GO30140 A-arm biomarker evaluable population (BEP).

[0082] Table 3

[0083]

[0084] Table 4 below depicts the demographic characteristics and known prognostic factors for HCC by molecular subtype in the BEP of the Phase III IMBrave 150.

[0085]

[0086] In some exemplary embodiments, effector T cells (T eff) and CD274 expression can be associated with a patient's response to a combination immunotherapy (e.g., atezolizumab (anti-PD-L1) and bevacizumab (anti-VEGF) combination therapy) and / or pathological response (e.g., major pathological response (MPR)), while regulatory T cells (T reg ) and effector T cells (T eff ), GPC3 expression, and AFP expression were associated with less clinical benefit of combined immunotherapy. Figure 2 As shown in (e), these genomic associations were differentially expressed in the three molecular subtypes of hepatocellular carcinoma (HCC). In fact, patients with the cholangiocarcinoma subtype in GO30140 group A had the highest number of effector T cells (T eff ) and CD274 expression and immune cell infiltration, the objective response rate (ORR) was the highest among all subtypes (49%, 18 / 37), much higher than the intention to treat (ITT) group (36%) of group A patients. In contrast, the ORR of the progenitor-like subtype was less than half of that of the intention to treat (ITT) group (17%, 2 / 12) ( Figure 5 (a) This may be due to the GPC3 and AFP genes associated with lower clinical activity of the atezolizumab (anti-PD-L1) and bevacizumab (anti-VEGF) combination therapy (HCC Master Biomarker Manuscript). The median progression-free survival (mPFS) of patients with the cholangioid subtype (9.5 months) was five times that of patients with the progenitor-like subtype (1.9 months) ( Figure 5 (b)). Patients with the hepatocellular-like subtype had a median objective response rate (ORR) of 29% (12 / 41) and had a similar median progression-free survival (mPFS) as patients with the cholangiogenic subtype (9.9 months).

[0087] The association between molecular subtypes and response and resistance to combination immunotherapy, such as atezolizumab (anti-PD-L1) and bevacizumab (anti-VEGF), was further validated in the PhIII randomized trial IMbrave150. Figure 6 (c) showed that in both cholangioid (HR, CI = 0.37, 0.15 to 0.91; P = 0.039) and hepatocellular (HR, CI = 0.41, 0.18 to 0.91; P = 0.029) subtypes, the combination of atezolizumab (anti-PD-L1) and bevacizumab (anti-VEGF) had a significantly longer overall survival (OS) benefit compared with sorafenib. In contrast, patients who exhibited the progenitor-like subtype had less benefit from the combination of atezolizumab (anti-PD-L1) and bevacizumab (anti-VEGF) in terms of progression-free survival (PFS) and overall survival (OS) compared with sorafenib ( Figure 6(c)). Both high expression of the oncofetal protein GPC3 in the progenitor-like group and the immune desert-like phenotype may contribute to resistance to the combination therapy of atezolizumab (anti-PD-L1) and bevacizumab (anti-VEGF). Interestingly, a recent single-cell RNA-seq study has shown that the oncofetal characteristics of hepatocellular carcinoma are expressed through fetal-like PLVAP + Endothelial cells and FOLR2 + Tumor-associated macrophages and TIGIT + Colocalization and interaction of Tregs orchestrate the immunosuppressive tumor microenvironment. Consistently, significantly higher conventional T cells (Tregs) were observed in the progenitor-like subtype of the GO30140 A group. reg ) and effector T cells (T eff ) ratio, and similar trends were observed in the other two datasets ( Fig.11 ).

[0088] In some exemplary embodiments, subtype-specific resistance to combination immunotherapy, such as resistance to the atezolizumab (anti-PD-L1) and bevacizumab (anti-VEGF) combination therapy associated with the progenitor-like subtype, can be overcome by developing subtype-specific therapeutic strategies to target subtype-specific vulnerabilities. For example, the addition of an anti-GPC3 / anti-CD3 bispecific antibody (ERY974) can overcome subtype-specific resistance to combination immunotherapy by inducing T cell infiltration and activation in progenitor-like tumors expressing GPC3, thereby eliciting strong anti-tumor activity. To evaluate GPC3 as a target for the progenitor-like subtype, the anti-GPC3 / anti-CD3 bispecific antibody (ERY974) was evaluated in a humanized NOG (huNOG) xenograft model generated with a hepatocellular carcinoma cell line. To select hepatocellular carcinoma cell lines for the model, 16 cell lines were classified based on hierarchical clustering of 288 classifier genes ( Fig.12 ). Among them, huH-1 and huH-7 share multiple molecular features as progenitor-like subtypes, including high expression of GPC3 15 ( Figure 6 (a)). Figure 6 (b) shows that high cell surface expression of GPC3 protein was confirmed by fluorescence activated cell sorting (FACS). The antibody binding capacity of ERY974 in huh-1 and huh-7 was 5.94x10 4 / cell and 4.85x 10 4 / cell. In contrast, Figure 6(c) GPC3 surface expression was previously below the limit of detection in SK-HEP-1. In vitro T cell-dependent cytotoxicity (TDCC) assays further confirmed that binding of ERY974 resulted in T cell-mediated cell killing in huh-1 and huh-7 lines, but not in SK-HEP-1 line ( Figure 6 (d)).

[0089] The anti-tumor activity and target inhibition pharmacodynamics of ERY974 were also investigated in vivo in huh-1 and huh-7 xenograft models. Since the anti-human CD3 arm of ERY974 does not cross-react with murine CD3, human CD34+ stem cells were injected intravenously to provide human T cells to the huh-1huNOG model. For the huh-7 model, T cells from healthy donors were expanded and activated ex vivo before entering the circulation. Figure 6 As shown in (e), ERY974 exhibited potent antitumor activity in both the progenitor-like huh-1 and huh-7huNOG models. The tumor growth inhibition index (TGI) was 109% in the huh-1 model and 105% in the huh-7 model. On the other hand, ERY974 did not exhibit antitumor activity in the bile-like SK-HEP-1 model.

[0090] To determine whether the antitumor efficacy of ERY974 in a progenitor-like xenograft model is associated with the recruitment and activity of T cells in the tumor, flow cytometry, CD3 and CD8 immunohistochemistry (IH) analysis, and nanostring gene expression were performed to examine T cell tumor infiltration and activation in huh-1 tumors after 3 days of administration of 5 mg / kg ERY974 ( Figure 6 (f)). Fluorescence activated cell sorting (FACS) analysis showed that the infiltration of CD4 and CD8 T cells in the tumor was greatly increased. In addition, the expression of T cell activation markers GZMB, PD1, and IFNγ was also significantly increased ( Figure 4 (g)). Figure 6 (h) and as shown in Table 5 below, histopathological evaluation of CD3 and CD8 IHC images further verified the T cell and myeloid cell infiltration and tumor cell killing after ERY974 treatment. In addition, the activation of infiltrating immune cells introduced by ERY974 was evident from the increased expression of characteristic genes representing T cells in different functional states, including effector T cells and exhausted T cells (Figure (i)).

[0091] Table 5 below depicts histopathological analysis of huh-1 / huNOG tumor sections.

[0092] Table 5

[0093]

[0094] *; -, not observed; 1, scattered or rare; 2, diffuse, low density or focal; 3, diffuse, high density

[0095] **; -, not observed; 1, mild; 2, moderate; 3, severe fo, focal; p, present

[0096] ***; -. Not observed; 1, mild; 2, moderate; 3, severe fo, focal; s, single cell

[0097] As noted, in some exemplary embodiments, cluster analysis can be applied to transcriptome data to identify three unique hepatocellular carcinoma (HCC) molecular subtypes that are closely associated with liver epithelial cell lineages and unique tumor-intrinsic characteristics and tumor microenvironment (TME) features. These three molecular subtypes were accurately reproduced in samples from Ph1bGO30140 and Ph 3IMbrave150 trials classified using a classifier (e.g., an ensemble learning model such as a random forest classifier). These three molecular subtypes showed different responses to combination immunotherapy such as atezolizumab (anti-PD-L1) plus bevacizumab (anti-VEGF) combination therapy, with the bile duct-like subtype having the best relief and the progenitor-like subtype showing the strongest resistance. To overcome the subtype-specific resistance associated with the progenitor-like subtype, combination immunotherapy can be administered with GPC3. In particular, in a humanized hepatocellular carcinoma xenograft mouse model, anti-GPC3 / CD3 bispecific antibody (ERY974) induced T cell infiltration and activation, thereby triggering a strong anti-tumor activity.

[0098] The association between hepatocellular carcinoma (HCC) molecular subtypes and hepatocyte lineages is not only evident at the transcriptional level but also validated by in situ protein expression of HepPar1 and Arg1, two mature hepatocyte markers that are often used to distinguish HCC from liver metastases from other organs. The hepatocyte-like subtype had the highest expression of mature hepatocyte marker genes and HepPar1 / Arg1 among the three subtypes, indicating that this subtype retains most of the molecular features of functional hepatocytes. This was further confirmed by the highly expressed liver detoxification CYP genes and proteins and fatty acid oxidation in this subtype. However, even in the hepatocyte-like subtype, adjacent cirrhotic tissue showed stronger HepPar1 / Arg1 staining than tumor tissue, suggesting that carcinogenic progression may have disrupted the differentiation and function of normal hepatocytes. Because liver damage caused by viral infection, excessive alcohol consumption, or steatosis almost always precedes the oncogenic transformation of HCC, the potential mechanism for the lineage association of the cell lineage-associated subtypes may be the dysregulation and disruption of hepatocyte differentiation, dedifferentiation, or transdifferentiation programs during liver injury repair. YES-associated protein (YAP) activation, a molecular signature of cholangioid hepatocellular carcinoma, was found to disrupt hepatocyte differentiation and shift them toward the cholangiocyte lineage. Although cell lineage tracing studies have shown that hepatocytes, rather than other epithelial cell types, are the sole origin of hepatocellular carcinoma in mouse models, the cell of origin may also be a contributor to the cell lineage heterogeneity between subtypes. A third potential mechanism for lineage heterogeneity may also be driven by tumorigenesis occurring in hepatocytes in different metabolic compartments of the liver lobule. For example, the high immune presence in cholangioids echoes the periportal concentrations of sinusoidal endothelial cells and innate and adaptive immune cells mediated by the CXCL9 cytokine gradient. Although the causal mechanisms of the lineage associations of molecular subtypes remain puzzling and require further validation in high-resolution single-cell genomic data, protein markers of known liver epithelial cell types (e.g., HepPar1, Arg1, CK19, or GPC3) can be used to design multiplex immunohistochemical assays for subtype classification.

[0099] As noted, the three molecular subtypes show different responses to combination immunotherapy such as atezolizumab (anti-PD-L1) plus bevacizumab (anti-VEGF) combination therapy. In particular, the cholangioid subtype is characterized by high immune cell infiltration and pre-existing immunity, which may explain the good response of patients with this subtype to anti-PD-L1 plus anti-VEGF combination therapy. In fact, in the GO30140 A group and the IMbrave150 Atezo+Bev group (HCC main biomarker paper), pre-existing immunity has been shown to be associated with a better response to combination immunotherapy. On the other hand, the progenitor cell-like subtype has a desert-like immune environment with high expression of oncofetal genes GPC3 and AFP and resistance to Atezo+Bev. In fact, compared with normal tissue, GPC3 is one of the top 10 significantly upregulated genes in TCGA hepatocellular carcinoma and is associated with a poor prognosis, further highlighting its importance as a new therapeutic target. Furthermore, the S2 subclass, an overlapping subclass of progenitor-like hepatocellular carcinoma, was associated with the worst prognosis, further highlighting the high unmet medical need and limited treatment options for this patient subgroup. To this end, ERY974, a bispecific antibody targeting GPC3 and CD3 T cells, demonstrated the ability to recruit and activate T cells from the periphery to GPC3-expressing tumors in a huNOG xenograft model and resulted in potent tumor cell killing. Thus, the addition of an anti-GPC3 / anti-CD3 bispecific antibody (ERY974) may overcome subtype-specific resistance to combination immunotherapy associated with the progenitor-like subtype.

[0100] Surveys of the mutational landscape identified TP53 and CTNNB1 mutations as being clearly associated with subtype. The hepatocyte-like subtype was enriched for CTNNB1 activating mutations, whereas the other two subtypes harbored more TP53 mutations. Activation of WNT / β-catenin signaling has previously been implicated in the immune desert phenotype and resistance to immunotherapy in hepatocellular carcinoma, as well as potential acquired resistance to ICIs in metastatic melanoma. However, in the IMbrave 150 study, similar survival benefits from combination immunotherapy (e.g., atezolizumab (anti-PD-L1) plus bevacizumab (anti-VEGF) combination therapy) were observed in patients carrying wild-type or mutant CTNNB1, suggesting that the addition of anti-VEGF may overcome resistance to ICIs such as atezolizumab mediated by WNT / β-catenin signaling (HCC Leading Biomarkers Paper). This hypothesis was further validated in a mouse model of ICI-resistant hepatocellular carcinoma driven by β-catenin activation, in which anti-PD-L1 plus anti-VEGF overcame resistance to ICI. This may also explain why the clinical activity of atezolizumab (anti-PD-L1) plus bevacizumab (anti-VEGF) combination therapy persisted in the hepatocyte-like subtype regardless of CTNNB1 alterations and lack of immune presence.

[0101] Fig.13 A flow chart illustrating an example of a process 1400 for hepatocellular carcinoma (HCC) subtype classification and treatment according to some exemplary embodiments is depicted. Figures 1 to 14 , the process 1400 can be performed at the digital pathology platform 110 , for example, by the analysis engine 115 .

[0102] At 1402, the analysis engine 115 can identify one or more features associated with the hepatic epithelial cell lineage for the hepatic epithelial cell lineage. As noted, hepatocellular carcinoma (HCC) is a highly heterogeneous disease with complex etiologies and diverse molecular and cellular dysfunctions. Individual hepatocellular carcinoma molecular subtypes, such as cholangioid subtypes, progenitor cell-like subtypes, and hepatocyte-like subtypes, can be identified and validated based on genetic features that are closely associated with different hepatic epithelial cell lineages. In addition, the association between hepatocellular carcinoma molecular subtypes and hepatic epithelial cell lineages can be further characterized by unique tumor-intrinsic features and tumor microenvironment (TME) features, wherein each hepatocellular carcinoma molecular subtype exhibits a unique combination of tumor-intrinsic features and tumor microenvironment features.

[0103] In some exemplary embodiments, individual hepatocellular carcinoma molecular subtypes can be identified and verified based on transcriptome data associated with different various hepatocellular carcinoma tumor samples. For example, bile duct-like, progenitor-like and hepatocyte-like subtypes can be identified by applying cluster analysis (such as non-negative matrix decomposition (NMF)), classifiers (such as random forest classifiers) to transcriptome data. In the case of cluster analysis, the analysis engine 115 can iteratively group transcriptome data to identify the most robust clustering pattern, which includes the best number of clusters, wherein the intra-cluster correlation between the members of each cluster is maximized. The resulting clusters corresponding to bile duct-like subtypes, progenitor-like subtypes and hepatocyte-like subtypes can show the largest co-phenotype correlation value (0.98).

[0104] At 1404, the analysis engine 115 may designate one or more features as representatives of molecular subtypes of hepatocellular carcinoma. In some exemplary embodiments, each molecular subtype of hepatocellular carcinoma (HCC) may be associated with different combinations of tumor cell intrinsic features and tumor microenvironment (TME) features. For example, in addition to the connection with a specific hepatic epithelial cell lineage, each subtype may exhibit different tumor cell intrinsic features, such as immunohistochemistry of cytochrome P450, expression levels of cytochrome P450, Hippo signaling pathway, expression levels of YES-associated protein (YAP), etc. Alternatively and / or in addition, each molecular subtype may also exhibit different tumor microenvironment (TME) features, including, for example, the number of fibroblast activation proteins in the matrix, vascular density, density of differentiation cluster 8 (CD8) in tumor epithelium, the number of MHCI+ tumor cells, density of differentiation cluster 8 (CD8) in tumor epithelium, density of PDL1+, density of activated T cells, density of exhausted T cells, etc.

[0105] At 1406, analysis engine 115 may receive a tumor sample from a patient. In some exemplary embodiments, analysis engine 115 may receive various data associated with a hepatocellular carcinoma (HCC) tumor sample from an imaging system 120 and / or client device 130. For example, in some cases, analysis engine 115 may receive transcriptome data associated with a hepatocellular carcinoma tumor sample. Alternatively and / or in addition, analysis engine 115 may receive one or more images (e.g., whole-slice images) of the hepatocellular carcinoma tumor sample.

[0106] At 1408, the analysis engine 115 may determine that the patient exhibits a molecular subtype associated with hepatocellular carcinoma based at least on one or more features detected in the patient's tumor sample. In some exemplary embodiments, the molecular subtype of hepatocellular carcinoma (HCC) exhibited by the patient may be determined based at least on one or more features present in the patient's hepatocellular carcinoma tumor sample. For example, in the case where the analysis engine 115 receives transcriptome data associated with a hepatocellular carcinoma tumor sample, the analysis engine 115 may apply a cluster analysis or a classifier to determine the molecular subtype of hepatocellular carcinoma (HCC) exhibited by the patient based at least on one or more genetic features present in the transcriptome data. Alternatively and / or in addition, the analysis engine 115 may receive one or more images (e.g., full-slice images) of a hepatocellular carcinoma tumor sample, in which case the molecular subtype of the patient's hepatocellular carcinoma may be determined by analyzing the one or more images to detect one or more of the tumor cell intrinsic features and / or tumor microenvironment features (TME).

[0107] At 1410, the analysis engine 115 may determine a treatment for hepatocellular carcinoma based at least on the molecular subtype of the patient. As noted, each molecular subtype of hepatocellular carcinoma may be associated with a unique immune landscape, which may give rise to a subtype-specific response to combination immunotherapies such as atezolizumab (anti-PD-L1) plus bevacizumab (anti-VEGF) combination therapy, etc. For example, the cholangioid subtype, and to a lesser extent the hepatocellular subtype, may be more responsive to atezolizumab (anti-PD-L1) plus bevacizumab (anti-VEGF) combination therapy, etc. In contrast, the progenitor-like subtype is associated with subtype-specific resistance to atezolizumab (anti-PD-L1) plus bevacizumab (anti-VEGF) combination therapy, etc.

[0108] Thus, treatment for a patient may be determined based at least on the molecular subtype present in the patient's hepatocellular carcinoma (HCC) tumor sample. In the case where a patient is identified as a molecular subtype of hepatocellular carcinoma (e.g., a cholangioid subtype or a hepatocellular subtype) that exhibits no subtype-specific resistance to combination immunotherapy, the analysis engine 115 may determine treatment including combination immunotherapy such as atezolizumab (anti-PD-L1) plus bevacizumab (anti-VEGF) combination therapy, etc. Alternatively, when a patient is identified as a molecular subtype of hepatocellular carcinoma that exhibits subtype-specific resistance to combination immunotherapy, the analysis engine 115 may determine to include additional therapy (e.g., a biological drug such as a GPC3 / CD3 bispecific antibody) in addition to combination immunotherapy such as atezolizumab (anti-PD-L1) plus bevacizumab (anti-VEGF) combination therapy to overcome subtype-specific resistance to combination immunotherapy.

[0109] Fig.14A block diagram illustrating an example of a computing system 1500 is depicted in accordance with some example embodiments. Figure 1 to 1 5. The computing system 1500 may be used to implement the digital pathology platform 110, the client device 130, and / or any components thereof.

[0110] like Fig.14 As shown, the computing system 1500 may include a processor 1510, a memory 1520, a storage device 1530, and an input / output device 1540. The processor 1510, the memory 1520, the storage device 1530, and the input / output device 1540 may be interconnected via a system bus 1550. The processor 1510 is capable of processing instructions for execution within the computing system 1500. Such executed instructions may implement, for example, one or more components of the digital pathology platform 110, the client device 130, etc. In some exemplary embodiments, the processor 1510 may be a single-threaded processor. Alternatively, the processor 1510 may be a multi-threaded processor. The processor 1510 is capable of processing instructions stored on the memory 1520 and / or the storage device 1530 to display graphical information for a user interface provided via the input / output device 1540.

[0111] Memory 1520 is a computer-readable medium, such as a volatile or non-volatile computer-readable medium, that stores information within computing system 1500. For example, memory 1520 may store a data structure representing a configuration object database. Storage device 1530 is capable of providing persistent storage for computing system 1500. Storage device 1530 may be a floppy disk device, a hard disk device, an optical disk device, a tape device, or other suitable persistent storage device. Input / output device 1540 provides input / output operations for computing system 1500. In some exemplary embodiments, input / output device 1540 includes a keyboard and / or a pointing device. In various specific implementations, input / output device 1540 includes a display unit for displaying a graphical user interface.

[0112] According to some exemplary embodiments, the input / output device 1540 may provide input / output operations for network devices. For example, the input / output device 1540 may include an Ethernet port or other networking port to communicate with one or more wired and / or wireless networks (e.g., a local area network (LAN), a wide area network (WAN), the Internet).

[0113] In some exemplary embodiments, computing system 1500 can be used to execute various interactive computer software applications that can be used to organize, analyze and / or store data in various formats. Alternatively, computing system 1500 can be used to execute any type of software application. These applications can be used to perform various functions, such as planning functions (e.g., generating, managing, editing electronic spreadsheet documents, word processing documents and / or any other objects, etc.), computing functions, communication functions, etc. The application can include various additional functions or can be an independent computing product and / or function. After activation within the application, the function can be used to generate a user interface provided via input / output device 1540. The user interface can be generated by computing system 1500 and presented to the user (e.g., on a computer screen monitor, etc.).

[0114] One or more aspects or features of the subject matter described herein can be implemented with digital electronic circuits, integrated circuits, specially designed ASICs, field programmable gate arrays (FPGAs) computer hardware, firmware, software, and / or combinations thereof. These various aspects or features can include a specific implementation in one or more computer programs, which are executable and / or interpretable on a programmable system, which includes at least one programmable processor (which can be dedicated or general, coupled to receive data and instructions from it and send data and instructions to it), a storage system, at least one input device, and at least one output device. A programmable system or computing system can include a client and a server. Typically, the client and the server are remotely arranged from each other, and generally interact through a communication network. The relationship between the client and the server is generated by means of a computer program running on each computer and a client-server relationship between each other.

[0115] These computer programs may also be referred to as programs, software, software applications, applications, components or codes, including machine instructions for programmable processors, and may be implemented in high-level procedural and / or object-oriented programming languages ​​and / or in assembly / machine languages. As used herein, the term "machine-readable medium" refers to any computer product, device and / or equipment (such as, for example, disks, optical disks, memories and programmable logic devices (PLDs)) for providing machine instructions and / or data to a programmable processor, including a machine-readable medium that receives machine instructions as machine-readable signals. The term "machine-readable signal" refers to any signal for providing machine instructions and / or data to a programmable processor. A machine-readable medium may store such machine instructions non-temporarily (such as, for example, a non-temporary solid-state memory or a magnetic hard drive or any equivalent storage medium). A machine-readable medium may store such machine instructions in a temporary manner (such as, for example, a processor cache or other random access memory associated with one or more physical processor cores) alternatively or additionally.

[0116] To provide interaction with a user, one or more aspects or features of the subject matter described herein may be implemented on a computer having a display device (such as, for example, a cathode ray tube (CRT) or a liquid crystal display (LCD) or a light emitting diode (LED) monitor for displaying information to a user) and a keyboard and a pointing device (such as, for example, a mouse or trackball, through which a user can provide input to the computer). Other types of devices may also be used to provide interaction with a user. For example, the feedback provided to the user may be any form of sensory feedback, such as, for example, visual feedback, auditory feedback, or tactile feedback; input from the user may be received in any form, including sound, voice, or tactile input. Other possible input devices include touch screens or other touch-sensitive devices, such as single-point or multi-point resistive or capacitive tracking pads, voice recognition hardware and software, optical scanners, optical pointers, digital image capture devices and associated interpretation software, and the like.

[0117] Example

[0118] The examples provided are:

[0119] 1. A computer-implemented method comprising:

[0120] for a hepatic epithelial cell lineage, identifying one or more characteristics associated with said hepatic epithelial cell lineage;

[0121] designating the one or more features as representative of a molecular subtype associated with hepatocellular carcinoma (HCC);

[0122] receiving a tumor sample from the patient; and

[0123] The patient is determined to exhibit a molecular subtype associated with hepatocellular carcinoma based on the one or more features detected within the tumor sample of the patient.

[0124] 2. The system of embodiment 1, wherein the one or more characteristics include genetic characteristics.

[0125] 3. The system of embodiment 2, wherein the operations further comprise:

[0126] A plurality of molecular subtypes associated with hepatocellular carcinoma (HCC) are identified based at least on transcriptome data associated with a plurality of hepatocellular carcinoma (HCC) tissue samples.

[0127] 4. The system of embodiment 3, wherein the plurality of subtypes are identified by applying cluster analysis to the transcriptome data to identify the number of subpopulations present within the transcriptome data.

[0128] 5. The system of embodiment 4, wherein cluster analysis is applied to identify one or more subpopulations associated with the greatest co-phenotypic correlation values.

[0129] 6. A system according to any one of embodiments 4 to 5, wherein the cluster analysis includes non-negative matrix factorization (NMF).

[0130] 7. A system according to any one of embodiments 4 to 6, wherein the clustering analysis includes one or more of connectivity-based clustering, centroid-based clustering, distribution-based clustering, density-based clustering, subspace-based clustering, group-based clustering, and graph-based clustering.

[0131] 8. The system of any one of embodiments 3 to 7, wherein the plurality of subtypes are identified and / or validated by applying a classifier to the transcriptome data.

[0132] 9. The system of embodiment 8, wherein the classifier comprises a random forest classifier.

[0133] 10. The system of any one of embodiments 1 to 9, wherein the one or more characteristics include at least one of tumor cell intrinsic characteristics and tumor microenvironment characteristics.

[0134] 11. The system of any one of embodiments 1 to 10, wherein the one or more features comprise immunohistochemistry of cytochrome P450, expression levels of cytochrome P450, Hippo signaling pathway, and / or expression levels of YES-associated protein (YAP).

[0135] 12. A system according to any one of embodiments 1 to 11, wherein the one or more characteristics include the number of fibroblast activation proteins in the matrix, the vascular density, the density of cluster of differentiation 8 (CD8) in the tumor epithelium, the number of MHCI+ tumor cells, the density of cluster of differentiation 8 (CD8) in the tumor epithelium, the density of PDL1+, the density of activated T cells, and / or the density of exhausted T cells.

[0136] 13. The system of any one of embodiments 1 to 12, wherein the molecular subtype associated with hepatocellular carcinoma comprises a cholangioid subtype, and wherein the hepatic epithelial cell lineage comprises a cholangiocyte.

[0137] 14. The system of any one of embodiments 1 to 13, wherein the molecular subtype associated with hepatocellular carcinoma comprises a hepatocyte-like subtype, and wherein the liver epithelial cell lineage comprises a hepatocyte.

[0138] 15. The system of any one of embodiments 1 to 14, wherein the molecular subtype associated with hepatocellular carcinoma comprises a progenitor-like subtype, and wherein the hepatic epithelial cell lineage comprises a bipotent progenitor cell.

[0139] 16. The system of any one of embodiments 1 to 15, wherein the operations further comprise:

[0140] A treatment for hepatocellular carcinoma (HCC) is determined based at least on the molecular subtype of the patient.

[0141] 17. The system of embodiment 16, wherein the treatment for hepatocellular carcinoma (HCC) comprises a combination immunotherapy based at least on whether the patient has a cholangio-like subtype or a hepatocellular-like subtype.

[0142] 18. The system of any one of embodiments 16 to 17, wherein the treatment for hepatocellular carcinoma (HCC) comprises a combination therapy of atezolizumab (anti-PD-L1) plus bevacizumab (anti-VEGF) based at least on whether the patient has a cholangioid subtype or a hepatocellular subtype.

[0143] 19. A system according to any one of embodiments 16 to 18, wherein treatment for hepatocellular carcinoma (HCC) includes one or more additional therapies based at least on the patient having a progenitor-like subtype to overcome subtype-specific resistance to combination immunotherapy associated with the progenitor-like subtype.

[0144] 20. The system of any one of embodiments 16 to 19, wherein the treatment for hepatocellular carcinoma (HCC) comprises a GPC3 / CD3 bispecific antibody in addition to the combination immunotherapy based at least on the patient having a progenitor-like subtype.

[0145] 21. A system according to any one of embodiments 1 to 20, wherein the one or more features include cancerous epithelial tissue, necrotic tissue and / or normal tissue present in the image of the tumor sample.

[0146] 22. A system according to any one of embodiments 1 to 21, wherein the one or more features include a growth pattern present in an image of the tumor sample.

[0147] 23. A system according to any one of embodiments 1 to 22, wherein the one or more features include one or more cancer epithelial cells, fibroblasts, endothelial cells, and normal cells present in the image of the tumor sample.

[0148] 24. A system according to any one of embodiments 1 to 23, wherein the one or more features include one or more hepatocellular carcinoma (HCC) hepatocellular carcinoma epithelial cells, hepatocellular carcinoma (HCC) carcinoma epithelial cells with Mallory hyaline bodies or globules, and hepatocellular carcinoma (HCC) hepatoblastoid carcinoma epithelial cells.

[0149] 25. The system of any one of embodiments 1 to 24, further comprising:

[0150] The patient's response to treatment and / or pathological response is determined based at least on the molecular subtype of the patient.

[0151] 26. A system comprising:

[0152] at least one data processor; and

[0153] At least one memory storing instructions that, when executed by the at least one data processor, result in operations including the method according to any one of embodiments 1 to 25.

[0154] 27. A non-transitory computer readable medium storing instructions that, when executed by at least one data processor, result in operations including those according to any one of embodiments 1 to 25.

[0155] In the above description and claims, phrases such as "at least one" or "one or more" may appear, followed by a list of combinations of elements or features. The term "and / or" may also appear in a list of two or more elements or features. Unless otherwise implicitly or explicitly contradicted by the context in which it is used, the phrase is intended to represent any element or feature listed alone, or any other described element or feature combined with any other described element or feature. For example, the phrase "at least one of A and B"; "one or more of A and B"; "A and / or B" are each intended to represent "single A, single B, or A and B together". A similar interpretation also applies to lists including three or more items. For example, the phrases "at least one of A and B and C"; "one or more of A, B and C" and "A, B and / or C" are each intended to represent "single A, single B, single C, A and B together, A and C together, B and C together, or A and B and C together". The use of the term "based on" above and in the claims is intended to represent "based at least in part", so that undescribed features or elements are also permissible.

[0156] Depending on the desired configuration, the subject matter described herein may be embodied in systems, devices, methods and / or articles. The embodiments described in the foregoing description do not represent all embodiments consistent with the subject matter described herein. Instead, they are only some examples consistent with aspects related to the described subject matter. Although some variations have been described in detail above, other modifications or additions are possible. In particular, in addition to those features and / or variations described herein, other features and / or variations may also be provided. For example, the above-mentioned specific implementations may be directed to various combinations and sub-combinations of the disclosed features and / or to combinations and sub-combinations of several further features disclosed above. In addition, the logical flows depicted in the drawings and / or described herein do not necessarily require the specific order or sequential order shown to achieve the desired results. Other specific implementations may be within the scope of the following claims.

Claims

1. A computer-implemented method comprising: for a hepatic epithelial cell lineage, identifying one or more characteristics associated with said hepatic epithelial cell lineage; designating the one or more features as representative of a molecular subtype associated with hepatocellular carcinoma (HCC); receiving a tumor sample from the patient; and The patient is determined to exhibit a molecular subtype associated with hepatocellular carcinoma based on the one or more features detected within the tumor sample of the patient.

2. The system of claim 1, wherein the one or more characteristics include genetic characteristics.

3. The system of claim 2, wherein the operations further comprise: A plurality of molecular subtypes associated with hepatocellular carcinoma (HCC) are identified based at least on transcriptome data associated with a plurality of hepatocellular carcinoma (HCC) tissue samples. 4 . The system of claim 3 , wherein a plurality of subtypes are identified by applying a cluster analysis to the transcriptome data to identify the number of subpopulations present within the transcriptome data.

5. The system of claim 4, wherein the cluster analysis is applied to identify one or more subpopulations associated with maximum co-phenotype correlation values.

6. The system according to any one of claims 4 to 5, wherein the cluster analysis comprises non-negative matrix factorization (NMF).

7. A system according to any one of claims 4 to 6, wherein the clustering analysis comprises one or more of connectivity-based clustering, centroid-based clustering, distribution-based clustering, density-based clustering, subspace-based clustering, group-based clustering and graph-based clustering.

8. The system according to any one of claims 3 to 7, wherein a plurality of subtypes are identified and / or validated by applying a classifier to the transcriptome data.

9. The system of claim 8, wherein the classifier comprises a random forest classifier.

10. The system according to any one of claims 1 to 9, wherein the one or more characteristics include at least one of tumor cell intrinsic characteristics and tumor microenvironment characteristics.

11. The system of any one of claims 1 to 10, wherein the one or more features comprise immunohistochemistry of cytochrome P450, expression levels of cytochrome P450, Hippo signaling pathway and / or expression levels of YES-associated protein (YAP).

12. A system according to any one of claims 1 to 11, wherein the one or more characteristics include the amount of fibroblast activation protein in the stroma, the vascular density, the density of cluster of differentiation 8 (CD8) in the tumor epithelium, the amount of MHCI+ tumor cells, the density of cluster of differentiation 8 (CD8) in the tumor epithelium, the density of PDL1+, the density of activated T cells, and / or the density of exhausted T cells.

13. The system of any one of claims 1 to 12, wherein the molecular subtype associated with hepatocellular carcinoma comprises a cholangioid subtype, and wherein the hepatic epithelial cell lineage comprises a cholangiocyte.

14. The system of any one of claims 1 to 13, wherein the molecular subtype associated with hepatocellular carcinoma comprises a hepatocyte-like subtype, and wherein the liver epithelial cell lineage comprises a hepatocyte.

15. The system of any one of claims 1 to 14, wherein the molecular subtype associated with hepatocellular carcinoma comprises a progenitor-like subtype, and wherein the hepatic epithelial cell lineage comprises a bipotent progenitor cell.

16. The system of any one of claims 1 to 15, wherein the operations further comprise: A treatment for hepatocellular carcinoma (HCC) is determined based at least on the molecular subtype of the patient.

17. The system of claim 16, wherein the treatment for hepatocellular carcinoma (HCC) comprises combination immunotherapy based at least on the patient having a cholangio-like subtype or a hepatocellular-like subtype.

18. The system of any one of claims 16 to 17, wherein the treatment for hepatocellular carcinoma (HCC) comprises a combination therapy of atezolizumab (anti-PD-L1) plus bevacizumab (anti-VEGF) based at least on the patient having a cholangioid subtype or a hepatocellular subtype.

19. A system according to any one of claims 16 to 18, wherein the treatment for hepatocellular carcinoma (HCC) includes one or more additional therapies based at least on the patient having a progenitor-like subtype to overcome subtype-specific resistance to combination immunotherapy associated with the progenitor-like subtype.

20. The system of any one of claims 16 to 19, wherein the treatment for hepatocellular carcinoma (HCC) comprises a GPC3 / CD3 bispecific antibody in addition to combination immunotherapy based at least on the fact that the patient has a progenitor-like subtype.

21. The system of any one of claims 1 to 20, wherein the one or more features include cancerous epithelial tissue, necrotic tissue and / or normal tissue present in the image of the tumor sample.

22. The system of any one of claims 1 to 21, wherein the one or more features include a growth pattern present in an image of the tumor sample.

23. The system of any one of claims 1 to 22, wherein the one or more features include one or more cancerous epithelial cells, fibroblasts, endothelial cells, and normal cells present in the image of the tumor sample.

24. The system of any one of claims 1 to 23, wherein the one or more features include one or more hepatocellular carcinoma (HCC) hepatocellular-like carcinoma epithelial cells, hepatocellular carcinoma (HCC) carcinoma epithelial cells with Mallory hyaline bodies or globules, and hepatocellular carcinoma (HCC) hepatoblastoid carcinoma epithelial cells.

25. The system according to any one of claims 1 to 24, further comprising: The patient's response to treatment and / or pathological response is determined based at least on the molecular subtype of the patient.

26. A system comprising: at least one data processor; as well as At least one memory storing instructions which, when executed by said at least one data processor, result in operations including the method according to any one of claims 1 to 25.

27. A non-transitory computer-readable medium storing instructions that, when executed by at least one data processor, result in operations including those according to any one of claims 1 to 25.

Images