WO2020092623A1 - Comprehensive characterization of immune landscape in gastrointestinal cancers and head and neck cancers via computational deconvolution - Google Patents

Abstract

Contemplated systems and methods are directed to classification of gastrointestinal (GI) tumors using omics data, and especially RNAseq data. Notably, GI tumors including non-colorectal tumors (e.g., gastroesophageal, head and neck, or biliary cancers) may be classified using immune-related geneset, tumor location, and consensus molecular subtype (CMS) classification to elucidate potentially therapeutically targetable features of these GI tumors.

Description

COMPREHENSIVE CHARACTERIZATION OF IMMUNE LANDSCAPE IN GASTROINTESTINAL CANCERS AND HEAD AND NECK CANCERS VIA COMPUTATIONAL DECONVOLUTION

[0001] This application claims priority to our co-pending U.S. Provisional Application No. with serial number 62/753,855, filed on October 31, 2018, the entire content of which is herein incorporated by reference.

Field of the Invention

[0002] The field of the invention is omics analysis of cancers, and especially as it relates to immune related gene sets and location of various gastrointestinal (GI) cancers.

Background of the Invention

[0003] The background description includes information that may be useful in understanding the present invention. It is not an admission that any of the information provided herein is prior art or relevant to the presently claimed invention, or that any publication specifically or implicitly referenced is prior art.

[0004] All publications and patent applications herein are incorporated by reference to the same extent as if each individual publication or patent application were specifically and individually indicated to be incorporated by reference. Where a definition or use of a term in an incorporated reference is inconsistent or contrary to the definition of that term provided herein, the definition of that term provided herein applies and the definition of that term in the reference does not apply.

[0005] Despite the relatively simple categorization of various cancers (typically by location of the tumor), there is a high degree of variability among cancers with respect to progression, immune therapy, and response to therapy. Unfortunately, there are no simple single markers that could help stratify a tumor into one or another class and several approaches have been undertaken to do so. With the availability of various omics data, analyses to arrive at a meaningful result are often complicated and time consuming.

[0006] Furthermore, immune contexture profoundly shapes tumorigenesis, tumor progression, and response to therapy. Traditional approaches for investigating the tumor microenvironment, such as immunohistochemistry and flow cytometry are often limited by relatively low throughput. More recently, computational methods using transcriptomic datasets have enabled comprehensive deconvolution of the immune landscape in solid tumors, particularly colorectal cancer (CRC). However, the immune contextures of other non-CRC gastrointestinal and head and neck cancers remain to be clarified.

[0007] Consensus molecular subtype (CMS) classification of CRC tumors has been used as a predictive tool for chemotherapeutic efficacy against metastatic colorectal cancer using specific drug treatments. Some of these studies established that patients with CMS 1 subtype CRC, and particularly those with a high microsatellite instability (MSI-H) genotype were the suitable candidates for immunotherapy, which led to the regulatory approval of the checkpoint inhibitors nivolumab and pembrolizumab. Unfortunately, MSI-H only represents a small fraction of patients with metastatic CRC (J Cancer Metastasis Treat 20l8;4:28). Moreover, these studies did not classify non-CRC gastrointestinal cancers.

[0008] Therefore, while numerous manners of omics data processing are known in the art, all or almost all of them, suffer from one or more disadvantages. Thus, there is still a need to provide improved analyses that allow for classification of non-CRC GI tumors, particularly with regard to the immune status of the tumor.

Summary of The Invention

[0009] The inventive subject matter is directed to methods of analyzing, classifying, and/or treating gastrointestinal (GI) tumors. In particular, while colorectal cancer (CRC) tumors are also included, the inventive subject matter was rendered to obtain analysis of non-CRC GI tumors.

[0010] More specifically, the contemplated subject matter includes a method of classifying a gastrointestinal (GI) tumor, including providing omics data of the GI tumor, assigning an immune-related group to the GI tumor, and assigning a consensus molecular subtype (CMS) to the GI tumor selected from CMS1, CMS2, CMS3, or CMS4.

[0011] In typical embodiments, the GI tumors in the database are first separated into one of four GI types: 1) colorectal (CRC) group, 2) gastroesophageal (GE) group, 3) head, neck, and oral (H&N) group, and biliary (other) group. [0012] The omics data for each GI tumor may be genomics data, transcriptomics data, or proteomics data. Preferably, the omics data is transcriptomics data. More preferably, the transcriptomics data is whole transcriptomics sequencing data (RNA-Seq). Most preferably, the RNA-Seq data have at least 200xl0⁶ reads per GI tumor.

[0013] For assigning the immune -related group to the GI tumor or GI tumor group, in typical embodiments, this designation includes unsupervised clustering. This unsupervised clustering renders a designation in Cluster 1 which is characterized as immunologically “hot,” or in Cluster 2 which is characterized as immunologically“cold.” More specifically, Cluster 1 is defined by an increase of natural killer (NK) cells and regulatory T cells (T-regs), and Cluster 2 is defined by an increase in eosinophils and fibroblasts.

[0014] In typical embodiments, the assigning of the immune-related group to the GI tumor includes analysis of the omics data using an immune -related geneset. More typically, the immune-related geneset includes genes selected from dendritic cells (DCs), T helper 1 cells (Thl), Helper T Cells, monocytes, natural killer (NK) CD56dim cells, mast cells, cytotoxic lymphocytes, immature dendritic cells (iDCs), eosinophils, CD8 T cells, regulatory T cells (Treg), T helper 2 cells (Th2), fibroblasts, NK CD56 bright cells, T helper 17 cells (Thl7), effector memory T cells (TEM), B cell lineage cells, neutrophils, T cells, endothelial cells, natural killer (NK) cells, T follicular helper cells (TFH), myeloid dendritic cells, central memory T cells (TCM), adipose dendritic cells (aDC), gamma delta T cells (Tgd), and/or plasmacytoid dendritic cells (pDCs).

[0015] In additional embodiments, the assigning of the immune-related group to the GI tumor includes quantifying or obtaining expression levels for a plurality of distinct genes, wherein the distinct genes are associated with respective distinct types of immune cells, and quantifying a level of expression for each of the distinct genes relative to respective reference ranges for Cluster 1 and/or Cluster 2.

In additional and/or alternative embodiments, a method of treating a gastrointestinal (GI) tumor, includes providing whole transcriptomic sequencing (RNA-Seq) data of the GI tumor, assigning an immune classification of immunologically hot or immunologically cold to the GI tumor, and assigning a colorectal consensus molecular subtype (CMS) to the GI tumor selected from CMS1, CMS2, CMS3, or CMS4, and treating the GI tumor based on whether the immune classification and the CMS type of the GI tumor. In particular embodiments, when the immune classification is immunologically hot (Cluster 1), the treating of the GI tumor comprises administering a treatment to the patient of immune checkpoint inhibitor therapy.

[0016] Various objects, features, aspects and advantages of the inventive subject matter will become more apparent from the following detailed description of preferred embodiments, along with the accompanying drawing figures in which like numerals represent like components.

Brief Description of The Drawing

[0017] Fig. 1A is an exemplary graph showing distribution of the indicated GI cancer type present in one set of whole transcriptomic sequence (RNA-Seq) data of gastrointestinal (GI) tumors used for analyses disclosed herein.

[0018] Fig. IB is an exemplary graph showing the distribution of age of the patient for each of the GI tumors of Fig. 1A after grouping of the GI tumors into one of four GI types— colorectal, head and neck, gastroesophageal (GE), and biliary (other).

[0019] Fig. 1C is an exemplary graph showing a log distribution of tumor mutational burden (TMB) for each of the GI tumors of Fig. 1A after grouping of the GI tumors into one of four GI types— colorectal, head and neck, gastroesophageal (GE), and biliary (other).

[0020] Fig. 2 depicts Z-scores for the indicated cancer types— head and neck (H&N), other (biliary), colorectal cancer (CRC), and gastroesophageal (GE).

[0021] Fig. 3 is an exemplary graph showing an immune-related geneset heatmap for the indicated cancer types— head and neck (H&N), other (biliary), colorectal cancer (CRC), and gastroesophageal (GE) relative to immune-related cluster type (Cluster 1 (hot) and Cluster 2 (cold), as indicated) and consensus molecular subtype (CMS).

[0022] Fig. 4 is an exemplary graph showing expression levels of the indicated checkpoint genes for the indicated cancer types of from left to right for each checkpoint gene, the cancer types are colorectal cancer (CRC), gastroesophageal (GE), other (biliary), and head and neck (H&N) as labeled for the first [0023] Fig. 5 is an exemplary table and graph depicting CMS classification for non-CRC GI tumors including head and neck (H&N) and gastroesophageal (GE) cancers as indicated.

Detailed Description

[0024] The inventors have now discovered that upper and lower gastrointestinal (GI) tumors are distinct in their tolerated immune-cell infiltration. Therefore, immune-oncological (IO) therapies may be tailored based on location to take advantage of the innate immune apparatus present in the tumor environment. Viewed from a different perspective, the inventors discovered that elucidating and comparing the immune environments in these cancers can reveal various actionable biological insights and potentially therapeutic opportunities, and as such serve to guide treatment of GI tumors— in particular, non-colorectal cancer (non-CRC) tumors.

[0025] Accordingly, the inventive subject matter is directed to methods of analyzing, classifying, and/or treating gastrointestinal (GI) tumors. Considering transcriptomic datasets have not been used to deconvolute the immune landscape in non-CRC GI tumors, the inventive subject matter was rendered to obtain the lacking analysis, classification, and guided treatment of non-CRC GI tumors. These presently disclosed methods for analyzing and classifying GI tumors include obtaining and/or using existing omics data of GI tumors to render an omics database of GI tumors.

[0026] In more specific embodiments, a GI tumor omics database may be obtained or an existing GI tumor database or combination of existing GI tumor databases may be used. A GI tumor database includes any omics database from which transcriptomics database can be derived. For example, a genomics, transcriptomics, or proteomics database may be used. In typical embodiments a whole transcriptomics sequencing (RNA-Seq) data of a set of GI tumors is used. In more typical embodiments, the RNA-Seq data includes at least 200xl0⁶ reads per GI tumor.

[0027] With a suitable GI database selected, each of the GI tumors in the database are first separated into one of four GI types— 1) colorectal (CRC), 2) gastroesophageal (GE), 3) head, neck, and oral (H&N), and 4) biliary (other) cancers. For example, each of 464 GI tumors in an exemplary GI database of whole transcriptomic sequencing (RNA-Seq) was grouped by tumor location as shown in Fig. 1A, with the patient’s age and tumor mutational burden (TMB) shown for each tumor group in Figs. IB and 1C, respectively. [0028] Notably, while colorectal consensus molecular subtyping (CMS) has been utilized to classify CRC tumors, the inventive subject matter includes methods of assigning a CMS type to non-CRC GI tumors. The four established CMS types include CMS1, CMS2, CMS3, and CMS4, using the CMS classifier as disclosed in Guinney et al., 2015, Nature Medicine, 21:1350-1356. More specifically, CMS1 is the microsatellite instability immune (MSI) subtype, distinguished by hypermutation, unstable microsatellite, and strong immune activation; CMS2 is the canonical subtype, distinguished by epithelial cells and marked WNT and MYC signaling activation; CMS3 is the metabolic subtype, distinguished by epithelial cells and evident metabolic dysregulation; and CMS4 is the mesenchymal subtype, distinguished by prominent transforming growth factor-beta activation, stromal invasion and angiogenesis. Accordingly, methods for classifying GI tumors include using omics data (e.g., RNA-Seq) to assign a CMS type to each tumor of the four groups.

[0029] While the CMS subtyping provides a level of characterization and information for non-CRC GI tumors, CMS subtyping may not necessarily provide an understanding of the local immune microenvironment of the non-CRC GI tumor. Accordingly, embodiments of the inventive subject matter may also include computational immune deconvolution of the GI tumors. To this end, omics data of GI tumors are analyzed using a comprehensive immune- related geneset resulting in immune-related cluster groups. This combination of CMS subtyping and immune deconvolution (e.g., immune-related clustering) advantageously provides a further stratification of non-CRC GI tumors.

[0030] Typically, the classifying of the GI tumor into an immune-related cluster group includes unsupervised clustering based on the RNA-Seq data of the GI tumor. For example, with reference to Figs. 2-5, unsupervised clustering analysis of each of the four GI tumor groups (CRC, GE, H&N, and biliary) provided two moderately distinct clusters of GI immune-cell activity scores using unsupervised clustering. These two groups were designated Cluster 1 and Cluster 2, where Cluster 1 is characterized as immunologic ally “hot” having an increase of natural killer (NK) cells and regulatory T cells (T-regs) and Cluster 2 is characterized as immunologically“cold” having an increase in eosinophils and fibroblasts. More typically, the increase of NK cells and T-regs in Cluster 1 and the increase in eosinophils and fibroblasts in Cluster 2 are both defined as an increase when the measured transcript data (RNA-Seq) exceeds + 2 standard deviations (SD) of a reference range for Cluster 1 or Cluster 2. Accordingly, each of the four GI groups can be assigned an immunologically“hot” or immunologically“cold” designation.

[0031] Preferably, the immune -related geneset includes genes selected from dendritic cells (DCs), T helper 1 cells (Thl), Helper T Cells, monocytes, natural killer (NK) CD56dim cells, mast cells, cytotoxic lymphocytes, immature dendritic cells (iDCs), eosinophils, CD8 T cells, regulatory T cells (Treg), T helper 2 cells (Th2), fibroblasts, NK CD56 bright cells, T helper 17 cells (Thl7), effector memory T cells (TEM), B cell lineage cells, neutrophils, T cells, endothelial cells, natural killer (NK) cells, T follicular helper cells (TFH), myeloid dendritic cells, central memory T cells (TCM), adipose dendritic cells (aDC), gamma delta T cells (Tgd), and/or plasmacytoid dendritic cells (pDCs).

[0032] As used herein, an established GI database includes an omics database of GI tumors from which transcriptomics (RNA-Seq) data is used for performing both the CMS subtyping of each GI tumor and immune deconvolution using the immune-related genesets. These two analytical methods are implemented independently using the same GI omics database. As such, the CMS subtyping analysis is performed in parallel with the immune deconvolution, and is therefore, a wholly separate analysis. View from another perspective, the immune deconvolution is not performed with a bias of any particular CMS subtype and the CMS subtyping is not performed with a bias for an immunologically hot or cold cluster group.

[0033] With respect to the grouping of the GI tumors into one of four groups— CRC, GE, H&N, or biliary, in typical embodiments, this grouping and separation of the GI tumors (e.g., the omics data for each GI tumor) occurs prior to the CMS subtyping and immune deconvolution analysis. As used herein, a GI tumor includes any cancer tumor along the gastrointestinal tract including head and neck cancers. Grouping of the GI tumors is based on anatomical location. As disclosed herein, typical embodiments group the GI tumors as CRC, GE, H&N, or biliary. Most typically, the CRC group includes colon, rectal, and anal tumors; the GE group includes stomach and esophageal tumors, the H&N group includes head, neck, thyroid, throat, and oral (mouth) tumors; and the biliary (other) group includes tumors of the small intestine, pancreas, gallbladder, liver, and intrahepatic bile duct.

[0034] The combination analysis of CMS subtyping and immune deconvolution of GI tumors as disclosed herein may be performed on any GI tumor omics database from which RNA-Seq data may be derived. Furthermore, for a specific tumor, the corresponding omics data for the specific tumor may be analyzed with respect to an established GI tumor database. To this end, as understood by one of skill in the art, the omics data may be comparatively incorporated using a classification model (e.g., k-nearest neighbor (k-NN) or nearest centroid classifier).

[0035] In particular, methods for classifying non-CRC GI tumors include using omics data to assign a CMS type and therefore the relevant features for each tumor and GI tumor group to obtain further understanding of potential targeting therapies for GE, H&N, and biliary cancers as well as more effectively eliminating therapies that may not be effective.

[0036] With reference to Figs. 3-5, the presently disclosed method of CMS subtyping and immune deconvolution was implemented with an exemplary GI tumor transcriptomics database. As shown and calculated herein, despite slightly higher average tumor mutational burden (TMB), colorectal cancer (CRC) was found to have“cold” or“colder” (e.g., less) immune-activity (e.g., lower expression of immune -related genes) than other GI tumor locations across almost all immune cell types. As such, TMB does not represent direct evidence of immunogenicity and does not accurately predict the dynamic immune response. More specifically, CRC cancers have significantly lower PDL1 and PDL2 (PDL1/2) expression, while head and neck (H&N)) cancers (including oral and throat cancers) have relatively high checkpoint (e.g., PDL1/2) expression. Checkpoint gene expression for each checkpoint gene of indoleamine-2, 3-dioxygenase (IDO), TIM3, LAG3, PDL1, PDL2, and CTLA4 is quantified in Fig. 4 for each of the CRC, GE, other, and H&N cancer types as indicated from left to right. Accordingly, H&N cancers showed“hotter” immune- activity (e.g., higher expression levels of immune -related genes) compared to CRC.

[0037] With continued reference to Figs. 3-5, upper GI tumors (H&N and GE) were significantly enriched in Cluster 1 defined by high natural killer (NK) cells and regulatory T cells (T-regs) (adj. p < l.lxlO ⁴). The upper GI tumors were also most often classified as the immune-active CMS1 subtype (adj. p <2.9xl0⁴). Conversely, CRC tumors were significantly associated with the other immune-type cluster defined by particularly high eosinophils and fibroblasts (adj. p < 7.3xl0 ⁷) and designated Cluster 2. Biliary tumors spanned both immune-type clusters (Clusters 1 and 2) and are significantly classified as the stromal-rich subtype CMS4 (adj. p <2.lxl0 ⁵). Eosinophils are more active across all GI locations than in other tumor types. Immature dendritic cells are often excluded from GI tumors, especially the GE location. Therefore, it should be appreciated that upper and lower GI tumors are distinct in their tolerated immune-cell infiltration. Consequently, and advantageously, immune-oncological therapies may be tailored based on Cluster group assignment to take advantage of the innate immune apparatus present.

[0038] It should still further be appreciated that contemplated methods and analyses of both CMS subtyping and immune deconvolution are useful individually and in combination for determining a suitable treatment where location may provide a contributing factor. As noted herein, upper and lower GI tumors are distinct in their tolerated immune cell infiltration. Immune therapies should therefore be tailored based on location to take advantage of the innate immune apparatus present. Specifically, upper GI cancers appear especially fit for checkpoint therapy despite having lower average TMB.

Examples

[0039] To elucidate the inventive subject matter, the inventors used a commercial GI data set containing 592 clinical samples of which the following omic assay data were available: 592 Germline and somatic variant calls (annotated with Clinvar) including 592 mutational signature profiles; 464 RNAseq; and 326 targeted mass-spec (~30 markers). Clinical covariates included 592 TMB values, 575 cancer types, 447 ICD10 codes, and 464 CMS types. A distribution of cancer types, patient age, and tumor mutational burden for the 464 GI tumors are shown in Figs.lA-lC. The immune infiltrate dataset was determined from 464 GI cancers from clinical set with RNAseq. The data were assigned to 4 groups with the indicated sample size (N): CRC: (CRC, anal) N=l78; GE: (stomach, esophageal), N=69; H&N (head, neck, thyroid, oral, and throat cancers), N=60; and biliary/other (small intestine, pancreas, gallbladder, liver, intrahepatic bile duct, etc.) N=l57.

[0040] A curated panel of 122 genes that accurately discriminated between 28 immune cell subpopulations was identified. A database of 1880 unselected tumor samples used to define a distribution of expression values for each of these 28 immune cell signatures. The study samples were then scored for their deviances within these distributions. In addition to immune cell deconvolution, all 464 GI patients were assigned a Colorectal Consensus Molecular Subtypes (CMS) type, regardless of whether they were CRC tumors. Significant enrichment for immune cell subpopulations between locations and CMS types was analyzed.

[0041] Using the above samples, an immune-related genesets heatmap was prepared and Figs.2-5 depict exemplary results. Here, row-scaled RNAseq TPMs were used for genes specific to immune-cell types (annotated on left). Locations and a k-means clustering (k=2) are annotated at the top. As can be taken from the heatmap, the location within the GI tract and cluster assignment (Cluster 1 and Cluster 2) are related. Fig. 3 provides immune scores. Here, inferred‘activity’ of each cell type is shown by comparing geneset expression to all clinical samples. Clustering is based off of this space. Cluster 1 is enriched for NK (natural killer cells), Treg (regulatory T cells) and T-helper cells, and Cluster 2 is enriched for Fibroblasts, Eosinophils, and Mast cells. Cluster 2 may also have a subset of immuno-active samples that are more enriched than Cluster 1 for iDC/DC (immature dendritric cells (iDC) and dendritic cells (DC) and Myeloid dendritic cells. Cluster 1 is also significantly enriched for H&N and GE cancers (Fishers exact one-sided, Benjmini-Hochberg (BH) procedure with adjusted false discovery rate (FDR)). Cluster 2 is significantly enriched for CRC. Table 1 below shows various comparisons of clusters versus location.

[0042] Table 1.

[0043] Fig. 2 is a summary based on GI locations. Here, the average inferred‘activity’ of each cell type within GI locations is shown (left), along with the percentage of times that reach statistically significantly high (middle) and low (right). As can be readily seen from the graph, eosinophils are especially high in all GI compared to the rest. H&N have higher NK infiltration than other GI types. iDCs are often excluded from the tumor in non-H&N tumors, and especially in GE tumors.

[0044] Finally, the inventors investigated whether or not there is an association between CMS and location enrichment. To that end, a one-sided Fishers exact tests for enrichment of CMS types and location was used, however, limited to those results that were significant (p<0.05) after BH FDR adjustment. As can be readily seen from the Table 2 below and Fig. 5, CMS1 is significantly enriched for H&N and GE, while CMS2 and CMS3 are significantly enriched for CRC, and CMS4 is significantly enriched for the biliary (‘other’) category.

[0045] Table 2

[0046] As used in the description herein and throughout the claims that follow, the meaning of“a,”“an,” and“the” includes plural reference unless the context clearly dictates otherwise. Also, as used in the description herein, the meaning of“in” includes“in” and“on” unless the context clearly dictates otherwise. Unless the context dictates the contrary, all ranges set forth herein should be interpreted as being inclusive of their endpoints, and open-ended ranges should be interpreted to include commercially practical values. Similarly, all lists of values should be considered as inclusive of intermediate values unless the context indicates the contrary.

[0047] Moreover, all methods described herein can be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context. The use of any and all examples, or exemplary language (e.g.“such as”) provided with respect to certain embodiments herein is intended merely to better illuminate the invention and does not pose a limitation on the scope of the invention otherwise claimed. No language in the specification should be construed as indicating any non-claimed element essential to the practice of the invention.

[0048] Groupings of alternative elements or embodiments of the invention disclosed herein are not to be construed as limitations. Each group member can be referred to and claimed individually or in any combination with other members of the group or other elements found herein. One or more members of a group can be included in, or deleted from, a group for reasons of convenience and/or patentability. When any such inclusion or deletion occurs, the specification is herein deemed to contain the group as modified thus fulfilling the written description of all Markush groups used in the appended claims.

[0049] It should be apparent to those skilled in the art that many more modifications besides those already described are possible without departing from the inventive concepts herein. The inventive subject matter, therefore, is not to be restricted except in the scope of the appended claims. Moreover, in interpreting both the specification and the claims, all terms should be interpreted in the broadest possible manner consistent with the context. In particular, the terms“comprises” and“comprising” should be interpreted as referring to elements, components, or steps in a non-exclusive manner, indicating that the referenced elements, components, or steps may be present, or utilized, or combined with other elements, components, or steps that are not expressly referenced. Where the specification claims refers to at least one of something selected from the group consisting of A, B, C ... and N, the text should be interpreted as requiring only one element from the group, not A plus N, or B plus N, etc.

Claims

CLAIMS What is claimed is:

1. A method of classifying a gastrointestinal (GI) tumor, comprising: providing omics data of the GI tumor; assigning an immune-related group to the GI tumor; and assigning a consensus molecular subtype (CMS) to the GI tumor selected from CMS1, CMS2, CMS3, or CMS4.

2. The method of claim 1, wherein the omics data is selected from genomics, transcriptomics, or proteomics.

3. The method of claim 2, wherein the transcriptomics is whole transcriptomics sequencing (RNA-Seq).

4. The method of claim 1, wherein the providing omics data of the GI tumor comprises separating the omics data for each tumor into one GI group selected from a colorectal cancer (CRC) group, a head and neck (H&N) group, a gastroesophageal (GE) group, or a biliary group.

5. The method of claim 4, wherein the CRC group includes colon, anal, and rectal tumors, the H&N group includes head, neck, throat, oral, and thyroid tumors, the GE group includes stomach and esophageal tumors, and the biliary group includes small intestine, pancreas, gallbladder, liver, and intrahepatic bile duct tumors.

6. The method of claim 1, wherein the assigning of the immune -related cluster group comprises unsupervised clustering based on the RNA-Seq data on the GI tumor.

7. The method of claim 1, wherein the assigning of the immune -related group comprises unsupervised clustering analysis.

8. The method of claim 1, wherein the at least one immune-related group is selected from Cluster 1 and Cluster 2, wherein Cluster 1 is defined by an increase of natural killer (NK) cells and regulatory T cells (T-regs) and Cluster 2 is defined by an increase in eosinophils and fibroblasts.

9. The method of claim 8, wherein the increase of NK cells and T-regs in Cluster 1 and the increase in eosinophils and fibroblasts in Cluster 2 are both defined as an increase when the increase exceeds + 2 standard deviations (SD) of a reference range for Cluster 1 or Cluster 2.

10. The method of any of claims 1-9, wherein the assigning the immune -related group to the GI tumor comprises using an immune -related geneset.

11. The method of any one of claims 1-10, wherein the assigning an immune -related group includes using an immune -related geneset comprised of genes selected from dendritic cells (DCs), T helper 1 cells (Thl), Helper T Cells, monocytes, natural killer (NK) CD56dim cells, mast cells, cytotoxic lymphocytes, immature dendritic cells (iDCs), eosinophils, CD8 T cells, regulatory T cells (Treg), T helper 2 cells (Th2), fibroblasts, NK CD56 bright cells, T helper 17 cells (Thl7), effector memory T cells (TEM), B cell lineage cells, neutrophils, T cells, endothelial cells, natural killer (NK) cells, T follicular helper cells (TFH), myeloid dendritic cells, central memory T cells (TCM), adipose dendritic cells (aDC), gamma delta T cells (Tgd), and/or plasmacytoid dendritic cells (pDCs).

12. The method of any one of claims 1-11, wherein the omics data of the GI tumor is RNA-Seq data having at least 200x10⁶ reads per GI tumor.

13. The method of claim 1, wherein the assigning the immune -related group to the GI tumor comprises:

quantifying or obtaining expression levels for a plurality of distinct genes, wherein the distinct genes are associated with respective distinct types of immune cells; and quantifying a level of expression for each of the distinct genes relative to respective reference ranges for Cluster 1 and/or Cluster 2.

14. A method of treating a gastrointestinal (GI) tumor, comprising: providing whole transcriptomic sequencing (RNA-Seq) data of the GI tumor; assigning an immune classification of immunologically hot or immunologic ally cold to the GI tumor; and assigning a colorectal consensus molecular subtype (CMS) to the GI tumor selected from CMS1, CMS2, CMS3, or CMS4; and treating the GI tumor based on whether the immune classification and the CMS type of the GI tumor.

15. The method of claim 14, wherein the treating of the GI tumor comprises administering an immune checkpoint inhibitor therapy when the immune classification is immunologically hot.

16. The method of claim 14, wherein the GI tumor is a non-colorectal cancer (non-CRC) GI tumor.

17. The method of any one of claims 14-16, wherein the assigning of the immune classification comprises unsupervised clustering based on the RNA-Seq data on the GI tumor.

18. The method of any one of claims 14-17, wherein immunologically hot is defined by an increase of natural killer (NK) cells and regulatory T cells (T-regs) and immunologically cold is defined by an increase in eosinophils and fibroblasts.

19. The method of claim 18, wherein the increase of NK cells and T-regs in Cluster 1 and the increase in eosinophils and fibroblasts in Cluster 2 are both defined as an increase when the increase exceeds + 2 standard deviations (SD) of a referenced range.

20. The method of any of claims 14-19, wherein the assigning of the immune classification comprises analysis with an immune -related geneset.

21. The method of any one of claims 14-19, wherein the assigning of the immune classification comprises analysis with an immune -related geneset, wherein the immune -related geneset includes genes selected from dendritic cells (DCs), T helper 1 cells (Thl), Helper T Cells, monocytes, natural killer (NK) CD56dim cells, mast cells, cytotoxic lymphocytes, immature dendritic cells (iDCs), eosinophils, CD8 T cells, regulatory T cells (Treg), T helper 2 cells (Th2), fibroblasts, NK CD56 bright cells, T helper 17 cells (Thl7), effector memory T cells (TEM), B cell lineage cells, neutrophils, T cells, endothelial cells, natural killer (NK) cells, T follicular helper cells (TFH), myeloid dendritic cells, central memory T cells (TCM), adipose dendritic cells (aDC), gamma delta T cells (Tgd), and/or plasmacytoid dendritic cells (pDCs).

22. The method of any one of claims 14-21, wherein the RNA-Seq data comprises 200xl0⁶ reads per GI tumor.