Methods for predicting lymphoma responsiveness to drugs and methods for treating lymphoma - Patents.com

Abstract

Provided herein is a method of predicting a lymphoma patient's responsiveness to a cancer treatment, comprising clustering the patients into patient subgroups using gene expression levels. Also provided herein is a method of treating a lymphoma patient based on a prediction of the lymphoma patient's responsiveness to a cancer treatment.

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority to U.S. Provisional Patent Application No. 63 / 331,725, filed April 15, 2022, the entire disclosure of which is incorporated herein by reference.

[0002] Sequence Listing This application contains a computer readable sequence listing submitted herewith in XML file format, the entire disclosure of which is incorporated herein by reference. The sequence listing XML file submitted herewith is entitled "14247-722-228_SEQLISTING.xml", was created on Mar. 29, 2023, and is 4,143 bytes in size.

[0003] 1. Provided herein are methods for predicting a lymphoma patient's responsiveness to a cancer treatment. Also provided herein are methods for treating a lymphoma patient based on a prediction of the lymphoma patient's responsiveness to a cancer treatment. [Background technology]

[0004] 2. Non-Hodgkin's lymphoma (NHL) is a diverse group of blood cancers that includes all types of lymphoma except Hodgkin's lymphoma. Types of NHL vary greatly in severity, ranging from indolent to very aggressive. Less aggressive non-Hodgkin's lymphoma allows for long-term survival, while aggressive non-Hodgkin's lymphoma can be rapidly fatal without treatment. They can form from either B cells or T cells. B-cell non-Hodgkin's lymphomas include Burkitt's lymphoma, chronic lymphocytic leukemia / small lymphocytic lymphoma (CLL / SLL), diffuse large B-cell lymphoma, follicular lymphoma, immunoblastic large cell lymphoma, precursor B-lymphoblastic lymphoma, and mantle cell lymphoma. T-cell non-Hodgkin's lymphomas include mycosis fungoides, anaplastic large cell lymphoma, and precursor T-lymphoblastic lymphoma. Prognosis and treatment depend on the stage and type of disease.

[0005] Diffuse large B-cell lymphoma (DLBCL) accounts for approximately one-third of non-Hodgkin's lymphomas. Some DLBCL patients are cured with conventional chemotherapy, while the rest die from the disease. Anticancer drugs can cause rapid and sustained depletion of lymphocytes, in some cases by directly inducing apoptosis in mature T and B cells. See Non-Patent Document 1.

[0006] Diffuse large B-cell lymphoma (DLBCL) can be classified into different molecular subtypes of germinal center B-cell-like DLBCL (GCB-DLBCL), activated B-cell-like DLBCL (ABC-DLBCL), and primary mediastinal B-cell lymphoma (PMBL) or unclassified type according to their genetic profiling patterns. These subtypes are characterized by distinct differences in survival rate, chemotherapy response, and signaling pathway dependency, especially the NF-κB pathway. See Non-Patent Document 2; Non-Patent Document 3; Non-Patent Document 4. Such differences have prompted the search for more effective and subtype-specific therapeutic strategies in DLBCL.

[0007] Current cancer treatments may generally include surgery, chemotherapy, hormone therapy, and / or radiation therapy to eradicate the patient's tumor cells (see, for example, Non-Patent Document 5). Recently, cancer treatments may also include biological therapy and immunotherapy. All of these approaches pose significant disadvantages to the patient. For example, surgery may be contraindicated or unacceptable to the patient due to the patient's health condition. In addition, surgery may not completely remove the tumor tissue. Radiation therapy is only effective if the neoplastic tissue is more sensitive to radiation than normal tissue. Radiation therapy can also often cause serious side effects. Hormonal therapy is rarely administered as a single agent. Although hormone therapy can be effective, it is often used to prevent or delay the recurrence of cancer after other treatments have removed the majority of the cancer cells. Biological and immunotherapies are limited in number and may cause side effects such as rashes and swelling, flu-like symptoms such as fever, chills, and fatigue, gastrointestinal problems, or allergic reactions.

[0008] In the context of DLBCL, treatment usually involves the combined administration of chemotherapy and antibody therapy. The most widely used treatment for DLBCL is a combination of the antibody rituximab (Rituxan) and chemotherapy drugs cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP), sometimes with the addition of etoposide (R-EPOCH). DLBCL also typically requires treatment soon after diagnosis, depending on how rapidly the disease can progress. In some patients, DLBCL relapses or becomes refractory after treatment. Several alternative treatments, which may include the use of lenalidomide, are currently being tested in clinical trials for newly diagnosed, relapsed, or refractory DLBCL patients. See Non-Patent Document 6.

[0009] Among non-Hodgkin's lymphomas, there are also types of indolent lymphoma. Indolent B-cell lymphomas include, for example, follicular lymphoma, small lymphocytic lymphoma; nodal marginal zone B-cell lymphoma, lymphoplasmacytic lymphoma, anaplastic large cell lymphoma, primary cutaneous type, and mycosis fungoides. The clinical course of patients with indolent B-cell lymphoma is characterized by, among other features, the risk of histological transformation (HT) into aggressive lymphoma (mainly DLBCL) and, less commonly, Burkitt's lymphoma (BL) or other types of aggressive lymphoma. See Non-Patent Document 7.

[0010] Due to the clinical and biological heterogeneity of lymphomas, particularly DLBCL, there is a great need for an effective method to classify DLBCL subtypes for administering specific treatments. DLBCL has traditionally been classified by cell of origin (COO) subcategory based on tumor gene expression profiles, including activated B cell (ABC) and germinal center B cell (GCB) subtypes. See Non-Patent Document 8; Non-Patent Document 9; Non-Patent Document 10. GCB and ABC subtypes have different pathogenic mechanisms that may affect the outcome of targeted therapy in DLBCL patients. See Non-Patent Document 11; Non-Patent Document 12; Non-Patent Document 13; Non-Patent Document 14; Non-Patent Document 15.

[0011] Recently, new classification models have focused on DNA alterations using tumor samples from patients treated with R-CHOP. See, for example, Non-Patent Document 16. However, a comprehensive integrated approach using transcriptome data from both newly diagnosed (nd) and relapsed / refractory (r / r) DLBCL has yet to be achieved. The present invention satisfies these and other needs. [Prior art documents] [Non-patent literature]

[0012] [Non-Patent Document 1] Stahnke et al.,Blood,2001,98:3066-3073

Non-licensed Document 2

Non-licensed Document 4

Non-licensed Document 5

Non-licensed Document 6

Non-licensed Document 7

Non-licensed literature 9

Non-licensed literature 10

Non-licensed Document 11

Non-licensed Document 12

Non-licensed Document 13

[0013] 3. In one aspect, provided herein is a method of predicting responsiveness of a lymphoma patient to a cancer treatment, the method comprising: (a) clustering reference lymphoma patients of a reference patient group into subgroups using an expression level of at least one gene in a reference biological sample of the reference lymphoma patients; (b) determining a subgroup to which the lymphoma patient belongs based on the expression level of at least one gene in the biological sample of the lymphoma patient; and (c) predicting the responsiveness of the lymphoma patient to a first cancer treatment based on the subgroup of lymphoma patients.

[0014] In some embodiments, the method further comprises administering to the lymphoma patient a second cancer treatment.

[0015] In some embodiments, step (a) of the method further comprises generating clustering information defining a relationship between expression levels of at least one gene in the reference biological sample, and reorganizing the heatmap representation based on the clustering information.

[0016] In some embodiments, step (a) of the method uses a hierarchical or non-hierarchical method, hi another aspect, step (a) uses the iClusterPlus method.

[0017] In some embodiments, the reference lymphoma patients are clustered into 2-12 subgroups.

[0018] In some embodiments, the reference lymphoma patients are clustered into seven subgroups.

[0019] In some embodiments, the method further comprises training a classifier model using the expression level of at least one gene in the reference biological sample.

[0020] In some embodiments, the at least one gene is selected from the genes in Table 1.

[0021] In some embodiments, the at least one gene comprises five or more genes in Table 1.

[0022] In some embodiments, the at least one gene includes all of the genes in Table 1.

[0023] In some embodiments, the classifier model is a grouped multinomial generalized linear model (GLM).

[0024] In some embodiments, the classifier model is a binary model.

[0025] In some embodiments, the method further comprises setting a threshold confidence level for at least one of the subgroups of step (a) to exclude patients providing clustering data of lower confidence from at least one subgroup.

[0026] In some embodiments, the lymphoma is selected from the group consisting of diffuse large B-cell lymphoma (DLBCL), indolent B-cell lymphoma, follicular lymphoma, small lymphocytic lymphoma, nodal marginal zone B-cell lymphoma, lymphoplasmacytic lymphoma, anaplastic large cell lymphoma, primary cutaneous lymphoma, mycosis fungoides, chronic lymphocytic leukemia, and mantle cell lymphoma.

[0027] In some embodiments, the lymphoma is DLBCL.

[0028] In some embodiments, the lymphoma is indolent B-cell lymphoma, nodal marginal zone B-cell lymphoma, mantle cell lymphoma, or chronic lymphocytic leukemia.

[0029] In some embodiments, the reference patients of the reference patient group are clustered into subgroups A1 to A7: (ii) subgroup A2 includes about 80% to about 90% patients with GCB DLBCL, about 0% to about 5% patients with ABC DLBCL, about 15% to about 25% patients with TME+ DLBCL, and about 25% to about 35% patients with DHITsig+ DLBCL; (iii) subgroup A3 includes about 40% to about 55% patients with GCB DLBCL, about 30% to about 40% patients with ABC DLBCL, about 10% to about 20% patients with TME+ DLBCL, and about 30% to about 40% patients with DHITsig+ DLBCL; (iv) subgroup A4 includes about 25% to about 35% of patients with GCB DLBCL, about 40% to about 50% of patients with ABC DLBCL, about 30% to about 40% of patients with TME+ DLBCL, and about 10% to about 20% of patients with DHITsig+ DLBCL; (v) subgroup A5 includes about 20% to about 40% of patients with GCB DLBCL, about 45% to about 65% of patients with ABC DLBCL, about 30% to about 40% of patients with TME+ DLBCL, and about 0% to about 10% of patients with DHITsig+ DLBCL; (vi) subgroup A6 includes about 25% to about 35% of patients with GCB DLBCL, about 40% to about 50% of patients with ABC DLBCL, about 30% to about 40% of patients with TME+ DLBCL, and about 0% to about 10% of patients with DHITsig+ DLBCL. and (vii) subgroup A7, which includes about 0% to about 10% of patients with GCB DLBCL, about 80% to about 90% of patients with ABC DLBCL, about 0% to about 10% of patients with TME+ DLBCL, and about 0% to about 15% of patients with DHITsig+ DLBCL.

[0030] In some embodiments, the first cancer treatment is a combination treatment with rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP).

[0031] In some embodiments, if the lymphoma patient is determined to belong to subgroup A1, the method includes predicting that the patient is unlikely to respond to the first cancer treatment.

[0032] In some embodiments, if the lymphoma patient is determined to belong to subgroup A2, the method includes predicting that the patient is unlikely to respond to the first cancer treatment.

[0033] In some embodiments, if the lymphoma patient is determined to belong to subgroup A3, the method includes predicting that the patient is unlikely to respond to the first cancer treatment.

[0034] In some embodiments, if the lymphoma patient is determined to belong to subgroup A4, the method includes predicting that the patient is unlikely to respond to the first cancer treatment.

[0035] In some embodiments, if the lymphoma patient is determined to belong to subgroup A5, the method includes predicting that the patient is unlikely to respond to the first cancer treatment.

[0036] In some embodiments, if the lymphoma patient is determined to belong to subgroup A6, the method includes predicting that the patient is unlikely to respond to the first cancer treatment.

[0037] In some embodiments, if the lymphoma patient is determined to belong to subgroup A7, the method includes predicting that the patient is unlikely to respond to the first cancer treatment.

[0038] In some embodiments, the second cancer treatment is R-CHOP.

[0039] In some embodiments, the second cancer treatment is not R-CHOP.

[0040] In some embodiments, the second cancer treatment is a bromodomain and extraterminal (BET) inhibitor or a cyclin-dependent kinase (CDK) inhibitor.

[0041] In one aspect, provided herein is a method of predicting a lymphoma patient's responsiveness to cancer treatment, the method comprising the steps of: (a) determining an expression level of at least one gene of Table 1 in a biological sample of the lymphoma patient; and (b) comparing the expression level of the at least one gene of step (a) with an expression level of at least one gene in a reference biological sample from a reference lymphoma patient, wherein if the reference lymphoma patient responds to the cancer treatment, and the expression level of the at least one gene in the biological sample is similar to the expression level of the at least one gene in the reference biological sample, indicates that the lymphoma patient is unlikely to respond to the cancer treatment.

[0042] In some embodiments, the at least one gene comprises five or more genes in Table 1.

[0043] In one aspect, provided herein is a method of predicting responsiveness of a lymphoma patient to a cancer treatment, the method comprising the steps of: (a) determining an expression level of at least one gene of Table 1 in a biological sample of the lymphoma patient; and (b) comparing the expression level of the at least one gene in the biological sample with: (i) an expression level of the at least one gene in a biological sample of a lymphoma patient that responds to the cancer treatment, and (ii) an expression level of the at least one gene in a biological sample of a lymphoma patient that does not respond to the cancer treatment, wherein if the expression level of (a) is similar to the expression level of (i), it indicates that the first lymphoma patient is more likely to respond to the cancer treatment; and if the expression level of (a) is similar to the expression level of (ii), it indicates that the first lymphoma patient is less likely to respond to the cancer treatment.

[0044] In one aspect, provided herein is a method of treating a lymphoma patient, the method comprising: (i) identifying a lymphoma patient likely to respond to a cancer treatment; and (ii) administering a cancer treatment to the lymphoma patient.

[0045] In one aspect, provided herein is a method of treating a lymphoma patient, the method comprising: (i) identifying a lymphoma patient who is unlikely to respond to a cancer treatment; and (ii) administering an alternative cancer treatment to the lymphoma patient.

[0046] In some embodiments, the cancer treatment is R-CHOP.

[0047] In some embodiments, the alternative cancer treatment is a BET inhibitor or a CDK inhibitor.

[0048] In some embodiments, the lymphoma is selected from the group consisting of DLBCL, indolent B-cell lymphoma, follicular lymphoma, small lymphocytic lymphoma, nodal marginal zone B-cell lymphoma, lymphoplasmacytic lymphoma, anaplastic large cell lymphoma, primary cutaneous lymphoma, mycosis fungoides, chronic lymphocytic leukemia, and mantle cell lymphoma.

[0049] In some embodiments, the lymphoma is DLBCL.

[0050] In some embodiments, the lymphoma is DLBCL, indolent B-cell lymphoma, follicular lymphoma, nodal marginal zone B-cell lymphoma, mantle cell lymphoma, or chronic lymphocytic leukemia.

[0051] In some embodiments, the expression levels of all of the genes in Table 1 are determined in (a) and compared in (b), as described herein.

[0052] In some embodiments, the biological sample is a tumor biopsy sample.

[0053] In some embodiments, determining the expression level of the at least one gene comprises detecting the presence or amount of at least one complex in the biological sample, wherein the presence or amount of the at least one complex is indicative of the expression level of the at least one gene.

[0054] In some embodiments, at least one complex is a hybridization complex.

[0055] In some embodiments, at least one of the complexes is detectably labeled.

[0056] In some embodiments, determining the expression level of the at least one gene comprises detecting the presence or amount of at least one reaction product in the biological sample, wherein the presence or amount of the at least one reaction product is indicative of the expression level of the at least one gene.

[0057] In some embodiments, at least one reaction product is detectably labeled.

[0058] In some embodiments, the reference lymphoma patient is a refractory DLBCL patient, a relapsed DLBCL patient, or a newly diagnosed DLBCL patient.

[0059] In some embodiments, the lymphoma patient is a refractory DLBCL patient, a relapsed DLBCL patient, or a newly diagnosed DLBCL patient.

[0060] In some embodiments, the lymphoma patient is a GCB DLBCL patient or an ABC DLBCL patient.

[0061] In some embodiments, the lymphoma patient is a DHITsig+ DLBCL patient or a DHITsig- DLBCL patient. [Brief description of the drawings]

[0062] [Figure 1][Figure 1A-E] Figure 1A-E show that unsupervised clustering was applied to a large cohort of patient-derived RNAseq data to identify biologically homogenous segments of DLBCL. Figure 1A shows a schematic of the data transformation, unsupervised clustering, and classifier training methods. Figure 1B shows a co-clustering frequency heat map that identified sample clusters that were consistently grouped together across repeated subsampling runs. Figure 1C shows the top 50 up- and down-regulated genes per cluster. Figure 1D shows the top 50 up- and down-regulated genes per cluster in the independent cohort MER. Figure 1E shows the top 50 up- and down-regulated genes per cluster in the independent cohort REMoDL-B. [Diagram 2] [Figures 2A-I] Figures 2A-2I show clinical outcomes stratified by cluster. Figure 2A shows event-free survival (EFS) of RCHOP-treated ABC-COO patients in the ROBUST subset of the Discovery cohort. Figure 2B shows EFS of RCHOP-treated patients in the MER replication cohort. Figure 2C shows progression-free survival (PFS) of RCHOP-treated patients in the REMoDL-B replication cohort. Figures 2D-2F show EFS / PFS of ROBUST (Figure 2D), MER (Figure 2E), and REMoDL-B (Figure 2F) cohorts classified as A7 / non-A7. Figures 2G-2I show forest plots of log odds ratios of belonging to A7 given the presence of various clinical factors in the ROBUST (Figure 2G), MER (Figure 2H), and REMoDL-B (Figure 2I) cohorts. [Diagram 3][Figure 3A-E] Figure 3A-E show the biological characteristics of the discovered subtypes. Figure 3A shows a scatter plot of the Discovery cohort in the space of Reddy COO score vs. TME26 score. Figure 3B shows a collection of curated DLBCL signatures showing cluster-discriminating signals. Figure 3C shows cluster-associated single nucleotide variants (SNVs) (ROBUST). Figure 3D shows cluster-associated copy number variants (CNVs) (ROBUST). Figure 3E shows representative immunohistochemistry (IHC) images of A6 and A7. [Figure 4] [Figure 4A-G] Figure 4A-Figure 4G show the biological characteristics of A7. Figure 4A shows significantly dysregulated hallmark pathways in A7 (Discovery) ranked by the p-values ​​shown with Normalized Enrichment Scores (NES). Figure 4B shows MYC gene expression by A7 status among cohorts. Figure 4C shows representative MYC staining in A7. Figure 4D shows copy number amplification / deletion frequency in A7. Figure 4E shows Western blot showing expression of TCF4 in ABC-like DLBCL cell lines. Figure 4F shows Western blot showing expression of MYC and TCF4 in TCF4 knockdown ABC-like DLBCL cell lines. GAPDH was used as a loading control for the Western blot. Figure 4G shows cell proliferation assay of control (shNT) and TCF4 knockdown (shTCF4) in ABC-like DLBCL cell lines. Error bars represent the SEM of technical triplicates (SEM<1 not shown). [Diagram 5] [Figure 5A-B] Figure 5A and Figure 5B show the clinical utility of A7. A7 was predictive of outcome in RCHOP-treated patients in both ROBUST (Figure 5A) and REMoDL-B (Figure 5B), but was less accurate in predicting outcome in R2CHOP (R-CHOP with lenalidomide) or RBCHOP (R-CHOP with bortezomib)-treated patients. [Figure 6][Figures 6A-D] Figures 6A-6D show that samples clustered in A8 have significantly lower alignments to coding regions (Figure 6A) and significantly higher proportions of intergenic reads (Figure 6B), ribosomal reads (Figure 6C), and unaligned reads (Figure 6D) than other samples (Discovery). [Figure 7] Figure 7 shows the confusion matrix of the classifier output on the training dataset (Discovery). [Figure 8] [Figure 8A-C] Figures 8A-8C show survival probability for Robust (Figure 8A), Mer (Figure 8B), and REMoDL-B (Figure 8C). A7 stratified risk within clinically defined International Prognostic Index (IPI) groups. [Figure 9] [Figure 9A-B] Figure 9A and Figure 9B show the significantly dysregulated pathways in each of the discovered clusters when comparing each cluster to all other clusters (Discovery) (Figure 9A). Figure 9B shows pathway enrichment scores that were largely consistent between the Discovery and MER datasets, with most pathways sharing directionality and significance (shown in red). Cluster A4 is an exception to this trend, with many pathways showing opposite directionality between Discovery and MER. [Figure 10] [Figure 10A-F] Figure 10A-Figure 10F show the copy number aberration prevalence in each cluster compared to the rest of the population (ROBUST+MER). Figure 10A shows A1 DLBCL vs. non-A1 (ROBUST+MER combination), Figure 10B shows A2 DLBCL vs. non-A2 (ROBUST+MER combination), Figure 10C shows A3 DLBCL vs. non-A3 (ROBUST+MER combination), Figure 10D shows A4 DLBCL vs. non-A4 (ROBUST+MER combination), Figure 10E shows A5 DLBCL vs. non-A5 (ROBUST+MER combination), and Figure 10F shows A6D DLBCL vs. non-A6 (ROBUST+MER combination). [Figure 11][FIG. 11A-F] FIG. 11A-FIG. 11F show the major immune types detected by multiple ion beam imaging (MIBI) in a portion of the ROBUST plus cohort (n=43). Cell abundance was expressed as a percentage of total nucleated cells within the field of view (FOV). FIG. 11A shows CD3 total T cells; FIG. 11B shows CD4 T cells; FIG. 11C shows CD8 T cells; FIG. 11D shows CD163 macrophages / monocytes; FIG. 11E shows CD68 macrophages; and FIG. 11F shows CD11c dendritic cells. [Figure 12] [Figure 12A-E] Figure 12A-E show the genomic characteristics of A7. Figure 12A shows A7-related genomic events and their associated variant allele frequency (VAF) and cancer cell fraction (CCF). A7-related CNAs tended to be highly clonogenic, with 100% CCF in most samples. On the other hand, most A7 mutation events were observed at least partially or exclusively among subclones. Figure 12B shows A7-related CNAs (MER). Figure 12C shows MYC expression by A7 status (MER). Figure 12D shows tumor purity by A7 status (ROBUST). Figure 12E shows tumor purity that showed near-zero correlation with MYC expression (ROBUST). [Figure 13] [Figures 13A-B] Figures 13A and 13B show TCF4 mRNA expression in patients (Figure 13A) or DLBCL cell lines (Figure 13B) that exhibited TCF4 copy number alterations (ROBUST). [Figure 14] [Figure 14A-B] Figure 14A and Figure 14B show a comparison of A7 and MCD in Novel clusters with LymphGen (NCI cohort). Figure 14A shows PFS of patients treated with immunochemotherapy, stratified by MCD and A7 status. Figure 14B shows a Sankey plot illustrating the co-occurrence of LymphGen clusters with A1-A7. [Figure 15][Figure 15A-D] Figure 15A-D show Sankey plots and associated confusion matrices of the discovered subtypes compared to the LymphGen classifier output in ROBUST (Figure 15A and Figure 15C) and MER (Figure 15B and Figure 15D). The MCD subtype (based on the co-occurrence of MYD88L265P and CD79B mutations as described in Schmitz et al., 2018, New England Journal of Medicine, 378(15), 1396-1407) was increased in A7 patients, and the EZB subtype (based on EZH2 mutations and BCL2 translocations as described in Schmitz et al., 2018, New England Journal of Medicine, 378(15), 1396-1407) was increased in GCB-like clusters A2 and A3. Although statistically significant associations existed between classification methods (Fisher p=0.0005 for ROBUST and p=0.001 for MER), there was a lot of heterogeneity and no clear one-to-one mapping between any of the subtypes. [Figure 16] Figure 16 shows the novel clusters compared to LymphGen (NCI cohort). PCA plots of the Discovery, MER, and REMoDL-B datasets after normalization did not reveal any dataset-specific differences. [Figure 17] [Figures 17A-C] Figures 17A-C show the mutation landscape (Chapuy genes), sorted by mutation number (Figure 17A), significance (corrected for gene length) (Figure 17B), and Chapuy diagram (for reference) (Figure 17C). [Figure 18][Figure 18A-D] Figure 18A-D show expression of proteins encoded by genes on chromosome 18. Figure 18A shows expression by copy number amplification on Chr18. Figure 18B shows Western blots showing expression of MTAP in DLBCL cell lines. Figure 18C shows Western blots showing expression of SDMA and PRMT5 in DLBCL cell lines. GAPDH, loading control. Figure 18D shows cell proliferation assay of control (shNT) and PRMT5 knockdown (sh PRMT5) in ABC-like DLBCL cell lines. Error bars represent SEM of technical triplicates. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0063] 5. 5.1.Definition As used herein, the term "lymphoma" includes, but is not limited to, Hodgkin's lymphoma, non-Hodgkin's lymphoma, diffuse large B-cell lymphoma, indolent B-cell lymphoma, follicular lymphoma, small lymphocytic lymphoma, nodal marginal zone B-cell lymphoma, lymphoplasmacytic lymphoma, anaplastic large cell lymphoma, primary cutaneous lymphoma, cutaneous T-cell lymphoma, cutaneous B-cell lymphoma, mycosis fungoides, mantle cell lymphoma, and chronic lymphocytic leukemia.

[0064] As used herein, unless otherwise specified, the terms "treat", "treating" and "treatment" refer to actions taken while a patient is afflicted with a particular cancer (such as certain types of lymphoma, e.g., DLBCL) that reduce the severity of the cancer or slow or delay the progression of the cancer.

[0065] The term "sensitivity" or "sensitized," when used in reference to cancer treatment, is a relative term that refers to the degree of effectiveness of a cancer treatment to alleviate or reduce the progression of a tumor or cancer being treated. For example, the term "increased sensitivity," when used in reference to treatment of a cell or tumor associated with a compound, refers to an increase in the effectiveness of the cancer treatment of at least about 5% or more.

[0066] As used herein, unless otherwise specified, the term "therapeutically effective amount" of a cancer treatment is an amount sufficient to provide a therapeutic benefit in the treatment or management of cancer, or an amount sufficient to delay or minimize one or more symptoms associated with the presence of cancer. A therapeutically effective amount of a compound refers to an amount of a therapeutic agent that alone or in combination with other therapies provides a therapeutic benefit in the treatment or management of cancer. The term "therapeutically effective amount" can encompass an amount that improves overall treatment, reduces or avoids the symptoms or causes of cancer, or enhances the therapeutic effect of another therapeutic agent. The term also refers to an amount of a compound sufficient to elicit a biological or medical response in a biomolecule (e.g., protein, enzyme, RNA, or DNA), cell, tissue, system, animal, or human that is desired by a researcher, veterinarian, physician, or clinician.

[0067] The term "responsiveness" or "responsive to," when used in reference to cancer treatment, refers to the degree of effectiveness of a treatment in reducing or relieving symptoms of the cancer being treated, e.g., DLBCL. For example, the term "increased responsiveness," when used in reference to treatment of a cell or subject, refers to an increased effectiveness in reducing or relieving symptoms of a disease compared to a reference treatment (e.g., of the same cell or subject, or of a different cell or subject), as measured using methods known in the art. In certain embodiments, the increased effectiveness is at least about 5%, at least about 10%, at least about 20%, at least about 30%, at least about 40%, or at least about 50%.

[0068] As used herein, the terms "effective subject response", "effective patient response" and "effective patient tumor response" refer to any increase in therapeutic benefit to a patient. An "effective patient tumor response" can be, for example, about a 5%, about 10%, about 25%, about 50%, or about 100% reduction in the rate of tumor progression. An "effective patient tumor response" can be, for example, about a 5%, about 10%, about 25%, about 50%, or about 100% reduction in the physical symptoms of cancer. An "effective patient tumor response" can also be, for example, about a 5%, about 10%, about 25%, about 50%, about 100%, about 200%, or more increase in patient response as measured by any suitable means, such as gene expression, cell count, assay results, tumor size, etc.

[0069] Improvement in cancer (e.g., DLBCL or its subtypes) or cancer-related disease can be characterized as a complete response or partial response. A "complete response" refers to the absence of clinically detectable disease by normalization of any previous abnormal radiographic examination, bone marrow, cerebrospinal fluid (CSF), or normalization of abnormal monoclonal protein measurements. A "partial response" refers to a reduction in all measurable tumor burden (i.e., the number of malignant cells present in the subject, or the volume of the measured tumor mass or the amount of abnormal monoclonal protein) of at least about 10%, about 20%, about 30%, about 40%, about 50%, about 60%, about 70%, about 80%, or about 90% in the absence of new lesions. The term "treatment" contemplates both complete and partial responses.

[0070] The term "likelihood" generally refers to an increase in the probability of an event. When used in relation to the efficacy of a patient's tumor response, the term "likelihood" generally contemplates an increase in the probability that the rate of tumor progression or the rate of tumor cell growth will be reduced. When used in relation to the efficacy of a patient's tumor response, the term "likelihood" generally refers to an increase in indicators such as mRNA or protein expression that may demonstrate the promotion of tumor treatment progression.

[0071] The term "predict" generally means to determine or state in advance. For example, when used to "predict" the effectiveness of a cancer treatment, the term "predict" can mean that the likelihood of the outcome of the cancer treatment can be determined at the time of occurrence, before the treatment begins, or before the treatment period has substantially begun.

[0072] As used herein, the term "source" refers to the origin of a sample when used in reference to a reference sample. For example, a sample taken from blood will have a reference sample also taken from blood. Similarly, a sample taken from bone marrow will have a reference sample also taken from bone marrow.

[0073] As used herein, the term "refractory" or "resistant" refers to a disorder, disease, or condition that has not responded to previous treatment, which may include one or more lines of treatment. In some embodiments, the disorder, disease, or condition has been previously treated with one, two, three, or four lines of treatment. In some embodiments, the disorder, disease, or condition has been previously treated with two or more lines of treatment and has less than a complete response (CR) to the most recent systemic therapy, including regimens.

[0074] As used herein, the term "recurrent" refers to a disorder, disease, or condition that has responded to treatment (e.g., achieved a complete response) and then progressed. Treatment can include one or more lines of therapy.

[0075] The term "expressed" or "expression" as used herein refers to transcription from a gene that gives an RNA nucleic acid molecule that is at least partially complementary to a region of one of the gene's two nucleic acid strands. The term "expressed" or "expression" as used herein also refers to translation from an RNA molecule that gives a protein, polypeptide, or portion thereof. A "biological marker" or "biomarker" is a substance whose detection indicates a particular biological state, such as the presence of certain cancers. In some embodiments, biomarkers can be determined individually. In other embodiments, several biomarkers can be measured simultaneously.

[0076] The terms "polypeptide" and "protein", used interchangeably herein, refer to a polymer consisting of a contiguous sequence of three or more amino acids linked via peptide bonds. The term "polypeptide" includes proteins, protein fragments, protein analogs, and oligopeptides. The term "polypeptide" as used herein may also refer to a peptide. The amino acids that make up a polypeptide may be naturally occurring or synthetic. A polypeptide may be purified from a biological sample. A polypeptide, protein, or peptide also includes modified polypeptides, proteins, and peptides, such as glycopolypeptides, glycoproteins, or glycopeptides; or lipopolypeptides, lipoproteins, or lipopeptides.

[0077] The terms "antibody," "immunoglobulin," or "Ig," as used interchangeably herein, encompass fully constructed antibodies and antibody fragments that retain the ability to specifically bind to an antigen. Antibodies provided herein include, but are not limited to, synthetic antibodies, monoclonal antibodies, polyclonal antibodies, recombinantly produced antibodies, multispecific antibodies (including bispecific antibodies), human antibodies, humanized antibodies, chimeric antibodies, intrabodies, single chain Fvs (scFvs) (including, e.g., monospecific, bispecific, etc.), camelized antibodies, Fab fragments, F(ab') fragments, disulfide-linked Fvs (sdFvs), anti-idiotypic (anti-Id) antibodies, and epitope-binding fragments of any of the above. In particular, antibodies provided herein include immunoglobulin molecules and immunologically active portions of immunoglobulin molecules, i.e., antigen-binding domains or molecules that comprise an antigen-binding site that immunospecifically binds to a CRBN antigen (e.g., one or more complementarity determining regions (CDRs) of an anti-CRBN antibody). The antibodies provided herein can be of any class (e.g., IgG, IgE, IgM, IgD, and IgA) or any subclass (e.g., IgG1, IgG2, IgG3, IgG4, IgA1, and IgA2) of immunoglobulin molecule. In some embodiments, the anti-CRBN antibody is a fully human antibody, such as a fully human monoclonal CRBN antibody. In certain embodiments, the antibodies provided herein are IgG antibodies or subclasses thereof (e.g., human IgG1 or IgG4).

[0078] The terms "antigen-binding domain," "antigen-binding region," "antigen-binding fragment," and similar terms refer to the portion of an antibody that contains the amino acid residues that interact with an antigen and confer specificity and affinity to the binder for the antigen (e.g., CDRs). The antigen-binding region can be derived from any animal species, including rodents (e.g., rabbit, rat, or hamster) and humans. In some embodiments, the antigen-binding region is of human origin.

[0079] The term "epitope" as used herein refers to a localized region on the surface of an antigen that can bind to one or more antigen-binding regions of an antibody, has antigenic or immunogenic activity in an animal, such as a mammal (e.g., human), and can elicit an immune response. An epitope with immunogenic activity is a portion of a polypeptide that elicits an antibody response in an animal. An epitope with antigenic activity is a portion of a polypeptide to which an antibody immunospecifically binds, as determined by any method known in the art, for example, by immunoassays described herein. An antigenic epitope is not necessarily immunogenic. Epitopes are usually composed of chemically active surface groupings of molecules, such as amino acids or sugar side chains, and have specific three-dimensional structural characteristics and specific charge characteristics. The region of a polypeptide that contributes to an epitope may be contiguous amino acids of the polypeptide, or an epitope may be two or more non-contiguous regions of a polypeptide taken together. An epitope may or may not be a three-dimensional surface feature of an antigen.

[0080] The terms "fully human antibody" and "human antibody" are used interchangeably herein to refer to an antibody that comprises a human variable region, and in some embodiments, a human constant region. In certain embodiments, these terms refer to an antibody that comprises variable and constant regions of human origin. The term "fully human antibody" includes antibodies having variable and constant regions that correspond to human germline immunoglobulin sequences as described by Kabat et al., Sequences of Proteins of Immunological Interest, USDepartment of Health and Human Services, NIH Publication No. 91-3242 (5th ed. 1991).

[0081] The phrase "recombinant human antibody" includes human antibodies prepared, expressed, generated, or isolated by recombinant means, e.g., antibodies expressed using a recombinant expression vector transfected into a host cell, antibodies isolated from a recombinant, recombinant human combinatorial antibody library, antibodies isolated from animals (e.g., mice or cows) that are transgenic and / or transchromosomal with human immunoglobulin genes (see, e.g., Taylor et al., Nucl. Acids Res., 1992, 20:6287-6295), or antibodies prepared, expressed, generated, or isolated by other means, including splicing of human immunoglobulin gene sequences to other DNA sequences. Such recombinant human antibodies can have variable and constant regions derived from human germline immunoglobulin sequences. See Kabat et al., Sequences of Proteins of Immunological Interest, USDepartment of Health and Human Services, NIH Publication No. 91-3242 (5th ed. 1991). However, in certain embodiments, such recombinant human antibodies have been subjected to in vitro mutagenesis (or, where animals transgenic for human Ig sequences are used, in vivo somatic mutagenesis) such that the amino acid sequences of the heavy and light chain variable regions of the recombinant antibodies are derived from and related to human germline heavy and light chain variable sequences, but are sequences that may not naturally exist within the in vivo human antibody germline repertoire.

[0082] The term "monoclonal antibody" refers to an antibody obtained from a homogeneous or substantially homogeneous population of antibodies, with each monoclonal antibody typically recognizing a single epitope on an antigen. In some embodiments, a "monoclonal antibody" as used herein is an antibody produced by a single hybridoma or other cell, which immunospecifically binds only to an epitope as determined, for example, by ELISA or other antigen-binding or competitive binding assays known in the art or in the examples provided herein. The term "monoclonal" is not intended to be limited to any particular method for making the antibody. For example, the monoclonal antibodies provided herein may be made by hybridoma methods such as those described in Kohler et al., Nature, 1975, 256:495-497, or isolated from phage libraries using techniques such as those described herein. Other methods for preparing clonal cell lines and the monoclonal antibodies expressed thereby are well known in the art. See, e.g., Short Protocols in Molecular Biology, Chapter 11 (Ausubel et al., eds., John Wiley and Sons, New York, 5th ed. 2002). Other exemplary methods for producing other monoclonal antibodies are provided in the Examples herein.

[0083] As used herein, "polyclonal antibody" refers to an antibody population generated in an immunogenic response to a protein with many epitopes, and thus includes a variety of different antibodies to the same or different epitopes within the protein. Methods for producing polyclonal antibodies are known in the art. See, for example, Short Protocols in Molecular Biology, Chapter 11 (Ausubel et al., eds., John Wiley and Sons, New York, 5th ed. 2002).

[0084] The term "level" refers to the amount, accumulation, or proportion of a molecule. The level can be expressed, for example, by the amount or rate of synthesis of messenger RNA (mRNA) encoded by a gene, the amount or rate of synthesis of a polypeptide or protein encoded by a gene, or the amount or rate of synthesis of a biomolecule accumulated in a cell or body fluid. The term "level" refers to the absolute amount of a molecule in a sample or the relative amount of a molecule determined under steady-state or non-steady-state conditions.

[0085] An "upregulated" mRNA is generally increased upon a given treatment or condition. A "downregulated" mRNA generally refers to a decrease in the expression level of an mRNA in response to a given treatment or condition. In some cases, the mRNA level may remain unchanged upon a given treatment or condition. An mRNA from a patient sample may be "upregulated" upon treatment compared to an untreated control. This upregulation may be, for example, an increase of about 5%, about 10%, about 20%, about 30%, about 40%, about 50%, about 60%, about 70%, about 80%, about 90%, about 100%, about 200%, about 300%, about 500%, about 1,000%, about 5,000%, or more of the comparative control mRNA level. Alternatively, an mRNA may be "downregulated" or expressed at a lower level in response to administration of a particular compound or other agent. A downregulated mRNA can be present, for example, at levels of about 99%, about 95%, about 90%, about 80%, about 70%, about 60%, about 50%, about 40%, about 30%, about 20%, about 10%, about 1%, or less of the comparative control mRNA level.

[0086] Similarly, the level of a polypeptide or protein biomarker from a patient sample may be increased upon treatment compared to an untreated control. This increase may be about 5%, about 10%, about 20%, about 30%, about 40%, about 50%, about 60%, about 70%, about 80%, about 90%, about 100%, about 200%, about 300%, about 500%, about 1,000%, about 5,000%, or more of the comparative control protein level. Alternatively, the level of a protein biomarker may be decreased in response to administration of a particular compound or other agent. This decrease may be, for example, at about 99%, about 95%, about 90%, about 80%, about 70%, about 60%, about 50%, about 40%, about 30%, about 20%, about 10%, about 1%, or less of the comparative control protein level.

[0087] As used herein, the terms "determining," "measuring," "evaluating," "assessing," and "assaying" generally refer to any form of measurement, including determining whether an element is present. These terms include quantitative and / or qualitative determinations. Assessment can be relative or absolute. "Assessing the presence of" can include determining the amount of something present, as well as determining whether it is present or not.

[0088] The terms "isolated" and "purified" refer to the isolation of a material (such as mRNA, DNA, or protein) such that the material represents a significant proportion of the sample in which it is found, i.e., a greater proportion than the material would typically be found in its native or non-isolated state. Typically, a significant proportion of a sample includes, for example, more than 1%, more than 2%, more than 5%, more than 10%, more than 20%, more than 50%, or more of the sample, usually up to about 90%-100%. For example, a sample of isolated mRNA may typically contain at least about 1% of total mRNA. Techniques for purifying polynucleotides are well known in the art and include, for example, gel electrophoresis, ion exchange chromatography, affinity chromatography, flow sorting, and sedimentation according to density.

[0089] As used herein, the term "bonded" refers to a direct or indirect connection. In the context of a chemical structure, "bond" (or "bonded") can refer to the presence of a chemical bond that directly connects two moieties (e.g., via a linking group or other intervening portion of a molecule) or that indirectly connects two moieties. The chemical bond may be a covalent bond, an ionic bond, a coordination complex, hydrogen bonding, van der Waals interactions, or hydrophobic stacking, or may exhibit properties of multiple types of chemical bonds. In certain cases, "bonded" includes embodiments in which the connection is direct and embodiments in which the connection is indirect.

[0090] The term "sample" as used herein relates to a material or mixture of materials, typically, though not necessarily, in fluid form, containing one or more components of interest. In some embodiments, the sample may be a biological sample. As used herein, "biological sample" refers to a sample obtained from a living subject, including samples of biological tissue or fluid origin, obtained, achieved, or collected in vivo or in situ. Biological samples also include samples of regions of a biological subject that contain precancerous or cancerous cells or tissues. Such samples may be, but are not limited to, organs, tissues, and cells isolated from a mammal. Exemplary biological samples include, but are not limited to, cell lysates, cell cultures, cell lines, tissues, oral tissues, gastrointestinal tissues, organs, organelles, bodily fluids, blood samples, urine samples, and skin samples. Preferred biological samples include, but are not limited to, whole blood, partially purified blood, PBMCs, and tissue biopsies.

[0091] As used herein, the term "analyte" refers to a known or unknown component of a sample.

[0092] As used herein, the term "capture agent" refers to an agent that binds to an mRNA or protein through an interaction sufficient to allow the agent to bind and enrich the mRNA or protein in a heterogeneous mixture.

[0093] Unless otherwise indicated, the term "pharmaceutical acceptable salts" as used herein includes non-toxic acid addition salts and base addition salts of the compounds to which it refers. Acceptable non-toxic acid addition salts include those derived from organic and inorganic acids known in the art, including, for example, hydrochloric acid, hydrobromic acid, phosphoric acid, sulfuric acid, methanesulfonic acid, acetic acid, tartaric acid, lactic acid, succinic acid, citric acid, malic acid, maleic acid, sorbic acid, aconitic acid, salicylic acid, phthalic acid, embonic acid, and enanthic acid. Compounds that are acidic in nature can form salts with various pharmaceutical acceptable bases. Bases that can be used to prepare pharmaceutical acceptable base addition salts of such acidic compounds are those that form non-toxic base addition salts, i.e., salts with pharmacologically acceptable cations, such as, but not limited to, alkali metal or alkaline earth metal salts (particularly calcium salts, magnesium salts, sodium salts, or potassium salts). Suitable organic bases include, but are not limited to, N,N-dibenzylethylenediamine, chloroprocaine, choline, diethanolamine, ethylenediamine, meglumaine (N-methylglucamine), lysine, and procaine.

[0094] As described herein, the term "second active agent" refers to any additional biologically active treatment. It is understood that the second active agent may be a hematopoietic growth factor, a cytokine, an anti-cancer drug, an antibiotic, a cox-2 inhibitor, an immunomodulatory agent, an immunosuppressant, a corticosteroid, a therapeutic antibody that specifically binds to a cancer antigen or a pharmacologically active variant, or a derivative thereof. Exemplary second active agents include, but are not limited to, HDAC inhibitors (e.g., panobinostat, romidepsin, or vorinostat), BCL2 inhibitors (e.g., venetoclax), BTK inhibitors (e.g., ibrutinib or acalabrutinib), mTOR inhibitors (e.g., everolimus), PI3K inhibitors (e.g., idelalisib), PKCβ inhibitors (e.g., enzastaurin), SYK inhibitors (e.g., fostamatinib), JAK2 inhibitors, and the like. agents (e.g., fedratinib, pacritinib, ruxolitinib, baricitinib, gandotinib, lestaurtinib, or momelotinib), Aurora A kinase inhibitors (e.g., alisertib), EZH2 inhibitors (e.g., tazemetostat, GSK126, CPI-1205, 3-deazaneplanocin A, EPZ005687, EI1, UNC1999, or sinefungin), BET inhibitors (e.g., birabresib or 4[2-(

[0036] Therapeutic agents include, but are not limited to, cyclopropylmethoxy)-5-(methanesulfonyl)phenyl]-2-methylisoquinolin-1(2H)-one), hypomethylating agents (e.g., 5-azacytidine or decitabine), chemotherapy (e.g., bendamustine, doxorubicin, etoposide, methotrexate, cytarabine, vincristine, ifosfamide, or melphalan), or epigenetic compounds (e.g., DOT1L inhibitors such as pinometostat, HAT inhibitors such as C646, agents, WDR5 inhibitors such as OICR-9429, HDAC6 inhibitors such as ACY-241, DNMT1 selective inhibitors such as GSK3484862, LSD-1 inhibitors such as compound C or seclidemstat, G9A inhibitors such as UNC0631, PRMT5 inhibitors such as GSK3326595, BRPF1B / 2 inhibitors such as OF-1, BRD9 / 7 inhibitors such as LP99, SUV420H1 / H2 inhibitors such as A-196,Menin-MLL inhibitors such as MI-503, CARM1 inhibitors such as EZM2302, BRD9 inhibitors such as dBrd9, Aiolos / Ikaros degradative cereblon E3 ligase regulator (CELMoD), CREBBp2 inhibitors, anti-CD79b antibodies, CD19 CAR-T, p53 inhibitors (Nutlin), Bcl6 inhibitors, CREBBp2 CELMoD, CD79b CELMoD, CD19 CELMoD, p53 (Nutlin) CELMoD, Bcl6 CELMoD, inhibitors of ligand-directed degradation (LDD) of CREBBP2, LDD inhibitors of CD79b, LDD inhibitors of CD19, LDD inhibitors of p53 (Nutlin), LDD inhibitors of Bcl6, LDD inhibitors of CK1a, LDD inhibitors of IRAK4 (e.g., MYD88 L265p lymphoma), MALT1 inhibitors such as JNJ-67856633, MAT2A inhibitors (e.g., for 9p21 deletions), anti-CD3 x anti-CD19 bispecific antibodies, and anti-CD3 x anti-CD20 bispecific antibodies.

[0095] The term "about" or "approximately" refers to an acceptable error for a particular value as determined by one of ordinary skill in the art, which depends in part on how the value is measured or determined. In certain embodiments, the term "about" or "approximately" refers to no more than 1, no more than 2, no more than 3, or no more than 4 standard deviations. In certain embodiments, the term "about" or "approximately" refers to within 50%, within 20%, within 15%, within 10%, within 9%, within 8%, within 7%, within 6%, within 5%, within 4%, within 3%, within 2%, within 1%, within 0.5%, or within 0.05% of a given value or range.

[0096] The use of the words "a" or "an" when used in conjunction with the term "comprising" in the claims and / or specification may mean "one," but is also consistent with the meaning of "one or more," "at least one," and "one or more."

[0097] As used herein, unless otherwise specified or indicated by context, the term "pre-treatment" used in accordance with the methods described herein refers to prior to administration of a treatment.

[0098] As used herein, the terms "patient" and "subject" refer to animals, such as mammals. In some embodiments, the patient is a human. In other embodiments, the patient is a non-human animal, such as a dog, cat, farm animal (e.g., a horse, pig, or donkey), chimpanzee, or monkey. In certain embodiments, the patient is a human with lymphoma (e.g., DLBCL) in need of treatment.

[0099] 5.2 Clustering method for lymphoma patients In one aspect, provided herein is a method of classifying lymphoma patients, the method comprising: (a) obtaining a sample from the lymphoma patient; (b) measuring the expression level of at least one gene in the sample; and (c) using the expression level of at least one gene in the sample to cluster the lymphoma patients into subgroups of patients with lymphoma. In some embodiments, the lymphoma patient is a DLBCL patient.

[0100] In another aspect, a method for classifying lymphoma patients is provided herein, the method comprising: (a) measuring the expression level of at least one gene in a sample of the lymphoma patient; and (b) using the expression level of at least one gene in the sample to cluster the lymphoma patients into subgroups of patients with lymphoma. In certain embodiments, the method further comprises obtaining a sample from the lymphoma patient. In some embodiments, the lymphoma patient is a DLBCL patient.

[0101] In some embodiments, the lymphoma is DLBCL. In some embodiments, the lymphoma is an indolent B-cell lymphoma. In some embodiments, the lymphoma is selected from the group consisting of follicular lymphoma, small lymphocytic lymphoma, nodal marginal zone B-cell lymphoma, lymphoplasmacytic lymphoma, anaplastic large cell lymphoma, primary cutaneous lymphoma, mycosis fungoides, chronic lymphocytic leukemia, and mantle cell lymphoma. In some embodiments, the lymphoma is a follicular lymphoma. In some embodiments, the lymphoma is a nodal marginal zone B-cell lymphoma. In some embodiments, the lymphoma is a mantle cell lymphoma. In some embodiments, the lymphoma is a chronic lymphocytic leukemia.

[0102] In some embodiments, the sample is obtained from a patient tissue that contains DLBCL cells. A more detailed description of the sample (e.g., biological sample) is provided in Section 5.7 below.

[0103] In some embodiments, the DLBCL patient is a newly diagnosed (nd) DLBCL patient. In some embodiments, the DLBCL patient is a relapsed / refractory (r / r) DLBCL patient. In some embodiments, the DLBCL patient is a newly diagnosed (nd) relapsed / refractory (r / r) DLBCL patient.

[0104] As illustrated in Example 1, set forth in Section 6.1 below, multiple patient datasets can be utilized in the classification / clustering method. Patient datasets include, but are not limited to, the screening cohort of the ROBUST clinical trial consisting of 1016 patients newly diagnosed with DLBCL (see Clinical Trial Number: NCT02285062; see, e.g., Nowakowski et al., J. Clin. Onco., 2021, 39(12), 1317-1328); a set of 192 commercially available newly diagnosed DLBCL patient samples with molecular profiling but no survival data (the "Commercial dataset", which when combined with the 1016 patient dataset is referred to as the "Discovery-2 dataset"); a set of 343 ndDLBCL patients from the Molecular Epidemiology Resource (MER) (the "ndMER dataset"; see Cerhan et al., J. Clin. Onco., 2021, 39(12), 1317-1328); al., Int. J. Epidermiol., 2017, 46(6):1753-1754i); a set of 928 patients from the REMoDL-B dataset (see clinical trial number: NCT01324596). A subset of the Discovery-2 dataset had outcome and clinical information available. The MER dataset cohorts are well characterized in terms of clinical outcomes and treatment. In some embodiments, the dataset is the Discovery-2 dataset. In some embodiments, the Discovery-2 dataset includes the Commercial dataset and the ROBUST clinical trial screening dataset.

[0105] In some embodiments, a dataset of lymphoma patients is generated by measuring gene expression levels in the samples. In some embodiments, the lymphoma patients are a newly diagnosed lymphoma patient cohort. In some embodiments, the lymphoma patients are a relapsed / refractory lymphoma patient cohort.

[0106] In some embodiments, the step of clustering lymphoma patients into subgroups comprises using a discovery dataset and one or more replication / validation datasets. In some embodiments, the clustering comprises a discovery dataset and a replication / validation dataset. In some embodiments, the discovery dataset is from samples of a discovery cohort. In some embodiments, the replication / validation dataset is from samples of a replication / validation cohort. In some embodiments, the discovery dataset is selected from the group consisting of a Discovery-2 dataset, an ndMER dataset, a ROBUST dataset, and a REMoDL-B dataset. In some embodiments, the discovery dataset is selected from the group consisting of a Discovery-2 dataset, an ndMER dataset, and a REMoDL-B dataset. In some embodiments, the replication / validation dataset is a Discovery-2 dataset or an ndMER dataset. In some embodiments, the replication / validation dataset is an ndMER dataset or a REMoDL-B dataset. In some embodiments, the replication / validation dataset comprises ndMER and REMoDL-B datasets. In some embodiments, the discovery dataset is a Discovery-2 dataset. In some embodiments, the discovery dataset is an ndMER dataset. In some embodiments, the discovery dataset is a ROBUST dataset. In some embodiments, the discovery dataset is a REMoDL-B dataset. In some embodiments, the Discovery-2 dataset comprises a ROBUST dataset and a Commercial dataset. In some embodiments, the discovery dataset is a combination of one or more datasets of DLBCL patients. In some embodiments, the discovery dataset is a combination of one or more datasets described herein.

[0107] In some embodiments, the step of clustering lymphoma patients into subgroups comprises: (i) normalizing the dataset; (ii) selecting at least one clustering feature; (iii) applying a clustering method using the at least one clustering feature. In some embodiments, the clustering step further comprises detecting outliers in the dataset and filtering out the outliers. In some embodiments, the clustering step further comprises evaluating a clustering result of the clustering method. In some embodiments, the clustering step is an unsupervised clustering method. In some embodiments, the clustering method is a hierarchical clustering method. In some embodiments, the clustering method is a non-hierarchical method. In some embodiments, the clustering method is a K-means clustering method. In some embodiments, the clustering method is a partitioning method. In some embodiments, the clustering method is a fuzzy clustering method. In some embodiments, the clustering method is density-based clustering. In some embodiments, the clustering method is model-based clustering. In one preferred embodiment, the clustering method is the iClusterPlus clustering method. In some embodiments, the clustering method is the Cluster of Cluster Analysis (COCA) clustering method.

[0108] In some embodiments, the single sample normalization (ssNorm) method is used to normalize the datasets prior to analysis. The practice of running ssNorm can benefit both the immediate short-term analysis and the long-term translatability of the results. In the short term, ssNorm provides a fixed normalization scheme that can be applied completely independently to individual samples. There is no need to add new batches of samples and re-normalize the datasets or combine the datasets for larger-scale analysis. The ssNorm method can also implicitly perform batch correction, thus eliminating the effects of gene-specific experiments that appear as systematic biased differences in raw expression space. The ssNorm method can properly align diverse datasets in a common space and does not show meaningful separation by dataset / batch when projected into PCA space. ssNorm also allows for the simple translation or parameterization of any classifier from one dataset to another - a classifier built on one ssNorm dataset can be directly applied to other ssNorm datasets without the need to reweight model parameters or set new thresholds.

[0109] In certain embodiments, DLBCL-specific housekeeping genes are used for normalization. In some embodiments, ISY1, R3HDM1, TRIM56, UBXN4, and / or WDR55 are used for normalization.

[0110] Clustering methods are sensitive to the input data, and the introduction of noisy or irrelevant features can degrade the performance of the algorithm. Both feature selection and feature engineering approaches can be utilized to reduce the dimensionality of a dataset while maintaining the representation of relevant biological activities.

[0111] In some embodiments, one or more (e.g., one, two, three, four, or more) features are selected as clustering features as input to the clustering method. In some embodiments, a subset of gene expression data is selected as the clustering features. In some embodiments, the subset of gene expression data is the gene expression data of the top 50%, 45%, 40%, 35%, 30%, 25%, 20%, 15%, 10%, or 5% of most expressed genes. In some embodiments, the subset of gene expression data is the gene expression data of the top 25% of most expressed genes.

[0112] Some types of derived feature scores that integrate multiple biologically relevant genes into one feature can also be used as clustering features. Gene Set Variation Analysis (GSVA) ​​is a Gene Set Enrichment (GSE) method that estimates pathway activity variation in a sample population in an unsupervised manner. GSVA improves the power of detecting subtle pathway activity changes in a sample population compared to corresponding methods. GSVA is a starting point for building pathway-centric models of biology, and GSVA can contribute to the current need for GSE methods for RNA-seq data. See Hanzelman et al., BMC Bioinformatics, 2013, 14:7. GSVA can be performed on 50 Hallmark pathway gene sets from MSigDB (Hallmark GSVA score). See, e.g., Liberzon et al., Cell Syst., 2015, 1:417:425. GSVA can be performed on 299 C1 set positional site band signature gene sets from MSigDB (C1 Positional GSVA score). See, e.g., Alhamdoosh et al., F1000Research, 2017, 6:2010. GSVA can also be performed on DLBCL-specific LM23 matrices derived from DCQ cell deconvolution methods (Cell type LM23 GSVA score). See, Althoum et al., Mol. Syst. Bio., 2014, 10:720.

[0113] In some embodiments, the Hallmark GSVA score is selected as the clustering feature. In some embodiments, the C1 Positional GSVA score is selected as the clustering feature. In some embodiments, the Cell type LM23 GSVA score is selected as the clustering feature. In some embodiments, the subset of gene expression data, the Hallmark GSVA score, the C1 Positional GSVA score, and the Cell type LM23 GSVA score are selected as the matrix of clustering features.

[0114] In some embodiments, the clustering method performs clustering on each feature matrix independently and combines the cluster features. In some embodiments, the clustering method performs clustering that combines cluster features from different matrices of features and clusters all the clustering features.

[0115] In some embodiments, the dataset is clustered into 2-20 clusters (K=2-20). In some embodiments, the dataset is clustered into 2-15 clusters (K=2-15). In some embodiments, the dataset is clustered into 2-12 clusters (K=2-12). In some embodiments, the dataset is clustered into 2-10 clusters (K=2-10). In some embodiments, the dataset is clustered into 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 clusters. In some embodiments, the number of clusters is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20. In some embodiments, the dataset is clustered into 7 clusters (K=7, clusters A1-A7). In some embodiments, the dataset is clustered into eight clusters (K=8). In some embodiments, the number of clusters is seven.

[0116] In some embodiments, the step of clustering further comprises evaluating the clustering results of the clustering method. In some embodiments, the clustering results of each number of clusters (K) are evaluated. In some embodiments, the clustering results are evaluated using the nbClust R package. In some embodiments, the clustering results are evaluated using a metric selected from the group consisting of silhouette statistics, gap statistics, and the proportion of variance explained by the clustering. In some embodiments, the clustering results are evaluated by a minimum cluster size. In order to have a sufficient sample size in each cluster to build a downstream classifier model that can robustly discriminate each cluster or subgroup, the clusters resulting from each tested clustering must have a minimum cluster size.

[0117] In some embodiments, the clustering method is the iClusterPlus clustering method and the number of clusters is seven.

[0118] In some embodiments, the clustering method is iClusterPlus clustering method, a subset of gene expression data, Hallmark GSVA score, C1 Positional GSVA score, and Cell type LM23 GSVA score are selected as the matrix of clustering features, and the number of clusters is 7 (clusters A1 to A7).

[0119] In some embodiments, in one or more subgroups, there are patients who provide low confidence data for a particular subgroup in terms of the selected clustering classifier. In some embodiments, patients who provide low confidence clustering data are filtered out. In some embodiments, patients who provide low confidence clustering data are excluded from the subgroup. In some embodiments, patients who provide low confidence clustering data are excluded from allocation to a subgroup. In some embodiments, patients who provide low confidence clustering data are excluded from subgroup A7.

[0120] In some embodiments, the method further comprises setting a threshold confidence for one or more subgroups to exclude patients providing lower confidence clustering data from the one or more subgroups. In some embodiments, patients providing low confidence clustering data are excluded from subgroup A7.

[0121] In some embodiments, the method for clustering lymphoma patients further comprises: (i) identifying at least one cluster classifier by training a classifier model using the clustering results of the discovery dataset; (ii) applying the at least one cluster classifier to the replication / validation dataset to classify the replication / validation dataset; (iii) clustering the replication / validation dataset using a clustering method; and (iv) comparing the classification results of the replication / validation dataset using the at least one cluster classifier with the clustering results of the replication / validation dataset using the clustering method. In some embodiments, the model is a grouped multinomial generalized linear model (GLM). In some embodiments, the model is a binary generalized linear model (GLM). In some embodiments, the model is a GLM using a least absolute shrinkage selection operator (LASSO). In some embodiments, if the classification results of the replication / validation dataset using the at least one cluster classifier are similar to the clustering results of the replication / validation dataset using the clustering method, it indicates that the at least one cluster classifier is effective for classifying the replication / validation dataset.

[0122] In some embodiments, the method for clustering lymphoma patients further comprises: (i) identifying at least one cluster classifier by training a classifier model using the clustering results of the discovery dataset; and (ii) applying the at least one cluster classifier to the replication / validation dataset to classify the replication / validation dataset. In some embodiments, the model is a grouped multinomial generalized linear model (GLM). In some embodiments, the model is a binary generalized linear model (GLM). In some embodiments, the model is a GLM using a least absolute shrinkage selection operator (LASSO).

[0123] In some embodiments, the lymphoma patient is 30 years or older, 35 years or older, 40 years or older, 45 years or older, 50 years or older, or 55 years or older, or 60 years or older, or 65 years or older, or 70 years or older at baseline. In some embodiments, the lymphoma patient is 70 years or older. In some embodiments, the lymphoma patient is 60 years or older. In some embodiments, the lymphoma patient is 30-35 years old, 35-40 years old, 40-45 years old, 45-50 years old, 50-55 years old, 55-60 years old, 60-65 years old, or 65-70 years old.

[0124] In certain embodiments, the at least one gene is selected from the genes in Table 1. In certain embodiments, the at least one gene includes one, two, three, four, five, or more genes in Table 1. In certain embodiments, the at least one gene includes all of the genes in Table 1.

[0125] In some embodiments, the expression level of at least one gene in the discovery dataset is used to train the classifier model. In some embodiments, one, two, three, four, five, or more of the genes in Table 1 are used to train the classifier model. In some embodiments, all of the genes in Table 1 are used to train the classifier model.

[0126] In some embodiments, the classifier model includes one, two, three, four, five or more genes from Table 1. In some embodiments, the classifier model includes all of the genes from Table 1. In some embodiments, the expression levels of all of the genes from Table 1 are determined for clustering. In some embodiments, the expression levels of all of the genes from Table 1 are determined and compared to determine whether a lymphoma patient will respond to a cancer treatment.

[0127] [Table 1]

[0128] [Table 2]

[0129] In some embodiments, the methods provided herein involve the use of an inhibitor of one or more of the following: ABHD10, ACO1, ACTN4, AGRP, AKAP13, ALDH1A1, ALG13, AMT, ANKZF1, AOAH, AP1G2, AP3S1, APRT, ARG1, ARHGDIA, ARHGEF7, ART4, ASH1L, ATIC, ATP6V1G2, ATP9B, BAZ2A, BLNK, BPNT1, BRIP1, BTF3, BUB3, C1QBP, C2, CACUL1, CADPS, CAPZB CARD11, CBX5, CCDC136, CCR2, CCT7, CCT8, CD37, CD46, CDC25A, CDK12, CENPW, CEP85L, CEP97, CFH, CHD2, CHI3L1, CHKA, CIB1, CKAP2, CLCN7, CLEC7A, CLIC1, CLK1, CMSS1, COG7, COL1A2, COL4A3, CORO 1A, COX6C, CPD, CRBN, CSF1, CUEDC2, CUX1, CXCL10, DDHD1, DDX58, DIMT1, DNAJA3, DNAJC10, DNMT1, DYNLT1, E2F1, EBAG9, EIF1AX, EIF2B5, EIF3I, EIF5A, ELF1, ELMO1, EMP3, ENO1, EPS15, ERGIC2, ERH ESD, EYS, FBXO46, FLNA, FOXP1, FPR1, FUBP1, FUS, GALM GAPDH, GATM, GBP5, GIMAP4, GJD3, GLUL, GNA13GNB2、GNS、GPR82、GPX1、GRIP1、HAMP、HEXIM1、HNMT、HNRNPA2B1、HNRNPU、HSD11B1L、HSP90AA1、HSPD1、HSPE1、IDS、IFI30、IFITM3、IKZF1、IL24、IMPDH2、IST1、ITGB2、JAK1、KIF14、KIF4A、KLHL14、KLHL23、LAP3、LATS1、LMO4、LONP2、LPAR6、LRCH4、LRRC37A2、LRRC37A3、LRRC59、LY9、LYRM1、MAP3K14、MAP4K2、MAPK14、MARS2、MAT2A、MAX、MCL1、MCM4、MGAT4A、MRPL43、MRPS15、MRPS9、MSH6、MVB12A、MVB12B、MVP、MYBL2、MYOF、NACA、NCAPG2、NCBP2、NCL、NFE2L2、NFKBIA、NRXN1、NUDT21、ORC6、P2RX7、PAICS、PARG、PARP9、PAX5、PCNA、PDE9A、PFKL、PGAM1、PIF1、PILRB、PKIA、PKM、PKN2、PLEKHF2、PLK1PMM2, PNP, POLA1, POLR2E, POM121, PPA1, PPIA, PPP1R9B, PRDM15, PRDX5, PRKCB, PRKCH, PRMT1, PRPF4, RPF6, PRRC2C, PSMC1, PSMD13, PSME2, ​​P TBP3, PTENP1, PTPRC, R3HDM1, RAB32, RAMP3, RANBP2, RBM3, RBM33, RBP1, RFC1, RNASEL, RNF38, RPL30, RPL32P3, RPRD2, RPS3, RPS4X, RRP9, S1 00A11, S100Z, S1PR2, SAAL1, SBNO1, SCAMP2, SCARB1, SCARB2, SDCBP, SELPLG, SENP7, SERPINA3, SERPINB1, SF1, SF3A1, SFPQ, SFXN3, SH3BGRL 3, SH3BP1, SLAMF8, SLC35D1, SMARCA4, SMARCC1, SMG5, SNORA21, SNORA71B, SNORD104, SNRPA, SNRPD1, SNTA1, SNX29, SOBP, SOGA1, SP140, SP3, SPEN, SRGN, SRSF6, STAG3, STK4, STMN1, SUMO2, SYNJ2, SYPL1, SYTL3, TAGLN2, TAP2, TARDBP, TBCA, TGDS, TGOLN2, THRAP3, TIMM10, TLK1, TLN1 , TMBIM4, TMEM223, TNFRSF1A, TONSL, TP53INP1, TPM3, TRA2B, TRAM1, TRAPPC12, TRIOBP, TRIP13, TRMT1L, TRNT1, TSPYL2, TSSK4, TUBA1B, TUB The present invention relates to a method for treating idiopathic pulmonary disease (LPD) comprising administering to a patient a therapeutically effective amount of at least one gene selected from the group consisting of A1C, TWIST1, UBE2B, UBE2D2, UBE2G1, UXT, VASH1, VAV1, VDAC1, WEE1, XRCC6, YTHDC1, YWHAE, ZBED5, ZBTB37, ZFAND4, ZFAND5, ZMAT1, ZNF101, ZNF107, ZNF146, ZNF207, ZNF318, ZNF367, ZNF480, and ZWINT, or at least one gene selected from this group (e.g., one, two, three, four, five, or more).

[0130] 5.3 Methods for predicting response and treating lymphoma patients In another aspect, provided herein is a method of predicting a lymphoma patient's responsiveness to a cancer treatment, the method comprising: (a) obtaining a reference biological sample from each patient of a reference patient group, the reference patient including a reference patient having lymphoma; (b) using gene expression levels in the reference biological samples to cluster or classify the reference patient group into subgroups of patients; (c) obtaining a biological sample from the lymphoma patient; (d) using gene expression levels in the lymphoma patient biological sample to determine which subgroup the lymphoma patient belongs to; and (e) predicting the lymphoma patient's responsiveness to a first cancer treatment.

[0131] In another aspect, a method for predicting the responsiveness of a lymphoma patient to a cancer treatment is provided herein, the method comprising: (a) clustering reference lymphoma patients of a reference patient group into subgroups using the expression level of at least one gene in a reference biological sample of the reference lymphoma patient; (b) determining the subgroup to which the lymphoma patient belongs based on the expression level of at least one gene in the biological sample of the lymphoma patient; and (c) predicting the responsiveness of the lymphoma patient to a first cancer treatment based on the subgroup of the lymphoma patient. In certain embodiments, the method further comprises obtaining a reference biological sample from a reference lymphoma patient of the reference patient group. In certain embodiments, the method further comprises obtaining a biological sample from a lymphoma patient.

[0132] In some embodiments, the method further comprises administering to the lymphoma patient a second cancer treatment.

[0133] In some embodiments, the sample is obtained from a subject's tissue that contains DLBCL cells. A more detailed description of samples (or biological samples) is provided in Section 5.7 below.

[0134] In some embodiments, the lymphoma patient is a DLBCL patient. In some embodiments, the DLBCL patient is a newly diagnosed (nd) DLBCL patient and a relapsed / refractory (r / r) DLBCL patient. In some embodiments, the DLBCL patient is a newly diagnosed (nd) DLBCL patient. In some embodiments, the DLBCL patient is a relapsed / refractory (r / r) DLBCL patient.

[0135] In some embodiments, classifying or clustering the reference patient group comprises generating clustering information defining a relationship between expression levels of at least one gene in the reference biological sample; and reorganizing the heatmap representation based on the clustering information. In some embodiments, classifying or clustering the reference patient group comprises using a clustering method as described herein in Section 5.2.

[0136] In some embodiments, the clustering step includes a discovery dataset and at least one replication / validation dataset. In some embodiments, the clustering step includes a discovery dataset and a replication / validation dataset. In some embodiments, the discovery dataset is from a sample of a discovery cohort. In some embodiments, the replication / validation dataset is from a sample of a replication / validation cohort. In some embodiments, the discovery dataset is selected from the group consisting of a Discovery-2 dataset, an ndMER dataset, a ROBUST dataset, and a REMoDL-B dataset. In some embodiments, the replication / validation dataset is selected from the group consisting of a Commercial dataset, an ndMER dataset, a ROBUST dataset, and a REMoDL-B dataset. In some embodiments, the discovery dataset is a Discovery-2 dataset. In some embodiments, the Discovery-2 dataset includes a Commercial dataset. In some embodiments, the Discovery-2 dataset includes a ROBUST dataset. In some embodiments, the Discovery-2 dataset includes a Commercial dataset and a ROBUST dataset. In some embodiments, the discovery dataset is an ndMER dataset. In some embodiments, the discovery dataset is a REMoDL-B dataset.

[0137] In some embodiments, the replication / validation dataset is selected from the group consisting of Discovery-2 dataset, ndMER dataset, ROBUST dataset, and REMoDL-B dataset. In some embodiments, the replication / validation dataset is selected from the group consisting of Commercial dataset, ndMER dataset, ROBUST dataset, and REMoDL-B dataset. In some embodiments, the replication / validation dataset is selected from the group consisting of ndMER dataset, ROBUST dataset, and REMoDL-B dataset. In some embodiments, the replication / validation dataset is an ndMER dataset or a REMoDL-B dataset. In some embodiments, the replication / validation dataset is an ndMER dataset. In some embodiments, the replication / validation dataset is a REMoDL-B dataset. In some embodiments, the replication / validation dataset is an ndMER dataset and a REMoDL-B dataset.

[0138] In some embodiments, the clustering step includes: (i) normalizing the dataset; (ii) selecting at least one clustering feature; and (iii) applying a clustering method using the at least one clustering feature. In some embodiments, the clustering step further includes detecting outliers in the dataset and removing the outliers. In some embodiments, the clustering step further includes evaluating a clustering result of the clustering method. In some embodiments, the clustering is an unsupervised clustering method. In some embodiments, the clustering method is a hierarchical clustering method. In some embodiments, the clustering method is a non-hierarchical approach. In some embodiments, the clustering method is a K-means clustering method. In some embodiments, the clustering method is a partitioning method. In some embodiments, the clustering method is a fuzzy clustering method. In some embodiments, the clustering method is density-based clustering. In some embodiments, the clustering method is model-based clustering. In some embodiments, the clustering method is an iClusterPlus clustering method. In some embodiments, the clustering method is the Cluster of Cluster Analysis (COCA) clustering method. In some embodiments, the single sample normalization (ssNorm) method is used to normalize the dataset prior to analysis.

[0139] In certain embodiments, DLBCL-specific housekeeping genes are used for normalization. In some embodiments, ISY1, R3HDM1, TRIM56, UBXN4, and WDR55 are used for normalization.

[0140] In some embodiments, one, two, three, four or more features are selected as clustering features to be input into the clustering method. In some embodiments, a subset of gene expression data is selected as the clustering features. In some embodiments, the subset of gene expression data is the gene expression data of the top 50%, 45%, 40%, 35%, 30%, 25%, 20%, 15%, 10%, or 5% of most expressed genes. In some embodiments, the subset of gene expression data is the gene expression data of the top 25% of most expressed genes.

[0141] In some embodiments, the Hallmark GSVA score is selected as the clustering feature. In some embodiments, the C1 Positional GSVA score is selected as the clustering feature. In some embodiments, the Cell type LM23 GSVA score is selected as the clustering feature. In some embodiments, a subset of the gene expression data, the Hallmark GSVA score, the C1 Positional GSVA score, and the Cell type LM23 GSVA score are selected as the matrix of clustering features.

[0142] In some embodiments, the clustering method performs clustering on each feature matrix independently and combines the cluster features. In some embodiments, the clustering method performs clustering that combines cluster features from different matrices of features and clusters all the clustering features.

[0143] In some embodiments, the dataset is clustered into 2-20 clusters (K=2-20). In some embodiments, the dataset is clustered into 2-15 clusters (K=2-15). In some embodiments, the dataset is clustered into 2-12 clusters (K=2-12). In some embodiments, the dataset is clustered into 2-10 clusters (K=2-10). In some embodiments, the dataset is clustered into 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 clusters. In some embodiments, the dataset is clustered into 8 clusters (K=8). In some embodiments, the dataset is clustered into 7 clusters (K=7, clusters A1-A7). In some embodiments, the number of clusters is 7.

[0144] In some embodiments, the clustering method is the iClusterPlus clustering method and the number of clusters is seven.

[0145] In some embodiments, the clustering method is iClusterPlus clustering method, a subset of gene expression data, Hallmark GSVA score, C1 Positional GSVA score, and Cell type LM23 GSVA score are selected as the matrix of clustering features, and the number of clusters is 7 (clusters A1 to A7).

[0146] The terms "cluster" and "subgroup" are used interchangeably throughout this disclosure.

[0147] In some embodiments, the reference patients of the reference patient group are clustered into 2-20 subgroups (K=2-20). In some embodiments, the reference patients of the reference patient group are clustered into 2-15 subgroups (K=2-15). In some embodiments, the reference patients of the reference patient group are clustered into 2-12 subgroups (K=2-12). In some embodiments, the reference patients of the reference patient group are clustered into 2-10 subgroups (K=2-10). In some embodiments, the reference patients of the reference patient group are clustered into 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 subgroups. In some embodiments, the reference patients of the reference patient group are clustered into 8 groups (K=8). In some embodiments, the reference patients of the reference patient group are clustered into 7 groups (K=7, clusters A1-A7).

[0148] In some embodiments, the step of clustering further comprises evaluating the clustering results of the clustering method. In some embodiments, the clustering results of each number (K) of clusters are evaluated. In some embodiments, the clustering results are evaluated using the nbClust R package. In some embodiments, the clustering results are evaluated using a metric selected from the group consisting of silhouette statistics, gap statistics, and the proportion of variance explained by clustering. In some embodiments, the clustering results are evaluated by a minimum cluster size. In order to have a sufficient sample size for each cluster to build a downstream classifier model that can robustly discriminate each cluster or subgroup, it is necessary to have a minimum cluster size for the clusters obtained from each cluster tested.

[0149] In some embodiments, in at least one of the subgroups, there are patients who provide low confidence clustering data of the particular subgroup in terms of the selected clustering classifier. In some embodiments, the patients who provide low confidence clustering data are filtered. In some embodiments, the patients who provide low confidence clustering data are excluded from the subgroup. In some embodiments, the patients who provide low confidence clustering data are excluded from subgroup A7.

[0150] In some embodiments, the method further comprises setting a threshold confidence level for at least one subgroup to exclude patients providing lower confidence levels from the at least one subgroup. In some embodiments, patients providing low confidence clustering data are excluded from subgroup A7.

[0151] In some embodiments, the method for predicting lymphoma patient responsiveness to cancer treatment further comprises: (i) identifying at least one cluster classifier by training a classifier model using the clustering results of the discovery dataset; (ii) applying the at least one cluster classifier to the replication / validation dataset to classify the replication / validation dataset; (iii) clustering the replication / validation dataset using a clustering method; and (iv) comparing the classification results of the replication / validation dataset using the at least one cluster classifier with the clustering results of the replication / validation dataset using the clustering method. In some embodiments, the model is a grouped multinomial generalized linear model (GLM). In some embodiments, the model is a GLM using a least absolute shrinkage selection operator (LASSO). In some embodiments, if the classification results of the replication / validation dataset using the at least one cluster classifier are similar to the clustering results of the replication / validation dataset using the clustering method, it indicates that the cluster classifier is effective for classifying the replication / validation dataset.

[0152] In some embodiments, the method for predicting the response of lymphoma patients to cancer treatment further comprises: (i) identifying at least a cluster classifier by training a classifier model using the clustering results of the discovery dataset; and (ii) applying at least one cluster classifier to the replication / validation dataset to classify the replication / validation dataset. In some embodiments, the model is a grouped multinomial generalized linear model (GLM). In some embodiments, the model is a GLM using a least absolute shrinkage selection operator (LASSO).

[0153] In certain embodiments, the at least one gene is selected from the genes in Table 1. In certain embodiments, the at least one gene includes one, two, three, four, five, or more genes in Table 1. In certain embodiments, the at least one gene includes all of the genes in Table 1.

[0154] In some embodiments, the expression level of at least one gene in the discovery dataset is used to train the classifier model. In some embodiments, one, two, three, four, five, or more of the genes in Table 1 are used to train the classifier model. In some embodiments, all of the genes in Table 1 are used to train the classifier model.

[0155] In some embodiments, the classifier model includes one, two, three, four, five, or more genes from Table 1. In some embodiments, the classifier model includes all of the genes identified in Table 1.

[0156] In some embodiments, the determining step applies a clustering method to determine which subgroup a lymphoma patient belongs to using gene expression levels in the lymphoma patient's biological sample.

[0157] In some embodiments, the predicting step is applied to a trained classifier model to predict the responsiveness of the lymphoma patient to the first cancer treatment. In some embodiments, the predicting step is applied to a trained GLM model to predict the responsiveness of the lymphoma patient to the first cancer treatment.

[0158] In some embodiments, the lymphoma is diffuse large B-cell lymphoma (DLBCL). In some embodiments, the lymphoma is an indolent B-cell lymphoma. In some embodiments, the lymphoma is selected from the group consisting of follicular lymphoma, small lymphocytic lymphoma, nodal marginal zone B-cell lymphoma, lymphoplasmacytic lymphoma, anaplastic large cell lymphoma, primary cutaneous lymphoma, mycosis fungoides, chronic lymphocytic leukemia, and mantle cell lymphoma. In some embodiments, the lymphoma is follicular lymphoma. In some embodiments, the lymphoma is nodal marginal zone B-cell lymphoma. In some embodiments, the lymphoma is mantle cell lymphoma. In some embodiments, the lymphoma is chronic lymphocytic leukemia.

[0159] In some embodiments, the first cancer treatment is a combination treatment with rituximab (Rituxan), cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP).

[0160] In some embodiments, the second cancer treatment is R-CHOP. In some embodiments, the second cancer treatment is not R-CHOP.

[0161] In some embodiments, the second cancer treatment is a bromodomain and extraterminal (BET) inhibitor or a cyclin-dependent kinase (CDK) inhibitor.

[0162] Bromodomains (BDs) are protein modules of approximately 110 amino acids that recognize acetylated lysines on histones and other proteins. The BET family is a subset of 46 bromodomain-containing proteins found only in the human genome. BET proteins are composed of four proteins, namely, bromodomain-containing protein 2 (BRD2), BRD3, BRD4, and bromodomain testis-specific protein (BRDT). See Cochran et al., Nature Reviews Drug Discovery, 2019, 18:609-628; Duan et al., MedChemComm, 2018, 9:1779-1802. Tool compounds exhibiting robust preclinical activity have led to progress in the development of BET inhibitors, primarily in oncological conditions including DLBCL. As chemical probes described for other BD targets, clinical BET inhibitors are acetyl-lysine mimetics with a heterocyclic core that occupies the BD pocket 15. Deregulation of BET proteins, particularly BRD4, has been implicated in the development of various diseases, especially cancer. See Duan et al., MedChemComm, 2018, 9:1779-1802.

[0163] In some embodiments, the BET inhibitor is selected from the group consisting of OTX015, MK-8628, CPI-0610, BMS-986158, ZEN003694, GSK2820151, GSK525762, INCB054329, INCB057643, ODM-207, RO6870810, BAY1238097, CC-90010, AZD5153, FT-1101, ABBV-075, ABBV-744, SF1126, GS-5829, and CPI-0610.

[0164] In some embodiments, the BET inhibitor is a bromodomain-containing protein 4 (BRD4) inhibitor. In some embodiments, the BRD4 inhibitor is selected from the group consisting of OTX015, TEN-010, GSK525762, CPI-0610, I-BET151, PLX51107, INCB0543294, ABBV-075, BI894999, BMS-986158, and AZD5153.

[0165] Cyclin-dependent kinases (CDKs) are a family of serine-threonine kinases that have been identified as gene products involved in cell cycle control. Close cooperation between CDKs, cyclins, and cell containing endogenous inhibitors (CKIs) is necessary for orderly cell cycle progression under control. Mammalian CDKs, cyclins, and CKIs play important roles in transcriptional regulation, epigenetics, DNA damage response and repair (DDR), stemness, metabolism, and other biological processes including angiogenesis, among others. See Sanchez-Martinez et al., Bioorg. Med. Chem. Lett., 2019, 29:126637.

[0166] In some embodiments, the CDK inhibitor is PD-0332991 (Ibrance® or Palbociclib), LEE011 (Ribociclib), LY2835219 (Verzenio® or Abemaciclib), G1T28 (Trilaciclib), G1T38 (Lerociclib), SHR-6390, Flavopiridol (Alvocidib), PHA848125 (Milciclib), BCD-115, MM-D37K, PF-06873600, TG-02 (SB-1317 or Zoriraciclib), C7001 (ICEC 0942), BEY-1107, XZP-3287 (Birociclib), BPI-16350, FCN-437, CYC-065, R-Roscovitine (CY-2 02 or Seliciclib), AT-7519, AGM-130 (Inditinib), FN-1501, SY-1365, AZD-4573, TP-1287, P-14 46A-05 (Voruciclib), BAY-1251152, SCH-727965 (MK-765 or Dinaciclib), BEBT-209, TQB-3616, BAY-1000394 (Roniciclib), BAY-1143572 (Atuveciclib), and AGM-925 (FLX-925).

[0167] 5.3.1. Double-Hit Signature (DHITsig) Classifier It is well documented that there is a subset of high-risk DLBCL patients characterized by so-called "double-hit" (DHIT) translocations in MYC and either BCL2 or BCL6. Patients who are DHIT+ tend to have worse survival than DHIT patients, and these patients are typically GCB with MYC+BCL2 translocations. DHIT status has been measured using DNA sequencing or FISH probes to determine translocation status.

[0168] Ennishi et al. derived a gene expression signature reflective of DHIT status to capture broader segments of the population with different clinical outcomes. See Ennishi et al., J. Clin. Oncol., 2018, 37:190-201. Using the 104 genes, parameterization, and methodology described in the manuscript, the DHITsig method was adapted for use in the single sample normalized RNAseq space.

[0169] The Ennishi method for calculating the DHITsig score is a variable importance weighted sum of log-likelihood ratios. The likelihood function is calculated by assuming a Gaussian mixture model of gene expression that defines two normal distributions of expression for DHITsig+ and DHITsig- samples for each signature gene. The gene list and variable weights from Ennishi et al. were used with the distribution parameters of the mixture model derived from the DESeq2 processed dataset.

[0170] 5.3.2. Identification and treatment of biological characteristics of clusters of patients In some embodiments, the reference patients of the reference patient group are clustered into subgroups A1 to A7; (i) subgroup A1 includes about 50% to about 60% of patients with germinal center B cell-like (GCB) DLBCL, about 30% to about 40% of patients with activated B cell-like (ABC) DLBCL, about 10% to about 20% of patients who are TME+ DLBCL patients, and about 30% to about 40% of patients who are DHITsig+ DLBCL patients; (ii) subgroup A2 includes about 80% to about 90% of patients with GCB DLBCL, about 0% to about 5% of patients with ABC DLBCL, about 15% to about 25% of patients who are TME+ DLBCL patients, and about 25% to about 35% of patients who are DHITsig+ DLBCL patients; (iii) subgroup A3 includes about 40% to about 55% of patients with GCB DLBCL, about 10% to about 20% of patients who are ABC DLBCL, and about 25% to about 35% of patients who are DHITsig+ DLBCL patients. (iv) subgroup A4 includes about 25% to about 35% of patients with GCB DLBCL, about 40% to about 50% of patients with ABC DLBCL, about 30% to about 40% of patients with TME+ DLBCL, and about 10% to about 20% of patients with DHITsig+ DLBCL; (v) subgroup A5 includes about 20% to about 40% of patients with GCB DLBCL, about 45% to about 65% of patients with ABC DLBCL, about 30% to about 40% of patients with TME+ DLBCL, and about 0% to about 10% of patients with DHITsig+ DLBCL; (vi) subgroup A6 includes about 25% to about 35% of patients with GCB DLBCL, about 40% to about 50% of patients with ABC DLBCL, about 30% to about 40% of patients with TME+ DLBCL, and about 0% to about 10% of patients with DHITsig+ DLBCL. and (vii) subgroup A7, which includes about 0% to about 10% of patients with GCB DLBCL, about 80% to about 90% of patients with ABC DLBCL, about 0% to about 10% of patients with TME+ DLBCL, and about 5% to about 15% of patients with DHITsig+ DLBCL.

[0171] In some embodiments, if the lymphoma patient is determined to belong to subgroup A1, the method includes predicting that the patient is unlikely to respond to the first cancer treatment. In some embodiments, if the lymphoma patient is determined to belong to subgroup A2, the method includes predicting that the patient is unlikely to respond to the first cancer treatment. In some embodiments, if the lymphoma patient is determined to belong to subgroup A3, the method includes predicting that the patient is unlikely to respond to the first cancer treatment. In some embodiments, if the lymphoma patient is determined to belong to subgroup A4, the method includes predicting that the patient is unlikely to respond to the first cancer treatment. In some embodiments, if the lymphoma patient is determined to belong to subgroup A5, the method includes predicting that the patient is unlikely to respond to the first cancer treatment. In some embodiments, if the lymphoma patient is determined to belong to subgroup A6, the method includes predicting that the patient is unlikely to respond to the first cancer treatment. In some embodiments, if the lymphoma patient is determined to belong to subgroup A7, the method includes predicting that the patient is unlikely to respond to the first cancer treatment.

[0172] In some embodiments, if the lymphoma patient is determined to belong to subgroup A7 activated B-cell-like (ABC) lymphoma patients, the method includes predicting that the patient is unlikely to respond to the first cancer treatment.

[0173] In some embodiments, if the lymphoma patient is determined to belong to any of the subgroups of patients predicted to be unlikely to respond to the first cancer therapy, the lymphoma patient is treated with a second cancer therapy. In some embodiments, the second cancer therapy is a BET inhibitor or a CDK inhibitor. In some embodiments, the second cancer therapy is a CDK inhibitor. In some aspects, the second cancer therapy is a BET inhibitor.

[0174] In some embodiments, if the lymphoma patient is determined to belong to subgroup A7, the second cancer treatment is a BET inhibitor or a CDK inhibitor. In some embodiments, if the lymphoma patient is determined to belong to subgroup A7, the second cancer treatment is a CDK inhibitor. In some embodiments, if the lymphoma patient is determined to belong to subgroup A7, the second cancer treatment is a BET inhibitor.

[0175] In some embodiments, if the lymphoma patient is determined to belong to any of the subgroups of patients predicted to be unlikely to respond to the first cancer treatment, the lymphoma patient is treated with a BCL2 inhibitor. In some embodiments, if the lymphoma patient is determined to belong to any of the subgroups of patients predicted to be unlikely to respond to the first cancer treatment, the lymphoma patient is treated with an agent that increases FAS expression. In some embodiments, if the lymphoma patient is determined to belong to any of the subgroups of patients predicted to be unlikely to respond to the first cancer treatment, the lymphoma patient is treated with an inhibitor of human leukocyte antigen (HLA) genes. In some embodiments, the second cancer treatment is an inhibitor of HLA-A. In some embodiments, the second cancer treatment is an inhibitor of HLA-B. In some embodiments, the second cancer treatment is an inhibitor of HLA-C. In some embodiments, the second cancer treatment is an inhibitor of HLA-E. In some embodiments, the second cancer treatment is an inhibitor of HLA-F. In some embodiments, if the lymphoma patient is determined to belong to any of the subgroups of patients predicted to be unlikely to respond to the first cancer treatment, the lymphoma patient is treated with a CD47 treatment. In some embodiments, if the lymphoma patient is determined to belong to any of the subgroups of patients predicted to be unlikely to respond to the first cancer treatment, the lymphoma patient is treated with an IDO inhibitor or an agent that depletes regulatory T cells. In some embodiments, the second cancer treatment is an IDO inhibitor. In some embodiments, the second cancer treatment is an agent that depletes regulatory T cells. In some embodiments, if the lymphoma patient is determined to belong to any of the subgroups of patients predicted to be unlikely to respond to the first cancer treatment, the lymphoma patient is treated with a histone deacetylase (HDAC) inhibitor. In some embodiments, if the lymphoma patient is determined to belong to any of the subgroups of patients predicted to be unlikely to respond to the first cancer treatment, the lymphoma patient is treated with a galectin-3 (Gal3) inhibitor.

[0176] In some embodiments, the lymphoma patient belongs to a subgroup based on mutation data for at least one feature listed in Table 6 or Table 7.

[0177] In another aspect, provided herein is a method of predicting a lymphoma patient's responsiveness to a cancer treatment, the method comprising: (a) obtaining a first biological sample from a first patient having lymphoma; (b) determining an expression level of one, two, three, four, five or more genes set forth in Table 1; (c) comparing the expression level of the one, two, three, four, five or more genes set forth in Table 1 in the first biological sample with the expression level of the same genes in a second biological sample of a second patient, wherein the second lymphoma patient will respond to the cancer treatment, and a similar expression of the one, two, three, four, five or more genes in the first biological sample to the expression level of the one, two, three, four, five or more genes in the second biological sample indicates that the lymphoma of the first patient will respond to treatment with the cancer treatment.

[0178] In another aspect, provided herein is a method of predicting responsiveness of a lymphoma patient to a cancer treatment, the method comprising: (a) obtaining a first biological sample from a first lymphoma patient; (b) determining expression of a gene or a particular subset of genes set forth in Table 1, or any combination thereof, in the first biological sample; and (c) comparing a gene expression profile of the genes or the subset of genes in the first biological sample to a gene expression profile of the genes or the subset of genes in (i) a biological sample of a lymphoma patient that is responsive to a drug, and (ii) a biological sample of a lymphoma patient that is not responsive to a cancer treatment. The method includes a step of comparing the gene expression of a subset of genes in a first biological sample with that of a second biological sample, wherein a gene expression profile of the gene or subset of genes in the first biological sample is similar to a gene expression profile of the gene or subset of genes in a biological sample of a lymphoma patient that responds to the cancer treatment indicates that the first lymphoma patient that responds to the cancer treatment is responsive to the cancer treatment, and a gene expression profile of the gene or subset of genes in the first biological sample is similar to a gene expression profile of the gene or subset of genes in a biological sample of a lymphoma patient that does not respond to the cancer treatment indicates that the first lymphoma patient does not respond to the cancer treatment.

[0179] In another aspect, a method for predicting the responsiveness of a lymphoma patient to cancer treatment is provided herein, the method comprising: (a) determining the expression level of at least one gene of Table 1 in a biological sample of a lymphoma patient; (b) comparing the expression level of the at least one gene of step (a) with the expression level of at least one gene in a reference biological sample of a reference lymphoma patient, and if the reference lymphoma patient responds to the cancer treatment and the expression level of the at least one gene in the biological sample is similar to the expression level of the at least one gene in the reference biological sample, it indicates that the lymphoma patient is unlikely to respond to the cancer treatment. In certain embodiments, the method further comprises obtaining a biological sample from the lymphoma patient. In certain embodiments, the at least one gene comprises five or more genes of Table 1.

[0180] In another aspect, provided herein is a method for predicting the responsiveness of a lymphoma patient to a cancer treatment, the method comprising: (a) determining an expression level of at least one gene of Table 1 in a biological sample of the lymphoma patient; (b) comparing the expression level of the at least one gene of step (a) with an expression level of at least one gene in a reference biological sample of a group of reference lymphoma patients, where if the reference lymphoma patient responds to the cancer treatment and the expression level of the at least one gene in the biological sample is similar to the expression level of the at least one gene in the reference biological sample, it indicates that the lymphoma patient is unlikely to respond to the cancer treatment. In certain embodiments, the method further comprises obtaining a biological sample from the lymphoma patient. In certain embodiments, the at least one gene comprises five or more genes of Table 1. In certain embodiments, the expression level of the at least one gene in the reference biological sample is the mean or median of the expression levels of the genes measured in the reference biological samples of the reference lymphoma patients.

[0181] In another aspect, a method for predicting the responsiveness of a lymphoma patient to a cancer treatment is provided herein, the method comprising: (a) determining an expression level of at least one gene of Table 1 in a biological sample of the lymphoma patient; and (b) comparing the expression level of the at least one gene in the biological sample with: (i) an expression level of at least one gene in a biological sample of a lymphoma patient who responds to the cancer treatment, and (ii) an expression level of at least one gene in a biological sample of a lymphoma patient who does not respond to the cancer treatment, whereby if the expression level of (a) is similar to the expression level of (i), it indicates that the lymphoma patient is more likely to respond to the cancer treatment; and if the expression level of (a) is similar to the expression level of (ii), it indicates that the lymphoma patient is less likely to respond to the cancer treatment. In certain embodiments, the method further comprises obtaining a biological sample from the lymphoma patient. In certain embodiments, the at least one gene comprises five or more genes of Table 1. In certain embodiments, the expression level of the at least one gene in a biological sample of a lymphoma patient non-responsive or responsive to cancer treatment is the mean expression or median expression of the expression level of the at least one gene measured in biological samples of lymphoma patients non-responsive or responsive to cancer treatment.

[0182] In certain embodiments, two expression levels of a gene are similar means that the expression level of one of the genes is within at most 10% (e.g., 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, or 10%) of the other expression level of the same gene. In certain embodiments, the two expression levels are measured on an absolute scale. In certain embodiments, one expression level is the expression level of the gene in a biological sample of a lymphoma patient and the other expression level is the expression level of the same gene in a reference biological sample of a reference lymphoma patient. In certain embodiments, one expression level is the expression level of the gene in a biological sample of a first lymphoma patient and the other expression level is the expression level of the same gene in a biological sample of a second lymphoma patient.

[0183] In some embodiments, the two expression levels are similar, and the expression level of one of the genes in the biological samples is within one standard deviation or standard error value of the expression level of the same gene in the biological samples of a group of subjects (e.g., a group of reference lymphoma patients, a group of lymphoma patients responding to cancer treatment, or a group of lymphoma patients not responding to cancer treatment). In some embodiments, the two expression levels of the two groups of subjects are similar means that the average value of the expression level of one of the genes in the biological samples of one group of subjects is within one standard deviation or standard error value of the average value of the expression level of the same gene in the biological samples of another group of subjects. In some embodiments, the two expression levels of the two groups of subjects are similar means that a hypothesis test, such as a Wilcoxon test or a t-test, fails to reject the null hypothesis of equality of the two means or two medians at a predefined significance level, where each mean or median is the average or median value of the expression level of the same gene in the biological samples of one of the two groups of subjects. In some embodiments, the group of subjects is lymphoma patients responding to cancer treatment. In some embodiments, the group of subjects are lymphoma patients who do not respond to cancer treatment. In some embodiments, the group of subjects is a reference group of lymphoma patients.

[0184] In some embodiments, the determining step of the methods described herein comprises determining the expression of all of the genes listed in Table 1. In some embodiments, the expression levels of all of the genes listed in Table 1 are determined and compared. In some embodiments, the determining step of the methods described herein comprises determining the expression of ABHD10, ACO1, ACTN4, AGRP, AKAP13, ALDH1A1, ALG13, AMT, ANKZF1, AOAH, AP1G2, AP3S1, APRT, ARG1, ARHGDIA, ARHGEF7, ART4, ASH1L, ATIC, ATP6V1G2, ATP9B, BAZ2A, BLNK, BPNT1, BRIP1, BTF3, BUB3, C1QBP, C2, CACUL1, CADPS, CAPZB CARD11, CBX5, CCDC136, CCR2, CCT7, CCT8, CD37, CD46, CDC25A, CDK12, CENPW, CEP85L, CEP97, CFH, CHD2, CHI3L1, CHKA, CIB1, CKAP2, CLCN7 , CLEC7A, CLIC1, CLK1, CMSS1, COG7, COL1A2, COL4A3, CORO1A, COX6C, CPD, CRBN, CSF1, CUEDC2, CUX1, CXCL10, DDHD1, DDX58, DIMT1, DNAJA3,DNAJC10, DNMT1, DYNLT1, E2F1, EBAG9, EIF1AX, EIF2B5, EIF3I, EIF5A, ELF1, ELMO1, E MP3, ENO1, EPS15, ERGIC2, ERH, ESD, EYS, FBXO46, FLNA, FOXP1, FPR1, FUBP1, FUS, GALM GAPDH、GATM、GBP5、GIMAP4、GJD3、GLUL、GNA13 GNB2, GNS, GPR82, GPX1, GRIP1, HAMP, HEXIM1, HNMT, HNRNPA2B1, HNRNPU. HSD11B1L, HSP90AA1, HSPD1, HSPE1, IDS, IFI30, IFITM3, IKZF1, IL24, IMP DH2, IST1, ITGB2, JAK1, KIF14, KIF4A, KLHL14, KLHL23, LAP3, LATS1, LMO4 LONP2, LPAR6, LRCH4, LRRC37A2, LRRC37A3, LRRC59, LY9, LYRM1, MAP3K14 MAP4K2, MAPK14, MARS2, MAT2A, MAX, MCL1, MCM4, MGAT4A, MRPL43, MRPS1 5. MRPS9, MSH6, MVB12A, MVB12B, MVP, MYBL2, MYOF, NACA, NCAPG2, NCBP2, N CL, NFE2L2, NFKBIA, NRXN1, NUDT21, ORC6, P2RX7, PAICS, PARG, PARP9, PAX 5. PCNA,PDE9A, PFKL, PGAM1, PIF1, PILRB, PKIA, PKM, PKN2, PLEKHF2, PLK1 PMM2, PNP, POLA1, POLR2E, POM121, PPA1, PPIA, PPP1R9B, PRDM15, PRDX5, PRKCB, PRKCH, PRMT1, PR PF4, RPF6, PRRC2C, PSMC1, PSMD13, PSME2, ​​PTBP3, PTENP1, PTPRC, R3HDM1, RAB32, RAMP3, RANBP2,RBM3, RBM33, RBP1, RFC1, RNASEL, RNF38, RPL30, RPL32P3, RPRD2, RPS3, RPS4X, RRP9, S100A11, S100Z, S1PR2, SAAL1, S BNO1, SCAMP2, SCARB1, SCARB2, SDCBP, SELPLG, SENP7, SERPINA3, SERPINB1, SF1, SF3A1, SFPQ, SFXN3, SH3BGRL3, SH3BP 1, SLAMF8, SLC35D1, SMARCA4, SMARCC1, SMG5, SNORA21, SNORA71B, SNORD104, SNRPA, SNRPD1, SNTA1, SNX29, SOBP, SOG A1, SP140, SP3, SPEN, SRGN, SRSF6, STAG3, STK4, STMN1, SUMO2, SYNJ2, SYPL1, SYTL3, TAGLN2, TAP2, TARDBP, TBCA, TGDS , TGOLN2, THRAP3, TIMM10, TLK1, TLN1, TMBIM4, TMEM223, TNFRSF1A, TONSL, TP53INP1, TPM3, TRA2B, TRAM1,TRAPPC12, TRIOBP, TRIP13, TRMT1L, TRNT1, TSPYL2, TSSK4, TUBA1B, TUBA1C, TWIST1, UBE2B, UBE2D2, UBE2G1, UXT, VASH1, VAV1, VD The method includes determining the expression of at least one gene (e.g., one, two, three, four, five, or more) selected from the group consisting of AC1, WEE1, XRCC6, YTHDC1, YWHAE, ZBED5, ZBTB37, ZFAND4, ZFAND5, ZMAT1, ZNF101, ZNF107, ZNF146, ZNF207, ZNF318, ZNF367, ZNF480, and ZWINT, or all of the genes in this group.

[0185] In some embodiments, the cancer treatment is a combination treatment with rituximab (Rituxan), cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP).

[0186] In another aspect, provided herein is a method of treating a lymphoma patient, the method comprising: (i) identifying a lymphoma patient likely to respond to a predicted cancer treatment using the predictive methods described herein above; and (ii) administering a cancer treatment to the lymphoma patient.

[0187] In another aspect, provided herein is a method of treating a lymphoma patient, the method comprising: (i) identifying a lymphoma patient who is unlikely to respond to a cancer treatment predicted using the predictive methods described herein above; and (ii) administering an alternative cancer treatment to the lymphoma patient. In some embodiments, the alternative cancer treatment is a BET inhibitor or a CDK inhibitor.

[0188] In addition, in some embodiments, all of the genes listed in Table 1 can be used as biomarkers to predict responsiveness to treatment in lymphoma (eg, DLBCL) patients.

[0189] In another aspect, the subgroups provided herein (e.g., A1-A7) can be characterized and / or identified based on the Bcl6 signature score, as shown in the Examples section below. Thus, the Bcl6 signature score can also be used as a method of classifying patients into one of eight subgroups for purposes of determining the patient's responsiveness to treatment.

[0190] In another aspect, mutational data was collected and interpreted in relation to the identified subgroups, as shown in the Examples section below and in Tables 6 and 7. Thus, in some embodiments, the mutational profile of each subgroup (or cluster), or a subset thereof, can also be used to identify the subgroups or classify patients into one of the subgroups for purposes of determining the patient's responsiveness to treatment.

[0191] The subgroups (or clusters) provided herein were also characterized based on the total number of different T cell populations (e.g., CD3, CD4, CD8, CD163, CD68, and / or CD11c cells), as shown in the Examples section below and in Figures 11A-11F. Thus, in yet another embodiment, the proportions of different T cell populations (e.g., CD3, CD4, CD8, CD163, CD68, and / or CD11c cells) can be used to identify subgroups or classify patients into one of the subgroups for purposes of determining the patient's responsiveness to treatment.

[0192] 5.4 Method of administration In some embodiments, the methods provided herein include administering a first cancer therapeutic compound to a lymphoma patient predicted to respond to the first cancer treatment. Also provided herein is a method of treating a patient who has previously been treated for cancer (e.g., DLBCL or a subtype thereof) but has not responded to the first cancer treatment (e.g., standard treatment). Also provided herein is a method of treating a patient who has not been previously treated. The present invention also encompasses methods of treating a patient regardless of the patient's age, although some diseases or disorders are more prevalent in certain age groups. The present invention further encompasses methods of treating patients who have undergone surgery to treat a disease or condition, as well as patients who have not undergone surgery. Cancer patients have heterogeneous clinical manifestations and varying clinical outcomes, so the treatment given to a patient may vary depending on the patient's prognosis. A skilled clinician can easily determine, without undue experimentation, the particular second-line agent, type of surgery, and type of non-drug-based standard treatment that can be effectively used to treat an individual cancer patient (e.g., DLBCL or a subtype thereof).

[0193] In certain embodiments, a therapeutically or prophylactically effective amount for treating cancer is from about 0.005 mg / day to about 1,000 mg / day, from about 0.01 mg / day to about 500 mg / day, from about 0.01 mg / day to about 250 mg / day, from about 0.01 mg / day to about 100 mg / day, from about 0.1 mg / day to about 100 mg / day, from about 0.5 mg / day to about 100 mg / day, from about 1 mg / day to about 100 mg / day, from about 0.01 mg / day to about 50 mg / day, from about 0.1 mg / day to about 50 mg / day, from about 0.5 mg / day to about 50 mg / day, from about 1 mg / day to about 50 mg / day, from about 0.02 mg / day to about 25 mg / day, or from about 0.05 mg / day to about 10 mg / day.

[0194] In certain embodiments, the therapeutically or prophylactically effective amount is about 0.1 mg / day, about 0.2 mg / day, about 0.5 mg / day, about 1 mg / day, about 2 mg / day, about 5 mg / day, about 10 mg / day, about 15 mg / day, about 20 mg / day, about 25 mg / day, about 30 mg / day, about 40 mg / day, about 45 mg / day, about 50 mg / day, about 60 mg / day, about 70 mg / day, about 80 mg / day, about 90 mg / day, about 100 mg / day, or about 150 mg / day.

[0195] In some embodiments, the recommended daily dose range for cancer treatment for the conditions described herein is within the range of about 0.1 mg / day to about 50 mg / day, preferably administered as a single daily dose or in divided doses throughout the day. In some embodiments, the dosage ranges from about 1 mg / day to about 50 mg / day. In some embodiments, the dosage ranges from about 0.5 mg / day to about 5 mg / day. In some embodiments, the specific dose per day is 0.1 mg / day, 0.2 mg / day, 0.5 mg / day, 1 mg / day, 2 mg / day, 3 mg / day, 4 mg / day, 5 mg / day, 6 mg / day, 7 mg / day, 8 mg / day, 9 mg / day, 10 mg / day, 11 mg / day, 12 mg / day, 13 mg / day, 14 mg / day, 15 mg / day, 16 mg / day, 17 mg / day, 18 mg / day, 19 mg / day, 20 mg / day, 21 mg / day, 22 mg / day, 23 mg / day, 24 mg / day, 25 mg / day, 26 mg / day, 27 mg / day, 28 mg / day, 29 mg / day, 30 mg / day, 31 mg / day, 32 mg / day, 33 mg / day, 34 mg / day, 35 mg / day, 36 mg / day, 37 mg / day, 38 mg / day, 39 mg / day, 40 mg / day, 41 mg / day, 42 mg / day, 43 mg / day, 44 mg / day, 45 mg / day, 46 mg / day, 47 mg / day, 48 mg / day, 49 mg / day, 50 mg / day, 51 mg / day, 52 mg / day, 53 mg / day, 54 mg / day, 55 mg / day, 56 mg / day, 57 mg / day, 58 mg / day, 59 mg / day, 60 mg / day, 61 mg / day, 62 mg / day, 63 mg / day, 64 mg / day, 65 mg / day, 66 mg / day, 67 mg / day, 68 mg / day, 69 mg / day, 70 g / day, 24 mg / day, 25 mg / day, 26 mg / day, 27 mg / day, 28 mg / day, 29 mg / day, 30 mg / day, 31 mg / day, 32 mg / day, 33 mg / day, 34 mg / day, 35 mg / day, 36 mg / day, 37 mg / day, 38 mg / day, 39 mg / day, 40 mg / day, 41 mg / day, 42 mg / day, 43 mg / day, 44 mg / day, 45 mg / day, 46 mg / day, 47 mg / day, 48 mg / day, 49 mg / day, or 50 mg / day.

[0196] In some embodiments, the recommended starting dose may be 0.5 mg / day, 1 mg / day, 2 mg / day, 3 mg / day, 4 mg / day, 5 mg / day, 10 mg / day, 15 mg / day, 20 mg / day, 25 mg / day, or 50 mg / day. In some embodiments, the recommended starting dose may be 0.5 mg / day, 1 mg / day, 2 mg / day, 3 mg / day, 4 mg / day, or 5 mg / day. The recommended starting dose may be gradually increased to 10 mg / day, 15 mg / day, 20 mg / day, 25 mg / day, 30 mg / day, 35 mg / day, 40 mg / day, 45 mg / day, or 50 mg / day.

[0197] In some embodiments, a therapeutically or prophylactically effective amount is from about 0.001 mg / kg / day to about 100 mg / kg / day, from about 0.01 mg / kg / day to about 50 mg / kg / day, from about 0.01 mg / kg / day to about 25 mg / kg / day, from about 0.01 mg / kg / day to about 10 mg / kg / day, from about 0.01 mg / kg / day to about 9 mg / kg / day, from about 0.01 mg / kg / day to about 8 mg / kg / day, from about 0.01 mg / kg / day to about 100 mg / kg / day, from about 0.01 mg / kg / day to about 50 mg / kg / day, from about 0.01 mg / kg / day to about 25 mg / kg / day, from about 0.01 mg / kg / day to about 10 mg / kg / day, from about 0.01 mg / kg / day to about 9 mg / kg / day, from about 0.01 mg / kg / day to about 8 mg / kg / day, day, about 0.01 mg / kg / day to about 7 mg / kg / day, about 0.01 mg / kg / day to about 6 mg / kg / day, about 0.01 mg / kg / day to about 5 mg / kg / day, about 0.01 mg / kg / day to about 4 mg / kg / day, about 0.01 mg / kg / day to about 3 mg / kg / day, about 0.01 mg / kg / day to about 2 mg / kg / day, or about 0.01 mg / kg / day to about 1 mg / kg / day.

[0198] In some embodiments, the administered dose can be expressed in units other than mg / kg / day. For example, the parenteral dose can be expressed in mg / m 2 The dosage can be expressed as mg / kg / day. One of skill in the art can easily determine the dosage from mg / kg / day to mg / m2, given the subject's height or weight or both. 2 It will be readily apparent how to convert this to a daily dose (see www.fda.gov). For example, a dose of 1 mg / kg / day for a 65 kg human is 38 mg / m 2 Approximately equal to / day.

[0199] In certain embodiments, the amount of cancer therapy administered is sufficient to provide a steady state plasma concentration of the compound in the range of about 0.001 μM to about 500 μM, about 0.002 μM to about 200 μM, about 0.005 μM to about 100 μM, about 0.01 μM to about 50 μM, about 1 μM to about 50 μM, about 0.02 μM to about 25 μM, about 0.05 μM to about 20 μM, about 0.1 μM to about 20 μM, about 0.5 μM to about 20 μM, or about 1 μM to about 20 μM.

[0200] In some embodiments, the amount of cancer therapy administered is sufficient to provide a steady state plasma concentration of the compound in the range of about 5 nM to about 100 nM, about 5 nM to about 50 nM, about 10 nM to about 100 nM, about 10 nM to about 50 nM, or about 50 nM to about 100 nM.

[0201] As used herein, the term "steady state plasma concentration" is the concentration reached after a period of administration of a cancer treatment provided herein. When steady state is reached, there are small peaks and valleys in the time-dependent curve of the plasma concentration of the cancer treatment.

[0202] In some embodiments, the amount of cancer therapy administered is an amount sufficient to provide a maximum plasma concentration (peak concentration) of the compound in the range of about 0.001 μM to about 500 μM, about 0.002 μM to about 200 μM, about 0.005 μM to about 100 μM, about 0.01 μM to about 50 μM, about 1 μM to about 50 μM, about 0.02 μM to about 25 μM, about 0.05 μM to about 20 μM, about 0.1 μM to about 20 μM, about 0.5 μM to about 20 μM, or about 1 μM to about 20 μM.

[0203] In some embodiments, the amount of cancer therapy administered is an amount sufficient to provide a minimum plasma concentration (trough concentration) of the compound in the range of about 0.001 μM to about 500 μM, about 0.002 μM to about 200 μM, about 0.005 μM to about 100 μM, about 0.01 μM to about 50 μM, about 1 μM to about 50 μM, about 0.01 μM to about 25 μM, about 0.01 μM to about 20 μM, about 0.02 μM to about 20 μM, about 0.02 μM to about 20 μM, or about 0.01 μM to about 20 μM.

[0204] In some embodiments, the amount of cancer therapy administered is sufficient to provide an area under the curve (AUC) of the compound in the range of about 100 ng×hr / mL to about 100,000 ng×hr / mL, about 1,000 ng×hr / mL to about 50,000 ng×hr / mL, about 5,000 ng×hr / mL to about 25,000 ng×hr / mL, or about 5,000 ng×hr / mL to about 10,000 ng×hr / mL.

[0205] In some embodiments, a lymphoma patient to be treated with one of the methods provided herein has not been treated with an anti-cancer therapy prior to administering standard of care (e.g., R-CHOP). In some embodiments, a lymphoma patient to be treated with one of the methods provided herein has been treated with an anti-cancer therapy (standard of care, e.g., R-CHOP) prior to administering a second therapy. In some embodiments, a lymphoma patient to be treated with one of the methods provided herein has developed drug resistance to a first cancer therapy.

[0206] Depending on the subtype of lymphoma to be treated (e.g., DLBCL) and the condition of the subject, the cancer therapy is administered by parenteral (e.g., intramuscular, intraperitoneal, intravenous, CIV, intracisternal injection or infusion, subcutaneous injection, or implant), inhalation, nasal, vaginal, rectal, sublingual, or topical (e.g., transdermal or topical) routes of administration. In some embodiments, the cancer therapy is formulated with pharma- ceutically acceptable excipients, carriers, adjuvants, and vehicles, alone or together, in appropriate dosage units appropriate for each route of administration.

[0207] In some embodiments, the cancer treatment is administered parenterally. In some embodiments, the cancer treatment is administered intravenously.

[0208] Depending on the stage of lymphoma to be treated and the condition of the subject, in some embodiments, the therapeutic compounds are administered by oral, parenteral (e.g., intramuscular, intraperitoneal, intravenous, CIV, intracisternal injection or infusion, subcutaneous injection, or implant), inhalation, nasal, vaginal, rectal, sublingual, or topical (e.g., transdermal or topical) routes of administration. In some embodiments, the therapeutic compounds, alone or together, are formulated with pharma- ceutically acceptable excipients, carriers, adjuvants, and vehicles in appropriate dosage units suitable for each route of administration.

[0209] In some embodiments, the therapeutic compound is administered orally. In some embodiments, the therapeutic compound is administered parenterally. In some embodiments, the therapeutic compound is administered intravenously.

[0210] In some embodiments, the therapeutic compound can be delivered as a single dose, e.g., as a single bolus injection, or as an oral capsule, tablet, or pill; or over time, e.g., as a continuous infusion over time or as divided bolus doses over time. In some embodiments, the cancer treatments described herein can be administered repeatedly as necessary, e.g., until the patient experiences stable disease or regression, or until the patient experiences disease progression or unacceptable toxicity.

[0211] In some embodiments, the therapeutic compound can be administered once a day (QD) or divided into multiple daily doses, such as twice a day (BID), three times a day (TID), and four times a day (QID). In some embodiments, administration can be continuous (i.e., every day for consecutive days, or every day), intermittent, e.g., cyclical (i.e., with days, weeks, or months of drug-free rest). As used herein, the term "daily" is intended to mean that the therapeutic compound is administered, e.g., once or more times each day for a period of time. The term "continuous" is intended to mean that the therapeutic compound is administered every day for an uninterrupted period of at least 7 days to 52 weeks. As used herein, the term "intermittent" or "intermittently" is intended to mean stopping and starting at regular or irregular intervals. In some embodiments, intermittent administration is administration 1-6 days per week, cyclical administration (e.g., daily administration for 2-8 consecutive weeks followed by a rest period of up to 1 week without administration), or every other day administration. As used herein, the term "cyclical" is intended to mean that the therapeutic compound is administered daily or consecutively, but with a rest period.

[0212] In some embodiments, the frequency of administration ranges from about daily administration to about monthly administration.

[0213] In some embodiments, the cancer treatment can be delivered as a single dose (e.g., a single bolus injection) or over time (e.g., continuous infusion over time or split bolus doses over time). In some embodiments, the compound can be administered repeatedly as needed, for example, until the patient experiences stable disease or regression, or until the patient experiences disease progression or unacceptable toxicity. For example, stable solid cancer disease generally means that the vertical diameter of a measurable lesion has not increased by more than 25% since the previous measurement. Therasse et al., J. Natl. Cancer Inst., 2000, 92(3):205-216. Stable disease or its disappearance is determined by methods known in the art, such as evaluation of patient symptoms, physical examination, and visualization of the tumor on imaging using X-ray, CAT, PET, MRI scan, or other commonly accepted evaluation modalities.

[0214] In some embodiments, the cancer treatment can be administered once a day (QD) or divided into multiple daily doses, such as twice a day (BID), three times a day (TID), and four times a day (QID). In some embodiments, administration can be continuous (i.e., every day for consecutive days, or every day), intermittent, e.g., cyclical (i.e., with days, weeks, or months of drug-free rest). As used herein, the term "daily" is intended to mean that the cancer treatment is administered, e.g., once or more times each day for a period of time. The term "continuous" is intended to mean that the cancer treatment is administered daily for an uninterrupted period of at least 10 days to 52 weeks. As used herein, the term "intermittent" or "intermittently" is intended to mean stopping and starting at regular or irregular intervals. For example, intermittent administration of the cancer treatment can be administration 1-6 days per week, cyclic administration (e.g., daily administration for 2-8 consecutive weeks followed by a rest period of up to 1 week of no administration), or administration every other day. As used herein, the term "cycle" is intended to mean that the cancer treatment is administered daily or continuously, but with a rest period. In some embodiments, the rest period is the same length as the treatment period. In some embodiments, the rest period has a different length than the treatment period. In some embodiments, the length of the cycle is 1 week, 2 weeks, 3 weeks, 4 weeks, 5 weeks, 6 weeks, 7 weeks, 8 weeks, 9 weeks, or 10 weeks. In some embodiments of the cycle, the cancer treatment is administered daily for a period of 1 day, 2 days, 3 days, 4 days, 5 days, 6 days, 7 days, 8 days, 9 days, 10 days, 15 days, 20 days, 25 days, or 30 days, followed by a rest period. In some embodiments, the cancer treatment is administered daily for 5 days of a 4-week cycle. In another specific embodiment, the cancer treatment is administered daily for 10 days of a four week cycle.

[0215] In some embodiments, the frequency of administration ranges from about daily administration to about monthly administration. In some embodiments, administration is once a day, twice a day, three times a day, four times a day, once every other day, twice a week, once a week, once every two weeks, once every three weeks, or once every four weeks. In some embodiments, the cancer treatment is administered once a day. In some embodiments, the cancer treatment is administered twice a day. In some embodiments, the cancer treatment is administered three times a day. In some embodiments, the cancer treatment is administered four times a day.

[0216] In some embodiments, the cancer treatment is administered once daily for 1 day to 6 months, 1 week to 3 months, 1 week to 4 weeks, 1 week to 3 weeks, or 1 week to 2 weeks. In some embodiments, the cancer treatment is administered once daily for 1 week, 2 weeks, 3 weeks, or 4 weeks. In some embodiments, the cancer treatment is administered once daily for 1 week. In some embodiments, the cancer treatment is administered once daily for 2 weeks. In some embodiments, the cancer treatment is administered once daily for 3 weeks. In yet another embodiment, the cancer treatment is administered once daily for 4 weeks.

[0217] 5.5 Combination Therapy One or more additional therapies, such as additional active ingredients or active agents, can be used in combination with the administration of the cancer treatment described herein to treat lymphoma patients (e.g., patients with DLBCL). In some embodiments, the one or more additional therapies can be administered before, simultaneously with, or after the administration of the compounds described herein. The administration of the cancer treatment described herein and the additional active agent ("second active agent") to the patient can be performed simultaneously or sequentially by the same or different administration routes. The suitability of a particular administration route used for a particular active agent depends on the active agent itself (e.g., whether it can be administered orally without being degraded before entering the bloodstream) and the condition of the lymphoma (e.g., DLBCL) being treated. The administration routes of the additional active agents or active ingredients are known to those skilled in the art. See, for example, the Physicians' Desk Reference.

[0218] In some embodiments, the cancer therapy described herein and an additional active agent are cyclically administered to a patient with lymphoma (e.g., DLBCL). Cycling therapy involves administering an active agent for a period of time, followed by a period of rest, and then repeating this sequential administration. Cycling therapy can reduce the development of resistance to one or more therapies, avoid or reduce the side effects of one therapy, and / or improve the efficacy of the treatment.

[0219] In some embodiments, one or more second active ingredients or agents can be used in the methods and compositions provided herein. The second active agent can be a macromolecule (e.g., a protein) or a small molecule (e.g., a synthetic inorganic, organometallic, or organic molecule). A variety of agents can be used, such as those described in U.S. Patent Application Serial No. 16 / 390,815 or U.S. Provisional Patent Application entitled "SUBSTITUTED 4-AMINOISOINDOLINE-1,3-DIONE COMPOUNDS AND SECOND ACTIVE AGENTS FOR COMBINED USE" (filed on even date herewith) (Attorney Docket No. 14247-390-888), each of which is incorporated herein by reference in its entirety. In some embodiments, exemplary second active agents include, but are not limited to, HDAC inhibitors (e.g., panobinostat, romidepsin, or vorinostat), BCL2 inhibitors (e.g., venetoclax), BTK inhibitors (e.g., ibrutinib or acalabrutinib), mTOR inhibitors (e.g., everolimus), PI3K inhibitors (e.g., idelalisib), PKCβ inhibitors (e.g., enzastaurin), SYK inhibitors (e.g., fostamatinib), JAK2 inhibitors (e.g., fedratinib, pacritinib, ruxolitinib, baricitinib, gandotinib, lestaurtinib, or momelotinib), Aurora A kinase inhibitors (e.g., acalcitinib, alpha-amycin ... risertib), EZH2 inhibitors (e.g., tazemetostat, GSK126, CPI-1205, 3-deazaneplanocin A, EPZ005687, EI1, UNC1999, or sinefungin), BET inhibitors (e.g., birabresib or 4[2-(cyclopropylmethoxy)-5-(methanesulfonyl)phenyl]-2-methylisoquinolin-1(2H)-one), hypomethylating agents (e.g., 5-azacytidine or decitabine), chemotherapy (e.g., bendamustine, doxorubicin, etoposide, methotrexate, cytarabine, vincristine, ifosfamide, or melphalan), or epigenetic compounds (e.g.,DOT1L inhibitors such as pinometostat, HAT inhibitors such as C646, WDR5 inhibitors such as OICR-9429, HDAC6 inhibitors such as ACY-241, DNMT1 selective inhibitors such as GSK3484862, LSD-1 inhibitors such as compound C or seclidemstat, G9A inhibitors such as UNC0631, PRMT5 inhibitors such as GSK3326595, BRPF1B / 2 inhibitors such as OF-1, BRD9 / 7 inhibitors such as LP99, SUV420H1 / H2 inhibitors such as A-196, Menin-MLL inhibitors such as MI-503, CARM1 inhibitors such as EZM2302, BRD9 inhibitors such as dBrd9, Aiolos / Ikaros degrading cereblon E3 ligase regulator (CELMoD), CREBBp2 inhibitors, anti-CD79b antibodies, CD19 These include CAR-T, p53 inhibitors (nutlins), Bcl6 inhibitors, CREBBp2 CELMoD, CD79b CELMoD, CD19 CELMoD, p53(nutlin) CELMoD, Bcl6 CELMoD, inhibitors of ligand-directed degradation (LDD) of CREBBP2, LDD inhibitors of CD79b, LDD inhibitors of CD19, LDD inhibitors of p53(nutlin), LDD inhibitors of Bcl6, LDD inhibitors of CK1a, LDD inhibitors of IRAK4 (e.g., in MYD88 L265p lymphoma), MALT1 inhibitors such as JNJ-67856633, MAT2A inhibitors (e.g., for 9p21 deletions), anti-CD3 x anti-CD19 bispecific antibodies, and anti-CD3 x anti-CD20 bispecific antibodies.

[0220] In some embodiments, the method further comprises administration of one or more of rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone, etoposide, bendamustine (Treanda), lenalidomide, or gemcitabine.In some embodiments, the method further comprises administration of one or more of rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone, etoposide, bendamustine (Treanda), or gemcitabine. In some embodiments, the treatment further comprises treatment with one or more of R-CHOP (rituximab with cyclophosphamide, doxorubicin, vincristine, and prednisone), R EPOCH (etoposide, rituximab with cyclophosphamide, doxorubicin, vincristine, and prednisone), stem cell transplant, bendamustine (Treanda) and rituximab, rituximab, lenalidomide and rituximab, or gemcitabine-based combinations. In some embodiments, the treatment further comprises treatment with one or more of R-CHOP (rituximab with cyclophosphamide, doxorubicin, vincristine, and prednisone), R EPOCH (etoposide, rituximab with cyclophosphamide, doxorubicin, vincristine, and prednisone), stem cell transplantation, bendamustine (Treanda) with rituximab, rituximab, or gemcitabine-based combinations. In certain embodiments, the second active agent is rituximab as provided in U.S. Provisional Patent Application No. 62 / 833,432.

[0221] In some embodiments, the second active agent used in the methods provided herein is a histone deacetylase (HDAC) inhibitor. In some embodiments, the HDAC inhibitor is panobinostat, romidepsin, or vorinostat, or a stereoisomer, mixture of stereoisomers, tautomer, isotopologue, or pharma- ceutically acceptable salt thereof.

[0222] In some embodiments, the second active agent used in the methods provided herein is a B-cell lymphoma 2 (BCL2) inhibitor. In some embodiments, the BCL2 inhibitor is venetoclax, or a tautomer, isotopologue, or pharma- ceutically acceptable salt thereof. In some embodiments, the BCL2 inhibitor is venetoclax.

[0223] In some embodiments, the second active agent used in the methods provided herein is a Bruton's tyrosine kinase (BTK) inhibitor. In some embodiments, the BTK inhibitor is ibrutinib or acalabrutinib, or a stereoisomer, mixture of stereoisomers, tautomer, isotopologue, or pharmaceutically acceptable salt thereof. In some embodiments, the BTK inhibitor is ibrutinib or a stereoisomer, mixture of stereoisomers, tautomer, isotopologue, or pharmaceutically acceptable salt thereof. In some embodiments, the BTK inhibitor is ibrutinib. In some embodiments, the BTK inhibitor is acalabrutinib or a stereoisomer, mixture of stereoisomers, tautomer, isotopologue, or pharmaceutically acceptable salt thereof. In some embodiments, the BTK inhibitor is acalabrutinib.

[0224] In some embodiments, the second active agent used in the methods provided herein is a mammalian target of rapamycin (mTOR) inhibitor. In some embodiments, the mTOR inhibitor is rapamycin or an analog thereof (also called a rapalog). In some embodiments, the mTOR inhibitor is everolimus or a stereoisomer, mixture of stereoisomers, tautomer, isotopologue, or pharma- ceutically acceptable salt thereof. In some embodiments, the mTOR inhibitor is everolimus.

[0225] In some embodiments, the second active agent used in the methods provided herein is a phosphoinositide 3-kinase (PI3K) inhibitor. In some embodiments, the PI3K inhibitor is idelalisib or a stereoisomer, mixture of stereoisomers, tautomer, isotopologue, or pharma- ceutically acceptable salt thereof. In some embodiments, the PI3K inhibitor is idelalisib.

[0226] In some embodiments, the second active agent used in the methods provided herein is a protein kinase C beta (PKCβ or PKC-β) inhibitor. In some embodiments, the PKCβ inhibitor is enzastaurin or a stereoisomer, mixture of stereoisomers, tautomer, isotopologue, or pharma- ceutically acceptable salt thereof. In some embodiments, the PKCβ inhibitor is enzastaurin. In some embodiments, the PKCβ inhibitor is a pharma- ceutically acceptable salt of enzastaurin. In some embodiments, the PKCβ inhibitor is the hydrochloride salt of enzastaurin. In some embodiments, the PKCβ inhibitor is the bis-hydrochloride salt of enzastaurin.

[0227] In some embodiments, the second active agent used in the methods provided herein is a spleen tyrosine kinase (SYK) inhibitor. In some embodiments, the SYK inhibitor is fostamatinib, or a tautomer, isotopologue, or pharmaceutically acceptable salt thereof. In some embodiments, the SYK inhibitor is fostamatinib. In some embodiments, the SYK inhibitor is a pharmaceutically acceptable salt of fostamatinib. In some embodiments, the SYK inhibitor is fostamatinib disodium hexahydrate.

[0228] In some embodiments, the second active agent used in the methods provided herein is a Janus kinase 2 (JAK2) inhibitor. In some embodiments, the JAK2 inhibitor is fedratinib, pacritinib, ruxolitinib, baricitinib, gandotinib, lestaurtinib, or momelotinib, or a stereoisomer, mixture of tautomers, tautomers, isotopologues, or pharmaceutically acceptable salt thereof.

[0229] In some embodiments, the JAK2 inhibitor is fedratinib, or a tautomer, isotope, or a pharma- ceutically acceptable salt thereof. In some embodiments, the JAK2 inhibitor is fedratinib.

[0230] In some embodiments, the JAK2 inhibitor is pacritinib, or a tautomer, isotopologue, or pharma- ceutically acceptable salt thereof. In some embodiments, the JAK2 inhibitor is pacritinib.

[0231] In some embodiments, the JAK2 inhibitor is ruxolitinib or a stereoisomer, mixture of stereoisomers, tautomer, isotopologue, or pharmaceutically acceptable salt thereof. In some embodiments, the JAK2 inhibitor is ruxolitinib. In some embodiments, the JAK2 inhibitor is a pharmaceutically acceptable salt of ruxolitinib. In some embodiments, the JAK2 inhibitor is ruxolitinib phosphate.

[0232] In some embodiments, the second active agent used in the methods provided herein is an Aurora A kinase inhibitor. In some embodiments, the Aurora A kinase inhibitor is alisertib, or a tautomer, isotopologue, or pharma- ceutically acceptable salt thereof. In some embodiments, the Aurora A kinase inhibitor is alisertib.

[0233] In some embodiments, the second active agent used in the methods provided herein is an enhancer of zeste homolog 2 (EZH2) inhibitor. In some embodiments, the EZH2 inhibitor is tazemetostat, GSK126, CPI-1205, 3-deazaneplanocin A (DZNep), EPZ005687, EI1, UNC1999, or sinefungin, or a stereoisomer, mixture of stereoisomers, tautomer, isotopologue, or pharma- ceutically acceptable salt thereof.

[0234] In some embodiments, the EZH2 inhibitor is tazemetostat, or a tautomer, isotope, or a pharma- ceutically acceptable salt thereof. In some embodiments, the EZH2 inhibitor is tazemetostat.

[0235] In some embodiments, the EZH2 inhibitor is GSK126 or a stereoisomer, mixture of stereoisomers, tautomer, isotopologue, or pharma- ceutically acceptable salt thereof. In some embodiments, the EZH2 inhibitor is GSK126 (also known as GSK-2816126).

[0236] In some embodiments, the EZH2 inhibitor is CPI-1205 or a stereoisomer, mixture of stereoisomers, tautomer, isotopologue, or pharma- ceutically acceptable salt thereof. In some embodiments, the EZH2 inhibitor is CPI-1205.

[0237] In some embodiments, the EZH2 inhibitor is 3-deazaneplanocin A. In some embodiments, the EZH2 inhibitor is EPZ005687. In some embodiments, the EZH2 inhibitor is EI1. In some embodiments, the EZH2 inhibitor is UNC1999. In some embodiments, the EZH2 inhibitor is sinefungin.

[0238] In some embodiments, the second active agent used in the methods provided herein is a hypomethylating agent, hi some embodiments, the hypomethylating agent is 5-azacytidine or decitabine, or a stereoisomer, mixture of stereoisomers, tautomer, isotopologue, or pharma- ceutically acceptable salt thereof.

[0239] In some embodiments, the hypomethylating agent is 5-azacytidine or a stereoisomer, mixture of stereoisomers, tautomer, isotopologue, or pharma- ceutically acceptable salt thereof. In some embodiments, the hypomethylating agent is 5-azacytidine.

[0240] In some embodiments, the hypomethylating agent is decitabine or a stereoisomer, mixture of stereoisomers, tautomer, isotopologue, or pharma- ceutically acceptable salt thereof. In some embodiments, the hypomethylating agent is decitabine.

[0241] In some embodiments, the second active agent used in the methods provided herein is chemotherapy.In some embodiments, chemotherapy is bendamustine, doxorubicin, etoposide, methotrexate, cytarabine, vincristine, ifosfamide, or melphalan, or their stereoisomers, mixtures of stereoisomers, tautomers, isotopologues, prodrugs, or pharmaceutically acceptable salts.

[0242] In certain embodiments, the second therapeutic agent is administered before, after, or simultaneously with the cancer treatment described herein. The cancer treatment described herein and the second therapeutic agent can be administered to the patient simultaneously or sequentially by the same or different routes of administration. The suitability of a particular route of administration used for a particular second drug or agent will depend on the second therapeutic agent itself (e.g., whether it can be administered orally or topically without being degraded before entering the bloodstream) and the subject being treated. Particular routes of administration of second drugs or agents or components are known to those of skill in the art. See, for example, The Merck Manual, 448 (17 th ed., 1999).

[0243] Any combination of the above therapeutic agents may be administered as appropriate for treating the disease or its symptoms. Such therapeutic agents may be administered in any combination simultaneously or as separate courses of treatment.

[0244] As used herein, the term "in combination" does not restrict the order in which therapies (e.g., prophylactic and / or therapeutic agents) are administered to a patient with a disease or disorder. The administration of a second active agent provided herein to a patient can occur simultaneously or sequentially by the same or different routes of administration. The suitability of a particular route of administration used for a particular active agent will depend on the active agent itself (e.g., whether it can be administered orally without being degraded prior to entering the bloodstream).

[0245] 5.6 Pharmaceutical Compositions In some embodiments, the cancer therapy provided herein and / or the additional active agents provided herein are formulated into a pharmaceutical composition, and the methods provided herein comprise administering the pharmaceutical composition comprising the cancer therapy to a lymphoma (e.g., DLBCL) patient.

[0246] In some embodiments, the pharmaceutical compositions provided herein comprise a therapeutically effective amount of one or more cancer treatments provided herein and a pharma- ceutically acceptable carrier, diluent, or excipient. In some embodiments, the compound is formulated as the sole pharma- ceutically active ingredient in the composition or is combined with other active ingredients.

[0247] In some embodiments, the cancer treatments provided herein can be formulated into pharmaceutical compositions suitable for different routes of administration, such as injection, sublingual and buccal, rectal, vaginal, ocular, otic, nasal, inhalation, aerosol, dermal, or transdermal. Typically, the compounds described above are formulated into pharmaceutical compositions using techniques and procedures well known in the art (see, for example, Ansel, Introduction to Pharmaceutical Dosage Forms, (7th ed. 1999)).

[0248] In some embodiments, the compositions comprise an effective concentration of one or more compounds or a pharma- ceutically acceptable salt mixed with a suitable pharmaceutical carrier or vehicle. In some embodiments, the concentration of the compound in the composition is effective to deliver an amount that, upon administration, treats, prevents, or ameliorates one or more symptoms and / or progression of lymphoma (e.g., DLBCL).

[0249] In some embodiments, the active compound is in an amount sufficient to exert a therapeutically useful effect in the absence of undesirable side effects on the treated patient. The therapeutically effective concentration is empirically determined by testing the compound in the in vitro and in vivo systems described herein, and then extrapolated from that concentration to the dosage administered to humans. The concentration of the active compound in the pharmaceutical composition depends on the absorption, tissue distribution, inactivation, and excretion rates of the active compound, the physicochemical properties of the compound, the administration schedule, and the dosage, as well as other factors known to those skilled in the art.

[0250] In some embodiments, the pharma- ceutically therapeutically active compounds and their salts are formulated and administered in unit dosage forms or multiple dosage forms. Unit dosage forms, as used herein, refer to physically discrete units suitable for human and animal subjects and individually packaged as known in the art. Each unit dosage contains a predetermined amount of therapeutically active compound sufficient to produce a desired therapeutic effect, together with the necessary pharmaceutical carrier, vehicle, or diluent. Examples of unit dosage forms include ampoules and syringes, as well as individually packaged tablets or capsules. A unit dosage form is administered in fractions or multiples thereof. A multiple dosage form is a plurality of identical unit dosage forms packaged in a single container to be administered in separate unit dosage forms. Examples of multiple dosage forms include vials, bottles of tablets or capsules, or bottles of pints or gallons. Thus, a multiple dosage form is a plurality of unit dosages that are not separated by packaging.

[0251] In some embodiments, the exact dosage and duration of treatment are a function of the disease being treated and are empirically determined using known testing protocols or by extrapolation from in vivo or in vitro test data.It should be noted that concentration and dosage values ​​may also vary depending on the severity of the condition being alleviated.Furthermore, it should be understood that for any particular subject, specific regimens will be adjusted over time according to individual needs and the professional judgment of the person administering or supervising the administration of the composition, and that the concentration ranges described herein are merely exemplary and are not intended to limit the scope or practice of the claimed compositions.

[0252] In some embodiments, solutions or suspensions used for parenteral, intradermal, subcutaneous, or topical application may contain any of the following components: a sterile diluent (such as water, saline, fixed oils, polyethylene glycols, glycerin, propylene glycol, dimethylacetamide, or other synthetic solvents), antibacterial agents (such as benzyl alcohol and methylparabens), antioxidants (such as ascorbic acid and sodium bisulfate), chelating agents (such as ethylenediaminetetraacetic acid (EDTA)), buffers (such as acetates, citrates, phosphates, and the like), and isotonicity adjusting agents (such as sodium chloride and dextrose). Parenteral preparations may be enclosed in ampoules, pens, disposable syringes, or single- or multi-dose vials made of glass, plastic, or other suitable material.

[0253] In some embodiments, sustained release preparations can also be prepared. Suitable examples of sustained release preparations include semipermeable matrices of solid hydrophobic polymers containing the compounds provided herein, which matrices are in the form of shaped articles, e.g., films or microcapsules. Examples of sustained release matrices include iontophoretic patches, polyesters, hydrogels (e.g., poly(2-hydroxyethyl-methacrylate) or poly(vinyl alcohol)), polylactides, copolymers of L-glutamic acid and ethyl-L-glutamic acid, non-degradable ethylene-vinyl acetate, degradable lactic acid-glycolic acid copolymers such as LUPRON DEPOT™ (injectable microspheres composed of lactic acid-glycolic acid copolymer and leuprolide acetate), and poly D-(-)-3-hydroxybutyric acid. Polymers such as ethylene-vinyl acetate and lactic acid-glycolic acid can release molecules for over 100 days, while certain hydrogels release proteins for shorter periods of time. If encapsulated compounds remain in the body for a long time, they may denature or aggregate as a result of exposure to moisture at 37°C, resulting in loss of biological activity and changes in their structure. Rational strategies for stabilization can be devised depending on the mechanism of action involved. For example, if the aggregation mechanism is found to be the formation of intermolecular disulfate bonds by thiodisulfide exchange, stabilization can be achieved by modification of sulfhydryl residues, lyophilization from acidic solutions, control of water content, use of appropriate additives, and development of specific polymer matrix compositions.

[0254] In some embodiments, anhydrous pharmaceutical compositions and dosage forms comprising the compounds provided herein are further included. The anhydrous pharmaceutical compositions and dosage forms provided herein can be prepared using anhydrous or low moisture ingredients and low moisture or low humidity conditions known to those skilled in the art. The anhydrous pharmaceutical composition can be prepared and stored such that its anhydrous nature is maintained. Thus, the anhydrous composition is packaged using materials known to prevent exposure to water so that it can be included in a suitable formulary kit. Examples of suitable packaging include, but are not limited to, hermetically sealed foils, plastics, unit dose containers (e.g., vials), blister packs, and strip packs.

[0255] In some embodiments, dosage forms or compositions can be prepared that contain active ingredient in the range of 0.001% to 100%, with the remainder consisting of non-toxic carriers. In some embodiments, the compositions contain about 0.005% to about 95% active ingredient. In some embodiments, the compositions contain about 0.01% to about 90% active ingredient. In some embodiments, the compositions contain about 0.1% to about 85% active ingredient. In some embodiments, the compositions contain about 0.1% to about 95% active ingredient.

[0256] In some embodiments, pharma- ceutically acceptable carriers used in parenteral preparations include aqueous vehicles, non-aqueous vehicles, antibacterial agents, isotonic agents, buffers, antioxidants, local anesthetics, suspending and dispersing agents, emulsifying agents, sequestering or chelating agents, and other pharma- ceutically acceptable substances.

[0257] In some embodiments, examples of aqueous vehicles include sodium chloride injection, Ringer's injection, isotonic dextrose injection, sterile water for injection, dextrose injection, and lactated Ringer's injection. Non-aqueous parenteral vehicles include fixed oils of vegetable origin, such as cottonseed oil, corn oil, sesame oil, and peanut oil. Antimicrobial agents in bacteriostatic or fungistatic concentrations should be added to parenteral formulations packaged in multi-dose containers, which include phenol or cresol, mercury, benzyl alcohol, chlorobutanol, methyl and propyl P-hydroxybenzoic acid esters, thimerosal, benzalkonium chloride, and benzethonium chloride. Isotonic agents include sodium chloride and dextrose. Buffers include phosphates and citrates. Antioxidants include sodium bisulfate. Local anesthetics include procaine hydrochloride. Suspending and dispersing agents include sodium carboxymethylcellulose, hydroxypropyl methylcellulose, and polyvinylpyrrolidone. Emulsifying agents include polysorbate 80 (TWEEN® 80). Sequestering or chelating agents of metal ions include EDTA. Pharmaceutical carriers also include ethyl alcohol, polyethylene glycol, propylene glycol for water miscible vehicles, and sodium hydroxide, hydrochloric acid, citric acid, or lactic acid for pH adjustment.

[0258] In some embodiments, the injectables are designed for local and systemic administration. Typically, a therapeutically effective dose is formulated to contain a concentration of at least about 0.1% w / w to about 90% w / w or more of the active compound relative to the tissue being treated, e.g., greater than 1% w / w. The active ingredient may be administered at once or divided into several smaller doses administered at intervals of time. It is understood that the exact dose and duration of treatment is a function of the tissue being treated and can be determined empirically using known testing protocols or by extrapolation from in vivo or in vitro test data. It is noted that concentration and dosage values ​​may also vary with the age of the individual being treated. Furthermore, it is understood that for any particular subject, the specific regimen should be adjusted over time according to the individual needs and the professional judgment of the person administering or supervising the administration of the formulation, and that the concentration ranges set forth herein are merely exemplary and are not intended to limit the scope or practice of the claimed formulations.

[0259] Also of interest herein are lyophilized powders, which can be reconstituted for administration as solutions, emulsions, and other mixtures, and can also be reconstituted and formulated as solids or gels.

[0260] In some embodiments, the sterile lyophilized powder is prepared by dissolving a compound provided herein or a pharma- ceutically acceptable salt thereof in a suitable solvent. In some embodiments, the solvent contains an excipient that improves the stability or other pharmacological components of the powder or reconstituted solution prepared from the powder. Excipients that can be used include, but are not limited to, dextrose, sorbital, fructose, corn syrup, xylitol, glycerin, glucose, sucrose, or other suitable agents. In some embodiments, the solvent contains a buffer, such as citrate, phosphate, or other buffer known to those skilled in the art. Subsequent sterile filtration of the solution followed by lyophilization under standard conditions known to those skilled in the art provides the desired formulation. Generally, the resulting solution is dispensed into vials for lyophilization. Each vial contains a single dose or multiple doses of the compound. The lyophilized powder can be stored under appropriate conditions from about 4°C to room temperature.

[0261] In some embodiments, the lyophilized formulation is suitable for reconstitution with a suitable diluent to a suitable concentration prior to administration. In some embodiments, the lyophilized formulation is stable at room temperature. In some embodiments, the lyophilized formulation is stable at room temperature for up to about 24 months. In some embodiments, the lyophilized formulation is stable at room temperature for up to about 24 months, up to about 18 months, up to about 12 months, up to about 6 months, up to about 3 months, or up to about 1 month. In some embodiments, the lyophilized formulation is stable when stored under accelerated conditions of 40° C. / 75% RH for up to about 12 months, up to about 6 months, or up to about 3 months.

[0262] In some embodiments, the lyophilized formulation is suitable for reconstitution with an aqueous solution for intravenous administration. In certain embodiments, the lyophilized formulations provided herein are suitable for reconstitution with water. In some embodiments, the reconstituted aqueous solution is stable at room temperature for up to about 24 hours after reconstitution. In some embodiments, the reconstituted aqueous solution is stable at room temperature for about 1-24 hours, 2-20 hours, 2-15 hours, 2-10 hours after reconstitution. In some embodiments, the reconstituted aqueous solution is stable at room temperature for up to about 20 hours, 15 hours, 12 hours, 10 hours, 8 hours, 6 hours, 4 hours, or 2 hours after reconstitution. In certain embodiments, the lyophilized formulation upon reconstitution has a pH of about 4-5.

[0263] The active ingredients provided herein can be administered by controlled release means or by delivery devices that are well known to those of ordinary skill in the art. Such examples include, but are not limited to, U.S. Pat. Nos. 3,845,770, 3,916,899, 3,536,809, 3,598,123, 4,008,719, 5,674,533, 5,059,595, 5,591,767, 5,120,548, 5,073,543, 5,639,476, 5,354,556, 5,639,480, 5,733,566, 5,739,108, 5,891,47 No. 4, U.S. Pat. No. 5,922,356, U.S. Pat. No. 5,972,891, U.S. Pat. No. 5,980,945, U.S. Pat. No. 5,993,855, U.S. Pat. No. 6,045,830, U.S. Pat. No. 6,087,324, U.S. Pat. No. 6,113,943, U.S. Pat. No. 6,197,350, U.S. Pat. No. 6,248,363, U.S. Pat. No. 6,264,970, U.S. Pat. No. 6,267,981, U.S. Pat. No. 6,376,461, U.S. Pat. No. 6,419,961, U.S. Pat. No. 6,589,548, U.S. Pat. No. 6,613,358, U.S. Pat. No. 6,699,500, and U.S. Pat. No. 6,740,634. Such dosage forms can be used to provide sustained or controlled release of one or more active ingredients, for example, by using hydropropylmethylcellulose, other polymer matrices, gels, permeable membranes, osmotic systems, multi-layer coatings, microparticles, liposomes, microspheres, or combinations thereof, providing desired release profiles at various rates.Suitable controlled release formulations known to those skilled in the art, including those described herein, can be easily selected for use with the active ingredients provided herein.

[0264] 5.7 Biological Samples In certain embodiments, various methods provided herein use a sample (e.g., a biological sample) from a patient with lymphoma (e.g., DLBCL). The patient may be male or female, and may be an adult, child, or infant. The sample may be analyzed during an active phase of lymphoma (such as DLBCL) or when the lymphoma (such as DLBCL) is inactive. In some embodiments, the sample is obtained from the patient prior to, concurrently with, and / or after administration of a treatment described herein. In some embodiments, the sample is obtained from the patient prior to administration of a treatment described herein. In certain embodiments, more than one sample may be obtained from the patient.

[0265] In certain embodiments, the sample comprises a subject's bodily fluid. Non-limiting examples of bodily fluids include blood (e.g., peripheral whole blood, peripheral blood), plasma, amniotic fluid, aqueous humor, bile, earwax, Cowper's fluid, pre-ejaculate, chyle, chyme, female ejaculate, interstitial fluid, lymph, menstruation, breast milk, mucus, pleural fluid, pus, saliva, sebum, semen, serum, sweat, tears, urine, vaginal fluid, vomit, water, feces, internal fluids including cerebrospinal fluid around the brain and spinal cord, synovial fluid around bone joints, intracellular fluid, which is the fluid inside cells, and vitreous humor, which is the fluid inside the eye. In some embodiments, the sample is a blood sample. Blood samples can be obtained using conventional techniques, for example as described in Innis et al, editors, PCR Protocols (Academic Press, 1990). White blood cells can be separated from blood samples using conventional techniques or commercially available kits, such as the RosetteSep kit (Stein Cell Technologies, Vancouver, Canada). Subpopulations of white blood cells, such as mononuclear cells, B cells, T cells, monocytes, granulocytes, or lymphocytes, can be further isolated using conventional techniques, such as magnetic activated cell sorting (MACS) (Miltenyi Biotec, Auburn, California) or fluorescence activated cell sorting (FACS) (Becton Dickinson, San Jose, California).

[0266] In some embodiments, the blood sample is about 0.1 mL to about 10.0 mL, about 0.2 mL to about 7 mL, about 0.3 mL to about 5 mL, about 0.4 mL to about 3.5 mL, or about 0.5 mL to about 3 mL. In some embodiments, the blood sample is about 0.3 mL, 0.4 mL, 0.5 mL, 0.6 mL, 0.7 mL, 0.8 mL, 0.9 mL, 1.0 mL, 1.5 mL, 2.0 mL, 2.5 mL, 3.0 mL, 3.5 mL, 4.0 mL, 4.5 mL, 5.0 mL, 6.0 mL, 7.0 mL, 8.0 mL, 9.0 mL, or 10.0 mL.

[0267] In some embodiments, the sample used in the methods comprises a biopsy (e.g., a tumor biopsy). A biopsy can be performed on any organ or tissue, such as skin, liver, lung, heart, colon, kidney, bone marrow, teeth, lymph nodes, hair, spleen, brain, breast, or other organ. In some embodiments, the sample used in the methods described herein comprises a tumor biopsy. The sample can be isolated from the subject using any biopsy technique known to those of skill in the art, such as open biopsy, close biopsy, core biopsy, incisional biopsy, excisional biopsy, or fine needle aspiration biopsy.

[0268] In some embodiments, the sample used in the methods provided herein is obtained from a subject before the patient is treated for lymphoma (e.g., DLBCL). In some embodiments, the sample is obtained from a patient when the subject is being treated for lymphoma (e.g., DLBCL). In some embodiments, the sample is obtained from a patient after the patient is treated for lymphoma (e.g., DLBCL). In various embodiments, the treatment comprises administering to the subject a compound described herein.

[0269] In certain embodiments, the sample comprises a plurality of cells. Such cells may include any type of cell, for example, stem cells, blood cells (e.g., peripheral blood mononuclear cells), lymphocytes, B cells, T cells, monocytes, granulocytes, immune cells, or tumor or cancer cells. In some embodiments, the tumor or cancer cells or tumor tissue include tumor biopsies or tumor explants. In some embodiments, T cells (T lymphocytes) include, for example, helper T cells (effector T cells or Th cells), cytotoxic T cells (CTLs), memory T cells, and regulatory T cells. In some embodiments, the cells used in the methods provided herein are CD3+, CD4+, CD8 ... + The number of T cells used in this method ranges from a single cell to approximately 10 9 B cells (B lymphocytes) can be a range of cells. In some embodiments, B cells (B lymphocytes) include, for example, plasma B cells, dendritic cells, memory B cells, B1 cells, B2 cells, marginal zone B cells, and follicular B cells. B cells can express immunoglobulins (antibodies, B cell receptors).

[0270] In some embodiments, specific cell populations can be obtained using a combination of commercially available antibodies (eg, Quest Diagnostic (San Juan Capistrano, Calif.); Dako (Denmark)).

[0271] In certain embodiments, the samples used in the methods provided herein are from affected tissues of lymphoma (e.g., DLBCL) patients. In certain embodiments, the number of cells used in the methods provided herein ranges from a single cell to about 10 9 In some embodiments, the number of cells used in the methods provided herein can range from about 1×10 4 , 5×10 4 , 1×10 5 , 5×10 5 , 1×10 6 , 5×10 6 , 1×10 7 , 5×10 7 , 1×108 , or 5 × 10 8 It is.

[0272] In some embodiments, the number and type of cells harvested from a subject can be monitored, for example, by measuring changes in morphology and cell surface markers using standard cell detection techniques such as flow cytometry, cell sorting, immunocytochemistry (e.g., staining with tissue-specific or cell marker-specific antibodies), fluorescence-activated cell sorting (FACS), magnetic activated cell sorting (MACS), by examining the morphology of the cells using light or confocal microscopy, and / or by measuring changes in gene expression using techniques well known in the art such as PCR and gene expression profiling. These techniques can also be used to identify cells that are positive for one or more specific markers. Fluorescence-activated cell sorting (FACS) is a well-known method for separating particles, including cells, based on the fluorescent properties of the particles (Kamarch, Methods Enzymol., 1987, 151:150-165). Laser excitation of the fluorescent moieties of individual particles generates a small charge that allows electromagnetic separation of positive and negative particles from a mixture. In some embodiments, the cell surface marker-specific antibodies or ligands are labeled with different fluorescent labels. The cells are processed through a cell sorter, allowing cells to be separated based on their ability to bind to the antibody used. The FACS sorted particles can be placed directly into individual wells of a 96- or 384-well plate for easy separation and cloning.

[0273] In certain embodiments, a subset of cells is used in the methods provided herein. Methods for sorting and separating specific populations of cells are well known in the art and can be based on cell size, morphology, or intracellular or extracellular markers. Such methods include, but are not limited to, flow cytometry, flow sorting, FACS, bead-based separation, e.g., magnetic cell sorting, size-based separation (e.g., sieves, arrays of obstacles, or filters), sorting on microfluidic devices, antibody-based separation, precipitation, affinity adsorption, affinity extraction, density gradient centrifugation, laser capture microdissection, and the like.

[0274] 5.8 Methods for detecting expression levels In some embodiments, the methods provided herein include measuring the expression level of at least one gene listed in Table 1. The expression level of the at least one gene can be determined by any method known in the art.

[0275] In some embodiments, the expression level of at least one gene is determined by measuring the mRNA levels of these genes. Several methods of detecting or quantifying mRNA levels are known in the art. Exemplary methods include, but are not limited to, Northern blot, ribonuclease protection assay, and PCR-based methods. The mRNA sequence can be used to prepare at least partially complementary probes. The probe can then be used to detect the mRNA sequence in the sample using any suitable assay, such as PCR-based methods, digital PCR (dPCR), Northern blotting, and dipstick assays.

[0276] In some embodiments, a nucleic acid assay can be prepared for testing immunomodulatory activity in a biological sample. The assay includes a solid support and at least one nucleic acid contacted with the support, the nucleic acid corresponding to at least a portion of an mRNA. The assay can also include a means for detecting changes in expression of the mRNA in the sample.

[0277] In some embodiments, the assay method can be varied depending on the type of mRNA information desired. Exemplary methods include, but are not limited to, Northern blot and PCR-based methods (e.g., RT-qPCR). Methods such as RT-qPCR can also accurately quantify the amount of mRNA in a sample.

[0278] Any suitable assay platform can be used to determine the presence of mRNA in a sample. For example, the assay can be in the form of a dipstick, a membrane, a chip, a disk, a test strip, a filter, a microsphere, a slide, a multi-well plate, or an optical fiber. In some embodiments, the assay system can have a solid support to which a nucleic acid corresponding to the mRNA is attached. The solid support can include, for example, plastic, silicon, metal, resin, glass, membrane, particle, precipitate, gel, polymer, sheet, sphere, polysaccharide, capillary, film, plate, or slide. The assay components can be prepared and packaged together as a kit for detecting mRNA.

[0279] In some embodiments, nucleic acid can be labeled as desired to generate a population of labeled mRNA. Generally, sample can be labeled using methods well known in the art (e.g., using DNA ligase, terminal transferase, or by labeling RNA backbone; see, for example, Ausubel, et al., Short Protocols in Molecular Biology, 3rd ed., Wiley & Sons 1995 and Sambrook et al., Molecular Cloning: A Laboratory Manual, Third Edition, 2001 Cold Spring Harbor, NY). In some embodiments, sample is labeled with fluorescent label. Exemplary fluorescent dyes include, but are not limited to, xanthene dyes, fluorescein dyes, rhodamine dyes, fluorescein isothiocyanate (FITC), 6 carboxyfluorescein (FAM), 6 carboxy-2',4',7',4,7-hexachlorofluorescein (HEX), 6 carboxy-4',5'dichloro-2',7'dimethoxyfluorescein (JOE or J), N,N,N',N'tetramethyl 6 carboxyrhodamine (TAMRA or T), 6 carboxyXrhodamine (ROX or R), 5 carboxyrhodamine 6G (R6G5 or G5), 6 carboxyrhodamine 6G (R6G6 or G6), and rhodamine 6G (R6G7 or G7). amine 110; cyanine dyes such as Cy3, Cy5, and Cy7 dyes; Alexa dyes such as Alexa-Fluor-555; coumarin, diethylaminocoumarin, umbelliferone; benzimide dyes such as Hoechst 33258; phenanthridine dyes such as Texas Red; ethidium dyes; acridine dyes; carbazole dyes; phenoxazine dyes; porphyrin dyes; polymethine dyes, BODIPY dyes, quinoline dyes, pyrene, chlorotriazinyl fluorescein, R110, eosin, JOE, R6G, tetramethylrhodamine, Lissamine, ROX, and naphthofluorescein.

[0280] In some embodiments, an mRNA assay method may include the steps of (1) obtaining a surface-bound target probe; (2) hybridizing a population of mRNA to the surface-bound probe under conditions sufficient to provide specific binding; (3) washing after hybridization to remove nucleic acids not bound by hybridization; and (4) detecting the hybridized mRNA. The reagents used in each of these steps and their conditions of use may vary depending on the particular application.

[0281] In some embodiments, hybridization can be performed under suitable hybridization conditions, the stringency of which can be varied as desired.Typical conditions are sufficient to create a probe / target complex on the solid surface between the complementary binding sites, i.e., between the surface-bound target probe and the complementary mRNA in the sample.In certain embodiments, stringent hybridization conditions can be used.

[0282] In some embodiments, hybridization is typically performed under stringent hybridization conditions. Standard hybridization techniques (e.g., under conditions sufficient to provide specific binding of target mRNA in sample to probe) are described in Kallioniemi et al., Science, 258:818-821 (1992) and WO 93 / 18186. Several guides on general techniques are available, for example, Tijssen, Hybridization with Nucleic Acid Probes, Parts I and II (Elsevier, Amsterdam 1993). For a description of suitable techniques for in situ hybridization, see Gall et al. Meth. Enzymol., 1981, 21:470-480; and Angerer et al. in Genetic Engineering: Principles and Methods (Setlow and Hollaender, Eds.) Vol. 7, pgs 43-65 (Plenum Press, New York 1985). Selection of appropriate conditions, including temperature, salt concentration, polynucleotide concentration, hybridization time, and stringency of washing conditions, will depend on the experimental design, including the source of the sample, the identity of the capture agent, and the expected degree of complementation, and can be determined as a matter of routine experimentation by one of ordinary skill in the art.

[0283] One of skill in the art will readily recognize that alternative but equivalent hybridization and washing conditions can be utilized to provide conditions of similar stringency.

[0284] After the mRNA hybridization procedure, the surface-bound polynucleotide is typically washed to remove unbound nucleic acid.Washing can be carried out using any convenient washing protocol as described above, and the washing conditions are typically stringent.Then, the hybridization of target mRNA to the probe is detected using standard techniques.

[0285] In some embodiments, other methods, such as PCR-based methods, can also be used to track gene expression. Examples of PCR methods can be found in the literature. An example of a PCR assay can be found in U.S. Patent No. 6,927,024, which is incorporated herein by reference in its entirety. An example of a RT-PCR method can be found in U.S. Patent No. 7,122,799, which is incorporated herein by reference in its entirety. Fluorescent in situ PCR is described in U.S. Patent No. 7,186,507, which is incorporated herein by reference in its entirety.

[0286] In some embodiments, real-time reverse transcription PCR (RT-qPCR) can be used to both detect and quantify RNA targets (Bustin, et al., Clin. Sci., 2005, 109:365-379). Quantitative results obtained by RT-qPCR are generally more informative than qualitative data. Thus, in some embodiments, RT-qPCR-based assays can be useful for measuring mRNA levels in cell-based assays. RT-qPCR methods are also useful for monitoring patient treatment. Examples of RT-qPCR-based methods can be found in U.S. Pat. No. 7,101,663, which is incorporated herein by reference in its entirety.

[0287] In contrast to conventional reverse transcriptase PCR and agarose gel analysis, real-time PCR provides quantitative results. An additional advantage of real-time PCR is that it is relatively easy and convenient to use. Instruments for real-time PCR, such as the Applied Biosystems 7500, are commercially available, as are reagents such as TaqMan Sequence Detection chemistry. For example, TaqMan® gene expression assays can be used according to the manufacturer's instructions. These kits are pre-formulated gene expression assays for rapid and reliable detection and quantification of human, mouse, and rat mRNA transcripts. For example, an exemplary PCR program is 50°C for 2 minutes, 95°C for 10 minutes, 40 cycles of 95°C for 15 seconds, then 60°C for 1 minute.

[0288] Data can be analyzed using, for example, 7500 Real-Time PCR System Sequence Detection software v1.3 using the comparative CT relative quantification calculation method to determine the cycle number at which the fluorescent signal associated with a particular amplicon accumulation exceeds a threshold (called CT). Using this method, the output is expressed as a fold change in expression level. In some embodiments, the threshold level can be selected to be automatically determined by the software. In some embodiments, the threshold is set above the baseline but low enough to fall within the exponential growth region of the amplification curve.

[0289] In some embodiments, the amount of RNA transcripts can be measured using techniques known to those of skill in the art. In some embodiments, the amount of one, two, three, four, five, or more RNA transcripts is measured using deep sequencing, such as ILLUMINA® RNASeq, ILLUMINA® Next Generation Sequencing (NGS), ION TORRENT™ RNA Next Generation Sequencing, 454™ Pyrosequencing, or sequencing by Oligo Ligation Detection (SOLID™). In other embodiments, the amount of multiple RNA transcripts is measured using microarrays and / or gene chips. In certain embodiments, the amount of one, two, three, or more RNA transcripts is determined by RT-PCR. In other embodiments, the amount of one, two, three, or more RNA transcripts is measured by RT-qPCR. Techniques for performing these assays are known to those of skill in the art. In yet other embodiments, NanoString (eg, the nCounter® miRNA Expression Assay from NanoString® Technologies) is used to analyze RNA transcripts.

[0290] In some embodiments, protein detection and quantification methods can be used to measure protein levels. Any suitable protein quantification method can be used. In some embodiments, antibody-based methods are used. Exemplary methods that can be used include, but are not limited to, immunoblotting (Western blot), enzyme-linked immunosorbent assay (ELISA), immunohistochemistry, flow cytometry, cytometric bead array, and mass spectrometry. Several types of ELISA are commonly used, including direct ELISA, indirect ELISA, and sandwich ELISA.

[0291] In some embodiments, protein levels are determined by immunohistochemistry (IHC). IHC is a clinical test that uses antibodies to test for specific antigens (markers) in tissue samples, a process that utilizes the principle of antibodies specifically binding to antigens in living tissue to detect antigens (e.g., proteins) within cells in tissue sections. Antibodies are usually linked to enzymes or fluorescent dyes. Typically, when an antibody binds to an antigen in a tissue sample, the enzyme or dye is activated and the antigen can then be observed under a microscope. IHC can be used to aid in the diagnosis of diseases such as cancer. IHC can also be used to help distinguish between different types of cancer. IHC can be used to image individual components within tissues by using appropriately labeled antibodies that specifically bind to target antigens in situ. IHC allows for visualization and documentation of high-resolution distribution and localization of specific cellular components within cells and in their appropriate histological context. There are multiple approaches and variations of the IHC method, but all the steps involved can generally be divided into two groups: sample preparation and sample staining. In some embodiments, IHC is based on immunostaining of thin tissue sections attached to individual glass slides. Multiple small sections can be placed on a single slide for comparative analysis, a format called tissue microarray. In other embodiments, IHC is performed using high-throughput sample preparation and staining.

[0292] Samples can be viewed either by light or fluorescence microscopy. In some embodiments, antigen detection in tissues can be performed using antibodies conjugated to enzymes (horseradish peroxidase) and can utilize chromogenic substrates that can be detected by light microscopy.

[0293] In some embodiments, the sample (e.g., patient tissue) is flash frozen in liquid nitrogen, isopentane, or dry ice. In other embodiments, the sample (e.g., patient tissue) is fixed in formaldehyde and embedded in paraffin wax (FFPE). In both of the above methods, the tissue or a portion of the tissue can be mounted on a slide prior to staining. In yet other embodiments, the IHC free-floating technique can be used, in which the entire IHC procedure is performed in liquid to increase antibody binding and penetration, and slide mounting is performed only upon completion of the experiment. IHC free-floating appears to be the most popular in neuroscience research. If analysis of the tissue by electron microscopy is required, the tissue can be embedded in an acrylate resin, such as glycol methacrylate (GMA), a technique called IHC resin.

[0294] In some embodiments, IHC can be performed using the methods described in the Examples section below.

[0295] 5.9 Kits In some embodiments, provided herein is a kit for predicting the responsiveness of a lymphoma patient to a cancer treatment, the kit comprising an agent for measuring gene expression levels in a biological sample of a lymphoma patient. In some embodiments, the kit further comprises an agent (or tool) for obtaining a sample from a subject. In some embodiments, the kit further comprises instructions on how to interpret or use the determined expression levels to predict whether a patient has a particular subtype of lymphoma (e.g., DLBCL).

[0296] In certain embodiments, the kit contains, in one or more other containers, one or more reagents necessary to perform the assays described herein. In certain embodiments, the kit contains a solid support and a means for detecting RNA or protein expression of at least one biomarker in a biological sample. Such kits can utilize, for example, a dipstick, a membrane, a chip, a disk, a test strip, a filter, a microsphere, a slide, a multi-well plate, or an optical fiber. The solid support of the kit can be, for example, a plastic, a silicon, a metal, a resin, a glass, a membrane, a particle, a precipitate, a gel, a polymer, a sheet, a sphere, a polysaccharide, a capillary, a film, a plate, or a slide.

[0297] In some embodiments, the kit includes, in one or more containers, components for performing RT-PCR, RT-qPCR, deep sequencing, or microarrays such as a NanoString assay. In some embodiments, the kit includes a solid support, a nucleic acid contacted with the support, the nucleic acid being complementary to at least 10, 20, 50, 100, 200, 350, or more bases of the mRNA, and a means for detecting expression of the mRNA in a biological sample.

[0298] In some embodiments, the kit contains, in one or more containers, components for performing an assay capable of determining one or more protein levels, such as flow cytometry, ELISA, or HIC.

[0299] Such kits may include materials and reagents necessary for the measurement of RNA or protein. In some embodiments, such kits include a microarray, the microarray including oligonucleotides and / or DNA and / or RNA fragments that hybridize to one or more of the genes shown in Table 1. In some embodiments, such kits may include primers for PCR of either the RNA products or cDNA copies of the RNA products of the genes or a subset of the genes or both. In some embodiments, such kits may include primers for PCR and probes for quantitative PCR. In some embodiments, such kits may include multiple primers and multiple probes, some of which have different fluorescent dyes to allow multiplexing of multiple products of one gene product or multiple gene products. In some embodiments, such kits may further include materials and reagents for making cDNA from RNA. In some embodiments, such kits may include antibodies specific for one or more of the genes shown in Table 1. Such kits may additionally include materials and reagents for isolating RNA and / or protein from a biological sample. In some embodiments, such kits may include materials and reagents for synthesizing cDNA from RNA isolated from a biological sample. In some embodiments, such kits may include a computer program product embodied in a computer readable medium for predicting whether a patient will respond to a compound described herein. In some embodiments, the kits may include a computer program product embodied in a computer readable medium with instructions.

[0300] In some embodiments, the antibody-based kit may include, for example: (1) a first antibody (which may or may not be attached to a solid support) that binds to a peptide, polypeptide, or protein of interest; and, optionally, (2) a second, different antibody that binds to either the peptide, polypeptide, or protein or the first antibody and is conjugated to a detectable label (e.g., a fluorescent label, a radioisotope, or an enzyme). The antibody-based kit may also include beads for performing immunoprecipitation. Each component of the antibody-based kit is generally placed in its own appropriate container. Thus, these kits generally include different containers appropriate for each antibody. In some embodiments, the antibody-based kit may include instructions for performing the assay and methods for interpreting and analyzing data obtained from performing the assay. In some embodiments, the kit includes instructions for predicting whether a patient with lymphoma (e.g., DLBCL) belongs to a particular subgroup of DLBCL (e.g., a high-risk subgroup of DLBCL).

[0301] In certain embodiments of the methods and kits provided herein, the solid phase support is used to purify proteins, label samples, or perform solid phase assays. Examples of solid phases suitable for performing the methods disclosed herein include beads, particles, colloids, single surfaces, tubes, multi-well plates, microtiter plates, slides, membranes, gels, and electrodes. In some embodiments, when the solid phase is a particulate material (e.g., beads), the particulate material is dispersed within the wells of a multi-well plate to allow parallel processing of the solid phase support.

[0302] The practice of the embodiments provided herein will employ, unless otherwise indicated, conventional techniques of molecular biology, microbiology, and immunology, which are within the skill of those of ordinary skill in the art, and which are fully explained in the literature. Examples of texts that are particularly suitable for reference include: Sambrook et al., Molecular Cloning: A Laboratory Manual (2d ed. 1989); Glover, ed., DNA Cloning, Volumes I and II (1985); Gait, ed., Oligonucleotide Synthesis (1984); Hames & Higgins, eds., Nucleic Acid Hybridization (1984); Hames & Higgins, eds., Transcription and Translation (1984); Freshney, ed., Animal Cell Culture: Immobilized Cells and Enzymes (IRL Press, 1986); Immunochemical Methods in Cell and Molecular Biology (Academic Press, London); Scopes, Protein Purification: Principles and Practice (Springer Verlag, NY, 2d ed. 1987); and Weir & Blackwell, eds., Handbook of Experimental Immunology, Volumes I-IV (1986).

[0303] From the foregoing, it will be understood that, although specific embodiments have been described herein for purposes of illustration, various modifications may be made without departing from the spirit and scope of what is provided herein. All references mentioned above are incorporated herein by reference in their entirety.

[0304] Certain embodiments of the present invention are illustrated by the following non-limiting examples.

[0305] 6. Embodiment The present invention provides the following non-limiting embodiments: 1. A method for predicting a lymphoma patient's response to cancer treatment comprising: (a) clustering reference lymphoma patients of a reference patient group into subgroups using the expression level of at least one gene in a reference biological sample of the reference lymphoma patients; (b) determining a subgroup to which the lymphoma patient belongs based on the expression level of at least one gene in a biological sample of the lymphoma patient; and (c) predicting the responsiveness of the lymphoma patient to the first cancer treatment based on the subgroup of the lymphoma patient. 2. The method of embodiment 1, further comprising administering to the lymphoma patient a second cancer treatment. 3. The method according to embodiment 1 or 2, wherein step (a) comprises generating clustering information defining a relationship between the expression levels of at least one gene in the reference biological sample, and reorganizing the heatmap representation based on the clustering information. 4. The method of any one of embodiments 1 to 3, wherein step (a) uses a hierarchical or non-hierarchical method. 5. The method of any one of embodiments 1 to 3, wherein step (a) uses the iClusterPlus method. 6. The method of any one of embodiments 1-5, wherein the reference lymphoma patients are clustered into 2-12 subgroups. 7. The method of embodiment 6, wherein the reference lymphoma patients are clustered into seven subgroups. 8. The method of any one of embodiments 1 to 7, further comprising training a classifier model using the expression level of at least one gene in a reference biological sample. 9. The method of embodiment 8, wherein the at least one gene is selected from the genes in Table 1, and optionally the at least one gene comprises five or more genes in Table 1. 10. The method of embodiment 9, wherein the at least one gene comprises all of the genes in Table 1. 11. The method of any one of embodiments 8 to 10, wherein the classifier model is a grouped multinomial generalized linear model (GLM). 12. The method of any one of embodiments 8 to 11, wherein the classifier model is a binary model. 13. The method of any one of embodiments 8 to 12, further comprising setting a threshold confidence level for at least one of the subgroups of step (a) to exclude patients providing lower confidence clustering data from at least one subgroup. 14. The method of any one of embodiments 1-13, wherein the lymphoma is selected from the group consisting of diffuse large B-cell lymphoma (DLBCL), indolent B-cell lymphoma, follicular lymphoma, small lymphocytic lymphoma, nodal marginal zone B-cell lymphoma, lymphoplasmacytic lymphoma, anaplastic large cell lymphoma, primary cutaneous lymphoma, mycosis fungoides, chronic lymphocytic leukemia, and mantle cell lymphoma. 15. The method of embodiment 14, wherein the lymphoma is DLBCL. 16. The method of embodiment 14, wherein the lymphoma is indolent B-cell lymphoma, nodal marginal zone B-cell lymphoma, mantle cell lymphoma, or chronic lymphocytic leukemia. 17. Reference patients in the reference patient group are clustered into subgroups A1 to A7; (i) subgroup A1, comprising: about 50% to about 60% of patients with germinal center B cell-like (GCB) DLBCL, about 30% to about 40% of patients with activated B cell-like (ABC) DLBCL, about 10% to about 20% of patients who are TME+ DLBCL, and about 30% to about 40% of patients who are DHITsig+ DLBCL; (ii) subgroup A2 comprising about 80% to about 90% of patients with GCB DLBCL, about 0% to about 5% of patients with ABC DLBCL, about 15% to about 25% of patients with TME+ DLBCL, and about 25% to about 35% of patients with DHITsig+ DLBCL; (iii) subgroup A3, comprising: about 40% to about 55% of patients with GCB DLBCL, about 30% to about 45% of patients with ABC DLBCL, about 40% to about 50% of patients with TME+ DLBCL, and about 20% to about 30% of patients with DHITsig+ DLBCL; (iv) subgroup A4, comprising: about 25% to about 35% of patients with GCB DLBCL, about 40% to about 50% of patients with ABC DLBCL, about 30% to about 40% of patients with TME+ DLBCL, and about 10% to about 20% of patients with DHITsig+ DLBCL; (v) subgroup A5 comprising: about 20% to about 40% of patients with GCB DLBCL, about 45% to about 65% of patients with ABC DLBCL, about 30% to about 40% of patients with TME+ DLBCL, and about 0% to about 10% of patients with DHITsig+ DLBCL; (vi) subgroup A6 comprises about 30% to about 40% of patients with GCB DLBCL, about 40% to about 50% of patients with ABC DLBCL, about 75% to about 95% of patients who are TME+ DLBCL, and about 0% to about 10% of patients who are DHITsig+ DLBCL; and (vii) The method of any one of embodiments 1-16, wherein subgroup A7 comprises about 0% to about 10% of patients with GCB DLBCL, about 80% to about 90% of patients with ABC DLBCL, about 0% to about 10% of patients who are TME+ DLBCL patients, and about 0% to about 15% of patients who are DHITsig+ DLBCL patients. 18. The method of any one of embodiments 1-17, wherein the first cancer treatment is a combination treatment with rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP). 19. The method according to any one of embodiments 1 to 18, comprising predicting that a lymphoma patient is unlikely to respond to the first cancer treatment if the patient is determined to belong to subgroup A1. 20. The method according to any one of embodiments 1 to 18, comprising predicting that the patient is unlikely to respond to the first cancer treatment if the lymphoma patient is determined to belong to subgroup A2. 21. The method according to any one of embodiments 1 to 18, comprising predicting that the patient is unlikely to respond to the first cancer treatment if the lymphoma patient is determined to belong to subgroup A3. 22. The method according to any one of embodiments 1 to 18, comprising predicting that the patient is unlikely to respond to the first cancer treatment if the lymphoma patient is determined to belong to subgroup A4. 23. The method according to any one of embodiments 1 to 18, comprising predicting that the patient is unlikely to respond to the first cancer treatment if the lymphoma patient is determined to belong to subgroup A5. 24. The method according to any one of embodiments 1 to 18, comprising predicting that the patient is unlikely to respond to the first cancer treatment if the lymphoma patient is determined to belong to subgroup A6. 25. The method according to any one of embodiments 1 to 18, comprising predicting that the patient is unlikely to respond to the first cancer treatment if the lymphoma patient is determined to belong to subgroup A7. 26. The method of any one of embodiments 2-25, wherein the second cancer treatment is R-CHOP. 27. The method of any one of embodiments 2-25, wherein the second cancer treatment is not R-CHOP. 28. The method of embodiment 27, wherein the second cancer treatment is a bromodomain and extraterminal (BET) inhibitor, or a cyclin-dependent kinase (CDK) inhibitor. 29. A method for predicting the response of a lymphoma patient to a cancer treatment, comprising: (a) determining an expression level of at least one gene of Table 1 in a biological sample of a lymphoma patient, optionally the at least one gene includes five or more genes of Table 1; (b) comparing the expression level of the at least one gene of step (a) with the expression level of the at least one gene in a reference biological sample of a reference lymphoma patient, wherein the reference lymphoma patient is responsive to the cancer treatment; and The method, wherein if the expression level of at least one gene in the biological sample is similar to the expression level of at least one gene in a reference biological sample, it indicates that the lymphoma patient is unlikely to respond to the cancer treatment. 30. A method for predicting the response of a lymphoma patient to a cancer treatment, comprising: (a) determining the expression level of at least one gene of Table 1 in a biological sample from a lymphoma patient; and (b) comparing the expression level of at least one gene in the biological sample with: (i) the expression level of at least one gene in biological samples from lymphoma patients who are responsive to the cancer treatment, and (ii) the expression level of at least one gene in biological samples from lymphoma patients who are not responsive to the cancer treatment; When the expression level of (a) is similar to the expression level of (i), it indicates that the first lymphoma patient is more likely to respond to the cancer treatment; and when the expression level of (a) is similar to the expression level of (ii), it indicates that the first lymphoma patient is less likely to respond to the cancer treatment. 31. A method of treating a patient with lymphoma comprising: (i) identifying a lymphoma patient likely to respond to cancer treatment according to the method of embodiment 30; and (ii) administering a cancer treatment to a lymphoma patient. 32. A method of treating a patient with lymphoma comprising: (i) identifying a lymphoma patient who is unlikely to respond to cancer treatment according to the method of embodiment 29 or 30; and (ii) administering to the lymphoma patient an alternative cancer treatment. 33. The method of any one of embodiments 30-32, wherein the cancer treatment is R-CHOP. 34. The method of embodiment 32, wherein the alternative cancer treatment is a BET inhibitor or a CDK inhibitor. 35. The method of any one of embodiments 28-34, wherein the lymphoma is selected from the group consisting of DLBCL, indolent B-cell lymphoma, follicular lymphoma, small lymphocytic lymphoma, nodal marginal zone B-cell lymphoma, lymphoplasmacytic lymphoma, anaplastic large cell lymphoma, primary cutaneous lymphoma, mycosis fungoides, chronic lymphocytic leukemia, and mantle cell lymphoma. 36. The method of embodiment 35, wherein the lymphoma is DLBCL. 37. The method of embodiment 35, wherein the lymphoma is DLBCL, indolent B-cell lymphoma, follicular lymphoma, nodal marginal zone B-cell lymphoma, mantle cell lymphoma, or chronic lymphocytic leukemia. 38. The method according to any one of embodiments 29 to 37, wherein the expression levels of all genes in table 1 are determined in (a) and compared in (b). 39. The method of any one of embodiments 1 to 38, wherein the biological sample is a tumor biopsy sample. 40. The method of any one of embodiments 1 to 39, wherein the step of determining the expression level of at least one gene comprises detecting the presence or amount of at least one complex in a biological sample, the presence or amount of the at least one complex being indicative of the expression level of the at least one gene. 41. The method of embodiment 40, wherein at least one complex is a hybridization complex. 42. The method of embodiment 40, wherein at least one complex is detectably labeled. 43. The method of any one of embodiments 1 to 39, wherein the step of determining the expression level of the at least one gene comprises detecting the presence or amount of at least one reaction product in the biological sample, wherein the presence or amount of the at least one reaction product indicates the expression level of the at least one gene. 44. The method of embodiment 43, wherein at least one reaction product is detectably labeled. 45. The method of any one of embodiments 1 to 44, wherein the reference lymphoma patient is a refractory DLBCL patient, a relapsed DLBCL patient, or a newly diagnosed DLBCL patient. 46. ​​The method of any one of embodiments 1 to 45, wherein the lymphoma patient is a refractory DLBCL patient, a relapsed DLBCL patient, or a newly diagnosed DLBCL patient. 47. The method of any one of embodiments 1-46, wherein the lymphoma patient is a GCB DLBCL patient or an ABC DLBCL patient. 48. The method of any one of embodiments 1 to 47, wherein the lymphoma patient is a DHITsig+ DLBCL patient or a DHITsig- DLBCL patient. EXAMPLES

[0306] 7. The following examples are carried out using routine and standard techniques well known to those skilled in the art, unless specifically described, and are for illustrative purposes only.

[0307] 7.1 Example 1: Data Overview 7.1.1 Methodology The Discovery cohort included the ROBUST clinical trial screening population (NCT02285062, (Nowakowski, et al., (2021), ROBUST: A Phase III Study of Lenalidomide Plus R-CHOP Versus Placebo Plus R-CHOP in Previously Untreated Patients With ABC-Type Diffuse Large B-Cell Lymphoma. Journal of Clinical Oncology.)) and a commercially available set of newly diagnosed patient samples (n=1208). The validation dataset included the MER observational cohort (n=343) (Cerhan, et al., (2017), Cohort profile: the lymphoma specialized program of research excellence (SPORE) molecular epidemiology resource (MER) cohort study. International journal of epidemiology, 46(6), 1753-1754i) and the REMoDL-B clinical trial (n=928 (Davies, et al., 2019)). For the analysis of clinical outcomes, only patients treated with R-CHOP were considered unless otherwise specified (or, in the MER dataset, patients treated as R-CHOP, including MR-CHOP, R-EPOCH, ER-CHOP, RAD-RCHOP, and a small number of patients treated with RCHOP / Zevalin). For comparison with the LymphGen clusters, the NCI dataset was used (Schmitz et al., 2018, New England Journal of Medicine, 378(15), 1396-1407).

[0308] 7.1.2 Unsupervised Clustering Clustering input data consisted of normalized RNAseq gene expression features and feature scores derived from gene expression data. Expression features were restricted to the most variable and most expressed genes in TPM space. The resulting features consisted of GSVA signature scores including MSigDB Hallmark and C1 pathways (Haenzelmann, SC (2013). GSVA: gene set variation analysis for microarray and RNA-Seq data. BMC Bioinformatics.), as well as cell type signatures (Danziger, SA (2019). ADAPTS: Automated deconvolution augmentation of profiles for tissue specific cells. PLoS One, 14(11)). The iClusterPlus method (Mo Q, SR (2021). iClusterPlus: Integrative clustering of multi-type genomic data. R package version 1.30.0) was applied to multiple choice subset data with K ranging from 2 to 12. This procedure was repeated 200 times, recording the cluster assignment in each case. The 200 runs were then summarized using a sample-pairwise co-clustering frequency matrix, calculated as the number of times two samples were assigned to the same cluster divided by the number of times the two samples appeared in the same run. This sample-pairwise matrix was then clustered using hierarchical clustering with Ward's method and 1 minus the co-cluster frequency as the distance metric to obtain one final clustering for each choice of K.

[0309] 7.1.3 Linear Model Classification A generalized linear model (GLM) classifier was trained on the discovery data, using consensus cluster labeling (A8 samples removed) as the gold standard. Several choices of the elastic net mixing parameter alpha were tested, aiming to maximize predictive performance and minimize model complexity. The regularization parameter λ was optimized using cross-fold validation and selected as the smallest value for which the misclassification rate was within one standard error of the smallest value.

[0310] 7.1.4 Normalization of RNAseq data The MER and ROBUST datasets were reference-normalized to a subset of Discovery data, called commercial samples, which was fixed as the reference population. To do so, a sample-wise scaling was applied to the TPM RNAseq data using the average of five housekeeping genes (ISY1, R3HDM1, TRIM56, UBXN4, and WDR55). After sample-level scaling, each gene was normalized to the reference population by subtracting the reference mean and dividing by the reference standard deviation. Finally, the reference was fixed so that all genes have a mean of 0 and variance of 1, and all other datasets were transformed to be gene-wise Z-scored to the reference population. Because the REMoDL-B dataset was an Illumina BeadArray rather than an RNAseq data, a self-standardization approach was applied using the housekeeping scaling step described above, followed by gene-wise scaling that explicitly sets each gene to have a mean of 0 and variance of 1. This self-standardization approach is suitable for large and representative patient cohorts such as those applied here, but may yield unexpected results in smaller or non-representative cohorts. The same self-standardization approach was applied to the NCI data set for evaluation of clusters on LymphGen calls.

[0311] The reference normalization approach puts all data into a unified numerical space with comparable expression levels (Figure 14A and Figure 14B), allowing for a portable model that can be trained on any dataset and directly applied to any other cohort without the need for reparameterization. Also, normalization is possible even for a single sample, without the need for a representative batch, and furthermore, the normalized data is not affected by the introduction of new samples. Existing classifiers such as the Reddy COO classifier (Reddy A, 2017, Genetic and Functional Drivers of Diffuse Large B Cell Lymphoma. Cell. 2017 Oct 5; 171(2): 481-494) and the TME26 classifier were fitted to the normalized gene expression space by reweighting the discrimination thresholds.

[0312] In fact, no significant batch effect due to dataset was observed in the normalized combined cohort of all datasets. We also verified that the normalization approach leaves relevant biological signals intact by comparing gene expression classifiers / signatures applied to the normalized data with orthogonal non-RNAseq data. These included comparison of Reddy COO classification with Hans IHC-based method, comparison of double-hit signature classifier (Ennishi,et al.,2019,Double-hit gene expression signature defines a distinct subgroup of germinal center B-cell-like diffuse large B-cell lymphoma.Journal of Clinical Oncology,37(3),190) with FISH calls, and comparison of cell type abundance GSVA scores with cell type marker densities from IHC and MIBI. All features derived from normalized RNAseq data were highly concordant with their corresponding non-RNAseq features.

[0313] 7.1.5 Cell-Type Signatures Cell type-specific signatures were generated from the LM22 matrix, which describes 22 functionally defined leukocyte types (Chen B, 2018, Profiling Tumor Infiltrating Immune Cells with CIBERSORT. Methods in molecular biology (Clifton, NJ), 1711, 243-259). This signature matrix was extended and tuned for DLBCL by adding cell types representing malignant DLBCL B cells and trained on purified cell populations. Benchmarking results of deconvolution with the extended signature matrix showed a high correlation between DLBCL-specific cell type abundance and tumor purity, and also confirmed the abundance of CD20+ cells measured by IHC. The addition of DLBCL-specific cell types also significantly reduced the estimated abundance of the unclassifiable "other" cell type population, which previously accounted for up to 40% of the estimated abundance.

[0314] 7.1.6 Sequencing ROBUST, MER, and commercial samples were sequenced at Expression Analysis, Inc. (Durham, HC, USA) following standard protocols. Genomic DNA and total RNA were simultaneously purified from formalin-fixed paraffin-embedded (FFPE) tissue sections using the Allprep DNA / RNA FFPE kit. RNAseq libraries (75PE, 50M) were constructed using the Illumina TruSeq RNA Access method.

[0315] WES libraries (200x for tumors and 100x for germline controls) were constructed using the Agilent SureSelectXT method with on-bead modifications from Fisher et al., 2011. WGS libraries (60x for tumors and 30x for germline) were prepared using the Swift Accel-NGS 2S Plus DNA Library Kit (#21024 or 21096, Swift) with a modification of the Ampure Bead cleanup step of the procedure.

[0316] 7.1.7 Data Processing Sequencing data were processed by an in-house cloud-based platform using the Sentieon implementation of GATK best practice using BWA-mem for alignment, and the Sentieon implementation of Mutect2 (tnhaplotyper). Variants were annotated with SnpEff using the dbnsfp database. For WGS data copy number abnormalities were called using Battenberg and structural variants were called using Manta. For WES data copy number abnormalities were called using Sclust. Structural variants were found to be under-represented in the WES data. RNA-seq data were aligned with STAR aligner and quantified with salmon.

[0317] 7.1.8 shRNA knockdown Doxycycline (Dox)-inducible shRNA constructs were generated by Cellecta (Mountain View, CA, USA) using the pRSITEP-U6Tet-(sh)-EF1-TetRep-2A-Puro plasmid. Briefly, 293F T cells were co-transfected with lentiviral packaging plasmid mix (Cellecta, Cat#CPCP-K2A) and pRSITEP-shRNA constructs. Viral particles were collected 48 and 72 hours after transfection and then concentrated with a Lenti-X concentrator (Takara Bio USA). For infection, cells were incubated overnight with concentrated viral supernatant in the presence of 8 μg / ml polybrene. Cells were then washed to remove polybrene. 48 hours after infection, cells were selected with puromycin (2 μg / mL) up to one week before the experiment. For knockdown experiments, cells were cultured at 1 × 10 5Cells were seeded at 1000 cells / ml and induced with 20 ng / ml Dox or DMSO vehicle control. On day 3 of Dox induction, cells were counted and refreshed with Dox or DMSO. For proliferation assays, 15,000 cells were seeded in 96-well U-bottom plates and cell viability was subsequently measured by CellTiter-Glo (Promega) for 5 consecutive days. The remaining cells were aliquoted at 5 × 10 5 Cells / ml were seeded and incubated for another 2 days. Cells were then harvested for Western blot and apoptosis assay. shRNA target sequences were: shNT: CAACAAGATGAAGAGCACCAA (SEQ ID NO: 1); shTCF4-13: GAGACTGAACGGCAATCTTTC (SEQ ID NO: 2); shTCF4-14: CACGAAATCTTCGGAGGACAA (SEQ ID NO: 3).

[0318] 7.1.9 Western blotting Cells were lysed in cell lysis buffer (50 mM TrisHCl pH 7.4, 250 mM NaCl, 0.5% Triton X100, 10% glycerol) supplemented with Halt protease / phosphatase inhibitor (Thermo scientific, 78443). Cell lysates were sonicated to break down nuclei and reduce viscosity caused by released genomic DNA. Protein concentration was measured by Bradford Protein Assay (Bio-Rad). Samples were diluted to equal concentrations and then diluted with NuPAGE LDS sample buffer and 2-mercaptoethanol (final concentration 1.25%), followed by boiling at 95°C for 5 min. Total cell lysates were dissolved in NuPAG 4-12% Bis-Tris Midi Protein Gels (Invitrogen), transferred to nitrocellulose membranes, and then blocked with Intercept® (TBS) blocking buffer (LI-COR). Proteins of interest were detected by overnight incubation at 4°C with primary antibodies listed below. After one wash with TBST, membranes were incubated with either IRDye 800CW goat anti-rabbit IgG or IRDye 680LT goat anti-mouse IgG secondary antibodies (1:10,000) for 1 hour at room temperature. After one wash with TBST, bands were visualized with an Odyssey Imaging System (LI-COR). Antibody information: TCF4 (Proteintech, 22337-1-AP), MYC (abcam, ab32072), GAPDH (Cell Signaling Technology, 2118L).

[0319] 7.2 Example 2: Unsupervised clustering of DLBCL patients Diffuse large B-cell lymphoma (DLBCL) is a heterogeneous and aggressive group of germinal center B-cell neoplasms and is the most common form of non-Hodgkin lymphoma (NHL). The International Prognostic Index (IPI) for DLBCL predicts survival outcomes for newly diagnosed DLBCL patients based on clinical risk factors. Patients with an IPI of 3-5 are considered intermediate to high risk and are often used to select patients for clinical trials due to unfavorable outcomes with the standard of care immunochemotherapy R-CHOP. However, the IPI does not provide biological insight to elucidate treatment opportunities for high-risk patients.

[0320] Molecular classification using cell of origin (COO) well describes the activated B-cell (ABC) subtype as being at higher risk of relapse with R-CHOP and with shorter survival than the germinal center B-cell (GCB) subtype. In two phase 3 randomized controlled trials, the PHOENIX and ROBUST studies, the combination of ibrutinib (Younes, et al., 2019) or lenalidomide (Nowakowski, et al., 2021) with R-CHOP did not demonstrate significant activity compared with R-CHOP alone in the high-risk ABC population. A more detailed investigation of the ABC arm treated with R-CHOP showed variable clinical outcomes, indicating that the underlying disease heterogeneity of COO prevents changes in practice. Classification of chromosomal rearrangements involving MYC, BCL2, and / or BCL6, so-called double-hit and triple-hit patients, consistently identified a high-risk subset of GCB patients, but there are no therapies approved specifically for this population.

[0321] More recently, analysis of genetic features including mutations and copy number has identified new patient clusters that were then constructed.COO(Chapuy,et al.,2018,Molecular subtypes of diffuse large B cell lymphoma are associated with distinct pathogenic mechanisms and outcomes.Nature medicine,24(5),679-690)(Schmitz et al.,2018,New England Journal of Medicine,378(15),1396-1407)(Wright,et al.,2020,A probabilistic classification tool for genetic subtypes of diffuse large B cell lymphoma with therapeutic implications.Cancer Cell,37(4),551-568)(Lacy,et al.,2020,targeted sequencing in DLBCL,molecular subtypes,and outcomes:a Haematological Malignancy Research Network report.Blood,135(20),1759-1771).Classification also incorporates the tumor microenvironment (TME26) (Risueno,et al.,2020,Leveraging gene expression subgroups to classify DLBCL patients and select for clinical benefit from a novel agent.Blood,135(13),1008-1018), (Kotlov,et al.,2021,Clinical and biological subtypes of B-cell lymphoma revealed by microenvironmental signatures.Cancer discovery,11(6),1468-1489), (Steen,et al.,2021,The landscape of tumor cell states and ecosystems in diffuse large B cell lymphoma.Cancer Cell). Despite these advances, a path to prospectively identify newly diagnosed high-risk DLBCL patients in a clinically feasible and practical manner with drug approval remains unrealized.

[0322] The present disclosure finds that robust clustering allows for the identification of biologically driven DLBCL patient subgroups, which will predict patient outcomes and inform therapeutic approaches. By defining the DLBCL subtype landscape in a robust and meaningful way, future treatments tailored to those classified subtypes can be developed. Identifying patients belonging to specific subtypes may lead to the identification of high-risk patient groups for whom current therapies (such as R-CHOP) are inappropriate. Investigating the biological basis of these groups may also help elucidate the mechanisms underlying high-risk subtypes.

[0323] This disclosure identifies biologically homogenous high-risk DLBCL patients by unsupervised clustering on transcriptome features of both tumor and non-tumor cells. We identified several homogenous clusters, including one high-risk cluster described by an extreme ABC phenotype with predominantly promoted MYC pathway and low immune infiltration. We developed a gene expression classifier that allowed replication of clinical and biological characteristics in independent cohorts. Overall, multiple randomized studies demonstrated the poor prognostic nature of cluster A7 and retrospective treatment-specific response profiles, suggesting that high-risk A7 may be used in clinical trials.

[0324] 7.2.1 Discovery and validation of novel clusters and development of gene expression classifiers Unsupervised clustering was performed on the Discovery cohort of gene expression-derived data from newly diagnosed DLBCL patients (n=1208, Table 2), followed by supervised classifier training to identify the discovered clusters in an independent dataset (Figure 1A). Unsupervised clustering yielded eight clusters with distinct molecular patterns (Figure 1B). These clusters have different degrees of association with the COO and TME26 classes (Risueno, et al., 2020), but none can be uniquely determined by them. Cluster A8 turned out to be a technical artifact cluster with poor alignment metrics (Figure 6A-6D), and cluster A8 was excluded from classifier training and further analysis.

[0325] A multinomial classifier was trained on the discovery dataset to generate a model for identifying each cluster in an independent validation cohort. Cross-validation results showed good performance of the classifier training method, with an accuracy of 93% in the training cohort and a sensitivity / positive predictive value of 81-98% within each cluster individually (Figure 7). The training data was normalized to a reference population, so the classifier could be directly applied to other datasets normalized to this space without the need to retrain parameters or thresholds. This classifier could be applied to any FFPE RNAseq sample normalized in the same way and would generate a class label for each case (i.e., there are no unclassified cases).

[0326] Application of the classifier to the independent cohorts MER (validation cohort 1, n=343) (Cerhan, et al., 2017) and REMoDL-B (validation cohort 2, n=928) (Davies, et al., 2019, Gene-expression profiling of bortezomib added to standard chemoimmunotherapy for diffuse large B-cell lymphoma (REMoDL-B): an open-label, randomised, phase 3 trial. The Lancet Oncology, 20(5), 649-662) identified seven biologically reproducible clusters containing differentially expressed genes above and below the top 50 within each cluster, which show reproducible expression patterns across the clusters (Figure 1C).

[0327] 7.2.2 Clinical outcomes and characteristics of high-risk cluster A7 Cluster discovery was performed without clinical outcome data to determine whether any clusters were associated with poor prognosis. The association of clusters with survival outcomes with R-CHOP and with prognostic features is shown in Figure 2A-2I. Cluster A7 represents the group of patients with increased ABC who responded poorest to RCHOP among the seven clusters, with prevalences of 19% in ROBUST, 13% in MER, and 11% in REMoDL-B. A7 status was significantly prognostic, with hazard ratios (95% confidence intervals) of A7 vs. non-A7 being 1.65 (1.08-2.51) in ROBUST (ABC only), 1.87 (1.17-3.00) in MER, and 2.00 (1.23-3.20) in REMoDL-B.

[0328] Although the ABC COO subtype was associated with elevated risk, the high-risk nature of A7 was not simply due to increased ABC. Even within the ABC-only population, A7 patients were at higher risk than non-A7 patients (Figures 2D-F). Association tests of A7 with known clinical prognostic factors showed no strong influence by IPI or its components, indicating that A7 could not be defined using clinical features (Figures 2G-I). Cox proportional hazards models showed that A7 status was a significant prognostic factor in both univariate models (p=0.027) and multivariate models combined with IPI (p=0.047), and that the A7+IPI model had a slightly better prognosis than IPI alone (ANOVA p=0.06, Figures 8A-C). A7 was strongly associated with both COO and TME26 (both, p<2.2e-16), but neither feature alone nor in combination was sufficient to uniquely identify A7. Using COO or TME26 scores as univariate predictors of A7 membership, both ROBUST and MER yielded predictive AUCs ranging from 0.82 to 0.86, with the optimized classifier achieving approximately 80% sensitivity and 70% specificity in classifying A7.

[0329] 7.2.3 Biological interpretation of the new clusters Each cluster was examined for differential biology in terms of single gene expression, DLBCL-specific pathways (Wright, et al., 2020), copy number aberrations, single nucleotide variants, and tumor microenvironment. Distinctions between clusters were identifiable by the lens of COO and TME26 (Figure 3A), but significant heterogeneity remained across these dimensions. Three clusters were notably extreme in COO-TME26 space: low TME GCB-gain cases found in A2, low TME ABC-gain cases found in A7, and high TME unclassified-gain cases found in A6.

[0330] The various DLBCL-associated pathways utilized by Wright et al. allowed deeper insight into the pathways contributing to each cluster from a tumor microenvironment, COO, oncogenic pathway, and metabolomic perspective (Figure 3A-E). Among the clearest signals were upregulation of various immune-related, JAK, and NFKB signatures in A6, upregulation of a GCB-associated signature (IRF4Dn-1) in A2 with increased GCB, a relative balance of tumor microenvironment and malignant process signatures in A5, and downregulation of PI3K, malignant process, and metabolic signatures in A3. Clusters A1 and A4 both showed low expression of MYC and G2M checkpoint pathways, but had less clear gene expression signals. High-risk cluster A7 showed upregulation of ABC-associated signatures (IRF4Up-7) and low expression TME signatures. A7 was associated with a higher risk of ABC subtype (p<2.2×10 16 ), and had the most extreme COO scores, even among ABC patients, according to the Reddy et al. score (data not shown). A7 was also characterized by upregulation of signatures such as G2M checkpoint, oxidative phosphorylation, mitotic spindle, and DNA repair, and low expression of p53 and TME signatures (Figure 4A). Further elucidation of the pathways that define the cluster is shown in Figures 9A and 9B.

[0331] Genomic features increased in A7 reflected the ABC-increased nature of the cluster, with increased prevalence of mutations such as ETV6, PIM1, and OSBPL10 (Figure 3C). In general, however, SNVs were not strongly associated with our clusters, which was not surprising as the clusters were derived from transcriptional features that could be derived from sources other than SNVs, such as copy number and epigenetic changes. Significantly increased CNAs for each cluster are shown in Figure 3D, and features associated with A7 include treatment-level copy number increases in chromosomes 3 and 18. CNA features for each cluster are shown in Figures 10A-F.

[0332] Immunohistochemistry data validated gene expression-derived patterns of immune infiltration across clusters. Compared to non-A7, there was a consistent reduction in CD3, CD4, and CD8 T cells, with no strong trends in CD163 monocytes / macrophages, CD68 macrophages, and CD11 dendritic cells (Figures S11A-S11F). For example, representative MIBI images of high TME26 cluster A6 indeed showed abundant CD4 / CD8 T cells, whereas low TME26 cluster A7 conversely showed few T cells and abundant CD20 B cells (Figure 3E).

[0333] 7.2.4 MYC dysregulation is a key component of high-risk cluster A7 biology and could be targeted by TCF4 To further define the biology specific to cluster A7, GSEA analysis was performed to identify pathways differentially expressed in A7. This cluster showed upregulation of MYC targeting signature, E2F targeting signature, and metabolic pathways such as G2M checkpoint and oxidative phosphorylation, as well as downregulation of immune and inflammatory signatures including TNFα, IL2, IL6, IFN-α, and IFN-γ signaling pathways (Figure 4A).

[0334] MYC gene expression was upregulated in both cluster A7 and non-A7 (Figure 4B) and was not driven by high tumor cellularity (Figure S12E). Protein expression quantified by IHC also showed that Myc protein levels were higher in A7 than in non-A7 (Figure 4C). In B-cell lymphomas, MYC translocation and MYC amplification are well known to promote MYC signaling, but in A7, neither MYC translocation nor MYC amplification was increased (p=0.99), suggesting that upregulation of MYC activity is driven by other mechanisms.

[0335] In A7, a significant increase in several treatment-level amplifications was observed, with chromosome 18q and 3q amplifications showing the highest prevalence. Interestingly, 18q12.2 harbors the gene TCF4, which encodes a basic helix-loop-helix (bHLH) transcription factor and has been reported to drive the expression of the MYC gene by binding to its enhancer [PMID:31217338]. A strong correlation between TCF4 gene expression and TCF4 copy number was observed in the Discovery and MER cohorts (Figures 13A and 13B). We then sought to characterize TCF4 function using DLBCL cell line models. Knockdown of TCF4 dramatically reduced MYC protein expression in TCF4-amplified cell lines (RIVA and U2932) but not in cell lines lacking TCF4 amplification (SU-DHL-2 and TMD8) (Figure 4F), suggesting that TCF4 amplification contributed to the overexpression of MYC in ABC DLBCL. In line with these observations, knockdown of TCF4 strongly inhibited cell proliferation in TCF4-amplified cell lines (RIVA and U2932), whereas induction of the same shRNA only slightly inhibited proliferation in cell lines lacking TCF4 amplification (SU-DHL-2 and TMD8) (Figure 4G). Taken together, amplification-dependent overexpression of TCF4 stimulated MYC expression and rendered ABC DLBCL dependent on overexpressed TCF4. TCF4 may be a potential therapeutic target for the A7 population.

[0336] 7.2.5 Clinical utility of cluster A7 To evaluate the utility of A7 as a predictive patient population, ROBUST and REMoDL-B patients were retrospectively stratified by A7 status vs. non-A7 status in (Figure 5A and Figure 5B). The results showed differences between control and experimental treatment groups in the A7 population in both the ROBUST and REMoDL-B studies (p=0.0088 and 0.16, respectively), suggesting that cluster A7 served as a more homogenous and reliable high-risk patient population for drug development. With deeper molecular and biological insights underpinning the tumor and TME of DLBCL beyond COO, the field was ripe for high-risk patient selection, creative trial designs, and targeted therapies to shift clinical practice. Here, we identified a high-risk patient segment for newly diagnosed DLBCL by unsupervised clustering of transcriptome data. Under the clinical high-risk behavior of A7, three biological characteristics were known to result in poor outcome: extreme ABC subtype, low immune infiltration (particularly low CD4 and CD8 T cells), and elevated MYC pathway.

[0337] Comparison with recently published molecular classifications indicates that cluster A7 has some unique features, but is not mutually exclusive with other clusters. The hallmark feature of reduced immune infiltration of A7 is shared with the "Depleted" segment (DP) (Kotlov, et al., 2021) and lymphoma ecotype 1 (LE1) (Steen, et al., 2021), which showed similar unfavorable survival characteristics.

[0338] Cluster A7 also shared an increase in previously defined features with genetic subtypes C5 and MCD, including amplifications of chromosomes 3p, 3q, and 18q, and mutations in PIM1, ETV6, and OSBPL10 (MYD88 as described in Schmitz et al., 2018, New England Journal of Medicine, 378(15), 1396-1407). L265Pand CD79B mutations). The co-occurrence of the MCD and A7 clusters was examined in the NCI dataset for which LymphGen calls are publicly available (Schmitz et al., 2018, New England Journal of Medicine, 378(15), 1396-1407). Approximately one-third of patients identified as A7 or MCD were also identified as the other, a statistically significant overlap (p=0.038). Although both A7 and MCD identified patient groups of similar size and risk, the majority of cases in each cluster represented a unique subset of high-risk patients not identified by the other method (Figures 14A and 14B).

[0339] MCD subtype (MYD88 as described in Schmitz et al., 2018, New England Journal of Medicine, 378(15), 1396-1407) L265P The EZB subtype (based on the co-occurrence of CD79B and CD79B mutations) was increased in A7 patients, whereas the EZB subtype (based on EZH2 mutations and BCL2 translocations described in Schmitz et al., 2018, New England Journal of Medicine, 378(15), 1396-1407) was increased in GCB-like clusters A2 and A3 (Figures 15A-D). Although there was a statistically significant association between the classification methods (Fisher p=0.0005 for ROBUST and Fisher p=0.001 for MER), there was a lot of heterogeneity and no clear one-to-one mapping between the subtypes. In addition, PCA plots of the Discovery, MER, and REMoDL-B datasets after normalization showed no dataset-specific differences (Figure 16). We also measured the mutation landscape (Chapuy genes) sorted by mutation number (Figure 17A), significance (corrected for gene length) (Figure 17B), and Chapuy diagram (for reference) (Figure 17C). Expression of proteins encoded by genes on chromosome 18 was also made available (Figures 18A-D).

[0340] Although increased MYC pathway expression is associated with poor survival in DLBCL (Savage,et al.,2009,MYC gene rearrangements are associated with a poor prognosis in diffuse large B-cell lymphoma patients treated with R-CHOP chemotherapy.Blood,114(17),3533-3537) (Barrans,et al.,2010,Rearrangement of MYC is associated with poor prognosis in patients with diffuse large B-cell lymphoma treated in the era of rituximab.Journal of clinical oncology,28(20),3360-3365), the mechanisms are different in GCB and ABC subtypes. In GCB, chromosomal rearrangements of MYC and BCL2 to the IG locus are the main drivers of overexpression of MYC and BCL2. In ABC tumors, MYC translocations are relatively rare, and MYC overexpression is not associated with translocation events (Xu-Monette, et al., 2015, Clinical features, tumor biology, and prognosis associated with MYC rearrangement and Myc overexpression in diffuse large B-cell lymphoma patients treated with rituximab-CHOP. Modern Pathology, 28(12), 1555-1573). It has also been shown that MYC expression is not affected by its copy number increase (Collinge, et al., 2021, The impact of MYC and BCL2 structural variants in tumors of DLBCL morphology and mechanisms of false-negative MYC IHC. Blood, 137(16), 2196-2208).A similar pattern was seen in A7, where there was no difference in MYC translocations or copy number gain between A7 and non-A7 cases, but both gene and protein expression of MYC was elevated in A7. We validated the idea that specific MYC regulators, such as TCF4, are amplified as part of 18q gain and are responsible for this increase. Data in ABC cell lines substantiated this relationship and indicated that TCF4 could serve as a therapeutic target in A7 (Figure 4E-G). Other MYC regulators may share similar functional effects.

[0341] Additional pathway alterations unique to A7 included upregulation of the G2M checkpoint, mitotic spindle checkpoint, and DNA repair pathways (Figure 4A), indicating cell cycle deregulation and DNA replication stress. These alterations, along with downregulation of the p53 pathway, likely led to rapid proliferation and genomic instability supported by uncontrolled growth and numerous copy number alterations (Figure 3D). Another key feature of A7 was the upregulation of the oxidative phosphorylation pathway, indicating altered energy metabolism by the tumor through utilization of oxidized substrates such as fatty acids in the hypoxic microenvironment. Both phenomena were reported as molecular hallmarks of DLBCL subsets (Monti, et al., 2012, Molecular profiling of diffuse large B-cell lymphoma identified robust subtypes including one characterized by host inflammatory response. Blood, 105(5), 1851-1861), (Caro, et al., 2012, Metabolic signatures uncover distinct targets in molecular subsets of diffuse large B cell lymphoma. Cancer Cell, 22(4), 547-560). These observations have led to novel therapeutic strategies and targets for this high-risk segment.

[0342] Indeed, the poor prognosis of A7 necessitated a different approach than R-CHOP. Here, an example of an R2-CHOP regimen was shown that significantly improved the outcome of A7 patients, even though R2-CHOP was not specifically designed for this population (Figure 5A). Lenalidomide is a cereblon modulator with dual effects of autonomous antiproliferation and immune-mediated cytotoxicity against tumor B cells (Garciaz,et al.,2016,Lenalidomide for the treatment of B-cell lymphoma.Expert opinion on investigational drugs,25(9),1103-1116). The immune-modulatory activity was particularly beneficial for "cold tumors" such as those represented by A7. Ibrutinib also performed well in the A7 cluster, suggesting that the extreme ABC biology of A7 interacted with BCR-modulators. These examples demonstrated the utility of the A7 genetic classification tool for use in patient selection and its potential clinical value in this high-risk patient segment.

[0343] Finally, the biomarkers used to identify A7 patients each had several attributes that made them attractive for practical implementation. First, gene expression tests were easy to implement and demonstrated feasibility in a clinical trial setting, as seen with ROBUST, with a turnaround time of 2.4 days. Second, gene expression tests used diagnostic FFPE tissue and did not require additional biopsies. Third, gene expression tests classified all patients as either A7 or non-A7, eliminating ambiguity for unclassifiable populations. The future of DLBCL was targeted treatment of molecularly defined patient segments with actionable assays for risk and treatment decisions, similar to the paradigm shift that occurred in AML with FLT3 and IDH2 inhibitors. This work, along with the recent wave of DLBCL classification tools, was a major step in that direction.

[0344] [Table 3]

[0345] [Table 4]

[0346] 7.3 Example 3: Classified subgroups A1 to A8 Integrated clustering identified eight subgroups of ndDLBCL patients (designated A1–A8). The resulting clusters were analyzed in the lens of various biological features, including gene signatures such as double hit gene (DHIT+) signature and TMD gene signature (TME+). The prevalence and biological features (e.g., COO type, DHITsig positive, TME gene signature positive, BCL2 / BCL6 translocation, and MYC translocation) in the replication dataset (MER dataset) are summarized in Table 3. The resulting identification of clusters predicted the likelihood of response to standard treatments (e.g., R-CHOP combination therapy), and rational targeted therapy based on cluster-specific biological features was proposed.

[0347] [Table 5]

[0348] This clustering method enabled the transcriptomic identification of eight patient subgroups (e.g., subgroups A1–A8). For example, among the identified patient groups, subgroup A7 was a high-risk subgroup and was underserved by standard R-CHOP therapy.

[0349] Patients in the high-risk subgroup A7 showed i) low expression of TME signatures including low levels of infiltrating immune cells; ii) high expression of malignant processes such as MYC targets and proliferation; iii) high expression of tumor metabolic signatures (e.g., ribosomal processes and oxidative phosphorylation); iv) a mixture of B cell lineage signatures; and v) upregulation of B cell transcription factors such as IRF4 and OCT-2.

[0350] 7.4 Example 4: Proportions of cells within subgroups The total cell counts of different T cells in DLBCL patients in a subset (n=46) of the Discovery-2 dataset were measured using Multiplexed Ion Beam Imaging (MIBI). Patients were identified in subgroups A7 and non-A7. Figures 11A-11F show the patient total cell counts in the different subgroups.

[0351] From the foregoing, it should be understood that, although specific embodiments have been described herein for purposes of illustration, various modifications may be made without departing from the spirit and scope of what is provided herein. All references mentioned above are incorporated herein by reference in their entirety.

Claims

1. A method for predicting the response of lymphoma patients to cancer treatment: (a) A step of clustering reference lymphoma patients in a reference patient group into subgroups using the expression level of at least one gene in a reference biological sample of the reference lymphoma patients; (b) The step of determining the subgroup to which the lymphoma patient belongs based on the expression level of the at least one gene in the biological sample of the lymphoma patient; and (c) A method comprising the step of predicting the response of the lymphoma patient to a first cancer treatment based on a subgroup of the lymphoma patient.

2. (i) step (a) includes generating clustering information that defines relationships between the expression levels of the at least one gene in the reference biological sample, and / or rearranging a heatmap representation based on the clustering information. (ii) The method according to claim 1, wherein step (a) uses a hierarchical method, a non-hierarchical method, or an iClusterPlus method.

3. The method according to claim 1, wherein the reference lymphoma patients are clustered into 2 to 12 subgroups or 7 subgroups.

4. The method according to claim 1, further comprising the step of training a classifier model using the expression levels of the at least one gene in the reference biological sample, wherein the at least one gene comprises five or more genes from Table 1, or all of the genes from Table 1.

5. The method according to claim 4, wherein the classifier model is a grouped multinomial generalized linear model (GLM), and optionally the classifier model is a binary model.

6. The method according to claim 4, further comprising the step of setting a threshold confidence level for at least one of the subgroups of step (a) and excluding patients from the at least one subgroup that provide clustering data with a lower confidence level.

7. The method according to claim 1, wherein the lymphoma is diffuse large B-cell lymphoma (DLBCL), slow B-cell lymphoma, follicular lymphoma, small lymphocytic lymphoma, nodal marginal zone B-cell lymphoma, lymphoplasmacytic lymphoma, anaplastic large cell lymphoma, primary cutaneous lymphoma, mycosis fungoides, chronic lymphocytic leukemia, or mantle cell lymphoma.

8. The reference patients in the aforementioned reference patient group are clustered into subgroups A1 to A7, and: (i) Subgroup A1 includes approximately 50% to 60% of patients with germinal center B-cell-like (GCB) DLBCL, approximately 30% to 40% of patients with activated B-cell-like (ABC) DLBCL, approximately 10% to 20% of patients with TME+ DLBCL, and approximately 30% to 40% of patients with DHITsig+ DLBCL; (ii) Subgroup A2 includes approximately 80% to 90% of patients with GCB DLBCL, approximately 0% to 5% of patients with ABC DLBCL, approximately 15% to 25% of patients with TME+ DLBCL, and approximately 25% to 35% of patients with DHITsig+ DLBCL; (iii) Subgroup A3 includes approximately 40% to 55% of patients with GCB DLBCL, approximately 30% to 45% of patients with ABC DLBCL, approximately 40% to 50% of patients with TME+ DLBCL, and approximately 20% to 30% of patients with DHITsig+ DLBCL; (iv) Subgroup A4 includes approximately 25% to 35% of patients with GCB DLBCL, approximately 40% to 50% of patients with ABC DLBCL, approximately 30% to 40% of patients with TME+ DLBCL, and approximately 10% to 20% of patients with DHITsig+ DLBCL; (v) Subgroup A5 includes approximately 20% to 40% of patients with GCB DLBCL, approximately 45% to 65% of patients with ABC DLBCL, approximately 30% to 40% of patients with TME+ DLBCL, and approximately 0% to 10% of patients with DHITsig+ DLBCL; (vi) Subgroup A6 includes approximately 30% to 40% of patients with GCB DLBCL, approximately 40% to 50% of patients with ABC DLBCL, approximately 75% to 95% of patients with TME+ DLBCL, and approximately 0% to 10% of patients with DHITsig+ DLBCL; and (vii) The method according to claim 1, wherein subgroup A7 comprises about 0% to about 10% of patients having GCB DLBCL, about 80% to about 90% of patients having ABC DLBCL, about 0% to about 10% of patients having TME+ DLBCL, and about 0% to about 15% of patients having DHITsig+ DLBCL.

9. The method according to claim 1, wherein the first cancer treatment is combination therapy with rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP).

10. (i) If the lymphoma patient is determined to belong to subgroup A1, the step of predicting that the patient is unlikely to respond to the first cancer treatment; (ii) If the lymphoma patient is determined to belong to subgroup A2, the step of predicting that the patient is unlikely to respond to the first cancer treatment; (iii) If the lymphoma patient is determined to belong to subgroup A3, the step of predicting that the patient is unlikely to respond to the first cancer treatment; (iv) If the lymphoma patient is determined to belong to subgroup A4, the step of predicting that the patient is unlikely to respond to the first cancer treatment; (v) If the lymphoma patient is determined to belong to subgroup A5, the step of predicting that the patient is unlikely to respond to the first cancer treatment; (vi) If the lymphoma patient is determined to belong to subgroup A6, the step of predicting that the patient is unlikely to respond to the first cancer treatment; (vii) If the lymphoma patient is determined to belong to subgroup A7, the method according to claim 1, further comprising the step of predicting that the patient is unlikely to respond to the first cancer treatment.

11. A method for predicting the response of lymphoma patients to cancer treatment: (a) A step of determining the expression level of at least one gene from Table 1 in a biological sample of a lymphoma patient, wherein the at least one gene comprises five or more genes from Table 1; (b) a step of comparing the expression level of the at least one gene in step (a) with the expression level of the at least one gene in a reference biological sample of a reference lymphoma patient, the step of which the reference lymphoma patient responds to the cancer treatment, and A method in which, if the expression level of the at least one gene in the biological sample is similar to the expression level of the at least one gene in the reference biological sample, the lymphoma patient is less likely to respond to the cancer treatment, and the cancer treatment is R-CHOP, at the discretion of the patient.

12. A method for predicting the response of lymphoma patients to cancer treatment: (a) the step of determining the expression level of at least one gene in Table 1 in a biological sample from a lymphoma patient; and (b) Comparing the expression level of the at least one gene in the biological sample with: (i) the expression level of the at least one gene in a biological sample of a lymphoma patient responding to the cancer treatment, and (ii) the expression level of the at least one gene in a biological sample of a lymphoma patient not responding to the cancer treatment, A method in which, if the expression level of (a) is similar to the expression level of (i), it indicates that the first lymphoma patient is likely to respond to the cancer treatment; and if the expression level of (a) is similar to the expression level of (ii), it indicates that the first lymphoma patient is unlikely to respond to the cancer treatment, wherein the cancer treatment is R-CHOP, at the discretion of the patient.

13. The method according to claim 12, further comprising the step of selecting lymphoma patients who are likely to respond to the cancer treatment for the cancer treatment.

14. The method according to claim 11, further comprising the step of selecting lymphoma patients who are unlikely to respond to the cancer treatment for an alternative cancer treatment, wherein the alternative cancer treatment is optionally a BET inhibitor or a CDK inhibitor.

15. The method according to claim 12, further comprising the step of selecting lymphoma patients who are unlikely to respond to the cancer treatment for an alternative cancer treatment, wherein the alternative cancer treatment is optionally a BET inhibitor or a CDK inhibitor.

16. The method according to claim 11, wherein the lymphoma is DLBCL, slow chronic B-cell lymphoma, follicular lymphoma, small lymphocytic lymphoma, nodal marginal zone B-cell lymphoma, lymphoplasmacytic lymphoma, anaplastic large cell lymphoma, primary cutaneous lymphoma, mycosis fungoides, chronic lymphocytic leukemia, or mantle cell lymphoma.

17. The method according to claim 12, wherein the lymphoma is DLBCL, slow chronic B-cell lymphoma, follicular lymphoma, small lymphocytic lymphoma, nodal marginal zone B-cell lymphoma, lymphoplasmacytic lymphoma, anaplastic large cell lymphoma, primary cutaneous lymphoma, mycosis fungoides, chronic lymphocytic leukemia, or mantle cell lymphoma.

18. The method according to claim 11, wherein the expression levels of all genes in Table 1 are determined in (a) and compared in (b).

19. The method according to claim 12, wherein the expression levels of all genes in Table 1 are determined in (a) and compared in (b).

20. A pharmaceutical composition comprising a second cancer treatment for use in a method for treating a lymphoma patient according to Claim 1, wherein the method comprises administering the composition to the lymphoma patient, and optionally the second cancer treatment is R-CHOP, not R-CHOP, a bromodomain and extraterminal (BET) inhibitor, or a cyclin-dependent kinase (CDK) inhibitor.

21. The method according to any one of claims 1 to 19 or the pharmaceutical composition according to claim 20, wherein the biological sample is a tumor biopsy sample.

22. (i) The step of determining the expression level of the at least one gene includes the step of detecting the presence or amount of at least one complex in the biological sample, wherein the presence or amount of the at least one complex indicates the expression level of the at least one gene, and optionally the at least one complex is a hybridization complex or the at least one complex is detectably labeled; or (ii) The method according to any one of claims 1 to 19 or the pharmaceutical composition according to claim 20, wherein the step of determining the expression level of the at least one gene includes the step of detecting the presence or amount of the at least one reaction product in the biological sample, the presence or amount of the at least one reaction product indicating the expression level of the at least one gene, and optionally the at least one reaction product is detectably labeled.

23. (i) The reference lymphoma patient is a patient with refractory DLBCL, a patient with relapsed DLBCL, or a newly diagnosed DLBCL patient; and / or (ii) The lymphoma patient is a patient with refractory DLBCL, a patient with relapsed DLBCL, or a newly diagnosed DLBCL patient, according to the method of any one of claims 1 to 19 or the pharmaceutical composition according to claim 20.

24. The method according to any one of claims 1 to 19 or the pharmaceutical composition according to claim 20, wherein the lymphoma patient is a GCB DLBCL patient, an ABC DLBCL patient, a DHITsig+ DLBCL patient, or a DHITsig-DLBCL patient.

Images