Method for predicting responsiveness of lymphoma to drug and method for treating lymphoma

Abstract

Provided herein are methods of predicting responsiveness of a lymphoma patient to cancer treatment. Also provided herein are methods of treating a lymphoma patient based on predicting the responsiveness of the lymphoma patient to a cancer treatment.

Description

Cross-reference to applications related to methods for predicting lymphoma response to drugs and methods for treating lymphoma.

[0001] This application claims the benefit of U.S. Provisional Patent Application No. 63 / 588,889, filed October 9, 2023, the contents of which are incorporated herein by reference in their entirety. Sequence List

[0002] This application contains an electronic sequence list, which has been submitted with this application in XML file format, the entire contents of which are incorporated herein by reference in their entirety. The sequence list XML file submitted with this application is named "14247-755-228_SEQ_LISTING.xml", was created on September 2, 2024, and is 4,153 bytes in size. Technical Field

[0003] This article provides a method for predicting the responsiveness of lymphoma patients to cancer treatment. It also provides a method for treating lymphoma patients based on this predicted responsiveness. Background Technology

[0004] Non-Hodgkin lymphoma (NHL) is a diverse group of blood cancers, encompassing any type of lymphoma other than Hodgkin lymphoma. The severity of NHL varies greatly among different types, ranging from indolent to highly aggressive. Less aggressive NHLs can be associated with longer survival, while more aggressive NHLs can be rapidly fatal if left untreated. They can be B-cell or T-cell-mediated. B-cell NHLs include Burkitt lymphoma, chronic lymphocytic leukemia / small lymphocytic lymphoma (CLL / SLL), diffuse large B-cell lymphoma, follicular lymphoma, immunoblastic large cell lymphoma, precursor B-cell lymphoblastic lymphoma, and mantle cell lymphoma. T-cell NHLs include mycosis fungoides, anaplastic large cell lymphoma, and precursor T-cell lymphoblastic lymphoma. Prognosis and treatment depend on the stage and type of the disease.

[0005] Diffuse large B-cell lymphoma (DLBCL) accounts for approximately one-third of all non-Hodgkin's lymphomas. While some DLBCL patients have been cured with conventional chemotherapy, the rest die from the disease. Anticancer drugs may achieve rapid and sustained depletion of lymphocytes by directly inducing apoptosis in mature T and B cells. See Stahnke et al., Blood, 2001, 98:3066-3073.

[0006] Diffuse large B-cell lymphoma (DLBCL) can be classified into different molecular subtypes based on its genomic profiling: germinal center B-cell-like DLBCL (GCB-DLBCL), activated B-cell-like DLBCL (ABC-DLBCL), and primary mediastinal B-cell lymphoma (PMBL) or unclassified types. These subtypes are characterized by significant differences in survival, chemotherapy responsiveness, and dependence on signal transduction pathways, particularly the NF-κB pathway. See Kim et al., Journal of Clinical Oncology, Proceedings of the 2007 ASCO Annual Meeting, Part I, Volume 25, Issue 18S (Supplement 20 June), 2007: 8082. See Bea et al., Blood, 2005; 106: 3183-3190; Ngo et al., Nature, 2011; 470:115-119. Such differences have prompted the search for more effective subtype-specific treatment strategies in DLBCL.

[0007] Generally, current cancer treatments may involve surgery, chemotherapy, hormone therapy, and / or radiation therapy to eradicate the neoplastic cells in the patient's body (see, for example, Stockdale, 1998, Medicine, Vol. 3, edited by Rubenstein and Federman, Chapter 12, Section IV). More recently, cancer treatments may also involve biotherapy or immunotherapy. All of these methods have significant drawbacks for patients. For example, surgery may be contraindicated or unacceptable to the patient due to their health condition. Furthermore, surgery may not completely remove tumor tissue. Radiation therapy is only effective when tumor tissue is more sensitive to radiation than normal tissue. In addition, radiation therapy can often cause serious side effects. Hormone therapy is rarely used as a single agent. While hormone therapy can be effective, it is often used to prevent or delay cancer recurrence after other treatments have removed most of the cancer cells. Biotherapy and immunotherapy are limited in number and can produce side effects such as rash or swelling, flu-like symptoms (including fever, chills, and fatigue), gastrointestinal problems, or allergic reactions.

[0008] In the context of DLBCL, treatment typically involves a combination 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), with the addition of etoposide (R-EPOCH) in some cases). Because the disease can progress rapidly, DLBCL usually requires immediate treatment after diagnosis. For some patients, DLBCL relapses or becomes refractory after treatment. Several alternative treatments are currently being tested in clinical trials for newly diagnosed, relapsed, or refractory DLBCL patients, some of which may involve the use of lenalidomide. See Czuczman et al., Clin Cancer Res., 2017, 23:4127-4137.

[0009] Indolent lymphomas also exist within non-Hodgkin's lymphomas. Indolent B-cell lymphomas include, for example, follicular lymphoma, small lymphocytic lymphoma; 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 lymphomas is characterized by several features, one of which is the risk of histological transformation (HT) into aggressive lymphoma, primarily DLBCL, and less commonly Burkitt lymphoma (BL) or other types of aggressive lymphoma. See Montoto et al., Journal of Clinical Oncology, 2011, 29:1827-1834.

[0010] Due to the clinical and biological heterogeneity of lymphomas, particularly DLBCL, there is an urgent need for effective methods to classify DLBCL subtypes for specific treatment. DLBCL is traditionally classified based on the cell of origin (COO) subtype of tumor gene expression profiles, including activated B cells (ABC) and germinal center B cells (GCB) subtypes. See Alizadeh et al., Nature, 2000, 403:503-511; Wright et al., Proc. Natl. Acad. Sci. USA, 2003, 100:9991-9996; Scott et al., Blood, 2014, 123:1214-1217. GCB and ABC subtypes have different pathogenic mechanisms and may influence outcomes for DLBCL patients in targeted therapy. See Nyman et al., Mod. Pathol. [Modern Pathology], 2009, 22:1094-1101; Hans et al., Blood, 2004, 103:275-282; Choi et al., Clin. Cancer Res., 2009, 15:5494-5502; Meyer et al., J Clin Oncol., 2011, 29:200-207; Natkunam et al., J Clin Oncol., 2008, 26:447-454.

[0011] Recently, new classification models have focused on DNA alterations using tumor samples from patients treated with R-CHOP. See, for example, Schmitz et al., N Engl J Med. [New England Journal of Medicine], 2018, 378(15):1396-1407. However, a comprehensive approach capable of utilizing transcriptomic data from both newly diagnosed (nd) and relapsed / refractory (r / r) DLBCL has not yet been realized. This invention addresses this and other needs. Summary of the Invention

[0012] In one aspect, this paper provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) clustering reference lymphoma patients into subgroups using the expression levels of at least one gene in reference biological samples of reference lymphoma patients; (b) determining the subgroup to which the lymphoma patient belongs based on the expression levels of the at least one gene in the biological samples 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.

[0013] In some embodiments, the method further includes administering a second cancer treatment to a lymphoma patient based on a subgroup of the lymphoma patient.

[0014] In some embodiments, the at least one gene comprises two or more genes, and step (a) includes generating clustering information that defines the relationship between the expression levels of two or more genes in these reference biological samples, and rearranging the heatmap display based on the clustering information.

[0015] In some embodiments, step (a) includes using a hierarchical method or a non-hierarchical method. In some embodiments, step (a) includes using the iClusterPlus method.

[0016] In some embodiments, these reference lymphoma patients are clustered into 2-12 subgroups. In some embodiments, these reference lymphoma patients are clustered into 7 subgroups.

[0017] In some embodiments, the method further includes training a classifier model using the expression level of at least one gene in a reference biological sample.

[0018] In some embodiments, the at least one gene is selected from the genes in Table 2, optionally including one, two, three, four, five, or more genes from Table 2. In some embodiments, the at least one gene includes all the genes in Table 2. In some embodiments, the at least one gene is selected from the genes in Table 3, optionally including one, two, three, four, five, or more genes from Table 3. In some embodiments, the at least one gene includes all the genes in Table 3. In some embodiments, the at least one gene is selected from the genes in Table 4, optionally including one, two, three, four, five, or more genes from Table 4. In some embodiments, the at least one gene includes all the genes in Table 4.

[0019] In some embodiments, the classifier model is a grouped multinomial generalized linear model (GLM). In some embodiments, the classifier model is a binary generalized linear model.

[0020] In some embodiments, the method further includes setting a threshold confidence level for at least one of these subgroups to exclude patients from at least one of these subgroups who provide clustering data with a lower confidence level.

[0021] 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, 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 DLBCL. In some embodiments, the lymphoma is indolent B-cell lymphoma, marginal zone B-cell lymphoma, mantle cell lymphoma, or chronic lymphocytic leukemia.

[0022] In some embodiments, these reference lymphoma patients are clustered into seven subgroups A1-A7, and subgroup A7 includes approximately 0% to approximately 10% of germinal center B-cell-like (GCB) DLBCL patients, approximately 80% to approximately 90% of activated B-cell-like (ABC) DLBCL patients, approximately 0% to approximately 10% of TME+ DLBCL patients, and approximately 5% to approximately 15% of DHITsig+ DLBCL patients.

[0023] In some embodiments, (i) subgroup A1 comprises approximately 50% to 60% of GCB DLBCL patients, approximately 30% to 40% of ABC DLBCL patients, approximately 10% to 20% of TME+ DLBCL patients, and approximately 30% to 40% of DHITsig+ DLBCL patients; (ii) subgroup A2 comprises approximately 80% to 90% of GCB DLBCL patients, approximately 0% to 5% of ABC DLBCL patients, approximately 15% to 25% of TME+ DLBCL patients, and approximately 25% to 35% of DHITsig+ DLBCL patients; (iii) subgroup A3 comprises approximately 40% to 55% of GCB DLBCL patients, approximately 30% to 45% of ABC DLBCL patients, approximately 40% to 50% of TME+ DLBCL patients, and approximately 20% to 30% of DHITsig+ DLBCL patients; (iv) Subgroup A4 comprises 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 comprises 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; and / or (vi) Subgroup A6 comprises 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.

[0024] In some embodiments, the method further includes predicting that lymphoma patients identified as belonging to subgroup A7 are unlikely to respond to primary cancer treatment.

[0025] In some embodiments, the first cancer treatment is a combination of rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP).

[0026] In some embodiments, the second cancer treatment is not R-CHOP.

[0027] In some embodiments, the second cancer treatment is a combination of lenalidomide, rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R2-CHOP). In some embodiments, the second cancer treatment is a bromodomain and superterminal (BET) inhibitor or a cyclin-dependent kinase (CDK) inhibitor.

[0028] On the other hand, this article provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) determining the expression level of at least one gene from Table 2, Table 3, or Table 4 in a biological sample from a lymphoma patient, optionally wherein the at least one gene comprises one, two, three, four, five, or more genes from Table 2, Table 3, or Table 4; and (b) 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 from a reference lymphoma patient who is responsive to the cancer treatment, and wherein 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, it indicates that the lymphoma patient may be responsive to the cancer treatment.

[0029] On the other hand, this article provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) determining the expression level of at least one gene from Table 2, Table 3, or Table 4 in a biological sample of a lymphoma patient, optionally wherein the at least one gene comprises one, two, three, four, five, or more genes from Table 2, Table 3, or Table 4; and (b) comparing the expression level of the at least one gene in the biological sample with: (i) the expression level of at least one gene in a biological sample of a lymphoma patient who has responded to cancer treatment and (ii) the expression level of at least one gene in a biological sample of a lymphoma patient who has not responded to 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 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.

[0030] On the other hand, this article provides a method for treating a lymphoma patient, the method comprising: (i) identifying a lymphoma patient who may respond to cancer treatment according to the method provided herein; and (ii) administering the cancer treatment to the lymphoma patient.

[0031] On the other hand, this article provides a method for treating lymphoma patients, which includes:

[0032] (i) Identify lymphoma patients who are unlikely to respond to cancer treatment using the methods provided herein; and (ii) administer alternative cancer treatment to the lymphoma patients.

[0033] In some embodiments, the cancer treatment is R-CHOP. In some embodiments, the alternative cancer treatment is not R-CHOP. In some embodiments, the alternative cancer treatment is R2-CHOP. In some embodiments, the alternative cancer treatment is a BET inhibitor or a CDK inhibitor.

[0034] In some embodiments, the lymphoma is selected from the group consisting of: DLBCL, indolent B-cell lymphoma, follicular lymphoma, small lymphocytic lymphoma, 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 DLBCL. In some embodiments, the lymphoma is indolent B-cell lymphoma, follicular lymphoma, marginal zone B-cell lymphoma, mantle cell lymphoma, or chronic lymphocytic leukemia. In some embodiments, the expression levels of all genes in Tables 2, 3, or 4 are determined in step (a) and compared in step (b).

[0035] In some embodiments, the biological sample and the reference biological sample are tumor biopsy samples.

[0036] In some embodiments, the expression level of the at least one gene is determined by the mRNA level of the at least one gene.

[0037] In some embodiments, the expression level of the at least one gene is determined by detecting the presence or amount of at least one complex in the biological sample or reference biological sample, wherein the presence or amount of the at least one complex indicates the expression level of the at least one gene. In some embodiments, the at least one complex is a hybridization complex or is detectably labeled.

[0038] In some embodiments, the expression level of the at least one gene is determined by detecting the presence or amount of at least one reaction product in the biological sample or reference biological sample, wherein the presence or amount of the at least one reaction product indicates the expression level of the at least one gene. In some embodiments, the at least one reaction product is detectably labeled.

[0039] In some embodiments, the reference lymphoma patient is a patient with refractory DLBCL, relapsed DLBCL, or newly diagnosed DLBCL. In some embodiments, the lymphoma patient is a patient with refractory DLBCL, relapsed DLBCL, or newly diagnosed DLBCL. In some embodiments, the lymphoma patient is a GCB DLBCL patient or an ABC DLBCL patient. In some embodiments, the lymphoma patient is a DHITsig+ DLBCL patient or a DHITsig- DLBCL patient. Attached Figure Description

[0040] Figures 1A-1E depict the application of unsupervised clustering to large cohorts of RNAseq data derived from patients to identify biologically homogeneous populations of DLBCL. Figure 1A illustrates the data transformation, unsupervised clustering, and classifier training methods. Figure 1B depicts a heatmap of co-cluster frequencies, identifying sample clusters that consistently group together in repeated secondary sampling runs. Figure 1C depicts the top 50 upregulated and downregulated genes for each cluster. Figure 1D depicts the top 50 upregulated and downregulated genes for each cluster in the independent cohort MER. Figure 1E depicts the top 50 upregulated and downregulated genes for each cluster in the independent cohort REMoDL-B.

[0041] Figures 2A-2I depict clinical outcomes stratified by cluster. Figure 2A depicts the event-free survival (EFS) of ABC-COO patients receiving RCHOP in the ROBUST subset of the Discovery cohort. Figure 2B depicts the EFS of patients receiving RCHOP in the MER replication cohort. Figure 2C depicts the progression-free survival (PFS) of patients receiving RCHOP in the REMoDL-B replication cohort. Figures 2D-2F depict the EFS / PFS of the ROBUST (Figure 2D), MER (Figure 2E), and REMoDL-B (Figure 2F) cohorts classified as A7 / non-A7. Figures 2G-2I depict forest plots of the log-odds ratios for A7 in the ROBUST (Figure 2G), MER (Figure 2H), and REMoDL-B (Figure 2I) cohorts in the presence of multiple clinical factors.

[0042] Figures 3A-3E depict the biological characteristics of the discovered subtypes. Figure 3A depicts a scatter plot of the Discovery cohort in the Reddy COO score versus the TME26 score space. Figure 3B depicts a selection of DLBCL features showing clustering differentiation signals. Figure 3C depicts cluster-associated single nucleotide variants (SNVs) (ROBUST). Figure 3D depicts cluster-associated copy number variants (CNVs) (ROBUST). Figure 3E depicts representative immunohistochemical (IHC) images of A6 and A7.

[0043] Figures 4A-4G depict the biological characteristics of A7. Figure 4A depicts the significantly dysregulated marker pathways in A7 (Discovery), sorted by p-value, showing the Normalized Enrichment Score (NES). Figure 4B depicts MYC gene expression by A7 status for each cohort. Figure 4C depicts representative MYC staining in A7. Figure 4D depicts copy number amplification / deletion frequencies in A7. Figure 4E depicts Western blots showing TCF4 expression in ABC-like DLBCL cell lines. Figure 4F depicts Western blots showing MYC and TCF4 expression in TCF4 knockdown ABC-like DLBCL cell lines. GAPDH was used as a loading control in these Western blots. Figure 4G depicts cell proliferation assays in ABC-like DLBCL cell lines with control (shNT) and TCF4 knockdown (shTCF4). Error bars represent three technical replicates of SEM (SEM < 1 is not shown).

[0044] Figures 5A-5B depict the clinical utility of A7. In ROBUST (Figure 5A) and REMoDL-B (Figure 5B), A7 was a predictor of outcomes for patients receiving RCHOP treatment, but a poor predictor of outcomes for patients receiving R2CHOP (lenalidomide in combination with R-CHOP) or RBCHOP (bortezomib in combination with R-CHOP).

[0045] Figures 6A-6D depict how, compared to other samples (Discovery), the samples clustered into A8 had significantly lower alignment rates with coding regions (Figure 6A), and significantly higher proportions of intergenic regions (Figure 6B), ribosomes (Figure 6C), and unaligned reads (Figure 6D).

[0046] Figure 7 depicts the confusion matrix of the classifier output in the training dataset (Discovery).

[0047] Figures 8A-8C depict the survival probabilities of Robust (Figure 8A), Mer (Figure 8B), and REMoDL-B (Figure 8C). A7 shows the stratified risk within the clinically defined International Prognostic Index (IPI) group.

[0048] Figures 9A-9B depict the pathways that are significantly imbalanced in each discovered cluster (Discovery) when each cluster is compared to all other clusters (Figure 9A). Figure 9B depicts the generally consistent pathway enrichment scores between the Discovery and MER datasets, where most pathways share directionality and significance (marked in red). Cluster A4 is an exception to this trend, with many pathways showing opposite directionality between Discovery and MER.

[0049] Figures 10A-10F depict the copy number distortion rate (ROBUST + MER) in each cluster compared to the rest of the population. Figure 10A depicts the A1 DLBCL compared to non-A1 (ROBUST + MER), Figure 10B depicts the A2 DLBCL compared to non-A2 (ROBUST + MER), Figure 10C depicts the A3 DLBCL compared to non-A3 (ROBUST + MER), Figure 10D depicts the A4 DLBCL compared to non-A4 (ROBUST + MER), Figure 10E depicts the A5 DLBCL compared to non-A5 (ROBUST + MER), and Figure 10F depicts the A6 DLBCL compared to non-A6 (ROBUST + MER).

[0050] Figures 11A-11F depict the major immune types detected by multiplex ion beam imaging (MIBI) in a portion of the ROBUST+ cohort (n = 43). Cell abundance is expressed as a percentage of the total number of nucleated cells within the field of view (FOV). Figure 11A depicts total CD3 T cells; Figure 11B depicts CD4 T cells; Figure 11C depicts CD8 T cells; Figure 11D depicts CD163 macrophages / monocytes; Figure 11E depicts CD68 macrophages; and Figure 11F depicts CD11c dendritic cells.

[0051] Figures 12A-12E depict the genomic characteristics of A7. Figure 12A depicts A7-related genomic events and their associated variant allele frequencies (VAF) and cancer cell fractions (CCF). A7-related CNAs tend to be highly clonal, with a CCF of 100% in most samples. On the other hand, most A7 mutation events are observed at least partially or even entirely in subclones. Figure 12B depicts A7-related CNAs (MER). Figure 12C depicts MYC expression in the A7 state (MER). Figure 12D depicts tumor purity (ROBUST) in the A7 state. Figure 12E depicts tumor purity (ROBUST) associated with near-zero MYC expression.

[0052] Figures 13A-13B depict TCF4 mRNA expression in patients with the indicated TCF4 copy number alteration (ROBUST) (Figure 13A) or in the DLBCL cell line (Figure 13B).

[0053] Figures 14A-14B depict a comparison of A7 and MCD in the NCI cohort. Figure 14A depicts the PFS of patients receiving immunochemotherapy stratified using MCD and A7 status. Figure 14B depicts a Sankey plot illustrating LymphGen clustering and the co-occurrence of A1-A7.

[0054] Figures 15A-15D depict Sankey plots and related confusion matrices of the subtypes identified in ROBUST (Figures 15A and 15C) and MER (Figures 15B and 15D) compared to the LymphGen classifier output. For A7 patients, the MCD subtypes (based on MYD88) L265P Co-occurrence with CD79B mutations (as described by Schmitz et al., 2018, New England Journal of Medicine, 378(15), 1396-1407) was enriched; while for GCB-like clusters A2 and A3, the EZB subtype (based on EZH2 mutations and BCL2 translocations, as described by Schmitz et al., 2018, New England Journal of Medicine, 378(15), 1396-1407) was enriched. Although there were statistically significant associations among the various classification methods (Fisher p = 0.0005 in ROBUST, p = 0.001 in MER), there was significant heterogeneity, and no clear one-to-one mapping existed between any subtypes.

[0055] Figure 16 depicts the novel clustering compared to LymphGen (NCI cohort). After normalization, the PCA plots of the Discovery, MER, and REMoDL-B datasets do not show dataset-specific differences.

[0056] Figures 17A-17C depict the mutation landscape (Chapuy gene), in the following order: by mutation count (Figure 17A), by significance (corrected for gene length) (Figure 17B), and by Chapuy plot (for reference) (Figure 17C).

[0057] Figures 18A-18D depict the expression of proteins encoded by genes on chromosome 18 (Chr18). Figure 18A depicts expression on Chr18 due to copy number amplification. Figure 18B depicts a Western blot showing MTAP expression in DLBCL cell lines. Figure 18C depicts a Western blot showing SDMA and PRMT5 expression in DLBCL cell lines. GAPDH was used as a loading control in these blots. Figure 18D depicts cell proliferation assays in ABC-like DLBCL cell lines with control (shNT) and PRMT5 knockdown (shPRMT5). Error bars represent three technical replicates of SEM. Detailed Implementation

[0058] 5.1 Definition

[0059] 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, 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.

[0060] As used herein and unless otherwise stated, the term “treat, treating, and treatment” refers to measures taken when a patient has the specified cancer (a specific type of lymphoma, such as DLBCL) that may reduce the severity of the cancer or delay or slow its progression.

[0061] When referring to cancer treatment, the terms "sensitivity" or "sensitive" are relative terms that refer to the degree of effectiveness of cancer treatment in reducing or decreasing the progression of a tumor or the cancer being treated. For example, when referring to cell or tumor treatments associated with a compound, the term "increased sensitivity" means an increase in the effectiveness of the cancer treatment by at least about 5% or more.

[0062] As used herein and unless otherwise stated, the term "therapeuticly effective amount" in cancer treatment is an amount sufficient to provide therapeutic benefit in treating or managing cancer or to delay or minimize one or more symptoms associated with the presence of cancer. A therapeutically effective amount of a compound means the amount of a therapeutic agent, alone or in combination with other therapies, that provides therapeutic benefit in treating or managing cancer. The term "therapeuticly effective amount" may encompass amounts that improve overall therapy, reduce or prevent symptoms or causes of cancer, or enhance the therapeutic efficacy of another therapeutic agent. The term also refers to the 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 being sought by researchers, veterinarians, physicians, or clinicians.

[0063] When referring to cancer treatment, the term "responsiveness" or "reactivity" refers to the degree of effectiveness of the treatment in alleviating or reducing symptoms of the cancer being treated, such as DLBCL. For example, when referring to treatment of cells or subjects, the term "increased responsiveness" refers to an increase in effectiveness in alleviating or reducing symptoms of the disease compared to a reference treatment (e.g., treatment of the same cells or subjects, or treatment of different cells or subjects), when measured using any method known in the art. In some embodiments, the increase in 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%.

[0064] As used herein, the terms “effective subject response,” “effective patient response,” and “effective patient tumor response” refer to any increase in the therapeutic benefit to a patient. An “effective patient tumor response” can be, for example, a decrease in the rate of tumor progression of approximately 5%, approximately 10%, approximately 25%, approximately 50%, or approximately 100%. An “effective patient tumor response” can be, for example, a reduction in the physical symptoms of cancer of approximately 5%, approximately 10%, approximately 25%, approximately 50%, or approximately 100%. An “effective patient tumor response” can also be, for example, an increase in the patient’s response of approximately 5%, approximately 10%, approximately 25%, approximately 50%, approximately 100%, or approximately 200% or more, as measured by any suitable means, such as gene expression, cell count, assay results, tumor size, etc.

[0065] Improvement in cancer (e.g., DLBCL or its subtypes) or cancer-related disease can be characterized as complete or partial remission. "Complete remission" means the absence of clinically detectable disease and the return to normalcy of any previously abnormal imaging findings, bone marrow and cerebrospinal fluid (CSF) measurements, or abnormal monoclonal protein measurements. "Partial remission" means a reduction of at least approximately 10%, approximately 20%, approximately 30%, approximately 40%, approximately 50%, approximately 60%, approximately 70%, approximately 80%, or approximately 90% in all measurable tumor burden (i.e., the number of malignant cells present in the subject, or the measurable tumor mass volume, or the number of abnormal monoclonal proteins) without the presence of new lesions. The term "treatment" includes both complete and partial remission.

[0066] The term "probability" generally refers to an increased probability of an event occurring. When referring to the effectiveness of a patient's tumor response, "probability" is often considered to be an increased probability that the rate of tumor development or tumor cell growth will decrease. When referring to the effectiveness of a patient's tumor response, "probability" can also often refer to an increase in multiple indicators, such as mRNA or protein expression, which may demonstrate increased progress in cancer treatment.

[0067] The term "prediction" generally implies something predetermined or known in advance. For example, when used to "predict" the effectiveness of cancer treatment, the term "prediction" can refer to the possibility that the outcome of cancer treatment can be determined at the start, before treatment begins, or before substantial progress occurs during treatment.

[0068] As used herein, the term "source" refers to the origin of the sample when referring to a reference sample. For example, a sample collected from blood will have a reference sample collected from blood in the same way. Similarly, a sample collected from bone marrow will have a reference sample collected from bone marrow in the same way.

[0069] As used herein, the terms “refractory” or “resistant” refer to an obstacle, disease, or condition that has not responded to prior treatment, which may include first-line or multiple lines of therapy. In some embodiments, the obstacle, disease, or condition has been previously treated with first-, second-, third-, or fourth-line therapy. In some embodiments, the obstacle, disease, or condition has been previously treated with second- or more lines of therapy and has not achieved complete remission (CR) with a most recent regimen containing systemic therapy.

[0070] As used in this article, the term “relapse” refers to a disorder, disease, or condition that responds to treatment (e.g., achieves complete remission) and then progresses. Treatment may include first-line or multiple-line therapies.

[0071] As used herein, the term "expressed" or "expressed" refers to an RNA nucleic acid molecule transcribed from a gene to produce a region at least partially complementary to one of the two nucleic acid strands of that gene. As used herein, the term "expressed" or "expressed" also refers to the translation from an RNA molecule to produce a protein, polypeptide, or a portion thereof. A "biological marker" or "biomarker" is a substance whose detection indicates a specific biological state (e.g., the presence of a certain type of cancer). In some embodiments, biomarkers may be identified individually. In other embodiments, several biomarkers may be measured simultaneously.

[0072] As used interchangeably herein, the terms "peptide" and "protein" refer to a polymer of three or more amino acids linked by peptide bonds and arranged in tandem. The term "peptide" includes proteins, protein fragments, protein analogs, oligopeptides, etc. As used herein, the term "peptide" may also refer to a peptide. The amino acids constituting a polypeptide may be of natural origin or synthetically produced. Polypeptides can be purified from biological samples. Polypeptides, proteins, or peptides further include modified polypeptides, proteins, and peptides, such as glycopeptides, glycoproteins, or glycopeptides; or lipopeptides, lipoproteins, or lipopeptides.

[0073] The terms “antibody,” “immunoglobulin,” or “Ig,” used interchangeably herein, include fully assembled antibodies and antibody fragments that retain the ability to specifically bind antigens. The antibodies described herein include, but are not limited to, synthetic antibodies, monoclonal antibodies, polyclonal antibodies, recombinant antibodies, multispecific antibodies (including bispecific antibodies), human antibodies, humanized antibodies, chimeric antibodies, intracellular antibodies, single-chain antibodies (scFv) (e.g., including monospecific, bispecific, etc.), camelified antibodies, Fab fragments, F(ab') fragments, disulfide-linked Fv (sdFv), anti-idiotype (anti-Id) antibodies, and epitope-binding fragments of any of the above antibodies. In particular, the antibodies described herein include immunoglobulin molecules and immunoglobulin molecules with immunoactive portions, i.e., molecules containing antigen-binding domains or antigen-binding sites that specifically bind CRBN antigens (e.g., one or more complementarity-determining regions (CDRs) of anti-CRBN antibodies). The antibodies provided herein can be immunoglobulin molecules belonging to any class (e.g., IgG, IgE, IgM, IgD, and IgA) or any subclass (e.g., IgG1, IgG2, IgG3, IgG4, IgA1, and IgA2). In some embodiments, the anti-CRBN antibody is fully human, such as a fully human monoclonal CRBN antibody. In some embodiments, the antibodies provided herein are IgG antibodies or their subclasses (e.g., human IgG1 or IgG4).

[0074] The terms "antigen-binding domain," "antigen-binding region," "antigen-binding fragment," and similar terms refer to a portion of an antibody that contains amino acid residues (e.g., CDRs) that interact with an antigen and confer specificity and affinity to that antigen. This antigen-binding region can be derived from any animal species, such as rodents (e.g., rabbits, rats, or hamsters) and humans. In some embodiments, the antigen-binding region is a human antigen-binding region.

[0075] As used herein, the term "epitope" refers to a localized region on the surface of an antigen that is capable of binding to one or more antigen-binding regions of an antibody and possesses antigenic or immunogenic activity in an animal (e.g., a mammal (e.g., a human)) and is capable of inducing an immune response. An immunogenic epitope is a portion of a polypeptide that induces an antibody response in an animal. An antigenically active epitope is a portion of a polypeptide that is specifically bound by an antibody, as determined by any method well known in the art (e.g., by an immunoassay described herein). An antigenic epitope does not necessarily need to be immunogenic. Epitopes typically consist of chemically active surface groups of a molecule, such as amino acids or sugar side chains, and have specific three-dimensional spatial structural features and specific charge features. The polypeptide region constituting an epitope can be a continuous plurality of amino acids of the polypeptide, or the epitope can be composed of two or more discontinuous regions of the polypeptide. An epitope may or may not be a three-dimensional surface feature of an antigen.

[0076] The terms “fully human antibody” and “human antibody” are used interchangeably herein and refer to an antibody that contains a human variable region and, in some embodiments, a human constant region. In specific embodiments, these terms refer to an antibody that contains both human-derived variable and constant regions. The term “fully human antibody” includes antibodies having variable and constant regions corresponding to human germline immunoglobulin sequences, as described in Kabat et al., Sequences of Proteins of Immunological Interest, US Department of Health and Human Services, NIH Publication No. 91-3242 (5th edition 1991).

[0077] The phrase “recombinant human antibody” includes human antibodies prepared, expressed, created, or isolated by recombinant means, such as antibodies expressed using recombinant expression vectors transfected into host cells; antibodies isolated from recombinant or combined human antibody libraries; antibodies isolated from transgenic and / or transchromosomally transfected animals (e.g., mice or cows) using human immunoglobulin genes (see, for example, Taylor et al., Nucl. Acids Res., 1992, 20:6287-6295); or antibodies prepared, expressed, created, or isolated by any other method involving splicing human immunoglobulin gene sequences into other DNA sequences. Such recombinant human antibodies may have variable and constant regions derived from human germline immunoglobulin sequences. See Kabat et al., Sequences of Proteins of Immunological Interest, US Department of Health and Human Services, NIH Publication No. 91-3242 (5th edition 1991). However, in some embodiments, such recombinant human antibodies have undergone in vitro mutagenesis (or, when using transgenic animals with human Ig sequences, in vivo somatic cell mutagenesis), and therefore the amino acid sequences of the heavy chain variable and light chain variable regions of these recombinant antibodies, although derived from and associated with human germline heavy chain variable and light chain variable sequences, may not be naturally present in the in vivo human antibody germline library.

[0078] The term “monoclonal antibody” refers to an antibody obtained from a homogeneous or substantially homogeneous group of antibodies, and each monoclonal antibody typically recognizes a single epitope on that antigen. In some embodiments, “monoclonal antibody” as used herein is an antibody produced by a single hybridoma or other cell, wherein the antibody binds only immunospecifically to an epitope as determined by, for example, an ELISA or other antigen-binding or competitive binding assay known in the art or in the examples provided herein. The term “monoclonal” is not limited to any particular method of preparing the antibody. For example, the monoclonal antibodies provided herein can be prepared by a hybridoma method as described in Kohler et al., Nature, 1975, 256:495-497, or isolated from a phage library using the techniques described herein. Other methods for preparing clonal cell lines and monoclonal antibodies expressed therefrom are well known in the art. See, for example, Short Protocols in Molecular Biology, Chapter 11 (edited by Ausubel et al., John Wiley and Sons, New York, 5th edition 2002). Other exemplary methods for generating other monoclonal antibodies are provided in the examples in this article.

[0079] As used herein, a “polyclonal antibody” refers to a group of antibodies generated in an immunogenic response to a protein having many epitopes, and therefore includes a variety of different antibodies targeting the same or different epitopes within that protein. Methods for generating polyclonal antibodies are known in the art. See, for example, *Short Protocols in Molecular Biology*, Chapter 11 (edited by Ausubel et al., John Wiley and Sons, New York, 5th ed., 2002).

[0080] The term "level" refers to the quantity, accumulation, or rate of molecules. Levels can be expressed, for example, by the amount or rate of synthesis of gene-encoded messenger RNA (mRNA), gene-encoded polypeptides or proteins, or the amount or rate of synthesis of biomolecules accumulating in cells or biological fluids. The term "level" also refers to the absolute or relative amount of molecules in a sample measured under steady-state or non-steady-state conditions.

[0081] “Upregulated” mRNAs are mRNAs whose levels are normally elevated under a given treatment or condition. “Downregulated” mRNAs generally refer to a decrease in mRNA expression levels in response to a given treatment or condition. In some cases, mRNA levels may remain unchanged under a given treatment or condition. When a patient receives treatment, mRNAs from that patient's sample may be “upregulated” compared to an untreated control. This upregulation may be an increase relative to control mRNA levels, 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. Alternatively, mRNAs may be “downregulated” or expressed at lower levels in response to the administration of certain compounds or other agents. The downregulated mRNA may be present, for example, at levels of approximately 99%, approximately 95%, approximately 90%, approximately 80%, approximately 70%, approximately 60%, approximately 50%, approximately 40%, approximately 30%, approximately 20%, approximately 10%, approximately 1%, or lower than the comparative control mRNA level.

[0082] Similarly, when a patient receives treatment, the levels of peptide or protein biomarkers from that patient's sample may be elevated compared to an untreated control. This elevation may be approximately 5%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 100%, 200%, 300%, 500%, 1,000%, 5,000%, or more of the comparative control protein level. Alternatively, in response to the administration of certain compounds or other agents, the levels of protein biomarkers may decrease. This decrease may be, for example, at levels of approximately 99%, 95%, 90%, 80%, 70%, 60%, 50%, 40%, 30%, 20%, 10%, or 1% or lower of the comparative control patient's protein level.

[0083] As used herein, the terms “determine,” “measure,” “evaluate,” “estimate,” and “determine” generally refer to any form of measurement and include determining whether an element exists. These terms include quantitative and / or qualitative determinations. Evaluations can be relative or absolute. “Evaluating existence” can include determining the quantity of something that exists, as well as determining whether it exists or not.

[0084] The terms “isolation” and “purification” refer to the separation of a substance (e.g., mRNA, DNA, or protein) such that the substance constitutes a substantial proportion of the sample in which it is present, that is, a proportion greater than that normally present in the substance in its native or unisolated state. Typically, a substantial proportion of the sample constitutes, for example, greater than 1%, greater than 2%, greater than 5%, greater than 10%, greater than 20%, greater than 50%, or more (typically up to about 90%-100% of the sample). For example, a sample containing isolated mRNA may typically contain at least about 1% total mRNA. Techniques for purifying polynucleotides are well known in the art and include, for example, gel electrophoresis, ion exchange chromatography, affinity chromatography, flow cytometry, and density-based sedimentation.

[0085] As used herein, the term "bonding" refers to direct or indirect attachment. In the context of chemical structure, "bonding" (or "attachment") can refer to the presence of a chemical bond that directly connects two parts or indirectly connects two parts (e.g., by linking a group or any other intermediate part of a molecule). A chemical bond can be a covalent bond, an ionic bond, a coordination complex, a hydrogen bond, a van der Waals interaction, or a hydrophobic stacking, or can exhibit the properties of a variety of chemical bonds. In some cases, "bonding" includes examples of both direct and indirect attachment.

[0086] As used herein, the term "sample" refers to a material or mixture of materials that is typically, but not necessarily, in liquid form and contains one or more target components. In some embodiments, a sample may be a biological sample. As used herein, a "biological sample" means a sample obtained from a living organism, including samples of biological tissue or fluid origin obtained, acquired, or collected in vivo or in situ. Biological samples also include samples taken from biological sites containing precancerous or cancerous cells or tissues. Such samples may be, but are not limited to, organs, tissues, and cells isolated from mammals. Exemplary biological samples include, but are not limited to, cell lysates, cell cultures, cell lines, tissues, oral tissues, gastrointestinal tissues, organs, organelles, biofluids, blood samples, urine samples, skin samples, and so on. Preferred biological samples include, but are not limited to, whole blood, partially purified blood, PBMCs, tissue biopsies, and so on.

[0087] As used in this article, the term "analyte" refers to a known or unknown component of a sample.

[0088] As used herein, the term “capture agent” refers to an agent that binds to mRNA or protein through interactions sufficient to allow the agent to bind to and concentrate mRNA or protein from a heterogeneous mixture.

[0089] As used herein and unless otherwise stated, the term "pharmaceutically acceptable salt" includes non-toxic acid and base addition salts of the compounds referred to by that term. 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, embolic acid, heptanoic acid, etc. Acidic compounds can form salts with a variety of pharmaceutically acceptable bases. Bases that can be used to prepare pharmaceutically acceptable base addition salts of such acidic compounds are those that form non-toxic base addition salts, i.e., salts containing pharmacologically acceptable cations, such as, but not limited to, alkali metal or alkaline earth metal salts (especially calcium, magnesium, sodium, or potassium salts). Suitable organic bases include, but are not limited to, N,N-dibenzylethylenediamine, chloroprocaine, choline, diethanolamine, ethylenediamine, meglumine (N-methylglucosamine), lysine, and procaine.

[0090] As described herein, the term "secondary activator" refers to any additional therapeutic agent with biological activity. It should be understood that a secondary activator can be a hematopoietic growth factor, cytokine, anticancer agent, antibiotic, Cox-2 inhibitor, immunomodulator, immunosuppressant, corticosteroid, therapeutic antibody that specifically binds to a cancer antigen, or a pharmacologically active mutant thereof, 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., veneclax), BTK inhibitors (e.g., ibrutinib or acomitinib), 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), and Aurora. A kinase inhibitors (e.g., alisertib), EZH2 inhibitors (e.g., tazemetostat, GSK126, CPI-1205, 3-deadenine A, EPZ005687, EI1, UNC1999, or sinefungin), BET inhibitors (e.g., birabresib or 4-[2-(cyclopropylmethoxy)-5-(methanesulfonyl)phenyl]-2-methylisoquinoline-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), selective DNMT1 inhibitors (such as GSK3484862), LSD-1 inhibitors (such as compound C or selidemstat), and G9A inhibitors (such as UNC) 0631), PRMT5 inhibitors (e.g., GSK3326595), BRPF1B / 2 inhibitors (e.g., OF-1), BRD9 / 7 inhibitors (e.g., LP99), SUV420H1 / H2 inhibitors (e.g., A-196), Menin-MLL inhibitors (e.g., MI-503), CARM1 inhibitors (e.g., EZM2302), BRD9 (e.g., inhibitor dBrd9), aiolos / ikaros cerebellar protein E3 ligase modulators (CELMoDs), CREBBp2 inhibitors, anti-CD79b antibodies, CD19 CAR-T, p53 (nutlins) inhibitors, Bcl6 inhibitors, CREBBp2 CELMoDs, CD79b CELMoDs, CD19 CELMoDs, p53 (nutlins) CELMoDs, Bcl6 CELMoDs, ligand-directed degradation (LDD) inhibitors of CREBBP2, LDD inhibitors of CD79b, LDD inhibitors of CD19, LDD inhibitors of p53 (nutlins), LDD inhibitors of Bcl6, LDD inhibitors of CK1a, LDD inhibitors of IRAK4 (e.g., in MYD88 L265p lymphoma), MALT1 inhibitors (e.g., JNJ-67856633), MAT2A inhibitors (e.g., for 9p21 deletion), anti-CD3 x anti-CD19 bispecific antibodies, and anti-CD3 x anti-CD20 bispecific antibodies.

[0091] As used herein, the term “about” means a variation of 20%, 15%, 10%, 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2%, 1%, 0.5%, 0.05%, or less of a given value or range.

[0092] In the claims and / or description, when used in conjunction with the term “comprising,” the use of the word “a or an” may mean “one,” but is also consistent with “one or more,” “at least one,” and “one or more than one.”

[0093] As used herein, unless otherwise indicated by the context, the term "pre-treatment" as used in accordance with the methods described herein means before the application of treatment.

[0094] As used herein, the terms "patient" and "subject" refer to an animal, such as a mammal. In some embodiments, the patient is a human being. 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 a particular embodiment, the patient is a person suffering from lymphoma requiring treatment (e.g., DLBCL).

[0095] 5.2 Methods for clustering lymphoma patients

[0096] In one aspect, this article provides a method for classifying lymphoma patients, comprising (a) measuring the expression level of at least one gene in a sample from a lymphoma patient; and (b) clustering lymphoma patients into a lymphoma patient subgroup using the expression level of the at least one gene in the sample. In some embodiments, the method further comprises obtaining samples from lymphoma patients.

[0097] In some embodiments, the lymphoma is DLBCL. In some embodiments, the lymphoma is indolent B-cell lymphoma. In some embodiments, the lymphoma is selected from the group consisting of: follicular lymphoma, small lymphocytic lymphoma, 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 marginal zone B-cell lymphoma. In some embodiments, the lymphoma is mantle cell lymphoma. In some embodiments, the lymphoma is chronic lymphocytic leukemia.

[0098] In some embodiments, the sample is obtained from tissue from a lymphoma patient containing DLBCL cells. A more detailed description of the sample (e.g., a biological sample) is provided below in Section 5.7.

[0099] 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 both a newly diagnosed (nd) and a relapsed / refractory (r / r) DLBCL patient. In some embodiments, the reference lymphoma patient is a refractory DLBCL patient, a relapsed DLBCL patient, or a newly diagnosed DLBCL patient. In some embodiments, the lymphoma patient is a GCB DLBCL patient or an ABC DLBCL patient.

[0100] As illustrated in Example 1 provided in Section 7.1 below, multiple patient datasets can be utilized in the classification / clustering methods disclosed herein. The patient datasets include, but are not limited to, a screening cohort from the ROBUST clinical trial containing 1016 newly diagnosed DLBCL patients (see clinical trial number: NCT02285062; see, for example, Nowakowski et al., J. Clin. Onco. [Journal of Clinical Oncology], 2021, 39(12), 1317-1328), a set of 192 commercially sourced newly diagnosed DLBCL patient samples with molecular profiling analysis but no survival data (“commercial dataset”, referred to as the “Discovery-2 dataset” when combined with the 1016 patient dataset), a set of 343 ndDLBCL patients from the Molecular Epidemiology Resource (MER) (“ndMER dataset”; see, Cerhan et al., Int. J. Epidermiol. [International Journal of Epidemiology], 2017, 46(6):1753-1754i), and a set of 928 patients from the REMoDL-B dataset (see clinical trial number: NCT01324596). A subset of the Discovery-2 dataset has available outcome and clinical information. The cohorts in the MER dataset 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 commercial datasets and the ROBUST clinical trial screening dataset.

[0101] In some embodiments, a dataset of lymphoma patients is generated by measuring gene expression levels in samples. In some embodiments, these lymphoma patients are a cohort of newly diagnosed lymphoma patients. In some embodiments, these lymphoma patients are a cohort of relapsed / refractory lymphoma patients.

[0102] In some embodiments, clustering lymphoma patients into subgroups includes using a discovery dataset and one or more replication / validation datasets. In some embodiments, clustering these lymphoma patients into subgroups includes using a discovery dataset and a replication / validation dataset. In some embodiments, the discovery dataset is derived from samples from a discovery cohort. In some embodiments, the replication / validation dataset is derived from samples from a replication / validation cohort. In some embodiments, the discovery dataset is selected from the group consisting of: the Discovery-2 dataset, the ndMER dataset, the ROBUST dataset, and the REMoDL-B dataset. In some embodiments, the discovery dataset is selected from the group consisting of: the Discovery-2 dataset, the ndMER dataset, and the REMoDL-B dataset. In some embodiments, the replication / validation dataset is either the Discovery-2 dataset or the ndMER dataset. In some embodiments, the replication / validation dataset is either the ndMER dataset or the REMoDL-B dataset. In some embodiments, the replication / validation dataset includes both the ndMER and REMoDL-B datasets. In some embodiments, the discovery dataset is the Discovery-2 dataset. In some embodiments, the discovery dataset is the ndMER dataset. In some embodiments, the discovery dataset is the ROBUST dataset. In some embodiments, the discovery dataset is the REMoDL-B dataset. In some embodiments, the Discovery-2 dataset includes the ROBUST dataset and commercial datasets. 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.

[0103] In some embodiments, clustering lymphoma patients into subgroups includes (i) standardizing 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 and removing outliers in the dataset. In some embodiments, the clustering step further includes evaluating the clustering results 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 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 a preferred embodiment, the clustering method is the iClusterPlus clustering method. In some embodiments, the clustering method is a Cluster-Cluster Analysis (COCA) clustering method.

[0104] In some embodiments, the dataset is standardized using the single-sample normalization (ssNorm) method prior to analysis. Performing ssNorm facilitates immediate short-term analysis and the long-term translatability of research results. In the short term, ssNorm provides a fixed normalization scheme that can be applied independently to individual samples. No re-normalization is required when adding new batches of samples or combining datasets for larger analyses. The ssNorm method also implicitly performs batch correction, eliminating gene-specific experimental effects that manifest as systematic, biased differences in the original expression space. The ssNorm method correctly aligns different datasets into a common space, without meaningful separation between datasets and batches when projected into the PCA space. ssNorm also allows for easy translation of any classifier or parameterization from one dataset to another – a classifier built on an ssNorm dataset can be directly applied to any other ssNorm dataset without reweighting model parameters or setting new thresholds.

[0105] In some embodiments, normalization is performed using DLBCL-specific housekeeping genes. In some embodiments, normalization is performed using the ISY1, R3HDM1, TRIM56, UBXN4, and / or WDR55 genes.

[0106] Clustering methods are sensitive to input data; introducing distracting or irrelevant features can degrade algorithm performance. Feature selection and feature engineering methods can both be used to reduce the dimensionality of datasets while maintaining the representativeness of relevant biological activities.

[0107] In some embodiments, one or more features (e.g., one, two, three, four, or more) are selected as clustering features and input into the clustering method. In some embodiments, a subset of gene expression data is selected as 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 the most expressed genes. In some embodiments, the subset of gene expression data is the gene expression data of the top 25% of the most expressed genes.

[0108] Derivative feature scores of certain types, which aggregate multiple biologically related genes into a single feature, can also be used as clustering features. Gene set variation analysis (GSVA) ​​is a gene set enrichment (GSE) method that estimates the variation in pathway activity in a sample population in an unsupervised manner. Compared to corresponding methods, GSVA is more capable of detecting subtle changes in pathway activity in a sample population. GSVA forms a starting point for building pathway-centric biological models and can meet the current needs of GSE methods for RNA-seq data. See Hanzelman et al., BMC Bioinformatics, 2013, 14:7. GSVA (marker GSVA score) can be performed on a set of 50 marker pathway genes from MSigDB. See, for example, Liberzon et al., Cell Syst., 2015, 1:417:425. GSVA (C1 position GSVA score) can be performed on a set of 299 C1 position cytoband feature genes from MSigDB. See, for example, Alhamdoosh et al., F1000Research, 2017, 6:2010. GSVA (cell type LM23 GSVA score) can also be performed on the DLBCL-specific LM23 matrix obtained by DCQ cell deconvolution. See Althoum et al., Mol.Syst. Bio., 2014, 10:720.

[0109] In some embodiments, the GSVA score of the marker is selected as the clustering feature. In some embodiments, the GSVA score of the C1 position is selected as the clustering feature. In some embodiments, the GSVA score of the cell type LM23 is selected as the clustering feature. In some embodiments, a matrix is ​​formed by selecting a subset of gene expression data, the GSVA score of the marker, the GSVA score of the C1 position, and / or the GSVA score of the cell type LM23 as the clustering features.

[0110] In some embodiments, the clustering method independently clusters each feature matrix and aggregates the cluster features. In some embodiments, the clustering method performs clustering, aggregates cluster features from different feature matrices, and clusters all cluster features.

[0111] 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). In some embodiments, the dataset is clustered into 8 clusters (K = 8). In some embodiments, the number of clusters is 7.

[0112] In some embodiments, the clustering step further includes evaluating the clustering results of the clustering method. In some embodiments, the clustering results are evaluated for each number of clusters (K). In some embodiments, the clustering results are evaluated using the nbClust R package. In some embodiments, the clustering results are evaluated using an index selected from a group consisting of contour statistics, gap statistics, and the percentage of variance explained by the clustering. In some embodiments, the clustering results are evaluated by a minimum cluster size. To ensure sufficient sample size in each cluster to build a downstream classifier model for robust identification of each cluster or subgroup, the clusters obtained from each test clustering activity need to have a minimum cluster size.

[0113] In some embodiments, the clustering method is the iClusterPlus clustering method, with a cluster size of 7.

[0114] In some embodiments, the clustering method is the iClusterPlus clustering method, which selects a subset of gene expression data, GSVA score of the marker, GSVA score of C1 position, and / or LM23 GSVA score of cell type as a matrix of clustering features, and the number of clusters is 7 (clusters A1-A7).

[0115] In some embodiments, for one or more subgroups, there are patients who give low-confidence-level data for a particular subgroup based on the selected clustering classifier. In some embodiments, patients who give low-confidence-level clustering data are filtered. In some embodiments, patients who give low-confidence-level clustering data are excluded from the subgroup. In some embodiments, patients who give low-confidence-level clustering data are excluded from the subgroup assigned to them. In some embodiments, patients who give low-confidence-level clustering data are excluded from subgroup A7.

[0116] In some embodiments, the method further includes setting a threshold confidence level for at least one of these subgroups to exclude patients who provide clustering data with lower confidence levels from the at least one or more subgroups. In some embodiments, patients who provide clustering data with lower confidence levels are excluded from subgroup A7.

[0117] In some embodiments, the method for clustering lymphoma patients further includes (i) identifying at least one clustering classifier by training a classifier model using the clustering results of a discovery dataset; (ii) applying the at least one clustering classifier to a replication / validation dataset to classify the replication / validation dataset; (iii) clustering the replication / validation dataset using a clustering method; and (iv) comparing the classification result of the replication / validation dataset using the at least one clustering classifier with the clustering result 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 the least absolute value shrinkage and selection operator (LASSO). In some embodiments, if the classification result of the replication / validation dataset using the at least one clustering classifier is similar to the clustering result of the replication / validation dataset using the clustering method, it indicates that the at least one clustering classifier can effectively classify the replication / validation dataset.

[0118] In some embodiments, the method for clustering lymphoma patients further includes (i) identifying at least one clustering classifier by training a classifier model using the clustering results of the discovery dataset; and (ii) applying the at least one clustering 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 the least absolute value shrinkage and selection operator (LASSO).

[0119] In some embodiments, the baseline age of the lymphoma patient is 30 years or older, 35 years or older, 40 years or older, 45 years or older, 50 years or older, 55 years or older, 60 years or older, 65 years or older, or 70 years or older. 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 to 35 years old, 35 to 40 years old, 40 to 45 years old, 45 to 50 years old, 50 to 55 years old, 55 to 60 years old, 60 to 65 years old, or 65 to 70 years old.

[0120] In some embodiments, the at least one gene is selected from the genes in Table 1. In some embodiments, the at least one gene comprises one, two, three, four, five, or more genes from Table 1. In some embodiments, the at least one gene comprises all the genes in Table 1.

[0121] In some embodiments, the expression levels of at least one gene in the discovery dataset are used when training the classifier model. In some embodiments, one, two, three, four, five, or more genes from Table 1 are used when training the classifier model. In some embodiments, all genes from Table 1 are used when training the classifier model.

[0122] 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 genes from Table 1. In some embodiments, the expression levels of all genes in Table 1 are determined for clustering. In some embodiments, the expression levels of all genes in Table 1 are determined and compared to determine whether lymphoma patients respond to cancer treatment.

[0123] Table 1. List of genes used in the resulting grouped polynomial generalized linear model (GLM)

[0124]

[0125] In some embodiments, the methods provided herein include at least one gene (e.g., one, two, three, four, five or more) selected from the group consisting 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, 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, PSMD1 3. PSME2, ​​PTBP3, PTENP1, PTPRC, R3HDM1, RAB32, RAMP3, RANBP2, RBM3, RBM33, RBP1, RFC1, RNASEL, RNF38, RPL30, RPL32P3, RPRD2, R PS3, RPS4X, RRP9, S100A11, S100Z, S1PR2, SAAL1, SBNO1, SCAMP2, SCARB1, SCARB2, SDCBP, SELPLG, SENP7, SERPINA3, SERPINB1, SF 1. SF3A1, SFPQ, SFXN3, SH3BGRL3, SH3BP1, SLAMF8, SLC35D1, SMARCA4, SMARCC1, SMG5, SNORA21, SNORA71B, SNORD104, SNRPA, SNRPD 1. SNTA1, SNX29, SOBP, SOGA1, SP140, SP3, SPEN, SRGN, SRSF6, STAG3, STK4, STMN1, SUMO2, SYNJ2, SYPL1, SYTL3, TAGLN2, TAP2, TAR DBP, TBCA, TGDS, TGOLN2, THRAP3, TIMM10, TLK1, TLN1, TMBIM4, TMEM223, TNFRSF1A, TONSL, TP53INP1, TPM3, TRA2B, TRAM1, TRAPPC1 2. TRIOBP, TRIP13, TRMT1L, TRNT1, TSPYL2, TSSK4, TUBA1B, TUBA1C, TWIST1, UBE2B, UBE2D2, UBE2G1, UXT, VASH1, VAV1, VDAC1, WEE1 ,

[0126] In some embodiments, the at least one gene is selected from the genes in Table 2. In some embodiments, the at least one gene comprises one, two, three, four, five, or more genes from Table 2. In some embodiments, the at least one gene comprises all the genes in Table 2.

[0127] In some embodiments, the expression levels of at least one gene in the discovery dataset are used when training the classifier model. In some embodiments, one, two, three, four, five, or more genes from Table 2 are used when training the classifier model. In some embodiments, all genes from Table 2 are used when training the classifier model.

[0128] In some embodiments, the classifier model includes one, two, three, four, five, or more genes from Table 2. In some embodiments, the classifier model includes all genes from Table 2. In some embodiments, the expression levels of all genes in Table 2 are determined for clustering. In some embodiments, the expression levels of all genes in Table 2 are determined and compared to determine whether lymphoma patients respond to cancer treatment.

[0129] Table 2: List of genes used in the resulting grouped polynomial generalized linear model (GLM)

[0130]

[0131] In some embodiments, the methods provided herein include at least one gene selected from the group consisting of (e.g., one, two, three, four, five or more) or all genes selected from the group consisting of: ABHD10, ATIC, BUB3, BLNK, CARD11, CD37, CDK12, CHKA, CKAP2, COL1A2, DNAJA3, EIF2B5, FBXO46, FOXP1, IDS, IKZF1, ITGB2, K LHL14, KLHL23, MCM4, MGAT4A, NCBP2, NCL, NFE2L2, ORC6, PLK1, PMM2, PRMT1, RAB32, RPL32P3, RRP9, SMARC C1, SNRPA, SYNJ2, TARDBP, TGDS, TLK1, TNFRSF1A, TRA2B, TRNT1, VASH1, ZNF101, ZNF107, ZNF146, and ZNF480.

[0132] In some embodiments, the at least one gene is selected from the genes in Table 3. In some embodiments, the at least one gene comprises one, two, three, four, five, or more genes from Table 3. In some embodiments, the at least one gene comprises all the genes in Table 3.

[0133] In some embodiments, the expression levels of at least one gene in the discovery dataset are used when training the classifier model. In some embodiments, one, two, three, four, five, or more genes from Table 3 are used when training the classifier model. In some embodiments, all genes from Table 3 are used when training the classifier model.

[0134] In some embodiments, the classifier model includes one, two, three, four, five, or more genes from Table 3. In some embodiments, the classifier model includes all genes from Table 3. In some embodiments, the expression levels of all genes in Table 3 are determined for clustering. In some embodiments, the expression levels of all genes in Table 3 are determined and compared to determine whether lymphoma patients respond to cancer treatment.

[0135] Table 3: List of genes used in the resulting grouped polynomial generalized linear model (GLM)

[0136]

[0137] In some embodiments, the methods provided herein include at least one gene (e.g., one, two, three, four, five or more) selected from the group consisting of the following groups or all genes selected from the group consisting of the following groups: ABHD10, ATIC, BLNK, CARD11, CD37, CDK12, CHKA, CKAP2, COL1A2, DNAJA3, FBXO46, FOXP1, IDS, IKZF1, ITGB2, KLHL23, MCM4, MGAT4A, NCL, NFE2L2, PMM2, PRMT1, RAB32, RRP9, SMARCC1, SNRPA, SYNJ2, TARDBP, TGDS, TLK1, TNFRSF1A, VASH1, ZNF101, ZNF107 and ZNF480.

[0138] In some embodiments, the at least one gene is selected from the genes in Table 4. In some embodiments, the at least one gene comprises one, two, three, four, five, or more genes from Table 4. In some embodiments, the at least one gene comprises all the genes in Table 4.

[0139] In some embodiments, the expression levels of at least one gene in the discovery dataset are used when training the classifier model. In some embodiments, one, two, three, four, five, or more genes from Table 4 are used when training the classifier model. In some embodiments, all genes from Table 4 are used when training the classifier model.

[0140] In some embodiments, the classifier model includes one, two, three, four, five, or more genes from Table 4. In some embodiments, the classifier model includes all genes from Table 4. In some embodiments, the expression levels of all genes in Table 4 are determined for clustering. In some embodiments, the expression levels of all genes in Table 4 are determined and compared to determine whether lymphoma patients are responding to cancer treatment.

[0141] Table 4: List of genes used in the resulting grouped polynomial generalized linear model (GLM)

[0142]

[0143] In some embodiments, the methods provided herein include at least one gene (e.g., one, two, three, four, five or more) selected from the group consisting of the following groups or all genes selected from the group consisting of the following groups: ATIC, BLNK, CARD11, CD37, CDK12, CKAP2, COL1A2, FOXP1, IDS, ITGB2, KLHL23, MCM4, MGAT4A, NCL, NFE2L2, PMM2, PRMT1, RAB32, SYNJ2, TLK1, TNFRSF1A, VASH1, ZNF101, ZNF107 and ZNF480.

[0144] 5.3 Methods and treatments for predicting responsiveness in lymphoma patients

[0145] On the other hand, this document provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) clustering reference lymphoma patients in a reference patient group into subgroups using the expression levels of at least one gene in a reference biological sample of a reference lymphoma patient; (b) determining the subgroup to which the lymphoma patient belongs based on the expression levels of the 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 some embodiments, the method further comprises obtaining a reference biological sample from a reference lymphoma patient in a reference patient group. In some embodiments, the method further comprises obtaining a biological sample from a lymphoma patient.

[0146] In some embodiments, the method further includes administering a second cancer treatment to a lymphoma patient based on a subgroup of the lymphoma patient.

[0147] In some embodiments, the sample is obtained from subject tissue containing DLBCL cells. A more detailed description of the sample (or biological sample) is provided below in Section 5.7.

[0148] In some embodiments, the lymphoma patient is a DLBCL patient. In some embodiments, the DLBCL patient is a newly diagnosed (nd) and 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. In some embodiments, the reference lymphoma patient is a refractory DLBCL patient, a relapsed DLBCL patient, or a newly diagnosed DLBCL patient. In some embodiments, the lymphoma patient is a GCB DLBCL patient or an ABC DLBCL patient. In some embodiments, the reference lymphoma patient is a GCB DLBCL patient or an ABC DLBCL patient.

[0149] In some embodiments, the at least one gene comprises two or more genes. In some embodiments, classifying or clustering a reference patient group includes generating clustering information that defines the relationship between the expression levels of two or more genes in a reference biological sample; and rearranging the heatmap display based on the clustering information.

[0150] In some embodiments, classifying or clustering a reference patient group includes using the clustering methods described in Section 5.2 of this document.

[0151] 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 a sample from a discovery queue. In some embodiments, the replication / validation dataset is a sample from a replication / validation queue. In some embodiments, the discovery 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 discovery dataset is the Discovery-2 dataset. In some embodiments, the Discovery-2 dataset includes a commercial dataset. In some embodiments, the Discovery-2 dataset includes the ROBUST dataset. In some embodiments, the Discovery-2 dataset includes both a commercial dataset and the ROBUST dataset. In some embodiments, the discovery dataset is the ndMER dataset. In some embodiments, the discovery dataset is the REMoDL-B dataset.

[0152] 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 either the ndMER dataset or the REMoDL-B dataset. In some embodiments, the replication / validation dataset is the ndMER dataset. In some embodiments, the replication / validation dataset is the REMoDL-B dataset. In some embodiments, the replication / validation dataset is both the ndMER dataset and the REMoDL-B dataset.

[0153] In some embodiments, the clustering step includes (i) standardizing 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 and removing outliers in the dataset. In some embodiments, the clustering step further includes evaluating the clustering results 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 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 some embodiments, the clustering method is the iClusterPlus clustering method. In some embodiments, the clustering method is a Cluster-Cluster Analysis (COCA) clustering method. In some embodiments, the dataset is standardized using a single-sample normalization (ssNorm) method before analysis.

[0154] In some embodiments, normalization is performed using DLBCL-specific housekeeping genes. In some embodiments, normalization is performed using the ISY1, R3HDM1, TRIM56, UBXN4, and WDR55 genes.

[0155] In some embodiments, one, two, three, four, or more features are selected as clustering features and input into the clustering method. In some embodiments, a subset of gene expression data is selected as 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 the most expressed genes. In some embodiments, the subset of gene expression data is the gene expression data of the top 25% of the most expressed genes.

[0156] In some embodiments, the GSVA score of the marker is selected as the clustering feature. In some embodiments, the GSVA score of the C1 position is selected as the clustering feature. In some embodiments, the GSVA score of the cell type LM23 is selected as the clustering feature. In some embodiments, a matrix is ​​formed by selecting a subset of gene expression data, the GSVA score of the marker, the GSVA score of the C1 position, and / or the GSVA score of the cell type LM23 as the clustering features.

[0157] In some embodiments, the clustering method independently clusters each feature matrix and aggregates the cluster features. In some embodiments, the clustering method performs clustering, wherein cluster features from different feature matrices are aggregated, and all cluster features are clustered.

[0158] 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.

[0159] In some embodiments, the clustering method is the iClusterPlus clustering method, with a cluster size of 7.

[0160] In some embodiments, the clustering method is the iClusterPlus clustering method, and a subset of gene expression data, GSVA score of the marker, GSVA score of C1 position, and / or LM23 GSVA score of cell type are selected as a matrix of clustering features, and the number of clusters is 7 (clusters A1-A7).

[0161] The terms “cluster” and “subgroup” are used interchangeably throughout this disclosure.

[0162] In some embodiments, reference patients in the reference patient group are clustered into 2-20 subgroups (K = 2-20). In some embodiments, reference patients in the reference patient group are clustered into 2-15 subgroups (K = 2-15). In some embodiments, these reference lymphoma patients are clustered into 2-12 subgroups (K = 2-12). In some embodiments, reference patients in the reference patient group are clustered into 2-10 subgroups (K = 2-10). In some embodiments, reference patients in 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, reference patients in the reference patient group are clustered into 8 groups (K = 8). In some embodiments, these reference lymphoma patients are clustered into 7 groups (K = 7, clusters A1-A7).

[0163] In some embodiments, the clustering step further includes evaluating the clustering results of the clustering method. In some embodiments, the clustering results are evaluated for each number of clusters (K). In some embodiments, the clustering results are evaluated using the nbClust R package. In some embodiments, the clustering results are evaluated using an index selected from a group consisting of contour statistics, gap statistics, and the percentage of variance explained by the clustering. In some embodiments, the clustering results are evaluated by a minimum cluster size. To ensure sufficient sample size in each cluster to build a downstream classifier model for robust identification of each cluster or subgroup, the clusters obtained from each test clustering activity need to have a minimum cluster size.

[0164] In some embodiments, for at least one subgroup, there are patients who provide low-confidence-level clustering data for a particular subgroup based on the selected clustering classifier. In some embodiments, patients who provide low-confidence-level clustering data are filtered. In some embodiments, patients who provide low-confidence-level clustering data are excluded from the subgroup. In some embodiments, patients who provide low-confidence-level clustering data are excluded from subgroup A7.

[0165] In some embodiments, the method further includes setting a threshold confidence level for at least one of these subgroups to exclude patients who give lower confidence levels from at least one of these subgroups. In some embodiments, patients who give low-confidence-level clustering data are excluded from subgroup A7.

[0166] In some embodiments, the method for predicting the responsiveness of lymphoma patients to cancer treatment further includes (i) training a classifier model using clustering results from a discovery dataset to identify at least one clustering classifier, (ii) applying the at least one clustering classifier to a replication / validation dataset to classify the replication / validation dataset, (iii) clustering the replication / validation dataset using a clustering method, and (iv) comparing the classification result of the replication / validation dataset using the at least one clustering classifier with the clustering result 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 the least absolute value shrinkage and selection operator (LASSO). In some embodiments, if the classification result of the replication / validation dataset using the at least one clustering classifier is similar to the clustering result of the replication / validation dataset using the clustering method, it indicates that the clustering classifier effectively classifies the replication / validation dataset.

[0167] In some embodiments, the method for predicting the responsiveness of lymphoma patients to cancer treatment further includes (i) training a classifier model using clustering results from a discovery dataset, identifying at least one clustering classifier, and (ii) applying the at least one clustering classifier to a 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 the least absolute value shrinkage and selection operator (LASSO).

[0168] In some embodiments, the at least one gene is selected from the genes in Table 1. In some embodiments, the at least one gene comprises one, two, three, four, five, or more genes from Table 1. In some embodiments, the at least one gene comprises all the genes in Table 1.

[0169] In some embodiments, the expression levels of at least one gene in the discovery dataset are used when training the classifier model. In some embodiments, one, two, three, four, five, or more genes from Table 1 are used when training the classifier model. In some embodiments, all genes from Table 1 are used when training the classifier model.

[0170] In some embodiments, the classifier model contains one, two, three, four, five or more genes from Table 1. In some embodiments, the classifier model contains all the genes from Table 1.

[0171] In some embodiments, the at least one gene is selected from the genes in Table 2. In some embodiments, the at least one gene comprises one, two, three, four, five, or more genes from Table 2. In some embodiments, the at least one gene comprises all the genes in Table 2.

[0172] In some embodiments, the expression levels of at least one gene in the discovery dataset are used when training the classifier model. In some embodiments, one, two, three, four, five, or more genes from Table 2 are used when training the classifier model. In some embodiments, all genes from Table 2 are used when training the classifier model.

[0173] In some embodiments, the classifier model includes one, two, three, four, five, or more genes from Table 2. In some embodiments, the classifier model includes all genes from Table 2.

[0174] In some embodiments, the at least one gene is selected from the genes in Table 3. In some embodiments, the at least one gene comprises one, two, three, four, five, or more genes from Table 3. In some embodiments, the at least one gene comprises all the genes in Table 3.

[0175] In some embodiments, the expression levels of at least one gene in the discovery dataset are used when training the classifier model. In some embodiments, one, two, three, four, five, or more genes from Table 3 are used when training the classifier model. In some embodiments, all genes from Table 3 are used when training the classifier model.

[0176] In some embodiments, the classifier model includes one, two, three, four, five, or more genes from Table 3. In some embodiments, the classifier model includes all genes from Table 3.

[0177] In some embodiments, the at least one gene is selected from the genes in Table 4. In some embodiments, the at least one gene comprises one, two, three, four, five, or more genes from Table 4. In some embodiments, the at least one gene comprises all the genes in Table 4.

[0178] In some embodiments, the expression levels of at least one gene in the discovery dataset are used when training the classifier model. In some embodiments, one, two, three, four, five, or more genes from Table 4 are used when training the classifier model. In some embodiments, all genes from Table 4 are used when training the classifier model.

[0179] In some embodiments, the classifier model includes one, two, three, four, five, or more genes from Table 4. In some embodiments, the classifier model includes all genes from Table 4.

[0180] In some embodiments, the determining step applies a clustering method to determine which subgroup the lymphoma patient belongs to based on the expression level of at least one gene in a biological sample from a lymphoma patient.

[0181] In some embodiments, the prediction step applies a trained classifier model to predict the responsiveness of a lymphoma patient to a first cancer treatment. In some embodiments, the prediction step applies a trained GLM model to predict the responsiveness of a lymphoma patient to a first cancer treatment.

[0182] In some embodiments, the lymphoma is diffuse large B-cell lymphoma (DLBCL). In some embodiments, the lymphoma is indolent B-cell lymphoma. In some embodiments, the lymphoma is selected from the group consisting of: DLBCL, indolent B-cell lymphoma, follicular lymphoma, small lymphocytic lymphoma, 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 marginal zone B-cell lymphoma. In some embodiments, the lymphoma is mantle cell lymphoma. In some embodiments, the lymphoma is chronic lymphocytic leukemia.

[0183] In some embodiments, the lymphoma is DLBCL. In some embodiments, the lymphoma is indolent B-cell lymphoma, marginal zone B-cell lymphoma, mantle cell lymphoma, or chronic lymphocytic leukemia.

[0184] In some embodiments, the first cancer treatment is a combination therapy using rituximab (Rituxan), cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP).

[0185] In some embodiments, the second or alternative cancer treatment is R-CHOP. In some embodiments, the second or alternative cancer treatment is not R-CHOP. In some embodiments, the second or alternative cancer treatment is a combination of lenalidomide, rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R2-CHOP).

[0186] In some embodiments, the second or alternative cancer treatment is a bromodomain and superterminal (BET) inhibitor or a cyclin-dependent kinase (CDK) inhibitor. A bromodomain (BD) is a protein module containing approximately 110 amino acids that recognizes acetylated lysine residues in histones and other proteins. The BET family is a subset of 46 bromodomain-containing proteins found only in the human genome. BET proteins consist of four proteins: bromodomain-containing protein 2 (BRD2), BRD3, BRD4, and bromodomain-containing 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 demonstrating potent preclinical activity have driven the development of BET inhibitors, primarily for oncology indications, including, for example, DLBCL. As chemical probes described for other BD targets, clinical BET inhibitors are acetyllysine analogs with a heterocyclic core occupying BD pocket 15. Dysregulation of BET proteins, especially BRD4, is associated with the development of various diseases, particularly cancer. See Duan et al., MedChemComm [Medicinal Chemistry Communications], 2018, 9:1779-1802.

[0187] 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.

[0188] In some embodiments, the BET inhibitor is a bromine-containing domain 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, BI 894999, BMS-986158, and AZD5153.

[0189] Cyclin-dependent kinases (CDKs) belong to the serine-threonine kinase family and have been identified as gene products involved in cell cycle control. Close cooperation among CDKs, cyclins, and cells containing endogenous inhibitors (CKIs) is essential for regulated, orderly cell cycle progression. Mammalian CDKs, cyclins, and CKIs play important roles in other biological processes, including, for example, transcriptional regulation, experimental embryology, DNA damage response and repair (DDR), stemness, metabolism, and angiogenesis. See Sanchez-Martinez et al., Bioorg. Med. Chem. Lett. [Bioorganic and Medicinal Chemistry Letters], 2019, 29:126637.

[0190] In some embodiments, the CDK inhibitors are selected from the group consisting of: PD-0332991 (Ibrance® or palbociclib), LEE011 (ribociclib), LY2835219 (Verzenio® or abecilib), 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-202 or Seliciclib), AT-7519, AGM-130 (Inditinib), FN-1501, SY-1365, AZD-4573, TP- 1287, P-1446A-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).

[0191] 5.3.1. Dual-Hit Feature (DHITsig) Classifier

[0192] A subset of patients at high risk of DLBCL exists, characterized by so-called "double-hit" (DHIT) translocations in MYC and BCL2 or BCL6. DHIT+ patients tend to have worse survival than DHIT- patients; these DHIT+ patients are typically GCB patients with MYC+BCL2 translocations. DHIT status is usually determined using DNA sequencing or FISH probes to ascertain the translocation status.

[0193] Ennishi et al. derived gene expression profiles reflecting DHIT status, encompassing a broader population with varying clinical outcomes. See Ennishi et al., J Clin Oncol. [Journal of Clinical Oncology], 2018, 37:190-201. They adapted the DHITsig method using 104 genes, parameterizations, and methods described in the manuscript and applied it to a single-sample normalized RNAseq space.

[0194] The Ennishi method used to compute the DHITsig score is a variable importance-weighted sum of log-likelihood ratios. The likelihood function is computed by assuming a Gaussian mixture model of gene expression, which defines two normal distributions of expression for DHITsig+ and DHITsig- samples of each characteristic gene. The gene list and variable weights from Ennishi et al., along with the mixture model distribution parameters derived from their DESeq2-processed dataset, were used.

[0195] 5.3.2. Identify the biological characteristics of patient clusters and administer treatment.

[0196] In some embodiments, these reference lymphoma patients were clustered into seven subgroups (A1-A7); and wherein: (i) subgroup A1 comprised approximately 50% to 60% of germinal center B-cell-like (GCB) DLBCL patients, approximately 30% to 40% of activated B-cell-like (ABC) DLBCL patients, approximately 10% to 20% of TME+ DLBCL patients, and approximately 30% to 40% of DHITsig+ DLBCL patients; (ii) subgroup A2 comprised approximately 80% to 90% of GCB DLBCL patients, approximately 0% to 5% of ABC DLBCL patients, approximately 15% to 25% of TME+ DLBCL patients, and approximately 25% to 35% of DHITsig+ DLBCL patients; (iii) subgroup A3 comprised approximately 40% to 55% of GCB DLBCL patients, approximately 30% to 45% of ABC DLBCL patients, and approximately 40% to 50% of TME+ DLBCL patients. (iv) Subgroup A4 includes approximately 25% to 35% of GCB DLBCL patients, approximately 40% to 50% of ABC DLBCL patients, approximately 30% to 40% of TME + DLBCL patients, and approximately 10% to 20% of DHITsig+DLBCL patients; (v) Subgroup A5 includes approximately 20% to 40% of GCB DLBCL patients, approximately 45% to 65% of ABC DLBCL patients, approximately 30% to 40% of TME + DLBCL patients, and approximately 0% to 10% of DHITsig+DLBCL patients; (vi) Subgroup A6 includes approximately 30% to 40% of GCB DLBCL patients, approximately 40% to 50% of ABC DLBCL patients, and approximately 20% to 30% of TME + DLBCL patients. Patients with DLBCL, approximately 75% to 95% of patients with TME+DLBCL and approximately 0% to 10% of patients with DHITsig+DLBCL; and (vii) subgroup A7 includes approximately 0% to 10% of patients with GCB DLBCL, approximately 80% to 90% of patients with ABC DLBCL, approximately 0% to 10% of patients with TME+DLBCL and approximately 5% to 15% of patients with DHITsig+DLBCL.

[0197] In some embodiments, when the lymphoma patient is determined to belong to subgroup A1, the method includes predicting that the patient is unlikely to respond to a first cancer treatment. In some embodiments, when the lymphoma patient is determined to belong to subgroup A2, the method includes predicting that the patient is unlikely to respond to a first cancer treatment. In some embodiments, when the lymphoma patient is determined to belong to subgroup A3, the method includes predicting that the patient is unlikely to respond to a first cancer treatment. In some embodiments, when the lymphoma patient is determined to belong to subgroup A4, the method includes predicting that the patient is unlikely to respond to a first cancer treatment. In some embodiments, when the lymphoma patient is determined to belong to subgroup A5, the method includes predicting that the patient is unlikely to respond to a first cancer treatment. In some embodiments, when the lymphoma patient is determined to belong to subgroup A6, the method includes predicting that the patient is unlikely to respond to a first cancer treatment. In some embodiments, when the lymphoma patient is determined to belong to subgroup A7, the method includes predicting that the patient is unlikely to respond to a first cancer treatment.

[0198] In some embodiments, when the lymphoma patient is determined to be an activated B-cell-like (ABC) lymphoma patient in subgroup A7, the method includes predicting that the patient is unlikely to respond to a first cancer treatment.

[0199] In some embodiments, when it is determined that the lymphoma patient belongs to any patient subgroup that is predicted to be unlikely to respond to the first cancer treatment, the lymphoma patient is treated with a second cancer treatment.

[0200] In some embodiments, the first cancer treatment is a combination of rituximab (Rituxan), cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP). In some embodiments, the second cancer treatment is not R-CHOP. In some embodiments, the second cancer treatment is a combination of lenalidomide, rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R2-CHOP). In some embodiments, the second cancer treatment is a BET inhibitor or a CDK inhibitor. In some embodiments, the second cancer treatment is a CDK inhibitor. In some embodiments, the second cancer treatment is a BET inhibitor.

[0201] In some embodiments, when the lymphoma patient is determined to belong to subgroup A7, the lymphoma patient is treated with a second cancer therapy containing R2-CHOP. In some embodiments, when the lymphoma patient is determined to belong to subgroup A7, the lymphoma patient is treated with a second cancer therapy containing a BET inhibitor or a CDK inhibitor. In some embodiments, when the lymphoma patient is determined to belong to subgroup A7, the lymphoma patient is treated with a second cancer therapy containing a CDK inhibitor. In some embodiments, when the lymphoma patient is determined to belong to subgroup A7, the lymphoma patient is treated with a second cancer therapy containing a BET inhibitor.

[0202] In some embodiments, when the lymphoma patient is determined to belong to any patient subgroup predicted to be unlikely to respond to the first cancer treatment, the lymphoma patient is treated with R2-CHOP. In some embodiments, when the lymphoma patient is determined to belong to any patient subgroup predicted to be unlikely to respond to the first cancer treatment, the lymphoma patient is treated with a BCL2 inhibitor. In some embodiments, when the lymphoma patient is determined to belong to any patient subgroup 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, when the lymphoma patient is determined to belong to any patient subgroup predicted to be unlikely to respond to the first cancer treatment, the lymphoma patient is treated with a human leukocyte antigen (HLA) gene inhibitor. In some embodiments, the second cancer treatment is an HLA-A inhibitor. In some embodiments, the second cancer treatment is an HLA-B inhibitor. In some embodiments, the second cancer treatment is an HLA-C inhibitor. In some embodiments, the second cancer treatment is an HLA-E inhibitor. In some embodiments, the second cancer treatment is an HLA-F inhibitor. In some embodiments, when a lymphoma patient is determined to belong to any patient subgroup predicted to be unlikely to respond to a first cancer treatment, the lymphoma patient receives CD47 treatment. In some embodiments, when a lymphoma patient is determined to belong to any patient subgroup predicted to be unlikely to respond to a 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, when a lymphoma patient is determined to belong to any patient subgroup predicted to be unlikely to respond to a first cancer treatment, the lymphoma patient is treated with a histone deacetylase (HDAC) inhibitor. In some embodiments, when a lymphoma patient is determined to belong to any patient subgroup predicted to be unlikely to respond to a first cancer treatment, the lymphoma patient is treated with a galactagogue-3 (Gal3) inhibitor.

[0203] In some embodiments, the lymphoma patient belongs to a subgroup based on mutation data of at least one feature listed in Figure 3C or Figure 3D.

[0204] On the other hand, this paper provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) obtaining a first biological sample from a first patient with lymphoma; (b) determining the expression levels of one, two, three, four, five or more genes in Table 1; and (c) comparing the expression levels of one, two, three, four, five or more genes in the first biological sample with the expression levels of the same genes in one or more second biological samples from one or more second patients, wherein one or more second lymphoma patients are responsive to cancer treatment, and wherein the expression of one, two, three, four, five or more genes in the first biological sample is similar to the expression levels of one, two, three, four, five or more genes in one or more second biological samples, indicating that the lymphoma in the first patient will be responsive to treatment with that cancer treatment.

[0205] On the other hand, this paper provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) obtaining a first biological sample from a first lymphoma patient; (b) determining the expression of genes or a subset of genes or any combination thereof shown in Table 1 in the first biological sample; and (c) comparing the gene expression profile of the genes or subset of genes in the first biological sample with (i) the gene expression profile of genes or subset of genes from a biological sample from a lymphoma patient who has responded to the drug and (ii) the gene expression profile of genes or subset of genes from a biological sample from a lymphoma patient who has not responded to the cancer treatment, wherein: the gene expression profile of the genes or subset of genes in the first biological sample is similar to the gene expression profile of the genes or subset of genes from a biological sample from a lymphoma patient who has responded to the cancer treatment, indicating that the first lymphoma patient will respond to the cancer treatment; and the gene expression profile of the genes or subset of genes in the first biological sample is similar to the gene expression profile of the genes or subset of genes from a biological sample from a lymphoma patient who has not responded to the cancer treatment, indicating that the first lymphoma patient will not respond to the cancer treatment.

[0206] On the other hand, this document provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method 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 the at least one gene of step (a) with the expression level of the at least one gene in a reference biological sample from a reference lymphoma patient, wherein the reference lymphoma patient is responsive to the cancer treatment, and wherein 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, it indicates that the lymphoma patient may be responsive to the cancer treatment. In some embodiments, the method further comprises obtaining a biological sample from a lymphoma patient. In some embodiments, the at least one gene comprises five or more genes of Table 1.

[0207] On the other hand, this document provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method 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 the at least one gene of step (a) with the expression level of at least one gene in a reference biological sample from a group of reference lymphoma patients who respond to cancer treatment, and 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 the reference biological sample, it indicates that the lymphoma patient is unlikely to respond to the cancer treatment. In some embodiments, the method further comprises obtaining a biological sample from a lymphoma patient. In some embodiments, the at least one gene comprises five or more genes of Table 1. In some embodiments, the expression level of the at least one gene in the reference biological sample is the mean or median of gene expression levels measured in the reference biological samples from these reference lymphoma patients.

[0208] On the other hand, this document provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) determining the expression level of at least one gene listed 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 at least one gene in a biological sample from a lymphoma patient who has responded to cancer treatment and (ii) the expression level of at least one gene in a biological sample from a lymphoma patient who has not responded to cancer treatment, wherein if the expression level of (a) is similar to the expression level of (i), it indicates that the 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 lymphoma patient is unlikely to respond to the cancer treatment. In some embodiments, the method further comprises obtaining a biological sample from a lymphoma patient. In some embodiments, the at least one gene comprises five or more genes listed in Table 1. In some embodiments, the expression level of at least one gene in a biological sample of a lymphoma patient who is unresponsive or responsive to cancer treatment is the average or median expression level of at least one gene measured in a biological sample of a lymphoma patient who is unresponsive or responsive to cancer treatment.

[0209] In some embodiments, the determining step of the method described herein includes determining the expression of all genes listed in Table 1. In some embodiments, the expression levels of all genes listed in Table 1 are determined and compared. In some embodiments, the determining step of the method described herein includes determining the expression of at least one gene (e.g., one, two, three, four, five or more) selected from the group consisting 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, EI F3I, EIF5A, ELF1, ELMO1, EMP3, ENO1, EPS15, ERGIC2, ERH, ESDEE'S, 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、PSMD1 3、PSME2、PTBP3、PTENP1、PTPRC、R3HDM1、RAB32、RAMP3、RANBP2、R BM3、RBM33、RBP1、RFC1、RNASEL、RNF38、RPL30、RPL32P3、RPRD2、R PS3, RPS4 1、SF3A1、SFPQ、SFXN3、SH3BGRL3、SH3BP1、SLAMF8、SLC35D1、SMARCA4、SMARCC1、SMG5、SNORA21、SNORA71B、SNORD104、SNRPA、SNRPD 1、SNTA1、SNX29、SOBP、SOGA1、SP140、SP3、SPEN、SRGN、SRSF6、ST AG3、STK4、STMN1、SUMO2、SYNJ2、SYPL1、SYTL3、TAGLN2、TAP2、TAR DBP、TBCA、TGDS、TGOLN2、THRAP3、TIMM10、TLK1、TLN1、TMBIM4、TMEM223、TNFRSF1A、TONSL、TP53INP1、TPM3、TRA2B、TRAM1、TRAPPC1 2、TRIOBP、TRIP13、TRMT1L、TRNT1、TSPYL2、TSSK4、TUBA1B、TUBA1C、TWIST1、UBE2B、UBE2D2、UBE2G1、UXT、VASH1、VAV1、VDAC1、WEE1 、XRCC6、YTHDC1、YWHAE、ZBED5、ZBTB37、ZFAND4、ZFAND5、ZMAT1、Z NF101、ZNF107、ZNF146、ZNF207、ZNF318、ZNF367、ZNF480、ZWINT。

[0210] On the other hand, this paper provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) obtaining a first biological sample from a first patient with lymphoma; (b) determining the expression levels of one, two, three, four, five or more genes in Table 2; and (c) comparing the expression levels of one, two, three, four, five or more genes in the first biological sample with the expression levels of the same genes in one or more second biological samples from one or more second patients, wherein one or more second lymphoma patients are responsive to cancer treatment, and wherein the similarity of the expression levels of one, two, three, four, five or more genes in the first biological sample to the expression levels of one, two, three, four, five or more genes in one or more second biological samples suggests that the lymphoma in the first patient may be responsive to treatment with that cancer treatment.

[0211] On the other hand, this paper provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) obtaining a first biological sample from a first lymphoma patient; (b) determining the expression of genes or a subset of genes or any combination thereof shown in Table 2 in the first biological sample; and (c) comparing the gene expression profile of the genes or subset of genes in the first biological sample with (i) the gene expression profile of genes or subset of genes from a biological sample from a lymphoma patient who has responded to the drug and (ii) the gene expression profile of genes or subset of genes from a biological sample from a lymphoma patient who has not responded to the cancer treatment, wherein: the gene expression profile of the genes or subset of genes in the first biological sample is similar to the gene expression profile of the genes or subset of genes from a biological sample from a lymphoma patient who has responded to the cancer treatment, indicating that the first lymphoma patient is likely to respond to the cancer treatment; and the gene expression profile of the genes or subset of genes in the first biological sample is similar to the gene expression profile of the genes or subset of genes from a biological sample from a lymphoma patient who has not responded to the cancer treatment, indicating that the first lymphoma patient is unlikely to respond to the cancer treatment.

[0212] On the other hand, this document provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) determining the expression level of at least one gene of Table 2 in a biological sample from a lymphoma patient; and (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 from a reference lymphoma patient, wherein the reference lymphoma patient is responsive to the cancer treatment, and wherein 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, it indicates that the lymphoma patient may be responsive to the cancer treatment. In some embodiments, the method further comprises obtaining a biological sample from a lymphoma patient. In some embodiments, the at least one gene comprises five or more genes of Table 2. In some embodiments, the at least one gene comprises all genes of Table 2.

[0213] On the other hand, this document provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) determining the expression level of at least one gene of Table 2 in a biological sample from a lymphoma patient; and (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 from a group of reference lymphoma patients who respond to cancer treatment, and 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 the reference biological sample, it indicates that the lymphoma patient may respond to the cancer treatment. In some embodiments, the method further comprises obtaining a biological sample from a lymphoma patient. In some embodiments, the at least one gene comprises five or more genes of Table 2. In some embodiments, the at least one gene comprises all genes of Table 2. In some embodiments, the expression level of the at least one gene in the reference biological sample is the mean or median of gene expression levels measured in the reference biological samples of these reference lymphoma patients.

[0214] On the other hand, this document provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) determining the expression level of at least one gene of Table 2 in a biological sample of 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 at least one gene in a biological sample of a lymphoma patient who has responded to cancer treatment and (ii) the expression level of at least one gene in a biological sample of a lymphoma patient who has not responded to cancer treatment, wherein if the expression level of (a) is similar to the expression level of (i), it indicates that the 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 lymphoma patient is unlikely to respond to the cancer treatment. In some embodiments, the method further comprises obtaining a biological sample from a lymphoma patient. In some embodiments, the at least one gene comprises five or more genes of Table 2. In some embodiments, the at least one gene comprises all genes of Table 2. In some embodiments, the expression level of at least one gene in a biological sample of a lymphoma patient who is unresponsive or responsive to cancer treatment is the average or median expression level of at least one gene measured in a biological sample of a lymphoma patient who is unresponsive or responsive to cancer treatment.

[0215] In some embodiments, the determining step of the method described herein includes determining the expression levels of all genes listed in Table 2. In some embodiments, the expression levels of all genes listed in Table 2 are determined and compared. In some embodiments, the determining step of the method described herein includes determining the expression levels of at least one gene (e.g., one, two, three, four, five or more) selected from the group consisting of: ABHD10, ATIC, BUB3, BLNK, CARD11, CD37, CDK12, CHKA, CKAP2, COL1A2, DNAJA3, EIF2B5, FBXO46, FOXP1, IDS, IKZF1, I TGB2, KLHL14, KLHL23, MCM4, MGAT4A, NCBP2, NCL, NFE2L2, ORC6, PLK1, PMM2, PRMT1, RAB32, RPL32P3, RRP9, SM ARCC1, SNRPA, SYNJ2, TARDBP, TGDS, TLK1, TNFRSF1A, TRA2B, TRNT1, VASH1, ZNF101, ZNF107, ZNF146, and ZNF480.

[0216] On the other hand, this paper provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) obtaining a first biological sample from a first patient with lymphoma; (b) determining the expression levels of one, two, three, four, five or more genes in Table 3; and (c) comparing the expression levels of one, two, three, four, five or more genes in the first biological sample with the expression levels of the same genes in one or more second biological samples from one or more second patients, wherein one or more second lymphoma patients are responsive to cancer treatment, and wherein the similarity of the expression levels of one, two, three, four, five or more genes in the first biological sample to the expression levels of one, two, three, four, five or more genes in one or more second biological samples suggests that the lymphoma in the first patient may be responsive to treatment with the cancer treatment.

[0217] On the other hand, this paper provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) obtaining a first biological sample from a first lymphoma patient; (b) determining the expression of genes or a subset of genes or any combination thereof shown in Table 3 in the first biological sample; and (c) comparing the gene expression profile of the genes or subset of genes in the first biological sample with (i) the gene expression profile of genes or subset of genes from a biological sample from a lymphoma patient who has responded to the drug and (ii) the gene expression profile of genes or subset of genes from a biological sample from a lymphoma patient who has not responded to the cancer treatment, wherein: the gene expression profile of the genes or subset of genes in the first biological sample is similar to the gene expression profile of the genes or subset of genes from a biological sample from a lymphoma patient who has responded to the cancer treatment, indicating that the first lymphoma patient is likely to respond to the cancer treatment; and the gene expression profile of the genes or subset of genes in the first biological sample is similar to the gene expression profile of the genes or subset of genes from a biological sample from a lymphoma patient who has not responded to the cancer treatment, indicating that the first lymphoma patient is unlikely to respond to the cancer treatment.

[0218] On the other hand, this document provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) determining the expression level of at least one gene of Table 3 in a biological sample from a lymphoma patient; and (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 from a reference lymphoma patient, wherein the reference lymphoma patient is responsive to the cancer treatment, and wherein 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, it indicates that the lymphoma patient may be responsive to the cancer treatment. In some embodiments, the method further comprises obtaining a biological sample from a lymphoma patient. In some embodiments, the at least one gene comprises five or more genes of Table 3. In some embodiments, the at least one gene comprises all genes of Table 3.

[0219] On the other hand, this document provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) determining the expression level of at least one gene of Table 3 in a biological sample from a lymphoma patient; and (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 from a group of reference lymphoma patients who respond to cancer treatment, and 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 the reference biological sample, it indicates that the lymphoma patient may respond to the cancer treatment. In some embodiments, the method further comprises obtaining a biological sample from a lymphoma patient. In some embodiments, the at least one gene comprises five or more genes of Table 3. In some embodiments, the at least one gene comprises all genes of Table 3. In some embodiments, the expression level of the at least one gene in the reference biological sample is the mean or median of gene expression levels measured in the reference biological samples of these reference lymphoma patients.

[0220] On the other hand, this document provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) determining the expression level of at least one gene of Table 3 in a biological sample of 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 at least one gene in a biological sample of a lymphoma patient who has responded to cancer treatment and (ii) the expression level of at least one gene in a biological sample of a lymphoma patient who has not responded to cancer treatment, wherein if the expression level of (a) is similar to the expression level of (i), it indicates that the 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 lymphoma patient is unlikely to respond to the cancer treatment. In some embodiments, the method further comprises obtaining a biological sample from a lymphoma patient. In some embodiments, the at least one gene comprises five or more genes of Table 3. In some embodiments, the at least one gene comprises all genes of Table 3. In some embodiments, the expression level of at least one gene in a biological sample of a lymphoma patient who is unresponsive or responsive to cancer treatment is the average or median expression level of at least one gene measured in a biological sample of a lymphoma patient who is unresponsive or responsive to cancer treatment.

[0221] In some embodiments, the determining step of the method described herein includes determining the expression of all genes listed in Table 3. In some embodiments, the expression levels of all genes listed in Table 3 are determined and compared. In some embodiments, the determining step of the method described herein includes determining the expression levels of at least one gene (e.g., one, two, three, four, five or more) or all genes selected from the group consisting of: ABHD10, ATIC, BLNK, CARD11, CD37, CDK12, CHKA, CKAP2, COL1A2, DNAJA3, FBXO46, FOXP1, IDS, IKZF1, ITGB2, KLHL23, MCM4, MGAT4A, NCL, NFE2L2, PMM2, PRMT1, RAB32, RRP9, SMARCC1, SNRPA, SYNJ2, TARDBP, TGDS, TLK1, TNFRSF1A, VASH1, ZNF101, ZNF107, and ZNF480.

[0222] On the other hand, this paper provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) obtaining a first biological sample from a first patient with lymphoma; (b) determining the expression levels of one, two, three, four, five or more genes in Table 4; and (c) comparing the expression levels of one, two, three, four, five or more genes in the first biological sample with the expression levels of the same genes in one or more second biological samples from one or more second patients, wherein one or more second lymphoma patients are responsive to cancer treatment, and wherein the similarity of the expression levels of one, two, three, four, five or more genes in the first biological sample to the expression levels of one, two, three, four, five or more genes in one or more second biological samples suggests that the lymphoma in the first patient may be responsive to treatment with that cancer treatment.

[0223] On the other hand, this paper provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) obtaining a first biological sample from a first lymphoma patient; (b) determining the expression of genes or a subset of genes or any combination thereof shown in Table 4 in the first biological sample; and (c) comparing the gene expression profile of the genes or subset of genes in the first biological sample with (i) the gene expression profile of genes or subset of genes from a biological sample from a lymphoma patient who has responded to the drug and (ii) the gene expression profile of genes or subset of genes from a biological sample from a lymphoma patient who has not responded to the cancer treatment, wherein: the gene expression profile of the genes or subset of genes in the first biological sample is similar to the gene expression profile of the genes or subset of genes from a biological sample from a lymphoma patient who has responded to the cancer treatment, indicating that the first lymphoma patient is likely to respond to the cancer treatment; and the gene expression profile of the genes or subset of genes in the first biological sample is similar to the gene expression profile of the genes or subset of genes from a biological sample from a lymphoma patient who has not responded to the cancer treatment, indicating that the first lymphoma patient is unlikely to respond to the cancer treatment.

[0224] On the other hand, this document provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) determining the expression level of at least one gene of Table 4 in a biological sample from a lymphoma patient; and (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 from a reference lymphoma patient, wherein the reference lymphoma patient is responsive to the cancer treatment, and wherein 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, it indicates that the lymphoma patient may be responsive to the cancer treatment. In some embodiments, the method further comprises obtaining a biological sample from a lymphoma patient. In some embodiments, the at least one gene comprises five or more genes of Table 4. In some embodiments, the at least one gene comprises all genes of Table 4.

[0225] On the other hand, this document provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) determining the expression level of at least one gene of Table 4 in a biological sample from a lymphoma patient; and (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 from a group of reference lymphoma patients who respond to cancer treatment, and 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 the reference biological sample, it indicates that the lymphoma patient may respond to the cancer treatment. In some embodiments, the method further comprises obtaining a biological sample from a lymphoma patient. In some embodiments, the at least one gene comprises five or more genes of Table 4. In some embodiments, the at least one gene comprises all genes of Table 4. In some embodiments, the expression level of the at least one gene in the reference biological sample is the mean or median of gene expression levels measured in the reference biological samples of these reference lymphoma patients.

[0226] On the other hand, this document provides a method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) determining the expression level of at least one gene of Table 4 in a biological sample of 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 at least one gene in a biological sample of a lymphoma patient who has responded to cancer treatment and (ii) the expression level of at least one gene in a biological sample of a lymphoma patient who has not responded to cancer treatment, wherein if the expression level of (a) is similar to the expression level of (i), it indicates that the 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 lymphoma patient is unlikely to respond to the cancer treatment. In some embodiments, the method further comprises obtaining a biological sample from a lymphoma patient. In some embodiments, the at least one gene comprises five or more genes of Table 4. In some embodiments, the at least one gene comprises all genes of Table 4. In some embodiments, the expression level of at least one gene in a biological sample of a lymphoma patient who is unresponsive or responsive to cancer treatment is the average or median expression level of at least one gene measured in a biological sample of a lymphoma patient who is unresponsive or responsive to cancer treatment.

[0227] In some embodiments, the determining step of the method described herein includes determining the expression of all genes listed in Table 4. In some embodiments, the expression levels of all genes listed in Table 4 are determined and compared. In some embodiments, the determining step of the method described herein includes determining the expression levels of at least one gene (e.g., one, two, three, four, five or more) or all genes selected from the group consisting of: ATIC, BLNK, CARD11, CD37, CDK12, CKAP2, COL1A2, FOXP1, IDS, ITGB2, KLHL23, MCM4, MGAT4A, NCL, NFE2L2, PMM2, PRMT1, RAB32, SYNJ2, TLK1, TNFRSF1A, VASH1, ZNF101, ZNF107, and ZNF480.

[0228] In some embodiments, determining the expression level of at least one gene includes detecting the mRNA level of at least one gene. In some embodiments, the mRNA level of at least one gene is detected by RNAseq. In some embodiments, the mRNA level of at least one gene is detected by a microarray.

[0229] In some embodiments, determining the expression level of at least one gene includes detecting the presence or amount of at least one complex in a biological sample. In some embodiments, the presence or amount of the at least one complex indicates the expression level of at least one gene. In some embodiments, the at least one complex is a hybridization complex or is detectably labeled.

[0230] In some embodiments, determining the expression level of at least one gene includes detecting the presence or amount of at least one reaction product in a biological sample. In some embodiments, the presence or amount of the at least one reaction product indicates the expression level of at least one gene. In some embodiments, the at least one reaction product is detectably labeled.

[0231] In some embodiments, similar expression levels of a gene means that one expression level of the gene varies within a maximum of 10% (e.g., 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, or 10%) of the other expression level of the same gene. In some embodiments, the two expression levels are measured on an absolute scale. In some embodiments, one expression level is the expression level of a gene from a biological sample from a lymphoma patient, and the other expression level is the expression level of the same gene from a reference biological sample from a reference lymphoma patient. In some embodiments, one expression level is the expression level of a gene from a biological sample from a first lymphoma patient, and the other expression level is the expression level of the same gene from a biological sample from a second lymphoma patient.

[0232] In some embodiments, similarity of two expression levels means that the expression level of a gene in a biological sample is within one standard deviation or standard error of the mean expression level of the same gene in biological samples from a group of subjects (e.g., a group of reference lymphoma patients, a group of lymphoma patients who responded to cancer treatment, or a group of lymphoma patients who did not respond to cancer treatment). In some embodiments, similarity of two expression levels between two groups of subjects means that the mean expression level of a gene in biological samples from one group of subjects is within one standard deviation or standard error of the mean expression level of the same gene in biological samples from another group of subjects. In some embodiments, similarity of two expression levels between two groups of subjects means that a hypothesis test (e.g., a Wilcoxon or t-test) fails to reject the null hypothesis that the two means or two medians are equal at a predetermined significance level, wherein each mean or median is the mean or median expression level of the same gene in biological samples from one of the two groups of subjects. In some embodiments, the group of subjects is a group of lymphoma patients who responded to cancer treatment. In some embodiments, the group of subjects is a group of lymphoma patients who did not respond to cancer treatment. In some embodiments, the group of subjects is a group of reference lymphoma patients.

[0233] In some embodiments, cancer treatment is a combination therapy using rituximab (MabThera), cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP).

[0234] On the other hand, this article provides a method for treating lymphoma patients, the method comprising: (i) identifying lymphoma patients who are predicted to be responsive to cancer treatment as described herein using the predictive methods described herein; and (ii) administering the cancer treatment to the lymphoma patient.

[0235] On the other hand, this article provides a method for treating lymphoma patients, comprising: (i) identifying lymphoma patients who are predicted to be unlikely to respond to cancer treatment as described herein using the predictive methods described herein; and (ii) administering alternative cancer treatment to the lymphoma patients.

[0236] In some embodiments, the alternative cancer treatment is a BET inhibitor or a CDK inhibitor.

[0237] In some embodiments, all the genes listed in Table 1 can be used as biomarkers to predict the responsiveness of lymphoma (e.g., DLBCL) patients to treatment.

[0238] In some embodiments, all the genes listed in Table 2 may be used as biomarkers to predict the responsiveness of lymphoma (e.g., DLBCL) patients to treatment.

[0239] In some embodiments, all the genes listed in Table 3 may be used as biomarkers to predict the responsiveness of lymphoma (e.g., DLBCL) patients to treatment.

[0240] In some embodiments, all the genes listed in Table 4 may be used as biomarkers to predict the responsiveness of lymphoma (e.g., DLBCL) patients to treatment.

[0241] On the other hand, the subgroups provided in this paper (e.g., A1-A7) can be characterized and / or identified based on the Bcl6 feature score, as shown in the Examples section below. Therefore, the Bcl6 feature score can also be used as a method to classify patients into one of eight subgroups to facilitate the determination of patient responsiveness to treatment.

[0242] On the other hand, as shown in the Examples section below and Figure 3C or Figure 3D, mutation data were collected and interpreted in the context of identified subgroups. Therefore, in some embodiments, the mutation spectrum of each subgroup (or cluster) or subset thereof can also be used to identify subgroups or classify patients into one of the subgroups to facilitate determination of patient responsiveness to treatment.

[0243] The subgroups (or clusters) provided in this article are also characterized based on the total count of different T cell populations (e.g., CD3, CD4, CD8, CD163, CD68, and / or CD11c cells), as shown in the Examples section below and Figures 11A-11F. Therefore, in another aspect, 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 them to facilitate determination of patient responsiveness to treatment.

[0244] 5.4 Application Method

[0245] In some embodiments, the methods provided herein include administering a first cancer treatment compound to a lymphoma patient who is expected to respond to a first cancer treatment. Methods are also provided herein for treating patients who have previously received cancer treatment (e.g., DLBCL or a subtype thereof) but are unresponsive to a first cancer treatment (e.g., standard therapy). Methods are also provided herein for treating previously untreated patients. The invention also covers methods for treating patients regardless of their age, although some diseases or disorders are more common in certain age groups. The invention further covers methods for treating patients who have and have not undergone surgery to treat a disease or condition. Because cancer patients have heterogeneous clinical presentations and different clinical outcomes, the treatment given to a patient may vary depending on his / her prognosis. Skilled clinicians will be able to easily identify specific second drugs, surgical procedures, and non-drug-based standard therapies that are effective for treating an individual patient with cancer (e.g., DLBCL or a subtype thereof) without requiring extensive experimentation.

[0246] In some embodiments, the therapeutic or preventative effective dose of the cancer treatment is about 0.005 mg / day to about 1,000 mg / day, about 0.01 mg / day to about 500 mg / day, about 0.01 mg / day to about 250 mg / day, about 0.01 mg / day to about 100 mg / day, about 0.1 mg / day to about 100 mg / day, about 0.5 mg / day to about 100 mg / day, about 1 mg / day to about 100 mg / day, about 0.01 mg / day to about 50 mg / day, about 0.1 mg / day to about 50 mg / day, about 0.5 mg / day to about 50 mg / day, about 1 mg / day to about 50 mg / day, about 0.02 mg / day to about 25 mg / day, or about 0.05 mg / day to about 10 mg / day.

[0247] In some embodiments, the effective dose for treatment or prevention 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.

[0248] In some embodiments, the recommended daily dose for cancer treatment of the conditions described herein ranges from about 0.1 mg / day to about 50 mg / day, preferably administered as a single dose once daily or in divided doses throughout the day. In some embodiments, the dose ranges from about 1 mg / day to about 50 mg / day. In some embodiments, the dose ranges from about 0.5 mg / day to about 5 mg / day. In some embodiments, the specific daily dose 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, or 50 mg / day.

[0249] In some embodiments, a starting dose of 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 is recommended. In some embodiments, the starting dose may be 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.

[0250] In some embodiments, the effective therapeutic or preventative dose is about 0.001 mg / kg / day to about 100 mg / kg / day, about 0.01 mg / kg / day to about 50 mg / kg / day, about 0.01 mg / kg / day to about 25 mg / kg / day, about 0.01 mg / kg / day to about 10 mg / kg / day, about 0.01 mg / kg / day to about 9 mg / kg / day, 0.01 mg / kg / day / kg / day to about 8 mg / kg / 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.

[0251] In some embodiments, the administered dose may also be expressed in units other than mg / kg / day. For example, the dose for parenteral administration may be expressed as mg / m³. 2 / day. Given the subject's height or weight, or both, those skilled in the art will readily know how to convert the dosage from mg / kg / day to mg / m². 2 / day (see www.fda.gov). For example, a dose of 1 mg / kg / day for a 65 kg person is approximately equal to 38 mg / m². 2 / sky.

[0252] In some embodiments, the amount of cancer treatment administered is sufficient to provide a plasma concentration of the compound in a steady state, the plasma concentration being 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.

[0253] In some embodiments, the amount of cancer treatment administered is sufficient to provide a plasma concentration of the compound in a steady state, the plasma concentration being 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.

[0254] As used herein, the term "plasma concentration at steady state" refers to the concentration reached after a period of time following administration of the cancer treatment described herein. Once a steady state is reached, small peaks and troughs will appear on the time-dependent curve of the plasma concentration of the cancer treatment.

[0255] In some embodiments, the amount of cancer treatment applied is sufficient to provide the maximum plasma concentration (peak concentration) of the compound, which ranges from 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.

[0256] In some embodiments, the amount of cancer treatment applied is sufficient to provide a minimum plasma concentration (trough concentration) of the compound, which ranges from 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.

[0257] In some embodiments, the amount of cancer treatment administered is sufficient to provide the area under the curve (AUC) of the compound, which ranges from about 100 ng. h / mL to approximately 100,000 ng hr / mL, from approximately 1,000 ng h / mL to approximately 50,000 ng hr / mL, from approximately 5,000 ng h / mL to approximately 25,000 ng hr / mL or from approximately 5,000 ng h / mL to approximately 10,000 ng hr / mL.

[0258] In some embodiments, lymphoma patients to be treated with one of the methods provided herein have not received anticancer therapy prior to administration of standard therapy (e.g., R-CHOP). In some embodiments, lymphoma patients to be treated with one of the methods provided herein have received anticancer therapy (standard therapy, e.g., R-CHOP) prior to administration of a second treatment. In some embodiments, lymphoma patients to be treated with one of the methods provided herein have developed resistance to a first cancer treatment.

[0259] Depending on the subtype of the lymphoma to be treated (e.g., DLBCL) and the subject's condition, the cancer treatment is administered via parenteral (e.g., intramuscular, intraperitoneal, intravenous, CIV, intracisional injection or infusion, subcutaneous injection or implantation), inhalation, nasal, vaginal, rectal, sublingual, or external (e.g., percutaneous or local) administration routes. In some embodiments, the cancer treatment is formulated into suitable dose units, alone or in combination with pharmaceutically acceptable excipients, carriers, adjuvants, and mediators suitable for each route of administration.

[0260] In some embodiments, the cancer treatment is administered parenterally. In some embodiments, the cancer treatment is administered intravenously.

[0261] Depending on the status of the lymphoma to be treated and the condition of the subject, in some embodiments, the therapeutic compound is administered via oral, parenteral (e.g., intramuscular, intraperitoneal, intravenous, CIV, intracisional injection or infusion, subcutaneous injection or implantation), inhalation, nasal, vaginal, rectal, sublingual, or external (e.g., percutaneous or local) administration routes. In some embodiments, the therapeutic compound is formulated alone or with pharmaceutically acceptable excipients, carriers, adjuvants, and mediators suitable for each route of administration into appropriate dose units.

[0262] 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.

[0263] In some embodiments, the therapeutic compound may be delivered as a single dose (e.g., a single bolus injection) or orally in capsules, tablets, or pills; or over time (e.g., continuous infusion over time or fractionated bolus injections over time). In some embodiments, the cancer treatment described herein may be repeated if necessary, for example, until the patient experiences disease stabilization or regression, or until the patient experiences disease progression or unacceptable toxicity.

[0264] In some embodiments, the therapeutic compound may be administered once daily (QD) or divided into multiple daily doses, such as twice daily (BID), three times daily (TID), and four times daily (QID). In some embodiments, the administration may be continuous (i.e., daily administration for several consecutive days or days) or intermittent, such as cyclical administration (i.e., including rest periods of several days, weeks, or months without drug use). As used herein, the term "daily" is intended to mean, for example, administration of the therapeutic compound once or more daily for a period of time. The term "continuous" is intended to mean daily administration of the therapeutic compound for an uninterrupted period of at least 7 days to 52 weeks. As used herein, the terms "intermittent" or "intermittently" are intended to mean stopping and starting at regular or irregular intervals. In some embodiments, intermittent administration is administration one to six days per week, cyclical administration (e.g., daily administration for two to eight consecutive weeks, followed by a week-long rest period without administration), or administration every other day. As used herein, the term "cyclical" is intended to mean daily or continuous administration of the therapeutic compound with rest periods.

[0265] In some embodiments, the administration frequency is in the range of approximately daily dose to approximately monthly dose.

[0266] In some embodiments, the cancer treatment may be delivered as a single dose (e.g., a single bolus) or over time (e.g., continuous infusion over time or fractionated bolus doses over time). In some embodiments, the compound may be repeatedly administered if necessary, for example, until the patient experiences disease stability or regression or until the patient experiences disease progression or unacceptable toxicity. For example, for solid cancers, disease stability generally means that the measurable vertical diameter of the lesion has not increased by 25% or more compared to the last measurement. Therasse et al., J Natl Cancer Inst., 2000, 92(3):205-216. Disease stability or non-disease stability is determined by methods known in the art, such as evaluation of patient symptoms, physical examination, and visualization of the tumor using X-ray, CAT, PET, MRI scan imaging, or other generally accepted evaluation modalities.

[0267] In some embodiments, the cancer treatment may be administered once daily (QD) or divided into multiple daily doses, such as twice daily (BID), three times daily (TID), and four times daily (QID). In some embodiments, administration may be continuous (i.e., daily administration for several consecutive days or days) or intermittent, such as cyclical administration (i.e., including rest periods of several days, weeks, or months without drug use). As used herein, the term "daily" is intended to mean, for example, administering cancer treatment once or more daily for a period of time. The term "continuous" is intended to mean an uninterrupted period of daily administration of cancer treatment for at least 10 days to 52 weeks. As used herein, the terms "intermittent" or "intermittently" are intended to mean stopping and starting at regular or irregular intervals. For example, intermittent administration of the cancer treatment may be administered one to six days per week, cyclical administration (e.g., daily administration for two to eight consecutive weeks, followed by a week-long rest period without administration), or every other day. As used herein, the term "cyclical" is intended to mean daily or continuous administration of cancer treatment with rest periods. In some embodiments, the rest period is the same length as the treatment period. In some embodiments, the rest period is different in length from the treatment period. In some embodiments, the cycle duration is 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 weeks. In some embodiments of the cycle, cancer treatment is administered daily for a period of 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, or 30 days, followed by a rest period. In some embodiments, cancer treatment is administered daily for a 5-day period within a 4-week cycle. In another specific embodiment, cancer treatment is administered daily for a 10-day period within a 4-week cycle.

[0268] In some embodiments, the frequency of administration ranges from approximately a daily dose to approximately a monthly dose. In some embodiments, administration is performed once daily, twice daily, three times daily, four times daily, every other day, twice weekly, once weekly, once every two weeks, once every three weeks, or once every four weeks. In some embodiments, cancer treatment is administered once daily. In some embodiments, cancer treatment is administered twice daily. In some embodiments, cancer treatment is administered three times daily. In some embodiments, cancer treatment is administered four times daily.

[0269] In some embodiments, cancer treatment is administered once daily for a duration ranging from one day to six months, from one week to three months, from one week to four weeks, from one week to three weeks, or from one week to two weeks. In some embodiments, cancer treatment is administered once daily for one week. In some embodiments, cancer treatment is administered once daily for two weeks. In some embodiments, cancer treatment is administered once daily for three weeks. In some embodiments, cancer treatment is administered once daily for four weeks.

[0270] 5.5 Combination Therapy

[0271] One or more additional therapies, such as additional active ingredients or agents, may be used in combination with the administration of the cancer treatments described herein to treat patients with lymphoma (e.g., patients with DLBCL). In some embodiments, one or more additional therapies may be administered before, simultaneously with, or after the administration of the compounds described herein. The cancer treatments described herein and additional active agents (“second active agents”) may be administered to a patient simultaneously or sequentially via the same or different routes of administration. The suitability of a particular route of administration for a particular active agent will depend on the active agent itself (e.g., whether it can be administered orally and does not break down before entering the bloodstream) and the condition of the lymphoma being treated (e.g., DLBCL). Routes of administration for additional active agents or ingredients are known to those skilled in the art. See, for example, Physicians' Desk Reference.

[0272] In some embodiments, the cancer treatments described herein and additional activator cycles are administered to patients with lymphoma (e.g., DLBCL). Cyclic therapy involves administering an activator for a period of time, followed by a rest period, and repeating this sequential administration. Cyclic therapy can reduce the development of resistance to one or more therapies, avoid or reduce the side effects of one of these therapies, and / or improve the efficacy of the treatment.

[0273] In some embodiments, one or more second active ingredients or agents may be used in the methods and compositions provided herein. The second active agent may be a macromolecule (e.g., a protein) or a small molecule (e.g., a synthetic inorganic molecule, organometallic molecule, or organic molecule). Various agents may be used, such as those described in U.S. Patent Application No. 16 / 390,815 or U.S. Provisional Application No. 14247-390-888, filed on the same day, entitled “SUBSTITUTED 4-AMINOISOINDOLINE-1,3-DIONE COMPOUNDS AND SECOND ACTIVEAGENTS FOR COMBINED USE [For use in combination of substituted 4-aminoisoindoline-1,3-dione compounds and second active agents], 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., pabisostat, romidesin, or vorinostat), BCL2 inhibitors (e.g., veneclax), BTK inhibitors (e.g., ibrutinib or acomitinib), mTOR inhibitors (e.g., everolimus), PI3K inhibitors (e.g., ederalipix), PKCβ inhibitors (e.g., enzatolin), SYK inhibitors (e.g., fattatinib), JAK2 inhibitors (e.g., fedrolinib, paretinib, ruxolitinib, baricitinib, gandutinib, letotinib, or mometinib), and Aurora. A kinase inhibitors (e.g., aliriteti), EZH2 inhibitors (e.g., tazestat, GSK126, CPI-1205, 3-deadenine A, EPZ005687, EI1, UNC1999, or cinnefenidine), BET inhibitors (e.g., piracetam or 4-[2-(cyclopropylmethoxy)-5-(methanesulfonyl)phenyl]-2-methylisoquinoline-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 (e.g., pinoxetine), HAT inhibitors (e.g., C646), WDR5 inhibitors (e.g., OICR-9429), HDAC6 inhibitors (e.g., ACY-241), DNMT1 selective inhibitors (e.g., GSK3484862), LSD-1 inhibitors (e.g., compound C or celidestat), G9A inhibitors (e.g., UNC 0631), PRMT5 inhibitors (e.g., GSK3326595), BRPF1B / 2 inhibitors (e.g., OF-1), BRD9 / 7 inhibitors (e.g., LP99), SUV420H1 / H2 inhibitors (e.g., A-196), Menin-MLL inhibitors (e.g., MI-503), CARM1 inhibitors (e.g., EZM2302), BRD9 (e.g., dBrd9 inhibitors), aiolos / ikaros cerebellar protein E3 ligase modulators (CELMoDs), CREBBp2 inhibitors, anti-CD79b antibodies, CD19 CAR-T, p53 (nutlins) inhibitors, Bcl6 inhibitors, CREBBp2 CELMoDs, CD79b CELMoDs, CD19 CELMoDs, p53 (nutlins) CELMoDs, Bcl6 CELMoDs, CREBBP2 ligand-directed degradation (LDD) inhibitors, CD79b LDD inhibitors, CD19 LDD inhibitors, p53 (nutlins) LDD inhibitors, Bcl6 LDD inhibitors, CK1a LDD inhibitors, IRAK4 LDD inhibitors (e.g., in MYD88 L265p lymphoma), MALT1 inhibitors (e.g., JNJ-67856633), MAT2A inhibitors (e.g., for 9p21 deletion), anti-CD3 x anti-CD19 bispecific antibodies, and anti-CD3 x anti-CD20 bispecific antibodies.

[0274] In some embodiments, these methods further include administering one or more of rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone, etoposide, bendamustine / Treanda, lenalidomide, or gemcitabine. In some embodiments, these methods further include administering one or more of rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone, etoposide, bendamustine / Treanda, or gemcitabine. In some embodiments, the treatment further includes treatment with one or more of the following: R-CHOP (rituximab + cyclophosphamide, doxorubicin, vincristine, and prednisone), R EPOCH (etoposide, rituximab + cyclophosphamide, doxorubicin, vincristine, and prednisone), R2-CHOP (lenalidomide, rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone), stem cell transplantation, bendamustine / Treanda + rituximab, rituximab, lenalidomide + rituximab, or gemcitabine-based combinations. In some embodiments, the treatment further comprises treatment with one or more of the following: R-CHOP (rituximab + cyclophosphamide, doxorubicin, vincristine, and prednisone), R EPOCH (etoposide, rituximab + cyclophosphamide, doxorubicin, vincristine, and prednisone), stem cell transplantation, bendamustine / Treanda + rituximab, rituximab, or gemcitabine-based combinations. In some embodiments, the second active agent is rituximab, as provided in U.S. Provisional Application No. 62 / 833,432.

[0275] 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 pabistat, romidesin, or vorinostat, or a stereoisomer, mixture of stereoisomers, tautomer, isotope, or pharmaceutically acceptable salt thereof.

[0276] 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 veneclade or its tautomer, isotope, or pharmaceutically acceptable salt. In some embodiments, the BCL2 inhibitor is veneclade.

[0277] 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 acutinib, or a stereoisomer, mixture of stereoisomers, tautomers, isotopes, or a pharmaceutically acceptable salt thereof. In some embodiments, the BTK inhibitor is ibrutinib, or a stereoisomer, mixture of stereoisomers, tautomers, isotopes, or a pharmaceutically acceptable salt thereof. In some embodiments, the BTK inhibitor is ibrutinib. In some embodiments, the BTK inhibitor is acutinib, or a stereoisomer, mixture of stereoisomers, tautomers, isotopes, or a pharmaceutically acceptable salt thereof. In some embodiments, the BTK inhibitor is acutinib.

[0278] 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 analogue thereof (also known as a rapamycin analogue). In some embodiments, the mTOR inhibitor is everolimus or a stereoisomer, mixture of stereoisomers, tautomer, isotope, or pharmaceutically acceptable salt thereof. In some embodiments, the mTOR inhibitor is everolimus.

[0279] In some embodiments, the second active agent used in the methods provided herein is a phosphoinositol 3-kinase (PI3K) inhibitor. In some embodiments, the PI3K inhibitor is ederaglilix or a stereoisomer, mixture of stereoisomers, tautomer, isotope, or pharmaceutically acceptable salt thereof. In some embodiments, the PI3K inhibitor is ederaglilix.

[0280] In some embodiments, the second active agent used in the methods provided herein is a protein kinase C β (PKCβ or PKC-β) inhibitor. In some embodiments, the PKCβ inhibitor is enzartolin or a stereoisomer, mixture of stereoisomers, tautomer, isotope, or pharmaceutically acceptable salt thereof. In some embodiments, the PKCβ inhibitor is enzartolin. In some embodiments, the PKCβ inhibitor is a pharmaceutically acceptable salt of enzartolin. In some embodiments, the PKCβ inhibitor is the hydrochloride salt of enzartolin. In some embodiments, the PKCβ inhibitor is a dihydrochloride salt of enzartolin.

[0281] 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 flotinib or its tautomer, isotope, or pharmaceutically acceptable salt. In one embodiment, the SYK inhibitor is flotinib. In some embodiments, the SYK inhibitor is a pharmaceutically acceptable salt of flotinib. In some embodiments, the SYK inhibitor is flotinib disodium hexahydrate.

[0282] 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 filtratinib, paritinib, ruxolitinib, baricitinib, gandutinib, letotinib, or mometinib, or a stereoisomer, mixture of stereoisomers, tautomer, isotope, or pharmaceutically acceptable salt thereof.

[0283] In some embodiments, the JAK2 inhibitor is filtratinib or its tautomer, isotope, or pharmaceutically acceptable salt. In some embodiments, the JAK2 inhibitor is filtratinib.

[0284] In some embodiments, the JAK2 inhibitor is paritinib or its tautomer, isotope, or pharmaceutically acceptable salt. In some embodiments, the JAK2 inhibitor is paritinib.

[0285] In some embodiments, the JAK2 inhibitor is ruxolitinib or a stereoisomer, mixture of stereoisomers, tautomer, isotope, 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.

[0286] 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 alistetine or its tautomer, isotope, or pharmaceutically acceptable salt. In some embodiments, the Aurora A kinase inhibitor is alistetine.

[0287] In some embodiments, the second active agent used in the methods provided herein is a Zeste homolog enhancer 2 (EZH2) inhibitor. In some embodiments, the EZH2 inhibitor is tazolidinyl, GSK126, CPI-1205, 3-deadenine A (DZNep), EPZ005687, EI1, UNC1999, or sinefenidone, or a stereoisomer, mixture of stereoisomers, tautomer, isotope, or pharmaceutically acceptable salt thereof.

[0288] In some embodiments, the EZH2 inhibitor is tazolidinyl or its tautomer, isotope, or pharmaceutically acceptable salt. In some embodiments, the EZH2 inhibitor is tazolidinyl.

[0289] In some embodiments, the EZH2 inhibitor is GSK126 or a stereoisomer, mixture of stereoisomers, tautomer, isotope, or pharmaceutically acceptable salt thereof. In some embodiments, the EZH2 inhibitor is GSK126 (also known as GSK-2816126).

[0290] In some embodiments, the EZH2 inhibitor is CPI-1205 or a stereoisomer, mixture of stereoisomers, tautomer, isotope, or pharmaceutically acceptable salt thereof. In some embodiments, the EZH2 inhibitor is CPI-1205.

[0291] In some embodiments, the EZH2 inhibitor is 3-deadenine A. In some embodiments, the EZH2 inhibitor is EPZ005687. In some embodiments, the EZH2 inhibitor is EI1. In one embodiment, the EZH2 inhibitor is UNC1999. In some embodiments, the EZH2 inhibitor is sinefenidine.

[0292] In some embodiments, the second active agent used in the methods provided herein is a hypomethylating agent. In some embodiments, the hypomethylating agent is 5-azacytidine or decitabine, or a stereoisomer, mixture of stereoisomers, tautomer, isotope, or pharmaceutically acceptable salt thereof.

[0293] In some embodiments, the hypomethylating agent is 5-azacytidine or a stereoisomer, mixture of stereoisomers, tautomer, isotope, or pharmaceutically acceptable salt thereof. In some embodiments, the hypomethylating agent is 5-azacytidine.

[0294] In some embodiments, the hypomethylating agent is decitabine or a stereoisomer, mixture of stereoisomers, tautomer, isotope, or pharmaceutically acceptable salt thereof. In one embodiment, the hypomethylating agent is decitabine.

[0295] In some embodiments, the second active agent used in the methods provided herein is chemotherapy. In some embodiments, the chemotherapy is bendamustine, doxorubicin, etoposide, methotrexate, cytarabine, vincristine, ifosfamide, or melphalan, or a stereoisomer, mixture of stereoisomers, tautomers, isotopes, prodrugs, or pharmaceutically acceptable salts thereof.

[0296] In some embodiments, the second therapeutic agent is administered before, after, or simultaneously with the cancer treatment described herein. The cancer treatment and the second therapeutic agent described herein may be administered to the patient simultaneously or sequentially via the same or different routes of administration. The suitability of a particular route of administration 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 and does not dissolve before entering the bloodstream) and the subject being treated. The particular route of administration of the second drug or agent or ingredient is known to those skilled in the art. See, for example, The Merck Manual, 448 (17th edition, 1999).

[0297] Any combination of the above-mentioned therapeutic agents suitable for treating these diseases or their symptoms may be administered. Such therapeutic agents may be administered simultaneously in any combination or as a separate treatment procedure.

[0298] As used herein, the term "combination" does not limit the order in which multiple therapies (e.g., prophylactic and / or therapeutic agents) are administered to a patient with a disease or disorder. A second active agent described herein may be administered to a patient simultaneously or sequentially via the same or different routes of administration. The suitability of a particular route of administration for a particular active agent will depend on the active agent itself (e.g., whether it can be administered orally and does not break down before entering the bloodstream).

[0299] 5.6 Pharmaceutical Compositions

[0300] In some embodiments, the cancer treatments and / or additional active agents provided herein are formulated in a pharmaceutical composition, and the methods provided herein include administering the pharmaceutical composition containing the cancer treatment to a patient with lymphoma (e.g., DLBCL).

[0301] In some embodiments, the pharmaceutical compositions provided herein comprise a therapeutically effective amount of one or more of the cancer treatments provided herein, as well as a pharmaceutically acceptable carrier, diluent, or excipient. In some embodiments, these compounds are formulated as the sole pharmaceutically active ingredient in the composition or in combination with other active ingredients.

[0302] 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, ocular, nasal, inhalation, nebulization, skin, or transdermal administration. Typically, the compounds described above can be formulated into pharmaceutical compositions using techniques and procedures well known in the art (see, for example, Ansel, Introduction to Pharmaceutical Dosage Forms, (7th edition, 1999)).

[0303] In some embodiments, the composition comprises an effective concentration of one or more compounds or pharmaceutically acceptable salts to be mixed with a suitable drug carrier or medium. In some embodiments, the concentration of the compounds in the composition is such that, upon administration, it effectively delivers an amount sufficient to treat, prevent, or improve one or more symptoms and / or progression of lymphoma (e.g., DLBCL).

[0304] In some embodiments, the active compound is present in an amount sufficient to exert a therapeutically useful effect without causing undesirable side effects in the treated patient. Therapeuticly effective concentrations are determined empirically by testing these compounds in the in vitro and in vivo systems described herein, from which the dosage for human use is deduced. The concentration of the active compound in a pharmaceutical composition depends on the absorption, tissue distribution, inactivation, and excretion rate of the active compound, the physicochemical characteristics of the compound, the dosing schedule and dosage, and other factors known to those skilled in the art.

[0305] In some embodiments, pharmaceutically therapeutically active compounds and their salts are formulated and administered in unit doses or multiple dose forms. As used herein, a unit dose form refers to a physically discrete unit suitable for human and animal subjects and individually packaged as known in the art. Each unit dose contains a predetermined amount of the therapeutically active compound associated with a desired drug carrier, medium, or diluent sufficient to produce the desired therapeutic effect. Examples of unit dose forms include ampoules and syringes, as well as individually packaged tablets or capsules. Unit dose forms are administered in fractions or multiples thereof. Multiple dose forms are multiple identical unit dose forms packaged in a single container for administration in separate unit dose forms. Examples of multiple dose forms include vials, tablet or capsule vials, or pints or gallons. Thus, multiple dose forms are multiples of unit doses that are not separately packaged.

[0306] In some embodiments, the precise dosage and duration of treatment vary depending on the disease being treated and are determined empirically using known testing protocols or by extrapolation from in vivo or in vitro testing data. It should be noted that concentration and dosage values ​​may also vary depending on the severity of the condition to be alleviated. It should be further understood that, for any particular subject, the specific dosing regimen is adjusted over time based on individual needs and the professional judgment of the person administering the composition or supervising its administration, and the concentration ranges set forth herein are merely exemplary and not intended to limit the scope or practice of the claimed compositions.

[0307] In some embodiments, solutions or suspensions for parenteral, intradermal, subcutaneous, or topical application may include any of the following components: a sterile diluent (e.g., water, saline solution, fixative oil, polyethylene glycol, glycerol, propylene glycol, dimethylacetamide, or other synthetic solvents), an antimicrobial agent (e.g., benzyl alcohol and methylparaben), an antioxidant (e.g., ascorbic acid and sodium bisulfate), a chelating agent (e.g., ethylenediaminetetraacetic acid (EDTA)), a buffer (e.g., acetate, citrate, and phosphate), and a tension-regulating agent (e.g., sodium chloride or glucose). Parenteral formulations may be encapsulated in ampoules, pens, disposable syringes, or single- or multi-dose vials made of glass, plastic, or other suitable materials.

[0308] In some embodiments, sustained-release formulations may also be prepared. Suitable examples of sustained-release formulations include a semi-permeable matrix of a solid hydrophobic polymer containing the compounds provided herein, in the form of a molded article, such as a membrane or microcapsule. Examples of sustained-release matrices include iontophoresis patches, polyesters, hydrogels (e.g., poly(2-hydroxyethyl-methacrylate) or poly(vinyl alcohol)), polylactide, copolymers of L-glutamic acid and L-glutamic acid ethyl ester, 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. While polymers such as ethylene-vinyl acetate and lactic acid-glycolic acid can release molecules for more than 100 days, some hydrogels release proteins for a shorter duration. When encapsulated compounds are retained in the body for extended periods, they may denature or aggregate due to exposure to moisture at 37°C, leading to loss of biological activity and potentially structural changes. Based on the mechanism of action involved, reasonable strategies can be designed to achieve stabilization. For example, if the aggregation mechanism is found to be the formation of intermolecular disulfate bonds through sulfur-disulfide exchange, stabilization can be achieved by modifying thiol residues, lyophilizing from acidic solutions, controlling moisture content, using appropriate additives, and developing specific polymer matrix compositions.

[0309] In some embodiments, anhydrous pharmaceutical compositions and dosage forms comprising the compounds provided herein are further covered. The anhydrous pharmaceutical compositions and dosage forms provided herein can be prepared using anhydrous or low-moisture components and low-moisture or low-humidity conditions, as known to those skilled in the art. Anhydrous pharmaceutical compositions can be prepared and stored such that their anhydrous properties are maintained. Therefore, anhydrous compositions are packaged using materials known to prevent exposure to water so that they can be included in suitable formulation kits. Examples of suitable packaging include, but are not limited to, sealing foil, plastics, unit-dose containers (e.g., vials), blister packs, and strip packs.

[0310] In some embodiments, dosage forms or compositions containing 0.001% to 100% of the active ingredient (the balance being made up by a non-toxic carrier) can be prepared. In some embodiments, the composition contains about 0.005% to about 95% of the active ingredient. In some embodiments, the composition contains about 0.01% to about 90% of the active ingredient. In some embodiments, the composition contains about 0.1% to about 85% of the active ingredient. In some embodiments, these compositions contain about 0.1% to about 75%-95% of the active ingredient.

[0311] In some embodiments, pharmaceutically acceptable carriers for parenteral formulations include aqueous media, non-aqueous media, antimicrobial agents, isotonic agents, buffers, antioxidants, local anesthetics, suspending and dispersing agents, emulsifiers, chelating agents, and other pharmaceutically acceptable substances.

[0312] In some embodiments, examples of aqueous media include sodium chloride injection, Ringer's injection, isotonic glucose injection, sterile water injection, and glucose and lactated Ringer's injection. Non-aqueous parenteral media include plant-derived fixed oils, such as cottonseed oil, corn oil, sesame oil, and peanut oil. Parenteral preparations packaged in multi-dose containers must contain antimicrobial agents at concentrations that inhibit bacteria or fungi, including phenols or cresols, mercury, benzyl alcohol, chlorobutanol, methylparaben and propylparaben, thimerosal, benzalkonium chloride, and benzyl chloride. Isotonic agents include sodium chloride and glucose. Buffers include phosphates and citrates. Antioxidants include sodium bisulfate. Local anesthetics include procaine hydrochloride. Suspensions and dispersants include sodium carboxymethyl cellulose, hydroxypropyl methylcellulose, and polyvinylpyrrolidone. Emulsifiers include polysorbate 80 (TWEEN® 80). Sequestering agents (or chelating agents) for metal ions include EDTA. Drug carriers also include ethanol, polyethylene glycol, and propylene glycol for water-miscible media, and sodium hydroxide, hydrochloric acid, citric acid, or lactic acid for pH adjustment.

[0313] In some embodiments, the injectable formulation is designed for both local and systemic administration. Typically, a therapeutically effective dose is formulated to contain at least about 0.1% w / w to about 90% w / w or higher concentrations (e.g., more than 1% w / w) of the active compound to the tissue to be treated. The active ingredient may be administered once or divided into many smaller doses administered at time intervals. It should be understood that the precise dose and duration of treatment vary depending on the tissue treated and can be determined empirically using known testing protocols or by extrapolation from in vivo or in vitro testing data. It should be noted that concentration and dose values ​​may also vary with the age of the individual being treated. It should be further understood that for any particular subject, the specific dosing regimen should be adjusted over time based on individual needs and the professional judgment of the person administering or supervising the administration of the formulation, and the concentration ranges set forth herein are merely exemplary and are not intended to limit the scope or practice of the claimed formulations.

[0314] Also of interest in this article are lyophilized powders, which can be reconstituted for application as solutions, emulsions, and other mixtures. They can also be reconstituted and formulated into solids or gels.

[0315] In some embodiments, sterile lyophilized powders are prepared by dissolving the compounds provided herein or their pharmaceutically acceptable salts in a suitable solvent. In some embodiments, the solvent contains excipients that improve the stability of the powder or a reconstituted solution prepared from the powder, or other pharmacological components. Excipients that may be used include, but are not limited to, dextrose, sorbitol, fructose, corn syrup, xylitol, glycerol, glucose, sucrose, or other suitable agents. In some embodiments, the solvent contains a buffer, such as citrate, phosphate, or other buffers known to those skilled in the art. The solution is then sterilely filtered and lyophilized under standard conditions known to those skilled in the art to obtain the desired formulation. Typically, the resulting solution is dispensed into vials for lyophilization. Each vial contains a single or multiple doses of the compound. The lyophilized powder can be stored under suitable conditions, such as from about 4°C to room temperature.

[0316] In some embodiments, the lyophilized formulation is adapted to be reconstituted to an appropriate concentration with a suitable diluent 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 under accelerated conditions at 40°C / 75% RH for up to about 12 months, up to about 6 months, or up to about 3 months.

[0317] In some embodiments, the lyophilized formulation is adapted to be reconstituted with an aqueous solution for intravenous administration. In some embodiments, the lyophilized formulation provided herein is adapted to be reconstituted 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 up to about 1-24, 2-20, 2-15, or 2-10 hours after reconstitution. In some embodiments, the reconstituted aqueous solution is stable at room temperature for up to about 20, 15, 12, 10, 8, 6, 4, or 2 hours after reconstitution. In some embodiments, the reconstituted lyophilized formulation has a pH of about 4 to 5.

[0318] The active ingredients described herein can be administered using controlled-release methods or delivery devices well known to those skilled in the art. Examples include, but are not limited to, those described in the following: U.S. Patent 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, and 5,891,474. The following numbers are included hereby cited: 5,922,356, 5,972,891, 5,980,945, 5,993,855, 6,045,830, 6,087,324, 6,113,943, 6,197,350, 6,248,363, 6,264,970, 6,267,981, 6,376,461, 6,419,961, 6,589,548, 6,613,358, 6,699,500, and 6,740,634. Dosage forms using, for example, hydroxypropyl methylcellulose, other polymer matrices, gels, permeable membranes, permeable systems, multilayer coatings, microparticles, liposomes, microspheres, or combinations thereof, can be used to provide sustained or controlled release of one or more active ingredients to provide different proportions of the desired release profile. Suitable controlled-release formulations known to those skilled in the art, including those described herein, can be readily selected for use with the active ingredients provided herein.

[0319] 5.7 Biological Samples

[0320] In some embodiments, the various methods provided herein use samples (e.g., biological samples) from patients with lymphoma (e.g., DLBCL). Patients may be male or female, and may be adults, children, or infants. Samples may be analyzed during the active phase or inactive phase of lymphoma (e.g., DLBCL). In some embodiments, samples are obtained from the patient before, during, and / or after administration of the treatments described herein. In some embodiments, samples are obtained from the patient before administration of the treatments described herein. In some embodiments, more than one sample may be obtained from the patient.

[0321] In some embodiments, the sample comprises bodily fluids from the subject. Non-limiting examples of bodily fluids include blood (e.g., peripheral whole blood, peripheral blood), plasma, amniotic fluid, aqueous humor, bile, earwax, Cooper's fluid, preejaculate fluid, chyle, chyme, female ejaculate fluid, interstitial fluid, lymph, menstrual blood, breast milk, mucus, pleural effusion, pus, saliva, sebum, semen, serum, sweat, tears, urine, vaginal lubricant, vomitus, water, feces, internal bodily fluids (including cerebrospinal fluid surrounding the brain and spinal cord), synovial fluid surrounding bones and joints, intracellular fluid (fluid within cells), and vitreous fluid (fluid in the eye). In some embodiments, the sample is a blood sample. Blood samples can be obtained using, for example, conventional techniques described in, *PCR Protocols*, edited by Innis et al. (Academic Press, 1990). White blood cells can be isolated from blood samples using conventional techniques or commercially available kits, such as the RosetteSep kit (Stein Cell Technologies, Vancouver, Canada). White blood cell subsets (e.g., monocytes, B cells, T cells, mononuclear leukocytes, granulocytes, or lymphocytes) can be further separated using conventional techniques, such as magnetically activated cell sorting (MACS) (Miltenyi Biotec, Auburn, CA) or fluorescence-activated cell sorting (FACS) (Becton Dickinson, San Jose, CA).

[0322] 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.

[0323] In some embodiments, the samples used in this method include biopsy material (e.g., tumor biopsy material). Biopsy material can be derived from any organ or tissue, such as skin, liver, lung, heart, colon, kidney, bone marrow, teeth, lymph nodes, hair, spleen, brain, breast, or other organs. In some embodiments, the samples used in the methods described herein include tumor biopsy material. Any biopsy technique known to those skilled in the art can be used to separate samples from a subject, such as open biopsy, closed biopsy, core biopsy, incisional biopsy, excisional biopsy, or fine-needle aspiration biopsy.

[0324] In some embodiments, the sample used in the methods provided herein is obtained from the subject prior to treatment for lymphoma (e.g., DLBCL). In some embodiments, the sample is obtained from the patient during treatment for lymphoma (e.g., DLBCL). In some embodiments, the treatment comprises administering the compound described herein to the subject.

[0325] In some embodiments, the sample comprises multiple cells. Such cells may include any type of cell, such as 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 includes tumor biopsy material or tumor explants. In some embodiments, T cells (T lymphocytes) include, for example, helper T cells (effective 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 cells. + T cells, for example, are detected by flow cytometry. The number of T cells used in the method can range from a single cell to approximately 10. 9Individual cells. In some embodiments, B cells (B lymphocytes) include, for example, plasma B cells, dendritic cells, memory B cells, B1 cells, B2 cells, limbic zone B cells, and follicular B cells. B cells may express immunoglobulins (antibodies, B cell receptors).

[0326] In some embodiments, a combination of commercially available antibodies (e.g., Quest Diagnostics, San Juan Capistrano, California; Dako, Denmark) can be used to obtain a specific cell population.

[0327] In some embodiments, the samples used in the methods provided herein are derived from diseased tissue from patients with lymphoma (e.g., DLBCL). In some embodiments, the number of cells used in the methods provided herein can range from a single cell to approximately 10. 9 Cells. In some embodiments, the number of cells used in the methods provided herein is approximately 1 × 10⁻⁶. 4 5 × 10 4 1×10 5 5 × 10 5 1 × 10 6 5 × 10 6 1 × 10 7 5 × 10 7 1 × 10 8 Or 5 × 10 8 .

[0328] In some embodiments, the number and type of cells collected from the 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), and magnetically activated cell sorting (MACS), by examining cell morphology using optical 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 fluorescence properties of particles (Kamarch, Methods Enzymol. [Enzymological Methods], 1987, 151:150-165). Laser excitation of the fluorescent portion in a single particle generates a small charge, thereby allowing the electromagnetic separation of positive and negative particles from a mixture. In some embodiments, cell surface marker-specific antibodies or ligands are labeled with different fluorescent labels. Cells are processed using a cell sorter, allowing them to be separated based on their ability to bind to the antibody used. FACS-sorted particles can be deposited directly into the individual wells of a 96-well or 384-well plate to facilitate separation and cloning.

[0329] In some embodiments, a subset of cells is used in the methods provided herein. Methods for sorting and separating specific cell populations 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 cytometry sorting, FACS, bead-based separation (e.g., magnetic cell sorting), size-based separation (e.g., sieves, barrier arrays, or filters), sorting in microfluidic devices, antibody-based separation, sedimentation, affinity adsorption, affinity extraction, density gradient centrifugation, laser capture microdissection, etc.

[0330] 5.8 Methods for detecting expression levels

[0331] In some embodiments, the methods provided herein include measuring the expression level of at least one gene listed in Table 1, Table 2, Table 3, or Table 4. The expression level of at least one gene can be determined by methods known in the art.

[0332] In some embodiments, the expression level of the at least one gene is determined by measuring the mRNA level of these genes. Several methods for detecting or quantifying mRNA levels are known in the art. Exemplary methods include, but are not limited to, RNA blotting, ribonuclease protection assays, PCR-based methods, RNA-seq, microarrays, etc. In some embodiments, the expression level of the at least one gene is determined by a PCR-based method. In some embodiments, the expression level of the at least one gene is determined by RNA-seq. In some embodiments, the expression level of the at least one gene is determined by a microarray.

[0333] The mRNA sequence can be used to prepare at least partially complementary probes. The probes can then be used to detect the mRNA sequence in the sample using any suitable assay, such as PCR-based methods, digital PCR (dPCR), RNA blotting, test strip assays, etc.

[0334] In some embodiments, a nucleic acid assay for testing immunomodulatory activity in biological samples can be prepared. The assay includes a solid support and at least one nucleic acid in contact with the support, wherein the nucleic acid corresponds to at least a portion of mRNA. The assay may also include means for detecting alterations in mRNA expression in the sample.

[0335] In some embodiments, the assay method can vary depending on the type of mRNA information desired. Exemplary methods include, but are not limited to, RNA blotting and PCR-based methods (e.g., RT-qPCR). Methods such as RT-qPCR can also accurately quantify the amount of mRNA in a sample.

[0336] The presence of mRNA in a sample can be determined using any suitable assay platform. For example, the assay can take the form of a test strip, membrane, chip, disk, test paper, filter, microsphere, glass slide, multiwell plate, or optical fiber. In some embodiments, the assay system may have a solid support on which the nucleic acid corresponding to the mRNA is attached. The solid support may include, for example, plastic, silicon, metal, resin, glass, membrane, particles, precipitate, gel, polymer, sheet, sphere, polysaccharide, capillary, membrane, disk, or glass slide. Assay components can be prepared and packaged together as a kit for detecting mRNA.

[0337] In some embodiments, if desired, nucleic acids can be labeled to prepare a population of labeled mRNAs. Typically, samples can be labeled using methods well known in the art (e.g., using DNA ligases, terminal transferases, or by labeling the RNA backbone, etc.; see, for example, Ausubel et al., *Short Protocols in Molecular Biology*, 3rd edition, Wiley & Sons, 1995, and Sambrook et al., *Molecular Cloning: A Laboratory Manual*, 3rd edition, 2001, Cold Spring Harbor, New York). In some embodiments, samples are labeled with fluorescent markers. Exemplary fluorescent dyes include, but are not limited to, xanthan 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-carboxy-X rhodamine (ROX or R), 5-carboxyrhodamine 6G (R6G5 or G5), 6-carboxyrhodamine 6G (R6G6 or G6), and rhodamine 110; anthocyanin dyes, such as Cy3, Cy5, and Cy7 dyes; Alexa dyes, such as Alexa-fluor-555; coumarin, diethylaminocoumarin, umbelliferone; benzylimine dyes, such as Hoechst 33258; phenanthridine dyes, such as Texas Red; acetyridine dyes; acridine dyes; carbazole dyes; phenoxazine dyes; porphyrin dyes; polyacetylenic dyes, BODIPY dyes, quinoline dyes, pyrene, fluorescein chlorotriazine, R110, eosin dyes, JOE, R6G, tetramethylrhodamine, lissamine, ROX, naphthol fluorescein, etc.

[0338] In some embodiments, the mRNA assay method includes the following steps: (1) obtaining a surface-bound target probe; (2) hybridizing a large amount of mRNA with the surface-bound probe under conditions sufficient to provide specific binding; (3) performing post-hybridization washing to remove unbound nucleic acids from the hybridization; and (4) detecting the hybridized mRNA. The reagents used in each of these steps and the conditions under which they are used may vary depending on the specific application.

[0339] In some embodiments, hybridization is performed under suitable hybridization conditions, the stringency of which can be varied as needed. Typical conditions are sufficient to generate probe / target complexes on a solid surface between complementary binding members (i.e., between the surface-bound target probe and the complementary mRNA in the sample). In some embodiments, stringent hybridization conditions may be employed.

[0340] In some embodiments, hybridization is performed under strict hybridization conditions. Standard hybridization techniques (e.g., under conditions sufficient to provide specific binding of the target mRNA to the probe in the sample) are described in Kallioniemi et al., Science, 258:818-821 (1992) and WO 93 / 18186. Several general technical guidelines are available, for example, Tijssen, Hybridization with Nucleic Acid Probes, Parts I and II (Elsevier, Amsterdam 1993). For a description of techniques applicable to in situ hybridization, see Gall et al., Meth. Enzymol, 1981, 21:470-480; and Angerer et al., Genetic Engineering: Principles and Methods (edited by Setlow and Hollaender), Vol. 7, pp. 43-65 (Plenum Press, New York 1985). The selection of appropriate conditions, including temperature, salt concentration, polynucleotide concentration, hybridization time, and the stringency of washing conditions, will depend on the experimental design, including the sample source, the identity of the trapping agent, the expected degree of complementarity, and can be determined as a routine experiment by those skilled in the art.

[0341] A person of ordinary skill will readily recognize that alternative but comparable hybridization and washing conditions can be used to provide conditions with similar strictness.

[0342] Following the mRNA hybridization procedure, surface-bound polynucleotides are typically washed to remove unbound nucleic acids. Any convenient washing protocol can be used, where the washing conditions are generally stringent, as described above. Hybridization of the target mRNA with the probe is then detected using standard techniques.

[0343] In some embodiments, other methods, such as PCR-based methods, may also be used to track the expression of these genes. Examples of PCR methods can be found in the literature. Examples of PCR assays can be found in U.S. Patent No. 6,927,024, which is incorporated herein by reference in its entirety. Examples of RT-PCR methods can be found in U.S. Patent No. 7,122,799, which is incorporated herein by reference in its entirety. A method for fluorescence in situ PCR is described in U.S. Patent No. 7,186,507, which is incorporated herein by reference in its entirety.

[0344] In some embodiments, real-time reverse transcription-PCR (RT-qPCR) can be used for the detection and quantification of RNA targets (Bustin et al., Clin. Sci. [Clinical Science], 2005, 109:365-379). Quantitative results obtained by RT-qPCR often provide more information than qualitative data. Therefore, in some embodiments, RT-qPCR-based assays can be used to measure mRNA levels during cell-based assays. RT-qPCR methods can also be used to monitor patient treatment. Examples of RT-qPCR-based methods can be found, for example, in U.S. Patent No. 7,101,663, which is incorporated herein by reference in its entirety.

[0345] Compared to conventional reverse transcriptase-PCR and agarose gel analysis, real-time PCR provides quantitative results. Another advantage of real-time PCR is its relative simplicity and ease of use. Instruments for real-time PCR (e.g., Applied Biosystems 7500) and reagents (e.g., TaqMan sequencing reagents) are commercially available. For example, TaqMan can be used according to the manufacturer's instructions. ® Gene expression assays. These kits are pre-formulated gene expression assay kits for the 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, followed by 60°C for 1 minute.

[0346] To determine the number of cycles in which the fluorescence signal associated with the accumulation of a specific amplicon exceeds a threshold (referred to as CT), data can be analyzed using, for example, the 7500 Real-Time PCR System Sequence Detection Software version 1.3, employing a comparative CT relative quantification method. 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 determined automatically by the software. In some embodiments, the threshold level is set above baseline but low enough to fall within the exponential growth region of the amplification curve.

[0347] In some embodiments, the amount of one or more RNA transcripts may be measured using techniques known to those skilled in the art. In some embodiments, deep sequencing (e.g., ILLUMINA® RNASeq, ILLUMINA® Next Generation Sequencing (NGS), ION TORRENT) may be used. TM Next-generation RNA sequencing, 454 TM Pyrosequencing or oligonucleotide ligation detection sequencing (SOLID) TMThe amounts of one, two, three, four, five, or more RNA transcripts are measured. In other embodiments, microarrays and / or gene chips are used to measure the amounts of multiple RNA transcripts. In some embodiments, the amounts of one, two, three, or more RNA transcripts are determined by RT-PCR. In other embodiments, the amounts of one, two, three, or more RNA transcripts are measured by RT-qPCR. The techniques for performing these assays are known to those skilled in the art. In yet another embodiment, RNA transcripts are analyzed using NanoString (e.g., the nCounter® miRNA expression assay provided by NanoString® Technologies, Inc.).

[0348] 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, Western blotting (protein blotting), enzyme-linked immunosorbent assay (ELISA), immunohistochemistry, flow cytometry, cell counting bead arrays, mass spectrometry, etc. Several types of commonly used ELISAs are included, including direct ELISA, indirect ELISA, and sandwich ELISA.

[0349] In some embodiments, protein levels are determined by immunohistochemistry (IHC). IHC is a laboratory test that uses antibodies to detect certain antigens (markers) in tissue samples. It is a method of detecting antigens (e.g., proteins) in cells of tissue sections by utilizing the principle that antibodies specifically bind to antigens in biological tissues. Antibodies are typically linked to enzymes or fluorescent dyes. Typically, when an antibody binds to an antigen in a tissue sample, the enzyme or staining agent 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. It can also be used to help provide differences between different types of cancer. IHC can be used to image discrete components in tissues by using appropriately labeled antibodies to specifically bind their target antigens in situ. IHC enables the visualization and recording of high-resolution distribution and localization of specific cellular components within cells and against their appropriate histological background. While there are various approaches and arrangements of IHC methods, the entire set of steps involved can generally be divided into two groups: sample preparation and sample staining. In some embodiments, IHC is based on the immunostaining of thin tissue sections attached to individual slides. Multiple small sections can be arranged on a single slide for comparative analysis; this form is called a tissue microarray. In other embodiments, IHC is performed using high-throughput sample preparation and staining.

[0350] Samples can be observed using an optical or fluorescence microscope. In some embodiments, antigen detection in tissues can be performed using an antibody conjugated to an enzyme (horseradish peroxidase) and a colorimetric substrate detectable by an optical microscope.

[0351] In some embodiments, the sample (e.g., tissue taken from a patient) has been flash-frozen in liquid nitrogen, isopentane, or dry ice. In other embodiments, the sample (e.g., tissue taken from a patient) has been fixed in formaldehyde and embedded in paraffin wax (FFPE). In both of these methods, the tissue or tissue section can be mounted on a slide prior to staining. In yet another embodiment, IHC free-floating technology can be used, in which the entire IHC process is performed in a liquid to increase antibody binding and penetration, and slide fixation is only performed at the end of the experiment. IHC free-floating appears to be the most popular method in neuroscience research. When tissue needs to be analyzed by electron microscopy, the tissue can be embedded in an acrylate resin, such as ethylene glycol methacrylate (GMA), a technique known as IHC resin technology.

[0352] In some embodiments, IHC can be performed using the methods described in the Examples section below.

[0353] 5.9 Reagent Kit

[0354] In some embodiments, this document provides a kit for predicting the responsiveness of lymphoma patients to cancer treatment, the kit comprising reagents for measuring gene expression levels in biological samples taken from lymphoma patients. In some embodiments, the kit comprises reagents (or tools) for collecting samples from subjects. In some embodiments, the kit comprises instructions on how to interpret or use the determined expression levels to predict whether a patient has a specific subtype of lymphoma (e.g., DLBCL).

[0355] In some embodiments, the kit contains one or more reagents required to perform one or more assays described herein in one or more other containers. In some embodiments, the kit comprises a solid support and means for detecting RNA or protein expression of at least one biomarker in a biological sample. Such kits may employ, for example, test strips, membranes, chips, discs, test strips, filters, microspheres, glass slides, multiwell plates, or optical fibers. The solid support of the kit may be, for example, plastic, silicon, metal, resin, glass, membrane, particles, precipitates, gels, polymers, sheets, spheres, polysaccharides, capillaries, films, discs, or glass slides.

[0356] In some embodiments, the kit contains components in one or more containers for performing RT-PCR, RT-qPCR, deep sequencing, or microarrays (e.g., NanoString assays). In some embodiments, the kit includes a solid support, nucleic acids in contact with the support (wherein the nucleic acids are complementary to at least 10, 20, 50, 100, 200, 350, or more bases of mRNA), and means for detecting mRNA expression in a biological sample.

[0357] In some embodiments, the kit contains components in one or more containers for performing assays that determine the levels of one or more proteins, such as flow cytometry, ELISA, or HIC.

[0358] In some embodiments, such a kit may include materials and reagents required for measuring RNA or protein. In some embodiments, such a kit includes a microarray comprising oligonucleotides and / or DNA and / or RNA fragments hybridizing to one or more genes in Table 1. In some embodiments, such a kit may include primers for PCR of a gene or subset of genes or the RNA product of both or a copy of cDNA of the RNA product. In some embodiments, such a kit may include primers for PCR and probes for quantitative PCR. In some embodiments, such a kit may include multiple primers and multiple probes, some of which have different fluorophores to allow multiplex detection of one or more gene products. In some embodiments, such a kit may further include materials and reagents for generating cDNA from RNA. In some embodiments, such a kit may include antibodies specific to one or more genes in Table 1. Such a kit may additionally include materials and reagents for isolating RNA and / or protein from a biological sample. In some embodiments, such a kit may include materials and reagents for synthesizing cDNA from RNA isolated from a biological sample. In some embodiments, such a kit may include a computer program product embedded in a computer-readable medium for predicting whether a patient will respond to the compounds described herein. In some embodiments, the kit may include a computer program product embedded in a computer-readable medium and instructions.

[0359] In some embodiments, such a kit may include materials and reagents required for measuring RNA or protein. In some embodiments, such a kit includes a microarray comprising oligonucleotides and / or DNA and / or RNA fragments hybridizing to one or more genes in Table 2. In some embodiments, such a kit may include primers for PCR of a gene or subset of genes or the RNA product of both or a copy of cDNA of the RNA product. In some embodiments, such a kit may include primers for PCR and probes for quantitative PCR. In some embodiments, such a kit may include multiple primers and multiple probes, some of which have different fluorophores to allow multiplex detection of one or more gene products. In some embodiments, such a kit may further include materials and reagents for generating cDNA from RNA. In some embodiments, such a kit may include antibodies specific to one or more genes in Table 2. Such a kit may additionally include materials and reagents for isolating RNA and / or protein from a biological sample. In some embodiments, such a kit may include materials and reagents for synthesizing cDNA from RNA isolated from a biological sample. In some embodiments, such a kit may include a computer program product embedded in a computer-readable medium for predicting whether a patient will respond to the compounds described herein. In some embodiments, the kit may include a computer program product embedded in a computer-readable medium and instructions.

[0360] In some embodiments, such a kit may include materials and reagents required for measuring RNA or protein. In some embodiments, such a kit includes a microarray comprising oligonucleotides and / or DNA and / or RNA fragments hybridizing to one or more genes listed in Table 3. In some embodiments, such a kit may include primers for PCR of a gene or subset of genes or the RNA product of both or a copy of the cDNA of the RNA product. In some embodiments, such a kit may include primers for PCR and probes for quantitative PCR. In some embodiments, such a kit may include multiple primers and multiple probes, some of which have different fluorophores to allow multiplex detection of one or more gene products. In some embodiments, such a kit may further include materials and reagents for generating cDNA from RNA. In some embodiments, such a kit may include antibodies specific to one or more genes listed in Table 3. Such a kit may additionally include materials and reagents for isolating RNA and / or protein from a biological sample. In some embodiments, such a kit may include materials and reagents for synthesizing cDNA from RNA isolated from a biological sample. In some embodiments, such a kit may include a computer program product embedded in a computer-readable medium for predicting whether a patient will respond to the compounds described herein. In some embodiments, the kit may include a computer program product embedded in a computer-readable medium and instructions.

[0361] In some embodiments, such a kit may include materials and reagents for measuring RNA or proteins. In some embodiments, such a kit includes a microarray comprising oligonucleotides and / or DNA and / or RNA fragments hybridizing to one or more genes listed in Table 4. In some embodiments, such a kit may include primers for PCR of an RNA product or a copy of the cDNA of a gene or subset of genes or both. In some embodiments, such a kit may include primers for PCR and probes for quantitative PCR. In some embodiments, such a kit may include multiple primers and multiple probes, some of which have different fluorophores to allow multiplex detection of one or more gene products. In some embodiments, such a kit may further include materials and reagents for generating cDNA from RNA. In some embodiments, such a kit may include antibodies specific to one or more genes listed in Table 4. Such a kit may additionally include materials and reagents for isolating RNA and / or proteins from a biological sample. In some embodiments, such a kit may include materials and reagents for synthesizing cDNA from RNA isolated from a biological sample. In some embodiments, such a kit may include a computer program product embedded in a computer-readable medium for predicting whether a patient will respond to the compounds described herein. In some embodiments, the kit may include a computer program product embedded in a computer-readable medium and instructions.

[0362] In some embodiments, antibody-based kits may include, for example: (1) a first antibody (attached or not attached to a solid support) that binds to a target peptide, polypeptide, or protein; and, optionally, (2) a second, different antibody that binds to 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). These antibody-based kits may also include beads for performing immunoprecipitation. Each component of an antibody-based kit is typically housed in its own suitable container. Therefore, these kits typically include different containers suitable for each antibody. In some embodiments, antibody-based kits may include instructions for performing the assay and methods for interpreting and analyzing the data generated from the assay. In some embodiments, the kit includes instructions for predicting whether a patient with lymphoma (e.g., DLBCL) belongs to a specific subgroup of DLBCL (e.g., a high-risk subgroup of DLBCL).

[0363] In some embodiments of the methods and kits provided herein, the solid support is used for protein purification, sample labeling, or solid-phase assays. Examples of solid phases suitable for implementing the methods disclosed herein include beads, particles, colloids, single surfaces, tubes, multiwell plates, microtiter plates, glass slides, membranes, gels, and electrodes. In some embodiments, when the solid phase is a particulate material (e.g., beads), it is distributed in the pores of a multiwell plate to allow for parallel processing of the solid support.

[0364] Unless otherwise stated, the embodiments provided herein will be practiced using conventional techniques of molecular biology, microbiology, and immunology, which are within the skill level of those skilled in the art. Such techniques are well described in the literature. Particularly relevant texts include: Sambrook et al., *Molecular Cloning: A Laboratory Manual* (2nd edition, 1989); Glover (ed.), *DNA Cloning*, Volumes I and II (1985); Gait (ed.), *Oligonucleotide Synthesis* (1984); Hames and Higgins (ed.), *Nucleic Acid Hybridization* (1984); Hames and Higgins (ed.), *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); and Scopes, *Protein Purification: Principles and Practice*. (Springer Verlag, New York, 2nd edition, 1987); and Weir and Blackwell, eds., Handbook of Experimental Immunology, Volumes I-IV (1986).

[0365] As should be understood from the foregoing, although specific embodiments have been described herein for illustrative purposes, various modifications may be made without departing from the spirit and scope of the content provided herein. All references mentioned above are incorporated herein by reference in their entirety.

[0366] Some embodiments of the present invention are illustrated by the following non-limiting examples. Embodiments

[0367] The present invention provides the following non-limiting embodiments:

[0368] 1. A method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising:

[0369] (a) Clustering reference lymphoma patients into subgroups using the expression levels of at least one gene in reference biological samples from reference lymphoma patients;

[0370] (b) Determine 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

[0371] (c) Based on a subgroup of lymphoma patients, predict the responsiveness of lymphoma patients to primary cancer treatment.

[0372] 2. The method as described in Example 1, further comprising administering a second cancer treatment to the lymphoma patient based on a subgroup of the lymphoma patient.

[0373] 3. The method as described in Example 1 or 2, wherein the at least one gene comprises two or more genes, and step (a) includes generating clustering information that defines the relationship between the expression levels of two or more genes in these reference biological samples, and rearranging the heatmap display based on the clustering information.

[0374] 4. The method as described in any one of Examples 1-3, wherein step (a) includes employing a layered method or a non-layered method.

[0375] 5. The method as described in any one of Examples 1-3, wherein step (a) includes employing the iClusterPlus method.

[0376] 6. The method as described in any one of Examples 1-5, wherein these reference lymphoma patients are clustered into 2-12 subgroups.

[0377] 7. The method as described in Example 6, wherein these reference lymphoma patients are clustered into 7 subgroups.

[0378] 8. The method as described in any one of Examples 1-7, wherein the method further comprises training a classifier model using the expression levels of at least one gene in the reference biological samples.

[0379] 9. The method as described in Examples 1-8, wherein the at least one gene is selected from the genes in Table 2, and optionally the at least one gene comprises one, two, three, four, five or more genes from Table 2.

[0380] 10. The method as described in Example 9, wherein the at least one gene comprises all the genes in Table 2.

[0381] 11. The method of any one of Examples 1-8, wherein the at least one gene is selected from the genes in Table 3, and optionally wherein the at least one gene comprises one, two, three, four, five or more genes in Table 3.

[0382] 12. The method as described in Example 11, wherein the at least one gene comprises all the genes in Table 3.

[0383] 13. The method of any one of Examples 1-8, wherein the at least one gene is selected from the genes in Table 4, and optionally wherein the at least one gene comprises one, two, three, four, five or more genes in Table 4.

[0384] 14. The method as described in Example 13, wherein the at least one gene comprises all the genes in Table 4.

[0385] 15. The method as described in any one of Examples 8-14, wherein the classifier model is a grouped multinomial generalized linear model (GLM).

[0386] 16. The method as described in any one of Examples 8-15, wherein the classifier model is a binary generalized linear model.

[0387] 17. The method of any one of Examples 6-16, wherein the method further comprises setting a threshold confidence level for at least one of these subgroups to exclude patients from at least one of these subgroups who give clustering data with a lower confidence level.

[0388] 18. The method of any one of Examples 1-17, 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, marginal zone B-cell lymphoma, lymphoplasmacytic lymphoma, anaplastic large cell lymphoma, primary cutaneous lymphoma, mycosis fungoides, chronic lymphocytic leukemia, and mantle cell lymphoma.

[0389] 19. The method as described in Example 18, wherein the lymphoma is DLBCL.

[0390] 20. The method as described in Example 18, wherein the lymphoma is an indolent B-cell lymphoma, a marginal zone B-cell lymphoma, a mantle cell lymphoma, or a chronic lymphocytic leukemia.

[0391] 21. The method as described in any one of Examples 6-20, wherein the reference lymphoma patients are clustered into 7 subgroups A1-A7, and subgroup A7 includes about 0% to about 10% of germinal center B-cell-like (GCB) DLBCL patients, about 80% to about 90% of activated B-cell-like (ABC) DLBCL patients, about 0% to about 10% of TME+ DLBCL patients and about 5% to about 15% of DHITsig+ DLBCL patients.

[0392] 22. The method as described in Example 21, wherein

[0393] (i) Subgroup A1 includes approximately 50% to 60% of GCB DLBCL patients, approximately 30% to 40% of ABC DLBCL patients, approximately 10% to 20% of TME+DLBCL patients and approximately 30% to 40% of DHITsig+DLBCL patients.

[0394] (ii) Subgroup A2 includes approximately 80% to 90% of GCB DLBCL patients, approximately 0% to 5% of ABC DLBCL patients, approximately 15% to 25% of TME+DLBCL patients and approximately 25% to 35% of DHITsig+DLBCL patients.

[0395] (iii) Subgroup A3 includes approximately 40% to 55% of GCB DLBCL patients, approximately 30% to 45% of ABC DLBCL patients, approximately 40% to 50% of TME+DLBCL patients and approximately 20% to 30% of DHITsig+DLBCL patients.

[0396] (iv) Subgroup A4 includes approximately 25% to 35% of GCB DLBCL patients, approximately 40% to 50% of ABC DLBCL patients, approximately 30% to 40% of TME+DLBCL patients and approximately 10% to 20% of DHITsig+DLBCL patients.

[0397] (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; and / or

[0398] (vi) Subgroup A6 includes approximately 30% to 40% of GCB DLBCL patients, approximately 40% to 50% of ABC DLBCL patients, approximately 75% to 95% of TME+DLBCL patients, and approximately 0% to 10% of DHITsig+DLBCL patients.

[0399] 23. The method of any one of Examples 1-22, wherein the method includes predicting that lymphoma patients identified as belonging to subgroup A7 are unlikely to respond to the first cancer treatment.

[0400] 24. The method as described in any one of Examples 1-23, wherein the first cancer treatment is a combination therapy of rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone (R-CHOP).

[0401] 25. The method as described in any one of Examples 2-24, wherein the second cancer treatment is not R-CHOP.

[0402] 26. The method as described in any one of Examples 2-24, wherein the second cancer treatment is a combination of lenalidomide, rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R2-CHOP).

[0403] 27. The method as described in Example 26, wherein the second cancer treatment is a bromine domain and superterminal (BET) inhibitor or a cyclin-dependent kinase (CDK) inhibitor.

[0404] 28. A method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising:

[0405] (a) In biological samples from lymphoma patients, determine the expression level of at least one gene from Table 2, Table 3 or Table 4, optionally wherein the at least one gene comprises one, two, three, four, five or more genes from Table 2, Table 3 or Table 4.

[0406] (b) The expression level of the at least one gene in step (a) is compared with the expression level of the at least one gene in a reference biological sample from a reference lymphoma patient who has responded to the cancer treatment, and

[0407] 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 the reference biological sample, it indicates that the lymphoma patient may be responding to the cancer treatment.

[0408] 29. A method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising:

[0409] (a) In biological samples from lymphoma patients, determining the expression level of at least one gene from Table 2, Table 3, or Table 4, optionally wherein the at least one gene comprises one, two, three, four, five, or more genes from Table 2, Table 3, or Table 4; and

[0410] (b) The expression level of the at least one gene in the biological sample is compared with: (i) the expression level of the at least one gene in biological samples from lymphoma patients who have responded to treatment for this cancer and (ii) the expression level of the at least one gene in biological samples from lymphoma patients who have not responded to treatment for this cancer.

[0411] If the expression level of (a) is similar to that of (i), it indicates that the patient with the first lymphoma may respond to the cancer treatment; and if the expression level of (a) is similar to that of (ii), it indicates that the patient with the first lymphoma is unlikely to respond to the cancer treatment.

[0412] 30. A method for treating a patient with lymphoma, the method comprising:

[0413] (i) Identifying lymphoma patients who may respond to cancer treatment according to the method described in Examples 28 or 29; and

[0414] (ii) administering the cancer treatment to the lymphoma patient.

[0415] 31. A method for treating a patient with lymphoma, the method comprising:

[0416] (i) Identifying lymphoma patients who are unlikely to respond to cancer treatment according to the method described in Examples 28 or 29; and

[0417] (ii) Administer alternative cancer treatment to the lymphoma patient.

[0418] 32. The method as described in Example 30, wherein the cancer treatment is R-CHOP.

[0419] 33. The method as described in Example 31, wherein the alternative cancer treatment is not R-CHOP.

[0420] 34. The method as described in Example 31, wherein the alternative cancer treatment is R2-CHOP.

[0421] 35. The method as described in Example 31, wherein the alternative cancer treatment is a BET inhibitor or a CDK inhibitor.

[0422] 36. The method of any one of Examples 28-35, wherein the lymphoma is selected from the group consisting of: DLBCL, indolent B-cell lymphoma, follicular lymphoma, small lymphocytic lymphoma, marginal zone B-cell lymphoma, lymphoplasmacytic lymphoma, anaplastic large cell lymphoma, primary cutaneous lymphoma, mycosis fungoides, chronic lymphocytic leukemia, and mantle cell lymphoma.

[0423] 37. The method as described in Example 36, wherein the lymphoma is DLBCL.

[0424] 38. The method as described in Example 36, wherein the lymphoma is an indolent B-cell lymphoma, follicular lymphoma, marginal zone B-cell lymphoma, mantle cell lymphoma, or chronic lymphocytic leukemia.

[0425] 39. The method as described in any one of Examples 28-38, wherein the expression levels of all genes in Table 2, Table 3 or Table 4 are determined in step (a) and compared in step (b).

[0426] 40. The method as described in any one of Examples 1-39, wherein the biological sample and the reference biological sample are tumor biopsy samples.

[0427] 41. The method of any one of Examples 1-40, wherein the expression level of the at least one gene is determined by the mRNA level of the at least one gene.

[0428] 42. The method of any one of Examples 1-41, wherein the expression level of the at least one gene is determined by detecting the presence or amount of at least one complex in the biological sample or reference biological sample, wherein the presence or amount of the at least one complex indicates the expression level of the at least one gene.

[0429] 43. The method as described in Example 42, wherein the at least one complex is a hybridization complex or is detectably labeled.

[0430] 44. The method of any one of Examples 1-41, wherein the expression level of the at least one gene is determined by detecting the presence or amount of at least one reaction product in the biological sample or reference biological sample, wherein the presence or amount of the at least one reaction product indicates the expression level of the at least one gene.

[0431] 45. The method as described in Example 44, wherein the at least one reaction product is detectably labeled.

[0432] 46. ​​The method as described in any one of Examples 1-45, wherein the reference lymphoma patient is a patient with refractory DLBCL, a patient with relapsed DLBCL, or a patient with newly diagnosed DLBCL.

[0433] 47. The method as described in any one of Examples 1-46, wherein the lymphoma patient is a patient with refractory DLBCL, a patient with relapsed DLBCL, or a patient with newly diagnosed DLBCL.

[0434] 48. The method as described in any one of Examples 1-47, wherein the lymphoma patient is a GCB DLBCL patient or an ABC DLBCL patient.

[0435] 49. The method as described in any one of Examples 1-48, wherein the lymphoma patient is a DHITsig+DLBCL patient or a DHITsig-DLBCL patient. Example

[0436] The examples below use standard techniques, which are common techniques well known to those skilled in the art, unless otherwise described in detail. These examples are intended to be illustrative only.

[0437] 7.1 Example 1: Materials and methods used in Examples 2-4

[0438] 7.1.1 Method

[0439] The Discovery cohort includes the ROBUST clinical trial screening population (NCT02285062, (Nowakowski et al., (2021), ROBUST: A Phase III Study of Lenalidomide Plus R-CHOP Versus PlaceboPlus R-CHOP in Previously Untreated Patients With ABC-Type Diffuse Large B-Cell Lymphoma [ROBUST: A Phase III Study of Lenalidomide Plus R-CHOP vs. Placebo Plus R-CHOP in Previously Untreated Patients with ABC-Type Diffuse Large B-Cell Lymphoma]. Journal of Clinical Oncology) plus a sample of newly diagnosed patients from commercial sources (n = 1208). The validation datasets 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, unless otherwise stated, only patients who received R-CHOP treatment were considered (or, in the MER dataset, patients who received R-CHOP class treatments, including a small number of patients who received MR-CHOP, R-EPOCH, ER-CHOP, RAD-RCHOP, and RCHOP / Zevalin). To compare with LymphGen clustering, the NCI dataset (Schmitz et al., 2018, New England Journal of Medicine, 378(15), 1396-1407) was used.

[0440] 7.1.2 Unsupervised Clustering

[0441] The clustered input data consists of normalized RNAseq gene expression features plus feature scores derived from the gene expression data. The expression features are limited to the genes with the most variation and the highest expression levels in the TPM space. The derived features include GSVA feature scores (Hänzelmann, S. C. (2013). GSVA: gene set variation analysis for microarray and RNA-Seq data [GSVA:微阵列和RNA-Seq数据的基因集变异分析]. BMC Bioinformatics [BMC生物信息学].) (including MSigDB signatures and C1 pathways) and cell type features (Danziger, S. A. (2019). ADAPTS: Automated deconvolution augmentation of profiles for tissue specific cells [ADAPTS:组织特异性细胞特征的自动去卷积增强]. PLoS One [公共科学图书馆·综合], 14(11)). The iClusterPlus method (Mo Q, S. R. (2021). iClusterPlus: Integrative clustering of multi-type genomic data [iClusterPlus:多类型基因组数据的整合聚类]. R package version 1.30.0 [R包1.30.0版]) was applied to the subset data, where the value of K was selected multiple times from 2 to 12. This process was repeated 200 times, and the clustering assignments for each time were recorded. Then, the 200 runs were summarized using a sample-pair-co-clustering frequency matrix, which was 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. Then, hierarchical clustering was used to cluster this sample-pair matrix, using the Ward method and 1 minus the co-clustering frequency as the distance metric, in order to obtain the final clustering results for each K value selection.

[0442] 7.1.3 Classification of Linear Models

[0443] Using the consensus clustering labels (with the A8 sample removed) as the gold standard, a generalized linear model (GLM) classifier was trained with the discovery data. Several choices of the elastic net mixing parameter α were tested, with the goal of maximizing the prediction performance and minimizing the model complexity. The regularization parameter λ was optimized by cross-validation and was chosen as the minimum value that made the classification error rate within one standard error of the minimum value.

[0444] 7.1.4 RNAseq Data Standardization

[0445] The MER and ROBUST datasets were reference-normalized to a subset of the Discovery data, termed the commercial samples, which were fixed as the reference population. For this purpose, sample-level scaling was applied to the TPM RNAseq data using the means of five housekeeper 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. Ultimately, the reference fixed all genes with a mean of 0 and a variance of 1, while all other datasets were converted to gene-level Z-scores relative to the reference population. Because the REMoDL-B dataset is an Illumina BeadArray dataset rather than RNAseq data, a self-normalization method using the aforementioned housekeeper gene scaling steps was applied, followed by gene-level scaling, explicitly setting each gene to a mean of 0 and a variance of 1. This self-normalization method is suitable for the large, representative patient cohorts used in this paper, but may produce unexpected results for small or unrepresentative cohorts. The same self-normalization method was applied to the NCI dataset to evaluate these clusters relative to LymphGen labels.

[0446] The reference normalization method places all data in a uniform numerical space with comparable expression levels (Fig. 14A-Fig. 14B) and allows transferable models to be trained on any dataset and directly applied to any other cohort without reparameterization. It even allows normalization of individual samples without requiring representative batches, and the normalized data is never 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, adapt to the normalized gene expression space by reweighting the decision threshold.

[0447] In practice, no significant batch effect was observed in the combined normalized cohort of all datasets. The normalization method was further validated by comparing gene expression classifiers / features applied to the normalized data with orthogonal non-RNAseq data, demonstrating that the normalization approach preserves relevant biological signals intact. These included comparing the Reddy COO classification with Hans' IHC-based approach, comparing the double-hit gene expression signature (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 labels, and comparing the cell type abundance GSVA score with cell type marker densities from IHC and MIBI. All features derived from the normalized RNAseq data were highly consistent with their corresponding non-RNAseq features.

[0448] 7.1.5 Cell type characteristics

[0449] Cell type-specific features were generated from an LM22 matrix describing 22 functionally defined leukocyte types (Chen B, 2018, Profiling Tumor Infiltrating Immune Cells with CIBERSORT. Methods in molecular biology (Clifton, NJ), 1711, 243–259). The feature matrix was enhanced and customized for DLBCL by adding another cell type representing malignant DLBCL B cells and trained on purified cell populations. Benchmarking the deconvolution results using the enhanced feature matrix identified a high correlation between the abundance of DLBCL-specific cell types and tumor purity, as well as the abundance of CD20+ cells as measured by IHC. Adding 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.

[0450] 7.1.6 Sequencing

[0451] ROBUST, MER, and commercial samples were sequenced according to standard protocols at Expression Analysis, Inc. (Durham, NAT, USA). The Allprep DNA / RNA FFPE kit was used for the simultaneous purification of genomic DNA and total RNA from formalin-fixed, paraffin-embedded (FFPE) tissue sections. RNA-seq libraries (75 PE, 50 M) were constructed using the Illumina TruSeq RNA-Access method.

[0452] WES libraries (tumor at 200x, germline control at 100x) were created using the Agilent SureSelectXT method proposed by Fisher et al. in 2011 (with bead modifications). WGS libraries (tumor at 60x, germline control at 30x) were prepared using the Swift Accel-NGS 2S Plus DNA Library Kit (#21024 or 21096, Swift) (with modifications to the Amure bead cleaning step in the procedure).

[0453] 7.1.7 Data Processing

[0454] Sequencing data was processed using an internal cloud-based platform. This platform ran a Sentieon implementation of GATK best practices (using BWA-mem for alignment) and a Sentieon implementation of Mutect2 (tnhaplotyper). Variants were annotated using the dbnsfp database via SnpEff. For WGS data, copy number aberrations were invoked using Battenberg, and structural variants were invoked using Manta. For WES data, copy number aberrations were invoked using Sclust. Poorly represented structural variants were observed in the WES data. RNA-seq data were aligned using the STAR alignment tool and quantified using Salmon.

[0455] 7.1.8 shRNA knockdown

[0456] The doxycycline (Dox)-inducible shRNA construct was generated by Cellecta (Mountain View, CA, USA) using the pRSITEP-U6Tet-(sh)-EF1-TetRep-2A-Puro plasmid. In short, 293FT cells were co-transfected with a lentiviral packaging plasmid mixture (Celecta, catalog number CPCP-K2A) and the pRSITEP-shRNA construct. Viral particles were collected at 48 and 72 hours post-transfection and then concentrated using Lenti-X concentrate (Takara Bio USA). At infection, cells were incubated overnight with the concentrated viral supernatant in the presence of 8 µg / ml polybrene. Cells were then washed to remove the polybrene. Cells were screened with puromycin (2 µg / ml) for more than one week after infection at 48 hours post-infection before experiments. For knockdown experiments, cells were selected at 1 × 10⁻⁶ cells / ml. 5 Cells were seeded at 100 cells / ml and induced with either 20 ng / ml Dox or DMSO as a control. On day 3 of Dox induction, cells were counted and the culture medium was refreshed with either Dox or DMSO. For proliferation assays, 15,000 cells were seeded into 96-well U-shaped plates, and cell viability was measured for 5 consecutive days using CellTiter-Glo (Promega). The remaining cells were seeded at 5 × 10⁶ cells / ml. 5 Cells were seeded at 100 cells / ml and incubated for 2 days. Cells were then harvested for Western blotting and apoptosis assays. The target sequences for shRNA were: shNT: CAACAAGATGAAGAGCACCAA (SEQ ID NO: 1); shTCF4-13: GAGACTGAACGGCAATCTTTC (SEQ ID NO: 2); shTCF4-14: CACGAAATCTTCGGAGGACAA (SEQ ID NO: 3).

[0457] 7.1.9 Western Blotting

[0458] Cells were lysed using a cell lysis buffer (50 mM TrisHCl pH 7.4, 250 mM NaCl, 0.5% Triton X100, 10% glycerol) supplemented with a Halt protease / phosphatase inhibitor (Thermo Scientific, 78443). The cell lysates were sonicated to break down the nuclei and reduce viscosity caused by the released genomic DNA. Protein concentration was measured using a Bradford protein assay (Bio-Rad). Samples were diluted to the same concentration and then added to NuPAGE LDS sample buffer and 2-mercaptoethanol (final concentration 1.25%), followed by boiling at 95°C for 5 min. Whole-cell lysates were separated on a NuPAG 4–12% Bis-Tris Midi protein gel (Invitrogen) and transferred to a nitrocellulose membrane, then blocked in Intercept® (TBS) blocking buffer (LI-COR). The target protein was detected by incubation overnight at 4°C with the primary antibodies listed below. After washing with 1X TBST, the membrane was incubated for 1 hour at room temperature with either IRDye 800CW goat anti-rabbit IgG or IRDye 680LT goat anti-mouse IgG secondary antibody (1:10,000). After washing with 1X TBST, the bands were visualized using an Odyssey imaging system (LI-COR). Antibody information: TCF4 (Proteintech, 22337-1-AP), MYC (abcam, ab32072), GAPDH (Cell Signaling Technology, 2118L).

[0459] 7.2 Example 2: Unsupervised clustering of DLBCL patients

[0460] Diffuse large B-cell lymphoma (DLBCL) is a heterogeneous and aggressive germinal center B-cell tumor and the most common form of non-Hodgkin lymphoma (NHL). The International Prognostic Index (IPI) for DLBCL is based on clinical risk factors to predict survival outcomes in newly diagnosed DLBCL patients. Patients with an IPI of 3–5 are considered intermediate to high-risk and are typically selected for clinical trials due to their poor prognosis with standard care such as immunochemotherapy (R-CHOP). However, the IPI does not provide biological insights into treatment opportunities for high-risk patients.

[0461] The molecular classification using cells of origin (COO) is well-described, with the activated B-cell (ABC) subtype exhibiting a higher risk of relapse and shorter survival with R-CHOP therapy compared to the germinal center B-cell (GCB) subtype. Two phase 3 randomized controlled trials (PHOENIX and ROBUST) failed to demonstrate that ibrutinib (Younes et al., 2019) or lenalidomide (Nowakowski et al., 2021) in combination with R-CHOP was more effective than R-CHOP alone in this high-risk ABC population. Further examination of the ABC groups receiving R-CHOP revealed variability in clinical outcomes, suggesting underlying disease heterogeneity in COO that prevents it from altering clinical practice. Classification involving chromosomal rearrangements of MYC, BCL2, and / or BCL6 (i.e., so-called “double-hit” and “triple-hit” patients) consistently identifies a subset of high-risk GCB patients, but no approved therapies currently exist for this population.

[0462] Recently, analyses of genetic characteristics, including mutations and copy numbers, have identified new patient clusters. These clusters are based on 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, Aprobabilistic 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.These classifications also include the tumor microenvironment (TME26) (Risueño 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 microenvironmentalsignatures. Cancer discovery, 11(6), 1468-1489), (Steen et al., 2021, The landscape of tumor cellstates and ecosystems in diffuse large B cell lymphoma. Cancer Cell). Despite these advances, a practical pathway for prospectively identifying newly diagnosed high-risk DLBCL patients in a drug-approved and clinically feasible manner has not yet been established.

[0463] This disclosure reveals that a robust clustering method can identify biologically driven subgroups of DLBCL patients, predict patient outcomes, and guide treatment approaches. By defining the profiles of DLBCL subtypes in a robust and meaningful manner, future treatments targeting those subtypes may advance. Identifying patients belonging to specific subtypes can further identify high-risk groups that do not respond well to current treatments (e.g., R-CHOP). Investigating the biological basis of those groups also helps elucidate the underlying mechanisms of these high-risk subtypes.

[0464] This disclosure identifies biologically homogeneous high-risk DLBCL patients through unsupervised clustering of transcriptomic features of tumor and non-tumor cells. Several homogeneous clusters were identified, including a high-risk cluster exhibiting an extreme ABC phenotype primarily driven by the MYC pathway and with low levels of immune infiltration. A gene expression classifier was developed that was able to replicate clinical and biological characteristics in independent cohorts. Overall, cluster A7 showed poor prognosis and retrospectively demonstrated treatment-specific response characteristics in multiple randomized trials, suggesting that high-risk A7 has the potential for clinical trials.

[0465] 7.2.1 Discovery and validation of new clusters and development of gene expression classifiers

[0466] Unsupervised clustering was performed on the Discovery cohort of gene expression-derived data from newly diagnosed DLBCL patients (n = 1208, Table 5), followed by supervised classifier training to identify clusters found in independent datasets (Fig. 1A). Unsupervised clustering yielded eight clusters with distinct molecular patterns (Fig. 1B). These clusters showed varying degrees of association with the COO and TME26 categories (Risueño et al., 2020), but none could be uniquely identified solely by these categories. Cluster A8 was found to be a technical error cluster with poor alignment metrics (Figs. 6A–6D) and was therefore excluded from classifier training and further analysis.

[0467] A multinomial classifier was trained on the discovery dataset to generate a model for identifying each cluster in the independent validation cohort. Cross-validation results showed that the classifier training method performed well, achieving 93% accuracy on the training cohort and sensitivity / positive prediction values ​​within each cluster ranging from 81% to 98% (Figure 7). Because the training data was normalized to the reference population, the classifier could be directly applied to other datasets normalized to that space without retraining parameters or thresholds. The classifier could be applied to any FFPE RNAseq samples normalized in the same way, generating a class label for each case (i.e., no cases were unclassified).

[0468] The classifier was applied 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, randomized, phase 3 trial [The Lancet Oncology, 20(5), 649-662]), identifying seven clusters with reproducible biological characteristics, including the top 50 upregulated / downregulated differentially expressed genes in each cluster, which showed reproducible expression patterns in each cluster (Figure 1C).

[0469] 7.2.2 Clinical outcomes and characteristics of high-risk cluster A7

[0470] Although the clustering findings were made in the absence of clinical outcome data, it was still investigated whether any clusters were associated with poor prognosis. Figures 2A-2I illustrate the association between clusters and survival outcomes in patients receiving R-CHOP, as well as their association with prognostic characteristics. Cluster A7 represents a group of patients enriched with the ABC subtypes who responded worst to R-CHOP among the seven clusters, with an incidence of 19%, 13%, and 11% in ROBUST, MER, and REMoDL-B, respectively. The A7 status had significant prognostic significance, with 95% confidence intervals (95% for A7 versus non-A7) of 1.65 (1.08–2.51), 1.87 (1.17–3.00), and 2.00 (1.23–3.20) in ROBUST (ABC only), MER, and REMoDL-B, respectively.

[0471] Although the ABC-COO subtype is associated with increased risk, the high risk of A7 cannot be simply attributed to its ABC enrichment. Even in the ABC-only population, patients with A7 have a higher risk than those without A7 (Figures 2D-2F). Association tests of A7 with known clinical prognostic factors showed that IPI or its components had no significant effect, suggesting that clinical features cannot be used to define A7 (Figures 2G-2I). Cox proportional hazards models showed that A7 status was a significant prognostic factor in both the univariate model (p = 0.027) and the multivariate model combined with IPI (p = 0.047), and the A7+IPI model was slightly more prognostic than IPI alone (ANOVA p = 0.06, Figures 8A-8C). Although A7 is closely associated with both COO and TME26 (both p < 2.2e-16), neither of these features alone or in combination is sufficient to uniquely identify A7. Using the COO or TME26 score as a univariate predictor of A7 eligibility, the predicted AUC in ROBUST and MER were both between 0.82 and 0.86, and the optimized classifier achieved approximately 80% sensitivity and 70% specificity when classifying A7.

[0472] 7.2.3 Biological Explanation of New Clusters

[0473] Each cluster underwent a differential biological examination in terms of single gene expression, DLBCL-specific pathways (Wright et al., 2020), copy number abnormalities, single nucleotide variants, and the tumor microenvironment. Distinctions between these clusters could be identified from the perspectives of COO and TME26 (Figure 3A), although significant heterogeneity remained after tracing these dimensions. Within the COO-TME26 space, three clusters were particularly extreme: low TME GCB-enriched cases found in A2, low TME ABC-enriched cases found in A7, and high TME unclassified enriched cases found in A6.

[0474] Wright et al. used multiple DLBCL-related pathways, allowing for a deeper understanding of the pathways contributing to each cluster from the perspectives of tumor microenvironment, COO, oncogenic pathways, and metabolomics (Figures 3A-3E). The most prominent signals included upregulation of multiple immune-related JAK and NFKB signatures in A6, upregulation of GCB-related signatures (IRF4Dn-1) in GCB enrichment in A2, 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 showed less pronounced gene expression signals, although both exhibited low expression of MYC and G2M checkpoint pathways. The high-risk cluster A7 showed upregulation of ABC-related signatures (IRF4Up-7) and low expression of TME signatures. A7 was highly enriched in ABC subtypes (p < 2.2 × 10⁻⁶). 16 Furthermore, according to Reddy et al.'s scores (data not shown), even in ABC patients, A7 had the most extreme COO score. It was also characterized by upregulation of multiple features (e.g., G2M checkpoint, oxidative phosphorylation, mitotic spindle, and DNA repair), and low expression of p53 and TME features (Figure 4A). Figures 9A-9B illustrate further description of the pathways defining the clustering.

[0475] The enriched genomic features in A7 reflect the ABC enrichment of this cluster, with increased incidence of mutations such as ETV6, PIM1, and OSBPL10 (Figure 3C). However, overall, SNVs did not correlate strongly with our clusters, which is not surprising, as these clusters are derived from transcriptional features that may originate from sources other than SNVs (e.g., copy number and epigenetic changes). Significantly enriched CNAs in each cluster are shown in Figure 3D, where A7-related features include increased arm-level copy numbers on chromosomes 3 and 18. CNA features for each cluster are shown in Figures 10A-10F.

[0476] Immunohistochemical data validated the immune infiltration patterns deduced from gene expression in each cluster. Compared to non-A7, CD3, CD4, and CD8 T cells were consistently reduced, while CD163 monocyte / macrophage, CD68 macrophage, and CD11 dendritic cells showed no clear trend (Fig. 11A–11F). For example, representative MIBI images of the high-TME26 cluster A6 indeed showed high abundance of CD4 / CD8 T cells, while the low-TME26 cluster A7 showed the opposite, exhibiting T cell deficiency and CD20 B cell abundance (Fig. 3E).

[0477] 7.2.4 MYC dysregulation is a key component of the biology of high-risk cluster A7 and can be targeted with TCF4. treatment

[0478] To further define the A7 cluster-specific biological characteristics, GSEA analysis was performed to identify differentially expressed pathways in A7. The cluster showed upregulation of MYC target signatures, E2F target signatures, 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).

[0479] Compared to the non-A7 cluster, MYC gene expression was also upregulated in the A7 cluster (Fig. 4B), and this was not driven by a high tumor cell composition (Fig. 12E). Protein expression quantification by IHC also showed that Myc protein levels were higher in A7 than in non-A7 clusters (Fig. 4C). Although MYC translocation and amplification are known to drive MYC signaling in B-cell lymphoma, neither MYC translocation nor MYC amplification was enriched in A7 (p = 0.99), suggesting that the upregulated MYC activity was driven by other mechanisms.

[0480] Significant enrichment of several arm-level amplifications was observed in A7, with amplifications of chromosomes 18q and 3q showing the highest incidence. Interestingly, 18q12.2 contains the gene TCF4, which encodes a basic helical-loop-helical (bHLH) transcription factor reported to drive MYC gene expression 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 (Fig. 13A–13B). An attempt was then made to characterize TCF4 function using a DLBCL cell line model. Knockdown of TCF4 significantly reduced MYC protein expression in TCF4-amplified cell lines (RIVA and U2932) but did not reduce MYC protein expression in those cell lines that did not undergo TCF4 amplification (SU-DHL-2 and TMD8) (Fig. 4F), suggesting that TCF4 amplification promotes MYC overexpression in ABC DLBCLs. Consistent with these observations, TCF4 knockdown strongly inhibited cell proliferation in TCF4-amplified cell lines (RIVA and U2932), while induction with the same shRNAs only slightly inhibited proliferation in cell lines that had not undergone TCF4 amplification (SU-DHL-2 and TMD8) (Figure 4G). In summary, TCF4 amplification-dependent overexpression stimulates MYC expression and makes ABC DLBCL dependent on overexpressed TCF4. TCF4 may be a potential therapeutic target for the A7 population.

[0481] 7.2.5 Clinical Utility of A7 Clustering

[0482] To assess the utility of A7 as a predictive patient population, ROBUST and REMoDL-B patients were retrospectively stratified according to A7 and non-A7 status (Figures 5A-5B). Results showed that both the ROBUST and REMoDL-B trials demonstrated differences between the control and experimental groups within the A7 population (p values ​​of 0.0088 and 0.16, respectively), indicating that the A7 cluster is a more homogeneous and reliable high-risk patient population for drug development. With the molecular and biological understanding of DLBCL tumors and TME extending beyond COO, the field is poised to transform clinical practice through high-risk patient selection, innovative trial design, and targeted therapies. Here, a high-risk patient population was identified in newly diagnosed DLBCL patients through unsupervised clustering of transcriptomic data. Behind the clinically high-risk behavior of A7 lie three known biological features leading to poor prognosis: extreme ABC subtypes, low immune infiltration (particularly low CD4 and CD8 T cell counts), and upregulated MYC pathway.

[0483] Comparison with recently published molecular classifications reveals that the A7 cluster possesses some unique characteristics, but is not mutually exclusive with other clusters. Immune infiltration exhaustion is a hallmark of A7, a feature shared with the “exhausted” population (DP) (Kotlov et al., 2021) and lymphoma ecotype 1 (LE1) (Steen et al., 2021), which exhibit similar unfavorable survival characteristics.

[0484] Cluster A7 also shared enrichment with genetic subtypes C5 and MCD for previously defined features, including amplifications of chromosomes 3p, 3q, and 18q, and mutations in PIM1, ETV6, and OSBPL10 (based on MYD88). L265P Co-occurrence with CD79B mutations, as described in Schmitz et al., 2018, New England Journal of Medicine, 378(15), 1396-1407. Co-occurrence of MCD and A7 clusters was investigated in the NCI dataset (Schmitz et al., 2018, New England Journal of Medicine, 378(15), 1396-1407), where the LymphGen tag was publicly available. Approximately one-third of patients identified as A7 or MCD were also identified as another type, which was a statistically significant overlap (p = 0.038). Although both A7 and MCD clusters 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 (Figs. 14A-14B).

[0485] For A7 patients, the MCD subtype (based on MYD88) L265P Co-occurrence with CD79B mutations (as described by Schmitz et al., 2018, New England Journal of Medicine, 378(15), 1396-1407) was enriched; while for GCB-like clusters A2 and A3, the EZB subtype (based on EZH2 mutations and BCL2 translocations, as described by Schmitz et al., 2018, New England Journal of Medicine, 378(15), 1396-1407) was enriched (Fig. 15A-Fig. 15D). Although there were statistically significant associations among the various classification methods (Fisher p = 0.0005 in ROBUST, p = 0.001 in MER), there was significant heterogeneity, and no clear one-to-one mapping was found between any subtypes. Furthermore, after standardization, the PCA plots of the Discovery, MER, and REMoDL-B datasets did not show dataset-specific differences (Fig. 16). Mutation profiles (Chapuy genes) were also measured, and sorted as follows: by mutation count (Fig. 17A), by significance (corrected for gene length) (Fig. 17B), and by Chapuy plot (for reference) (Fig. 17C). The expression of proteins encoded by genes on chromosome 18 was also assessed (Figs. 18A–18D).

[0486] Upregulated MYC pathway 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), although the mechanisms differ between GCB and ABC subtypes. In GCB, chromosomal rearrangements of the MYC and BCL2 to IG loci are the main drivers of MYC and BCL2 overexpression. 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 diffuselarge B-cell lymphoma patients treated with rituximab-CHOP. Modern Pathology, 28(12), 1555-1573).The study also showed that MYC expression was 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 found in A7, with no difference in MYC translocation or copy number increase between A7 and non-A7 cases, but both gene and protein expression of MYC were elevated in A7. The study examined the concept that certain MYC regulators (such as TCF4, which is amplified as part of the 18q increase) are responsible for this increase. Data from the ABC cell line confirmed this link and indicated that TCF4 acts as a therapeutic target for A7 (Fig. 4E-Fig. 4G). Other MYC regulators may also share similar functional effects.

[0487] Other pathway changes specific to A7 include upregulation of the G2M checkpoint, mitotic spindle checkpoint, and DNA repair pathway (Figure 4A), indicating cell cycle dysregulation and DNA replication stress. Combined with downregulation of the p53 pathway, these changes may lead to rapid proliferation and genomic instability, supported by uncontrolled growth and significant copy number alterations (Figure 3D). Another important feature of A7 is the upregulation of oxidative phosphorylation, suggesting that in a hypoxic microenvironment, tumors alter energy metabolism by utilizing oxidizable substrates, such as fatty acids. Both of these phenomena have been reported as molecular markers of subsets of DLBCL (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), and 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 new therapeutic strategies and targets for this high-risk population.

[0488] In fact, the poor prognosis of A7 necessitates a different approach than R-CHOP. An example of the R2-CHOP regimen is presented here, which significantly improved outcomes in A7 patients, although R2-CHOP was not specifically designed for this population (Figure 5A). Lenalidomide is a cerebellar protein modulator with a dual action – autonomous antiproliferative effect on tumor B cells and immune-mediated cytotoxicity (Garciaz et al., 2016, Lenalidomide for the treatment of B-cell lymphoma. Expert opinion on investigational druggs, 25(9), 1103-1116). This immunomodulatory activity is particularly beneficial for “cold tumors” (such as those represented by A7). Ibrutinib also performed well in the A7 cluster, suggesting that the extreme ABC biological characteristics of A7 interact with BCR modulators. These examples demonstrate the utility of A7 gene classifier tools in patient selection and the potential clinical value of this high-risk patient population.

[0489] Finally, these biomarkers for identifying A7 patients possess several attractive attributes from a practical application perspective. First, the gene expression assay is easy to administer and has demonstrated its feasibility in clinical trial settings, as seen in ROBUST, with a turnaround time of 2.4 days. Second, it uses diagnostic FFPE tissue, eliminating the need for additional biopsies. Third, it categorizes all patients as A7 or non-A7, avoiding the ambiguity of unclassifiable populations. Future DLBCL treatment, similar to the paradigm shift in AML with FLT3 and IDH2 inhibitors, involves using practical assays to determine risk and treatment, enabling targeted therapy to patient populations defined at the molecular level. This work, along with the recent emergence of DLBCL classification tools, marks significant progress in this direction.

[0490] Table 5. Demographic and clinical characteristics of the discovery and validation cohort.

[0491]

[0492] 7.3 Example 3: Classified subgroups A1 to A8

[0493] Comprehensive clustering identified eight subgroups (designated A1-A8) of ndDLBCL patients. The resulting clusters were analyzed by examining various biological characteristics, including genetic features such as double-hit (DHIT+) and TMD (TME+) traits. The incidence and biological characteristics (e.g., COO type, DHITsig positivity, TME trait positivity, BCL2 / BCL6 translocation, and MYC translocation) of the replication dataset (MER dataset) are summarized in Table 6. The resulting cluster identification can predict the likelihood of response to standard therapy (e.g., R-CHOP combination therapy) and recommend appropriate targeted therapies based on cluster-specific biological characteristics.

[0494] Table 6. Occurrence and biological characteristics of duplicated datasets (MER datasets)

[0495]

[0496] This clustering approach allows for transcriptomic identification of eight patient subgroups (e.g., subgroups A1 to A8). For example, within the identified patient groups, subgroup A7 is a high-risk subgroup for which standard R-CHOP therapy is ineffective.

[0497] Patients in this high-risk subgroup A7 showed i) low expression of TME signatures, including low levels of immune cell infiltration; 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) mixed B-cell lineage signatures; and v) upregulation of B-cell transcription factors (e.g., IRF4 and OCT-2).

[0498] 7.4 Example 4: Cell proportions in subgroups

[0499] Total T cell counts in different T cells were measured in a subset of DLBCL patients from the Discovery-2 dataset (n = 46) using multiplex ion beam imaging (MIBI). These patients were from the identified A7 and non-A7 subgroups. Figures 11A-11F show the total cell counts from patients from the different subgroups.

[0500] As should be understood from the foregoing, although specific embodiments have been described herein for illustrative purposes, various modifications may be made without departing from the spirit and scope of the content provided herein. All references mentioned above are incorporated herein by reference in their entirety.

Claims

1. A method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) Clustering reference lymphoma patients into subgroups using the expression levels of at least one gene in reference biological samples from reference lymphoma patients; (b) Determine 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; (c) Based on a subgroup of lymphoma patients, predict the responsiveness of the lymphoma patient to the first cancer treatment.

2. The method of claim 1, further comprising administering a second cancer treatment to the lymphoma patient based on a subgroup of the lymphoma patient.

3. The method of claim 1 or 2, wherein the at least one gene comprises two or more genes, and step (a) comprises generating clustering information defining the relationship between the expression levels of two or more genes in these reference biological samples, and rearranging the heatmap display based on the clustering information.

4. The method of any one of claims 1-3, wherein step (a) includes employing a hierarchical method or a non-hierarchical method.

5. The method of any one of claims 1-3, wherein step (a) includes using the iClusterPlus method.

6. The method of any one of claims 1-5, wherein these reference lymphoma patients are clustered into 2-12 subgroups.

7. The method of claim 6, wherein these reference lymphoma patients are clustered into 7 subgroups.

8. The method of any one of claims 1-7, wherein the method further comprises training a classifier model using the expression levels of the at least one gene in the reference biological samples.

9. The method of any one of claims 1-8, wherein the at least one gene is selected from the genes in Table 2, and optionally wherein the at least one gene comprises one, two, three, four, five or more genes in Table 2.

10. The method of claim 9, wherein the at least one gene comprises all the genes in Table 2.

11. The method of any one of claims 1-8, wherein the at least one gene is selected from the genes in Table 3, and optionally wherein the at least one gene comprises one, two, three, four, five or more genes in Table 3.

12. The method of claim 11, wherein the at least one gene comprises all the genes in Table 3.

13. The method of any one of claims 1-8, wherein the at least one gene is selected from the genes in Table 4, optionally wherein the at least one gene comprises one, two, three, four, five or more genes in Table 4.

14. The method of claim 13, wherein the at least one gene comprises all the genes in Table 4.

15. The method of any one of claims 8-14, wherein the classifier model is a grouped multinomial generalized linear model (GLM).

16. The method of any one of claims 8-15, wherein the classifier model is a binary generalized linear model.

17. The method of any one of claims 6-16, wherein the method further comprises setting a threshold confidence level for at least one of the subgroups to exclude patients who give clustering data with lower confidence levels from at least one of the subgroups.

18. The method of any one of claims 1-17, 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, marginal zone B-cell lymphoma, lymphoplasmacytic lymphoma, anaplastic large cell lymphoma, primary cutaneous lymphoma, mycosis fungoides, chronic lymphocytic leukemia, and mantle cell lymphoma.

19. The method of claim 18, wherein the lymphoma is DLBCL.

20. The method of claim 18, wherein the lymphoma is an indolent B-cell lymphoma, a marginal zone B-cell lymphoma, a mantle cell lymphoma, or a chronic lymphocytic leukemia.

21. The method of any one of claims 6-20, wherein the reference lymphoma patients are clustered into seven subgroups A1-A7, and subgroup A7 includes about 0% to about 10% of germinal center B-cell-like (GCB) DLBCL patients, about 80% to about 90% of activated B-cell-like (ABC) DLBCL patients, about 0% to about 10% of TME+ DLBCL patients, and about 5% to about 15% of DHITsig+ DLBCL patients.

22. The method of claim 21, wherein (i) subgroup A1 comprises approximately 50% to 60% of GCB DLBCL patients, approximately 30% to 40% of ABC DLBCL patients, approximately 10% to 20% of TME+DLBCL patients, and approximately 30% to 40% of DHITsig+DLBCL patients; (ii) subgroup A2 comprises approximately 80% to 90% of GCB DLBCL patients, approximately 0% to 5% of ABC DLBCL patients, approximately 15% to 25% of TME+DLBCL patients, and approximately 25% to 35% of DHITsig+DLBCL patients; (iii) subgroup A3 comprises approximately 40% to 55% of GCB DLBCL patients, approximately 30% to 45% of ABC DLBCL patients, approximately 40% to 50% of TME+DLBCL patients, and approximately 20% to 30% of DHITsig+DLBCL patients; (iv) Subgroup A4 comprises 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 comprises 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; and / or (vi) Subgroup A6 comprises 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.

23. The method of any one of claims 1-22, wherein the method includes predicting that lymphoma patients identified as belonging to subgroup A7 are unlikely to respond to the first cancer treatment.

24. The method of any one of claims 1-23, wherein the first cancer treatment is a combination therapy of rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone (R-CHOP).

25. The method of any one of claims 2-24, wherein the second cancer treatment is not R-CHOP.

26. The method of any one of claims 2-24, wherein the second cancer treatment is a combination of lenalidomide, rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R2-CHOP).

27. The method of claim 26, wherein the second cancer treatment is a bromine domain and superterminal (BET) inhibitor or a cyclin-dependent kinase (CDK) inhibitor.

28. A method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) In a biological sample from a lymphoma patient, determine the expression level of at least one gene from Table 2, Table 3, or Table 4, optionally wherein the at least one gene comprises one, two, three, four, five, or more genes from Table 2, Table 3, or Table 4; (b) Compare the expression level of the at least one gene from step (a) with the expression level of the at least one gene in a reference biological sample from a reference lymphoma patient who has responded to the cancer treatment, and wherein 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, it indicates that the lymphoma patient may have responded to the cancer treatment.

29. A method for predicting the responsiveness of lymphoma patients to cancer treatment, the method comprising: (a) In a biological sample from a lymphoma patient, determine the expression level of at least one gene from Table 2, Table 3, or Table 4, optionally wherein the at least one gene comprises one, two, three, four, five, or more genes from Table 2, Table 3, or Table 4; and (b) compare 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 from a lymphoma patient who has responded to treatment for the cancer and (ii) the expression level of the at least one gene in a biological sample from a lymphoma patient who has not responded to treatment for the cancer, wherein if the expression level in (a) is similar to the expression level in (i), it indicates that the first lymphoma patient is likely to respond to treatment for the cancer; and if the expression level in (a) is similar to the expression level in (ii), it indicates that the first lymphoma patient is unlikely to respond to treatment for the cancer.

30. A method for treating a patient with lymphoma, the method comprising: (i) Identifying lymphoma patients who may respond to cancer treatment according to the method of claim 28 or 29; (ii) administering the cancer treatment to the lymphoma patient.

31. A method for treating a patient with lymphoma, the method comprising: (i) Identifying lymphoma patients who are unlikely to respond to cancer treatment, according to the method of claim 28 or 29; (ii) administering alternative cancer treatment to the lymphoma patient.

32. The method of claim 30, wherein the cancer treatment is R-CHOP.

33. The method of claim 31, wherein the alternative cancer treatment is not R-CHOP.

34. The method of claim 31, wherein the alternative cancer treatment is R2-CHOP.

35. The method of claim 31, wherein the alternative cancer treatment is a BET inhibitor or a CDK inhibitor.

36. The method of any one of claims 28-35, wherein the lymphoma is selected from the group consisting of: DLBCL, indolent B-cell lymphoma, follicular lymphoma, small lymphocytic lymphoma, marginal zone B-cell lymphoma, lymphoplasmacytic lymphoma, anaplastic large cell lymphoma, primary cutaneous lymphoma, mycosis fungoides, chronic lymphocytic leukemia, and mantle cell lymphoma.

37. The method of claim 36, wherein the lymphoma is DLBCL.

38. The method of claim 36, wherein the lymphoma is an indolent B-cell lymphoma, follicular lymphoma, marginal zone B-cell lymphoma, mantle cell lymphoma, or chronic lymphocytic leukemia.

39. The method of any one of claims 28-38, wherein the expression levels of all genes in Table 2, Table 3 or Table 4 are determined in step (a) and compared in step (b).

40. The method of any one of claims 1-39, wherein the biological sample and the reference biological sample are tumor biopsy samples.

41. The method of any one of claims 1-40, wherein the expression level of the at least one gene is determined by the mRNA level of the at least one gene.

42. The method of any one of claims 1-41, wherein the expression level of the at least one gene is determined by detecting the presence or amount of at least one complex in the biological sample or reference biological sample, wherein the presence or amount of the at least one complex indicates the expression level of the at least one gene.

43. The method of claim 42, wherein the at least one complex is a hybridization complex or is detectably labeled.

44. The method of any one of claims 1-41, wherein the expression level of the at least one gene is determined by detecting the presence or amount of at least one reaction product in the biological sample or reference biological sample, wherein the presence or amount of the at least one reaction product indicates the expression level of the at least one gene.

45. The method of claim 44, wherein the at least one reaction product is detectably labeled.

46. ​​The method of any one of claims 1-45, wherein the reference lymphoma patient is a patient with refractory DLBCL, a patient with relapsed DLBCL, or a newly diagnosed patient with DLBCL.

47. The method of any one of claims 1-46, wherein the lymphoma patient is a patient with refractory DLBCL, a patient with relapsed DLBCL, or a patient with newly diagnosed DLBCL.

48. The method of any one of claims 1-47, wherein the lymphoma patient is a GCB DLBCL patient or an ABCDLBCL patient.

49. The method of any one of claims 1-48, wherein the lymphoma patient is a DHITsig+DLBCL patient or a DHITsig-DLBCL patient.

Images