Methods for selecting cancer patients for immunotherapy using subclonal mutation analysis

Abstract

The present disclosure provides methods for predicting whether a patient diagnosed with cancer will benefit from treatment with an immune checkpoint inhibitor based on subclonal mutation levels.

Description

Atty. Dkt. No.: 642631-0101 (pmda24-027) METHODS FOR SELECTING CANCER PATIENTS FOR IMMUNOTHERAPY USING SUBCLONAL MUTATION ANALYSIS CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of and priority to U.S. Provisional Appl. No. 63 / 569,978 filed March 26, 2025, which is incorporated herein by reference in its entirety for any and all purposes. TECHNICAL FIELD

[0002] The present technology relates generally to methods for predicting whether a patient diagnosed with cancer will benefit from treatment with an immune checkpoint inhibitor based on the level of subclonal tumor mutations determined using clonal structure identification through pairwise penalization (herein referred to as “CliPP”). STATEMENT OF GOVERNMENT SUPPORT

[0003] This invention was made with government support under W81XWH-22-1-0258 awarded by the Medical Research and Development Command, and CA268380 awarded by the National Institutes of Health. The government has certain rights in the invention. BACKGROUND

[0004] The following description of the background of the present technology is provided simply as an aid in understanding the present technology and is not admitted to describe or constitute prior art to the present technology.

[0005] Mutations in tumor cells accumulate over time, from the mutations present in a single cancer-initiating cell to the subsequent myriad of mutations present across a mature tumor. As tumors keep accumulating mutations, they are comprised of heterogeneous subpopulations of cells, i.e., (sub)clones, with additional and distinct mutations from those that were inherited from the ancestral cancer cell. SUMMARY OF THE PRESENT TECHNOLOGY

[0006] In one aspect, the present disclosure provides a method for selecting a cancer patient for treatment with an immune checkpoint inhibitor includes: (a) detecting a level of tumor subclonal mutations at or above a predetermined threshold in a biological sample obtained from the cancer patient; and (b) administering to the cancer patient the immune checkpoint inhibitor.Atty. Dkt. No.: 642631-0101 (pmda24-027)

[0007] Detecting the level of tumor subclonal mutations may include using clonal structure identification through pairwise penalization. The predetermined threshold may be determined for a cancer type using a recursive partitioning survival tree model analyzing at least ten biological samples obtained from different patients with the cancer type. Detecting the level of tumor subclonal mutations may include applying, as input to an objective function, at least a metric of cellular prevalence of a given nucleotide variation in the biological sample and a corresponding proportion of cancer cells amongst all cells associated with the biological sample. Detecting the level of tumor subclonal mutations may include detecting one or more clusters of cellular prevalence data relating to the biological sample, across a plurality of nucleotide variations of the biological sample, that satisfy a criterion of homogeneity.

[0008] In another aspect, the present disclosure provides a method for prolonging survival of a cancer patient including administering to the cancer patient an immune checkpoint inhibitor, wherein the tumor subclonal mutations level in a biological sample obtained from the cancer patient are at or above a predetermined threshold determined by comparing survival outcome to tumor subclonal mutations level in a cohort of cancer patients that have the same type of cancer as the cancer patient.

[0009] The predetermined threshold may be determined using clonal structure identification through pairwise penalization. The predetermined threshold may be determined for a cancer type using a recursive partitioning survival tree model analyzing at least ten biological samples obtained from different patients with the cancer type.

[0010] In any of the above aspects, the cancer patient may have a solid tumor with a tumor mutation burden (TMB) of less than 10. The cancer patient may have one or more of prostate cancer, a pheochromocytoma tumor, a thymoma tumor, chromophobe renal cell carcinoma, IDH1-mutant low grade glioma, prostate adenocarcinoma, adrenocortical adenocarcinoma, pancreatic adenocarcinoma, sarcoma, renal cell carcinoma, renal papillary cell carcinoma, ovarian cancer, liver hepatocellular carcinoma, endometrioid uterine cancer, endometrioid cervical cancer, colorectal cancer, head and neck squamous cell carcinoma, lung adenocarcinoma, lung squamous cell carcinoma, acute myeloid leukemia, thyroid carcinoma, and melanoma. The level of tumor subclonal mutations may be determined via next- generation sequencing or microarray. The level of tumor subclonal mutations may be determined via targeted next-generation sequencing.Atty. Dkt. No.: 642631-0101 (pmda24-027)

[0011] In any of the above aspects, the immune checkpoint inhibitor may include a programmed cell death protein 1 (PD-1) inhibitor, a programmed death-ligand 1 (PD-L1) inhibitor, a cytotoxic T-lymphocyte associated protein 4 (CTLA-4) inhibitor, or any combination thereof.

[0012] In any of the above aspects, the immune checkpoint inhibitor may include one or more of ipilimumab, tremelimumab, cadonilimab, zalifrelimab, pembrolizumab, nivolumab, cemiplimab, sintilimab, tislelizumab, toripalimab, camrelizumab, geptanolimab, toripalimab, zimberelimab, penpulimab, serplulimab, prolgolimab, balstilimab, retifanlimab, cadonilimab, pucotenlimab, sasanlimab, cetrelimab, tebotelimab, pidilizumab, dostarlimab, atezolizumab, durvalumab, avelumab, sugemalimab, or envafolimab.

[0013] In any of the above aspects, the methods may additionally include sequentially, simultaneously, or separately administering to the cancer patient an effective amount of a radiation therapy.

[0014] In any of the above aspects, the methods may additionally include sequentially, simultaneously, or separately administering to the cancer patient an effective amount of a chemotherapeutic agent. The chemotherapeutic agent may include one or more of alkylating agents, topoisomerase inhibitors, endoplasmic reticulum stress inducing agents, antimetabolites, mitotic inhibitors, nitrogen mustards, nitrosoureas, alkyl sulfonates, platinum agents, taxanes, vinca agents, anti-estrogen drugs, aromatase inhibitors, ovarian suppression agents, cytostatic alkaloids, cytotoxic antibiotics, antimetabolites, endocrine / hormonal agents, and bisphosphonate therapy agents. The chemotherapeutic agent may include one or more of cyclophosphamide, fluorouracil (or 5-fluorouracil or 5-FU), methotrexate, edatrexate (10- ethyl-10-deaza-aminopterin), thiotepa, carboplatin, cisplatin, taxanes, paclitaxel, protein- bound paclitaxel, docetaxel, vinorelbine, tamoxifen, raloxifene, toremifene, fulvestrant, gemcitabine, irinotecan, ixabepilone, temozolomide, topotecan, vincristine, vinblastine, eribulin, mutamycin, capecitabine, anastrozole, exemestane, letrozole, leuprolide, abarelix, buserlin, goserelin, megestrol acetate, risedronate, pamidronate, ibandronate, alendronate, denosumab, zoledronate, trastuzumab, tykerb, anthracyclines (e.g., daunorubicin and doxorubicin), cladribine, midostaurin, bevacizumab, oxaliplatin, melphalan, etoposide, mechlorethamine, bleomycin, microtubule poisons, annonaceous acetogenins, chlorambucil, ifosfamide, streptozocin, carmustine, lomustine, busulfan, dacarbazine, temozolomide, altretamine, 6-mercaptopurine (6-MP), cytarabine, floxuridine, fludarabine, hydroxyurea, pemetrexed, epirubicin, idarubicin, SN-38, ARC, NPC, campothecin, 9-nitrocamptothecin, 9-Atty. Dkt. No.: 642631-0101 (pmda24-027) aminocamptothecin, rubifen, gimatecan, diflomotecan, BN80927, DX-8951f, MAG-CPT, amsacrine, etoposide phosphate, teniposide, azacitidine (Vidaza), decitabine, accatin III, 10- deacetyltaxol, 7-xylosyl-10-deacetyltaxol, cephalomannine, 10-deacetyl-7-epitaxol, 7- epitaxol, 10-deacetylbaccatin III, 10-deacetyl cephalomannine, streptozotocin, nimustine, ranimustine, bendamustine, uramustine, estramustine, mannosulfan, camptothecin, exatecan, lurtotecan, lamellarin D9-aminocamptothecin, amsacrine, ellipticines, aurintricarboxylic acid, HU-331, or combinations thereof.

[0015] In any of the preceding embodiments of the methods disclosed herein, the methods further comprise sequentially, simultaneously, or separately administering to the cancer patient an additional anti-cancer therapy, optionally wherein the anti-cancer therapy comprises one or more of chemotherapy, targeted therapy, immunotherapy, radiation therapy, or surgery, optionally wherein the targeted therapy comprises a VEGF / VEGFR inhibitor, EGF / EGFR inhibitor, PARP inhibitor, or a combination thereof or optionally wherein the immunotherapy comprises an immune checkpoint inhibitor therapy.

[0016] In any of the above aspects, the biological sample may include plasma, blood, serum, or biopsied tissue. In some embodiments, the cancer patient has castration-resistant prostate cancer. In some embodiments, the biological sample may have a tumor mutation burden below a threshold, wherein the tumor mutation burden is calculated as the number of mutations per million bases. The threshold may be 10. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] FIGS.1A-1E: Subclonal reconstruction and CliPP model overview. FIG.1A is a visual representation of tumor evolution with genotype circles indicating representative somatic single-nucleotide variants (SNVs). Dots a-c represent SNVs. FIG.1B is an example of the cellular prevalence (CP) distribution of SNVs characterized by one clonal cluster (A) and two subclonal clusters (B and C). FIG.1C shows input data and the core regularization function for the CliPP model, a computational tool for subclonal reconstruction. FIG.1D illustrates how CliPP infers clusters of SNVs and CP values using a penalty function that varies over a penalty parameter λ. The y-axis corresponds to CP estimates (0 to 1) connecting the nearest CP estimates over 100 evenly spaced λ values (0.01 to 0.25). The vertical dotted line indicates the λ used in CliPP for finding the solution. Somatic SNV clusters distinguishing parental cancerous cells from new cancerous cells converge for homogeneity pursuit in CPs. FIG.1E illustrates the objectives for utilizing the CliPP-basedAtty. Dkt. No.: 642631-0101 (pmda24-027) tumor subclonal reconstruction results from 9,212 tumor samples across 38 cancer types in both the Cancer Genome Atlas (TCGA) and International Cancer Genome Consortium Pan- cancer Analysis of Whole Genomes (ICGC-PCAWG or PCAWG).

[0018] FIGS.2A-2F: CliPP benchmarking using simulation data and real patient cohorts. FIG.2A shows paired violin and boxplots representing total error score distributions comparing CliPP and PhyloWGS performance on simulated samples. These scores reflect combined metrics for accuracy in estimating the number of clusters, fraction of clonal mutations, and cellular prevalence. Subplots illustrate error score variations across four features (purity, somatic copy number alterations (CNV) rate, read depth and number of mutation clusters). A lower score indicated better performance. FIG.2B compares CliPP performance to PhyloWGS performance on simulated samples. Violin plots paired with boxplots display the total error score distributions. FIG.2C is a pairwise concordance correlation coefficient (CCC) heatmap of the estimated fraction of clonal mutations by 11 individual reconstruction methods, including CliPP, and the consensus calls on the PCAWG dataset (n=1,548). Color intensity reflects CCC values. FIG.2D is a graph of computational times for subclonal reconstruction on 2,146 PCAWG WGS samples and 9,843 TCGA WES samples. The vertical dashed line corresponds to samples with 5,000 SNVs. Labels indicate average computing times for 5 samples with about 5,000 SNVS, CliPP (1 minute) and PyClone-VI (4 minutes). Other methods’ timings: PhylogicNDT (37 minutes), PyClone (2,765 minutes), and PhyloWGS (2,450 minutes). FIG.2E illustrates computational time comparison relative to CliPP in fold change on PCAWG samples with ≤ 5,000 SNVs. Each dot represents one of the 35 samples where all five methods were applied. FIG.2F Projected time comparison between CliPP and Pyclone-VI in TCGA samples categorized into four quartiles based on the number of SNVs. Open dots correspond to the median time per quartile for each method. Error bars depict the 95% confidence intervals. CliPP’s error bars are too small to show at this scale.

[0019] FIGS.3A-3C: The subclonal landscape of 7,708 tumors in TCGA. FIG.3A illustrates coverage and read depth differences between whole genome sequencing (WGS) and whole-exome sequencing (WES) data. SNVs observed in samples sequenced by both methods are shown in red for cross-platform subclonal reconstruction comparison. FIG.3B includes violin plots displaying inter-rater agreement per sample of subclonal reconstruction results by CliPP from WES in TCGA and WGS in PCAWG, ordered by median agreement score (black bar). Cancer types with higher TMB (shown in red in top panel) tend to haveAtty. Dkt. No.: 642631-0101 (pmda24-027) higher agreement. All samples exhibited fair or better level of agreement. FIG.3C illustrates the distributions of the subclonal mutation load (sML) across 32 cancer types. The number of tumor samples for each cancer type is indicated above each violin plot. Median values for each cancer type are represented by black bars, while the median across all samples is represented by a red dotted line. Additionally, the sigmoid plot depicts the distributions of TMB in all tumor samples, and the stacked bar plots show the number of clusters for each sample estimated by CliPP across all tumor samples. Violin plots are color- coded in accordance with the Pearson correlation coefficients of sML versus TMB.

[0020] FIGS.4A-4E: Biological and clinical implications of sML in TCGA. FIG 4A is a Forest plot of hazard ratios (HRs) and 95% confidence intervals (CIs) from multivariate Cox proportional hazard models for Overall Survival (OS) or Progression-Free Interval (PFI) in TCGA for binarized (low / high) sML. Models are adjusted for age, sML (high versus low), and sex if applicable. Cancer types are ordered by median tumor mutational burden (TMB), for which a sigmoid plot is shown on top. The number of tumor samples for each cancer type is shown above each sigmoid plot. FIG.4B is a Forest plot of HRs and 95% CIs of multivariate Cox proportional hazard models for OS or PFI in TCGA for continuous sML. Models are adjusted for age, sML as a continuous variable, and sex if applicable. Cancer types are ordered by median TMB. FIG.4C is a Forest plot of HRs and 95% CIs of multivariate Cox proportional hazard models for OS or PFI in TCGA for continuous sML and TMB. Models are adjusted for age, sML as a continuous variable, TMB, and sex if applicable. An interaction term between sML and TMB is included if at least marginally significant (α=0.1). Cancer types are ordered by median TMB. FIG.4D is a Forest plot of HR and 95% CI of multivariate Cox proportional hazard model for PFI in thyroid cancer in TCGA. The model is adjusted as FIG.4C. FIG.4E includes Kaplan-Meier curves of PFI for TCGA thyroid cancer samples which are categorized first by TMB alone, and then by TMB and sML. P-values are obtained by log-rank tests. FIG.4F is a scatter plot of Shannon Index versus sML in TCGA, where the color variations represent samples with different numbers of subclones.

[0021] FIGS.5A-5J: Additional validation with non-TCGA patient cohorts and a data-driven hypothesis. FIG.5A includes boxplots with paired data points showing tumor mutational burden (TMB) distribution for all PCAWG and TCGA samples, along with their esophageal cancer and prostate adenocarcinoma subsets and the mCRPC samples in the clinical trial NCT02113657 (unfiltered). The dashed line represents the TMB thresholdAtty. Dkt. No.: 642631-0101 (pmda24-027) commonly used for receiving immunotherapy. FIG.5B includes Kaplan-Meier curves depicting Overall Survival (OS) in esophageal adenocarcinoma patients from TCGA (left) and PCAWG (right), divided by high and low sML. FIG.5C is a graph of hazard ratio and 95% confidence interval for the significant terms in a Cox proportional hazard model for the PCAWG ESAD samples. The first model (A) was fit with age, sex, TMB (as a continuous variable) and sML (as a continuous variable), with only continuous sML remaining significant. The second model (B) was fit with age, sex, TMB (dichotomized), and sML (dichotomized), with both sML and TMB showing significant effects. FIG.5D is a schematic of the design of the mCRPC clinical trial (NCT02113657). Patients with radiographic / clinical progression-free survival (rcPFS) > 6 months were classified as favorable; patients with rcPFS < 6 months and OS < 12 months were considered unfavorable; all other patients were indeterminate. FIG.5E is a boxplot showing distributions of sML in the favorable versus the unfavorable group. FIG.5F is a scatter plot showing a weak, non- significant negative correlation between sML and TMB (log10 transformed). FIG.5G shows clinical outcomes in patients stratified by sML=0.5. Alluvial plot comparing the sML category assignment (high vs low) and annotated treatment response (favorable, indeterminate, unfavorable), and the corresponding Kaplan-Meier curves of rcPFS. P-values were obtained by log-rank tests. FIG.5H is a graph of overall survival in patients stratified by sML=0.5. FIG.5I is a boxplot showing distributions of CD8 density (cells / mm2) in high versus low sML patient groups (n=9 vs.12). P-value of two-sided Wilcoxon rank sum test is shown. FIG.5J illustrates a data-driven hypothesis to summarize the implications of the CliPP-based large-scale subclonal analysis on cancer evolution. Significance levels are denoted as follows: *P < 0.05, **P < 0.01 and ***P < 0.001.

[0022] FIGS.6A-6B: Overview of the TCGA data analysis. FIG.6A is a CONSORT diagram for data processing TCGA data. FIG.6B provides acronyms for 38 cancer types analyzed herein.

[0023] FIG.7 is a flow chart showing the general scheme for an expansive clinical study validating CliPP to classify patients with cancer for alternative target therapies as compared to standard of care treatment.

[0024] FIG.8 is a block diagram of an example of a system to perform subclonal mutation detection.Atty. Dkt. No.: 642631-0101 (pmda24-027)

[0025] FIGS.9A-9F: Association of sML with response to ICB in clinical trial cohorts. FIG.9A provides boxplots of TMB distributions for all PCAWG and TCGA samples, prostate cancer samples from PCAWG and TCGA, and for castration-resistant metastatic prostate cancer (mCRPC) samples from clinical trials NCT02113657 and NCT02703623. Dashed line represents the TMB=10. FIG.9B provides a Kaplan-Meier (KM) plot showing rcPFS in patients from the ipilimumab (IPI) monotherapy stratified by sML, with log-rank test P-value shown. FIG.9C provides a KM plot showing failure-free survival (FFS) among patients from the combination abiraterone acetate, prednisone, and apalutamide (ARPi) and ipilimumab therapy (ARPi+IPI therapy), stratified by sML, with log-rank test P-value shown. FIG.9D provides a Forest plot depicting hazard ratios for Cox proportional hazard (PH) models for patients from FIGS.9B and 9C. Displayed hazard ratios are from the final selected model, after performing bi-direction stepwise variable selection, with age, sex, sML and TMB as the base model, and sML × TMB interaction, Programmed cell death ligand 1 (PD-L1) density, purity, ploidy, and coverage as candidate predictors. FIG.9E provides a boxplot quantifying distributions of CD8 T-cell density (cells / mm2) from immunohistochemistry staining in high versus low sML patient groups (n=8 vs.13) from NCT02113657. P-value of the two-sided Wilcoxon rank sum test is shown. FIG.9F provides a boxplot quantifying distributions of PD-L1 density from immunohistochemical assay in high versus low sML patient groups (n=8 vs.13) from IPI monotherapy. P-value of the two-sided Wilcoxon rank sum test is shown. Significance levels are denoted as follows: *P < 0.05, ** P < 0.01 and *** P < 0.001. DETAILED DESCRIPTION

[0026] It is to be appreciated that certain aspects, modes, embodiments, variations, and features of the present methods are described below in various levels of detail in order to provide a substantial understanding of the present technology. It is to be understood that the present disclosure is not limited to particular uses, methods, reagents, compounds, compositions, or biological systems, which can, of course, vary. It is also to be understood that the terminology used herein is for the purpose of describing particular embodiments only, and is not intended to be limiting.

[0027] In practicing the present methods, many conventional techniques in molecular biology, protein biochemistry, cell biology, immunology, microbiology, and recombinant DNA are used. See, e.g., Sambrook and Russell eds. (2001) Molecular Cloning: A Laboratory Manual, 3rd edition; the series Ausubel et al. eds. (2007) Current Protocols inAtty. Dkt. No.: 642631-0101 (pmda24-027) Molecular Biology; the series Methods in Enzymology (Academic Press, Inc., N.Y.); MacPherson et al. (1991) PCR 1: A Practical Approach (IRL Press at Oxford University Press); MacPherson et al. (1995) PCR 2: A Practical Approach; Harlow and Lane eds. (1999) Antibodies, A Laboratory Manual; Freshney (2005) Culture of Animal Cells: A Manual of Basic Technique, 5th edition; Gait ed. (1984) Oligonucleotide Synthesis; U.S. Patent No. 4,683,195; Hames and Higgins eds. (1984) Nucleic Acid Hybridization; Anderson (1999) Nucleic Acid Hybridization; Hames and Higgins eds. (1984) Transcription and Translation; Immobilized Cells and Enzymes (IRL Press (1986)); Perbal (1984) A Practical Guide to Molecular Cloning; Miller and Calos eds. (1987) Gene Transfer Vectors for Mammalian Cells (Cold Spring Harbor Laboratory); Makrides ed. (2003) Gene Transfer and Expression in Mammalian Cells; Mayer and Walker eds. (1987) Immunochemical Methods in Cell and Molecular Biology (Academic Press, London); and Herzenberg et al. eds (1996) Weir’s Handbook of Experimental Immunology. Methods to detect and measure levels of polypeptide gene expression products (i.e., gene translation level) are well-known in the art and include the use of polypeptide detection methods such as antibody detection and quantification techniques. (See also, Strachan & Read, Human Molecular Genetics, Second Edition. (John Wiley and Sons, Inc., NY, 1999)). Definitions

[0028] Unless defined otherwise, all technical and scientific terms used herein generally have the same meaning as commonly understood by one of ordinary skill in the art to which this technology belongs. As used in this specification and the appended claims, the singular forms “a”, “an” and “the” include plural referents unless the content clearly dictates otherwise. For example, reference to “a cell” includes a combination of two or more cells, and the like. Generally, the nomenclature used herein and the laboratory procedures in cell culture, molecular genetics, organic chemistry, analytical chemistry and nucleic acid chemistry and hybridization described below are those well-known and commonly employed in the art.

[0029] As used herein, the term “about” in reference to a number is generally taken to include numbers that fall within a range of 1%, 5%, or 10% in either direction (greater than or less than) of the number unless otherwise stated or otherwise evident from the context (except where such number would be less than 0% or exceed 100% of a possible value).Atty. Dkt. No.: 642631-0101 (pmda24-027)

[0030] As used herein, the “administration” of an agent or drug to a subject includes any route of introducing or delivering to a subject a compound to perform its intended function. Administration can be carried out by any suitable route, including but not limited to, orally, intranasally, intrathecally, parenterally (intravenously, intramuscularly, intraperitoneally, or subcutaneously), rectally, intrathecally, intraocularly, intradermally, transmucosally, iontophoretically, or topically. Administration includes self-administration and the administration by another.

[0031] As used herein, an “alteration” of a gene or gene product (e.g., a marker gene or gene product) refers to the presence of a mutation or mutations within the gene or gene product, e.g., a mutation, which affects the quantity or activity of the gene or gene product, as compared to the normal or wild-type gene. The genetic alteration can result in changes in the quantity, structure, and / or activity of the gene or gene product in a cancer tissue or cancer cell, as compared to its quantity, structure, and / or activity, in a normal or healthy tissue or cell (e.g., a control). For example, an alteration which is associated with cancer, or predictive of responsiveness to intraoperative analgesics, can have an altered nucleotide sequence (e.g., a mutation), amino acid sequence, chromosomal translocation, intra-chromosomal inversion, copy number, expression level, protein level, protein activity, in a cancer tissue or cancer cell, as compared to a normal, healthy tissue or cell. Exemplary mutations include, but are not limited to, point mutations (e.g., silent, missense, or nonsense), deletions, insertions, inversions, linking mutations, duplications, translocations, inter- and intra-chromosomal rearrangements. Mutations can be present in the coding or non-coding region of the gene.

[0032] As used herein, the terms “amplify” or “amplification” with respect to nucleic acid sequences, refer to methods that increase the representation of a population of nucleic acid sequences in a sample. Nucleic acid amplification methods are well known to the skilled artisan and include ligase chain reaction (LCR), ligase detection reaction (LDR), ligation followed by Q-replicase amplification, PCR, primer extension, strand displacement amplification (SDA), hyperbranched strand displacement amplification, multiple displacement amplification (MDA), nucleic acid strand-based amplification (NASBA), two- step multiplexed amplifications, rolling circle amplification (RCA), recombinase- polymerase amplification (RPA) (TwistDx, Cambridge, UK), transcription mediated amplification, signal mediated amplification of RNA technology, loop-mediated isothermal amplification of DNA, helicase-dependent amplification, single primer isothermal amplification, and self- sustained sequence replication (3SR), including multiplex versions or combinations thereof. Copies ofAtty. Dkt. No.: 642631-0101 (pmda24-027) a particular nucleic acid sequence generated in vitro in an amplification reaction are called “amplicons” or “amplification products.”

[0033] As used herein, the terms “cancer” or “tumor” are used interchangeably and refer to the presence of cells possessing characteristics typical of cancer-causing cells, such as uncontrolled proliferation, immortality, metastatic potential, rapid growth and proliferation rate, and certain characteristic morphological features. Cancer cells are often in the form of a tumor, but such cells can exist alone within an animal, or can be a non-tumorigenic cancer cell. As used herein, the term “cancer cells” includes precancerous (e.g., benign), malignant, pre-metastatic, metastatic, and non-metastatic cells. Cancers of virtually every tissue are known to those of skill in the art, including solid tumors such as carcinomas, sarcomas, glioblastomas, melanomas, etc., and circulating cancers such as leukemias. Examples of cancer include, but are not limited to, ovarian cancer, breast cancer, colon cancer, lung cancer, prostate cancer, gastric cancer, pancreatic cancer, cervical cancer, ovarian cancer, liver cancer, bladder cancer, cancer of the urinary tract, thyroid cancer, renal cancer, carcinoma, melanoma, head and neck cancer, esophageal adenocarcinoma, and brain cancer. The phrase “cancer burden” or “tumor burden” refers to the quantity of cancer cells or tumor volume in a subject. Reducing cancer burden accordingly may refer to reducing the number of cancer cells, or the tumor volume in a subject. The term “cancer cell” refers to a cell that exhibits cancer-like properties, e.g., uncontrollable reproduction, resistance to anti- growth signals, ability to metastasize, and loss of ability to undergo programmed cell death (e.g., apoptosis) or a cell that is derived from a cancer cell, e.g., clone of a cancer cell.

[0034] A “composition” is intended to mean a combination of active agent and another compound or composition, inert (for example, a nanoparticle, detectable agent, or label) or active, such as an adjuvant, diluent, binder, stabilizer, buffers, salts, lipophilic solvents, preservative, adjuvant or the like and include carriers, such as pharmaceutically acceptable carriers. In some embodiments, the carrier (such as the pharmaceutically acceptable carrier) comprises, or consists essentially of, or yet further consists of a nanoparticle, such as a polymeric nanoparticle carrier or a lipid nanoparticle that can be used alone or in combination with another carrier, such as an adjuvant or solvent. Carriers also include pharmaceutical excipients and additives proteins, peptides, amino acids, lipids, and carbohydrates (e.g., sugars, including monosaccharides, di-, tri, tetra-oligosaccharides, and oligosaccharides; derivatized sugars such as alditols, aldonic acids, esterified sugars and the like; and polysaccharides or sugar polymers), which can be present singly or in combination,Atty. Dkt. No.: 642631-0101 (pmda24-027) comprising alone or in combination 1-99.99% by weight or volume. Exemplary protein excipients include serum albumin such as human serum albumin (HSA), recombinant human albumin (rHA), gelatin, casein, and the like. Representative amino acid components, which can also function in a buffering capacity, include alanine, arginine, glycine, arginine, betaine, histidine, glutamic acid, aspartic acid, cysteine, lysine, leucine, isoleucine, valine, methionine, phenylalanine, aspartame, and the like. Carbohydrate excipients are also intended within the scope of this technology, examples of which include but are not limited to monosaccharides such as fructose, maltose, galactose, glucose, D-mannose, sorbose, and the like; disaccharides, such as lactose, sucrose, trehalose, cellobiose, and the like; polysaccharides, such as raffinose, melezitose, maltodextrins, dextrans, starches, and the like; and alditols, such as mannitol, xylitol, maltitol, lactitol, xylitol sorbitol (glucitol) and myoinositol. A composition as disclosed herein can be a pharmaceutical composition. A “pharmaceutical composition” is intended to include the combination of an active agent with a carrier, inert or active, making the composition suitable for diagnostic or therapeutic use in vitro, in vivo or ex vivo.

[0035] As used herein, a "control" is an alternative sample used in an experiment for comparison purpose. A control can be "positive" or "negative." For example, where the purpose of the experiment is to determine a correlation of the efficacy of a therapeutic agent for the treatment for a particular type of disease, a positive control (a compound or composition known to exhibit the desired therapeutic effect) and a negative control (a subject or a sample that does not receive the therapy or receives a placebo) are typically employed.

[0036] As used herein, the phrase “derived” means isolated, purified, mutated, or engineered, or any combination thereof. For example, a cell derived from a subject refers to the cell isolated from a biological sample obtained from the subject, and is optionally engineered.

[0037] “Detecting” as used herein refers to determining the presence of a mutation or alteration in a nucleic acid of interest in a sample. Detection does not require the method to provide 100% sensitivity. Analysis of nucleic acid markers can be performed using techniques known in the art including, but not limited to, sequence analysis, and electrophoretic analysis. Non-limiting examples of sequence analysis include Maxam-Gilbert sequencing, Sanger sequencing, capillary array DNA sequencing, thermal cycle sequencing (Sears et al., Biotechniques, 13:626-633 (1992)), solid-phase sequencing (Zimmerman et al., Methods Mol. Cell Biol, 3:39-42 (1992)), sequencing with mass spectrometry such as matrix-Atty. Dkt. No.: 642631-0101 (pmda24-027) assisted laser desorption / ionization time-of-flight mass spectrometry (MALDI-TOF / MS; Fu et al., Nat. Biotechnol, 16:381-384 (1998)), and sequencing by hybridization. Chee et al., Science, 274:610-614 (1996); Drmanac et al., Science, 260:1649-1652 (1993); Drmanac et al., Nat. Biotechnol, 16:54-58 (1998). DNA microarrays may also be used for sequencing. Other examples of sequencing analysis are described herein. Non-limiting examples of electrophoretic analysis include slab gel electrophoresis such as agarose or polyacrylamide gel electrophoresis, capillary electrophoresis, and denaturing gradient gel electrophoresis. Additionally, next generation sequencing methods can be performed using commercially available kits and instruments from companies such as the Life Technologies / Ion Torrent PGM or Proton, the Illumina HiSEQ or MiSEQ, Oxford Nanopore, and the Roche / 454 next generation sequencing system.

[0038] “Detectable label” as used herein refers to a molecule or a compound or a group of molecules or a group of compounds used to identify a nucleic acid or protein of interest. In some embodiments, the detectable label may be detected directly. In other embodiments, the detectable label may be a part of a binding pair, which can then be subsequently detected. Signals from the detectable label may be detected by various means and will depend on the nature of the detectable label. Detectable labels may be isotopes, fluorescent moieties, colored substances, and the like. Examples of means to detect detectable labels include but are not limited to spectroscopic, photochemical, biochemical, immunochemical, electromagnetic, radiochemical, or chemical means, such as fluorescence, chemifluorescence, or chemiluminescence, or any other appropriate means.

[0039] As used herein, the term “effective amount” refers to a quantity sufficient to achieve a desired therapeutic and / or prophylactic effect, e.g., an amount which results in the prevention of, or a decrease in a disease or condition described herein or one or more signs or symptoms associated with a disease or condition described herein. In the context of therapeutic or prophylactic applications, the amount of a composition administered to the subject will vary depending on the composition, the degree, type, and severity of the disease and on the characteristics of the individual, such as general health, age, sex, body weight and tolerance to drugs. The skilled artisan will be able to determine appropriate dosages depending on these and other factors. The compositions can also be administered in combination with one or more additional therapeutic compounds. In the methods described herein, the therapeutic compositions may be administered to a subject having one or more signs or symptoms of a disease or condition described herein. As used herein, aAtty. Dkt. No.: 642631-0101 (pmda24-027) "therapeutically effective amount" of a composition refers to composition levels in which the physiological effects of a disease or condition are ameliorated or eliminated. A therapeutically effective amount can be given in one or more administrations.

[0040] As used herein, the term “excipient” refers to a natural or synthetic substance formulated alongside the active ingredient of a medication, included for the purpose of long- term stabilization, bulking up solid formulations, or to confer a therapeutic enhancement on the active ingredient in the final dosage form, such as facilitating drug absorption, reducing viscosity, or enhancing solubility.

[0041] As used herein, the term “expression” refers to the process by which polynucleotides are transcribed into mRNA and / or the process by which the transcribed mRNA is subsequently being translated into peptides, polypeptides, or proteins. If the polynucleotide is derived from genomic DNA, expression can include splicing of the mRNA in a eukaryotic cell. The expression level of a gene can be determined by measuring the amount of mRNA or protein in a cell or tissue sample. In one aspect, the expression level of a gene from one sample can be directly compared to the expression level of that gene from a control or reference sample. In another aspect, the expression level of a gene from one sample can be directly compared to the expression level of that gene from the same sample following administration of the compositions disclosed herein. The term “expression” also refers to one or more of the following events: (1) production of an RNA template from a DNA sequence (e.g., by transcription) within a cell; (2) processing of an RNA transcript (e.g., by splicing, editing, 5’ cap formation, and / or 3’ end formation) within a cell; (3) translation of an RNA sequence into a polypeptide or protein within a cell; (4) post-translational modification of a polypeptide or protein within a cell; (5) presentation of a polypeptide or protein on the cell surface; and (6) secretion or presentation or release of a polypeptide or protein from a cell. The level of expression of a polypeptide can be assessed using any method known in art, including, for example, methods of determining the amount of the polypeptide produced from the host cell. Such methods can include, but are not limited to, quantitation of the polypeptide in the cell lysate by ELISA, Coomassie blue staining following gel electrophoresis, Lowry protein assay and Bradford protein assay.

[0042] “Next-generation sequencing or NGS” as used herein, refers to any sequencing method that determines the nucleotide sequence of either individual nucleic acid molecules (e.g., in single molecule sequencing) or clonally expanded proxies for individual nucleic acid molecules in a high throughput parallel fashion (e.g., greater than 103, 104, 105or moreAtty. Dkt. No.: 642631-0101 (pmda24-027) molecules are sequenced simultaneously). NGS includes whole genome sequencing (WGS), whole exome sequencing, and targeted sequencing. In one embodiment, the relative abundance of the nucleic acid species in the library can be estimated by counting the relative number of occurrences of their cognate sequences in the data generated by the sequencing experiment. Next generation sequencing methods are known in the art, and are described, e.g., in Metzker, M. Nature Biotechnology Reviews 11:31-46 (2010).

[0043] The terms “polynucleotide”, “nucleic acid” and “oligonucleotide” are used interchangeably and refer to a polymeric form of nucleotides of any length, either deoxyribonucleotides or ribonucleotides or analogs thereof, in modified or unmodified form. Polynucleotides can have any three-dimensional structure and may perform any function, known or unknown. The following are non-limiting examples of polynucleotides: a gene or gene fragment (for example, a probe, primer, EST, or SAGE tag), exons, introns, messenger RNA (mRNA), transfer RNA, ribosomal RNA, ribozymes, cDNA, recombinant polynucleotides, branched polynucleotides, plasmids, vectors, isolated DNA of any sequence, isolated RNA of any sequence, nucleic acid probes and primers. A polynucleotide can comprise modified nucleotides, such as methylated nucleotides and nucleotide analogs. If present, modifications to the nucleotide structure can be imparted before or after assembly of the polynucleotide. The sequence of nucleotides can be interrupted by non-nucleotide components. A polynucleotide can be further modified after polymerization, such as by conjugation with a labeling component. Unless otherwise specified or required, any embodiment of this disclosure that is a polynucleotide encompasses both the double-stranded form and each of two complementary single-stranded forms known or predicted to make up the double-stranded form. A polynucleotide is composed of a specific sequence of four nucleotide bases: adenine (A); cytosine (C); guanine (G); thymine (T); and uracil (U) for thymine when the polynucleotide is RNA. Thus, the term “polynucleotide sequence” is the alphabetical representation of a polynucleotide molecule. This alphabetical representation can be input into databases in a computer having a central processing unit and used for bioinformatics applications such as functional genomics and homology searching. Polynucleotides include, without limitation, single- and double-stranded DNA, DNA that is a mixture of single- and double-stranded regions, single- and double-stranded RNA, RNA that is mixture of single- and double-stranded regions, and hybrid molecules comprising DNA and RNA that may be single-stranded or, more typically, double-stranded or a mixture of single- and double-stranded regions. In addition, polynucleotide refers to triple-strandedAtty. Dkt. No.: 642631-0101 (pmda24-027) regions comprising RNA or DNA or both RNA and DNA. The term polynucleotide also includes DNAs or RNAs containing one or more modified bases and DNAs or RNAs with backbones modified for stability or for other reasons.

[0044] As used herein, the term “overall survival” or “OS” means the observed length of life from the start of treatment to death or the date of last contact.

[0045] As used herein, the term “progression-free survival” or “PFS” means the observed length of time from the start of treatment until disease progression or worsening.

[0046] As used herein, the term “radiographic / clinical PFS” or “rcPFS” means the observed length of time from the start of treatment until radiographic progression, symptoms, initiation of a new treatment, or death occurred. rcPFS may be used as a surrogate indicating overall survival.

[0047] As used herein, the term “failure-free survival” or “FFS” means the observed length of time from a specific event (e.g., diagnosis, treatment start) until a patient experiences a failure event, such as disease progression, relapse, or death. FFS may be used as an endpoint for a clinical trial.

[0048] As used herein, the term “primer” refers to an oligonucleotide, which is capable of acting as a point of initiation of nucleic acid sequence synthesis when placed under conditions in which synthesis of a primer extension product which is complementary to a target nucleic acid strand is induced, i.e., in the presence of different nucleotide triphosphates and a polymerase in an appropriate buffer (“buffer” includes pH, ionic strength, cofactors etc.) and at a suitable temperature. One or more of the nucleotides of the primer can be modified for instance by addition of a methyl group, a biotin or digoxigenin moiety, a fluorescent tag or by using radioactive nucleotides. A primer sequence need not reflect the exact sequence of the template. For example, a non-complementary nucleotide fragment may be attached to the 5′ end of the primer, with the remainder of the primer sequence being substantially complementary to the strand. The term primer as used herein includes all forms of primers that may be synthesized including peptide nucleic acid primers, locked nucleic acid primers, phosphorothioate modified primers, labeled primers, and the like. The term “forward primer” as used herein means a primer that anneals to the anti-sense strand of dsDNA. A “reverse primer” anneals to the sense-strand of dsDNA.Atty. Dkt. No.: 642631-0101 (pmda24-027)

[0049] As used herein, “primer pair” refers to a forward and reverse primer pair (i.e., a left and right primer pair) that can be used together to amplify a given region of a nucleic acid of interest.

[0050] “Probe” as used herein refers to nucleic acid that interacts with a target nucleic acid via hybridization. A probe may be fully complementary to a target nucleic acid sequence or partially complementary. The level of complementarity will depend on many factors based, in general, on the function of the probe. A probe or probes can be used, for example to detect the presence or absence of a mutation in a nucleic acid sequence by virtue of the sequence characteristics of the target. Probes can be labeled or unlabeled, or modified in any of a number of ways well known in the art. A probe may specifically hybridize to a target nucleic acid. Probes may be DNA, RNA, or an RNA / DNA hybrid. Probes may be oligonucleotides, artificial chromosomes, fragmented artificial chromosome, genomic nucleic acid, fragmented genomic nucleic acid, RNA, recombinant nucleic acid, fragmented recombinant nucleic acid, peptide nucleic acid (PNA), locked nucleic acid, oligomer of cyclic heterocycles, or conjugates of nucleic acid. Probes may comprise modified nucleobases, modified sugar moieties, and modified internucleotide linkages. A probe may be used to detect the presence or absence of a target nucleic acid. Probes are typically at least about 10, 15, 20, 25, 30, 35, 40, 50, 60, 75, 100 nucleotides or more in length.

[0051] As used herein, the terms “polypeptide,” “peptide,” and “protein” are used interchangeably herein to mean a polymer comprising two or more amino acids joined to each other by peptide bonds or modified peptide bonds, i.e., peptide isosteres. Polypeptide refers to both short chains, commonly referred to as peptides, glycopeptides, or oligomers, and to longer chains, generally referred to as proteins. Polypeptides may contain amino acids other than the 20 gene-encoded amino acids. Polypeptides include amino acid sequences modified either by natural processes, such as post-translational processing, or by chemical modification techniques that are well known in the art.

[0052] As used herein, a “sample” or “biological sample” refers to a body fluid or a tissue sample isolated from a subject. In some cases, a biological sample may consist of or comprise whole blood, platelets, red blood cells, white blood cells, plasma, sera, urine, feces, epidermal sample, vaginal sample, skin sample, cheek swab, sperm, amniotic fluid, cultured cells, bone marrow sample, tumor biopsies, aspirate and / or chorionic villi, cultured cells, endothelial cells, synovial fluid, lymphatic fluid, ascites fluid, interstitial or extracellular fluid and the like. The term "sample" may also encompass the fluid in spaces between cells,Atty. Dkt. No.: 642631-0101 (pmda24-027) including gingival crevicular fluid, bone marrow, cerebrospinal fluid (CSF), saliva, mucus, sputum, semen, sweat, urine, or any other bodily fluids. Samples can be obtained from a subject by any means including, but not limited to, venipuncture, excretion, ejaculation, massage, biopsy, needle aspirate, lavage, scraping, surgical incision, or intervention or other means known in the art. A blood sample can be whole blood or any fraction thereof, including blood cells (red blood cells, white blood cells or leukocytes, and platelets), serum and plasma.

[0053] As used herein, the term “separate” therapeutic use refers to an administration of at least two active ingredients at the same time or at substantially the same time by different routes.

[0054] As used herein, the term “sequential” therapeutic use refers to administration of at least two active ingredients at different times. More particularly, sequential use refers to the whole administration of one of the active ingredients before administration of the other or others commences. It is thus possible to administer one of the active ingredients over several minutes, hours, or days before administering the other active ingredient or ingredients. There is no simultaneous treatment in this case.

[0055] As used herein, the term “simultaneous” therapeutic use refers to the administration of at least two active ingredients by the same route and at the same time or at substantially the same time.

[0056] As used herein, a “sample” refers to a substance that is being assayed for the presence of a mutation in a nucleic acid of interest. Processing methods to release or otherwise make available a nucleic acid for detection are well known in the art and may include steps of nucleic acid manipulation. A biological sample may be a body fluid or a tissue sample. In some cases, a biological sample may consist of or comprise blood, plasma, sera, urine, feces, epidermal sample, vaginal sample, skin sample, cheek swab, sperm, amniotic fluid, cultured cells, bone marrow sample, tumor biopsies, aspirate and / or chorionic villi, cultured cells, and the like. Fresh, fixed, or frozen tissues may also be used. In one embodiment, the sample is preserved as a frozen sample or as formaldehyde- or paraformaldehyde-fixed paraffin-embedded (FFPE) tissue preparation. For example, the sample can be embedded in a matrix, e.g., an FFPE block or a frozen sample. Whole blood samples of about 0.5 to 5 ml collected with ethylenediaminetetraacetic acid (EDTA), acid- citrate-dextrose solution (ACD) or heparin as anti-coagulant are suitable.Atty. Dkt. No.: 642631-0101 (pmda24-027)

[0057] As used herein, the term “subclonal mutation” means a mutation that is present in the DNA of a subset of tumor cells in a tumor sample or biopsy.

[0058] As used herein, the term “subclonal reconstruction” means reconstituting the subclonal structure from sequencing data, including number of (sub)clones, size of subclones in terms of fraction of cancer cells, and genotype of the subclones, as well as their phylogenetic relationships. Subclonal reconstruction uses read counts and variant allele frequencies (VAF) to segregate groups of variants according to their cellular proportions within tumors. The computational strategies for subclonal reconstruction encompass a range of statistical and mathematical frameworks.

[0059] As used herein, the terms “subject,” “individual,” or “patient” are used interchangeably and refer to an individual organism, a vertebrate, or a mammal and may include humans, non-human primates, rodents, and the like (e.g., which is to be the recipient of a particular treatment, or from whom cells are harvested). In certain embodiments, the individual, patient, or subject is a human.

[0060] “Substantially” or “essentially” means nearly totally or completely, for instance, 95% or greater of some given quantity. In some embodiments, “substantially” or “essentially” means 95%, 96%, 97%, 98%, 99%, 99.5%, or 99.9%.

[0061] As used herein, the term “therapeutic agent” is intended to mean a compound that, when present in an effective amount, produces a desired therapeutic effect on a subject in need thereof.

[0062] “Treating” or “treatment” as used herein covers the treatment of a disease or disorder described herein, in a subject, such as a human, and includes: (i) inhibiting a disease or disorder, i.e., arresting its development; (ii) relieving a disease or disorder, i.e., causing regression of the disorder; (iii) slowing progression of the disorder; and / or (iv) inhibiting, relieving, or slowing progression of one or more symptoms of the disease or disorder. Therapeutic effects of treatment include, without limitation, inhibiting recurrence of disease, alleviation of symptoms, diminishment of any direct or indirect pathological consequences of the disease, preventing metastases, decreasing the rate of disease progression, amelioration or palliation of the disease state, and remission or improved prognosis. By “treating a cancer” is meant that the symptoms associated with the cancer are, e.g., alleviated, reduced, cured, or placed in a state of remission.Atty. Dkt. No.: 642631-0101 (pmda24-027)

[0063] The term “tumor mutation burden” or “TMB” as used herein means the number of genetic mutations found in the coding region of DNA of tumor cells in a tumor sample or biopsy per million bases. A “high TMB” is defined as 10 or more mutations per million bases in the coding region of DNA.

[0064] It is also to be appreciated that the various modes of treatment of disorders as described herein are intended to mean “substantial,” which includes total but also less than total treatment, and wherein some biologically or medically relevant result is achieved. The treatment may be a continuous prolonged treatment for a chronic disease or a single, or few time administrations for the treatment of an acute condition.

[0065] Pharmaceutically acceptable salts of compounds described herein are within the scope of the present technology and include acid or base addition salts which retain the desired pharmacological activity and is not biologically undesirable (e.g., the salt is not unduly toxic, allergenic, or irritating, and is bioavailable). When the compound of the present technology has a basic group, such as, for example, an amino group, pharmaceutically acceptable salts can be formed with inorganic acids (such as hydrochloric acid, hydroboric acid, nitric acid, sulfuric acid, and phosphoric acid), organic acids (e.g., alginate, formic acid, acetic acid, benzoic acid, gluconic acid, fumaric acid, oxalic acid, tartaric acid, lactic acid, maleic acid, citric acid, succinic acid, malic acid, methanesulfonic acid, benzenesulfonic acid, naphthalene sulfonic acid, and p-toluenesulfonic acid) or acidic amino acids (such as aspartic acid and glutamic acid). When the compound of the present technology has an acidic group, such as for example, a carboxylic acid group, it can form salts with metals, such as alkali and earth alkali metals (e.g., Na+, Li+, K+, Ca2+, Mg2+, Zn2+), ammonia or organic amines (e.g., dicyclohexylamine, trimethylamine, triethylamine, pyridine, picoline, ethanolamine, diethanolamine, triethanolamine) or basic amino acids (e.g., arginine, lysine and ornithine). Such salts can be prepared in situ during isolation and purification of the compounds or by separately reacting the purified compound in its free base or free acid form with a suitable acid or base, respectively, and isolating the salt thus formed.

[0066] Those of skill in the art will appreciate that compounds of the present technology may exhibit the phenomena of tautomerism, conformational isomerism, geometric isomerism, and / or stereoisomerism. As the formula drawings within the specification and claims can represent only one of the possible tautomeric, conformational isomeric, stereochemical or geometric isomeric forms, it should be understood that the present technology encompasses any tautomeric, conformational isomeric, stereochemical and / or geometric isomeric forms ofAtty. Dkt. No.: 642631-0101 (pmda24-027) the compounds having one or more of the utilities described herein, as well as mixtures of these various different forms.

[0067] Early pan-cancer studies identified an association of tumor total mutation load with patient outcomes, and later with response to immunotherapy. In June 2020, the US Food and Drug Administration (FDA) approved the first histology-agnostic programmed cell death (PD-1) therapy for patients with solid tumors exhibiting a tumor mutation burden (herein “TMB,” calculated as the number of mutations per million bases) of at least 10 (deemed “high-TMB”).

[0068] Herein, the inventors recognized that, as most cancers present less than 10 mutations per million bases, including 88% (6,766 / 7,708) of The Cancer Genome Atlas (TCGA) tumors, it is useful to evaluate the potential clinical impact of the different subclonal structures across cancers with low, moderate, and high TMB. In particular, evaluating clinical outcome using subclonal reconstruction may be useful for treating those 88% of cancer patients with low and moderate TMB (defined herein as those having less than 10 mutations per million bases). A hurdle in realizing such efforts has been the computational burden and poor scalability of existing methods for the reconstruction of subclonal structures using sequencing data generated from heterogeneous bulk tumor samples. Another limitation is the availability of many different panels which vary in the genes targeted, number of genes tested, and genomic space sequenced, which serves as a potential source of variation and a barrier to establishing a single standardized technique.

[0069] The present technology includes systems and methods for analyzing subclonal mutations of tumors, and using that analysis to predict a patient’s prognosis, future response to immunotherapy, and survival outcome. The technology is relevant to many different cancer types, and has demonstrated efficacy of this prediction in 16 different cancer types. These predictions can be used to select cancer patients for treatment with immunotherapy that are predicted to benefit from this treatment.

[0070] The systems and methods herein reconstruct and analyze subclonal architecture using any of the methods disclosed herein. Subclonal reconstruction uses read counts and variant allele frequencies (VAF) to segregate groups of variants according to their cellular proportions within tumors. The computational strategies for subclonal reconstruction encompass a range of statistical and mathematical frameworks. Many subclonal inference methods are based on a Dirichlet Process, including PyClone, PhyloSub, PhyloWGS,Atty. Dkt. No.: 642631-0101 (pmda24-027) CTPsingle, and DPclust. These methods typically employ Markov chain Monte Carlo (MCMC) techniques for estimation, which makes them relatively accurate but computationally heavy.

[0071] In contrast, Variational Bayesian-based Methods, including PyClone-VI, PhylogicNDT, Sci-Clone, Bayclone, Ccube, and others, model the data as a mixture of distributions and employ variational Bayesian techniques to estimate the composition. These methods prioritize computational efficiency and aim to maintain a high degree of accuracy in estimating clonal compositions. However, a key disadvantage of these methods is their reliance on certain assumptions and approximations, which can lead to inaccuracies in modeling highly complex and genetically heterogeneous systems, especially when dealing with sparse or low-quality data. Other approaches include deconvolution with smoothing splines, and the use of a Hidden Markov Model with Poisson and Binomial emission models, exemplified by CloneHD, among others. Together, these methods provide advanced analytical techniques for deciphering the complex genetic structure of tumors.

[0072] Another method for subclonal reconstruction is clonal structure identification through pairwise penalization (herein referred to as “CliPP”), which is an approach based on a regularized maximum likelihood framework. The systems and methods herein can be implemented with a statistical framework using penalized regression to analyze subclonal mutations to provide substantially faster processing while maintaining accuracy comparable to more computationally costly methods. Specifically, the systems and methods may use a regularized likelihood modeling framework imposing a pairwise penalty in the form of smoothly clipped absolute deviation (SCAD) to the parameter estimation per data point. The framework may not require pre-ordering of coefficients. The analysis provides clonal structure identification through pairwise penalization (herein referred to as “CliPP”). CliPP can be implemented with parallel computing of matrix multiplications to further increase process speed.

[0073] CliPP outputs subclonal structure, including the total number of clusters, the total number of single nucleotide variants (SNVs) in each cluster, the estimated cellular prevalence (CP) for each cluster, and the mutation assignment. This output can serve as the basis for predicting a patient’s prognosis, future response to immunotherapy, and survival outcome to develop precision treatment strategies. CliPP addresses some of the challenges of other methods of subclonal reconstruction in a way that improves computational speed per sample by 100-1000-fold as compared to other methods while maintaining comparable accuracy.Atty. Dkt. No.: 642631-0101 (pmda24-027) Methods For Detecting Subclonal Mutation Load in a Cancer Patient

[0074] Currently, there is no universally approved biomarker available to predict immunotherapy response among cancers presenting less than 10 mutations per million bases, which account for the majority (about 88%) of cancers. Therefore, an accurate stratification of cancers with less than 10 mutations per million bases (herein referred to as a low tumor mutation burden or TMB) for predicted immunotherapy response is useful, facilitating personalized therapeutic interventions. Herein, subclonal mutation load (“sML”), determined by subclonal reconstruction of tumor samples, are used to stratify cancers for prediction of immunotherapy response.

[0075] An obstacle to stratifying cancers based on sML is the extraordinary complexity of subclonal reconstruction. Previous attempts to determine sML have been hampered by the computational burden and poor scalability of existing methods for reconstruction of subclonal structures using sequencing data generated from heterogeneous bulk tumor samples. Because of this burden, there none of these existing methods can stratify patients’ prognostic outcomes. Emerging evidence suggests distinct evolutionary paths and immune environments between cancers with high versus low TMB. Selection appears to be much stronger in low mutational load cancers, while only high mutational load cancers show a tendency towards focal loss of heterozygosity of HLA-I as a mechanism of immune evasion. Instead of using TMB, which can only identify immunotherapy candidates in about 12% of cancers, sML can identify immunotherapy candidates across all cancers (including high TMB and low TMB cancers). Herein, sML, a surrogate measure for the latency time between the most recent common ancestor and the latest subclonal event, serves as a robust prognostic and therapeutic biomarker. This biomarker is useful across low- and high- TMB cancers, and can identify patients for immunotherapy among low- to moderate-TMB cancers that would otherwise be missed if only TMB was considered.

[0076] The diagnostic methods of the present technology involve determining sML levels in a biological sample obtained from the subject using CliPP. The CliPP Model

[0077] Conventional methods to address subclonal reconstruction with mixture models, as implemented via Dirichlet processes, are computationally costly, for example when the underlying subclonal structure is complex. CliPP is a statistical framework that addresses this task using penalized regression, drastically reducing the computational burden whileAtty. Dkt. No.: 642631-0101 (pmda24-027) maintaining high accuracy. CliPP ensures sparsity by introducing a pairwise penalty on parameter estimation without the need to pre-order coefficients, an advancement over preceding methods such as fused LASSO and CARDS algorithms. The full list of mathematical notations is provided in Table 1. Table 1 Parameter Definition^^ Tumor purity, takes value between 0 and 1^^ Cancer cell fraction (CCF), takes value between 0 and 1^^ Cellular prevalence (CP), is equal to ^^^^, takes value between 0 and 1^^^^ Number of reads observed with variant alleles covering SNV ^^^^^^ Total number of reads covering SNV ^^^^^^^^ total copy number for tumor cell covering SNV ^^^^^^^^ total copy number for normal cell covering SNV ^^^^^^^^ Copy number of the major allele covering SNV ^^SNV-specific copy number covering SNV ^^, also known asmultiplicity.^^ Tuning parameter that controls the degree of the penalization^^ Total number of SNVs^^ Average read depth for the sample^^ Expected proportion of the variant allele

[0078] Each sample was assumed to be composed of two populations of cells with respect to the ^^-th single-nucleotide variant (SNV): one population containing normal cells or cancer cells without the SNV, and the other population of cancer cells harboring this SNV. The infinite sites assumption was adopted to assume that there is a single origin for each SNV.Within cancer cells, SNV ^^ may occur in one or a few cell populations (FIG. 1A), therefore^^^ was introduced to denote the cancer cell fraction (CCF) of SNV ^^, i.e., the fraction ofcancer cells that carry SNV ^^. A clonal SNV exhibits a CCF of 1.0, indicating 100%presence in tumor cells. The cellular prevalence (CP) of SNV ^^ was defined as ^^^, so that=^^^^^ , where ^^ denotes tumor purity, i.e., the proportion of cancer cells among all cells.Given ^^ SNVs, assume that there exist ^^ groups of SNVs presenting distinct CCFs, where ^^ is significantly smaller than ^^. SNVs from a common cancer cell population or subclonewere expected to have identical CP (or CCF), and correspondingly share the same parameter^^ (or ^^). An established model was followed for observed variant read counts, and theAtty. Dkt. No.: 642631-0101 (pmda24-027) observed number of variant reads ^^^at each SNV was assumed to follow a binomial distribution ^^^^^^^^^^^^^^^^(^^^,^^^), where the total number of reads ^^^follow a Poisson distribution ^^^^^^^^^^^^^^(^^). The expected proportion of the variant allele ^^^was expressed as:

[0079] Hence the observed log-likelihood function, after canceling some constant, follows: ^^(^^) = ∑ே^ୀ^ ^^^^ log ^^ (^^^) + (^^^ − ^^^) log൫1 − ^^(^^^)൯൧ (2)

[0080] ^^^^was used to denote the SNV-specific copy number at the ^^-th SNV. This measurement is distinct from allele-specific copy number when the SNV does not occur in all tumor cells. ^^^ேand ^^^்were used to denote the total copy number for normal and tumor cells, respectively.

[0081] Input data. CliPP uses information of ^^ே,and ^^ as its Parameters ^^^்,^^^், and ^^ can be obtained using CNA-based deconvolution methods such as Allele-Specific Copy number Analysis of Tumors (ASCAT) and ABSOLUTE, whereas ^^^ேis typically set as 2. Parameter ^^^^is not directly observed or estimatable using existing CNA software tools. Following a current convention that proved to be effective, ^^^^was assumed as:

[0082] where^⋅^rounds ^^ to the nearest positive integer. Here, usingas the bound enforces the logical constraint that even if an SNV happened before any copy number event, the number of alleles carrying said SNV cannot exceed the total number of major alleles .

[0083] Output data. CliPP provides the inferred subclonal structure, including the total number of clusters, the total number of SNVs in each cluster, the estimated CP for each cluster, and the mutation assignment, i.e., cluster ID for each mutation. This output can then serve as the basis for various downstream analysis including the inference of phylogenetic trees.

[0084] Parameter estimation using a regularized likelihood. The CPs,=  (^^^,  … ,  ^^ே)may be estimated by increasing the corresponding likelihood for one SNV at a time. However, CliPP may also be used to identify a homogeneous and sparse structure, i.e.,Atty. Dkt. No.: 642631-0101 (pmda24-027) clusters, of the CPs across all SNVs. Penalized estimation is a canonical tool to achieve homogeneity detection and parameter estimation simultaneously. Therefore, a pairwise penalty was introduced to seek such homogeneity in Φ. To facilitate computation, a normalapproximation of the binomial random variable, i.e.,∼ ^^൫0,  ^^^(1 − ^^^)൯ wasemployed. This yielded a standard form of penalized objective function with loss ^^^(^^,  ^^),and the estimator was obtained by reducing it over Φ, given a tuning parameter ^^ > 0 thatcontrols the degree of the penalization:where ^^^ప is the estimated proportion of variant allele at the ^^ -th SNV, and ^^ఒ(⋅) denotes asparsity pursuing penalty to identify the homogeneity structure in Φ, such as LASSO, SCAD, and MCP. Here, a focus was the SCAD penalty which is defined as

[0085] ^^ was set to 3.7. These concave penalties offered sparsity similar to the L1 penalty, allowing them to automatically yield sparse estimates. The SCAD-penalized estimators possess the oracle property in identifying sparsity structure of the data. One can choose another penalty instead of SCAD, such as LASSO.

[0086] Solving the objective function in Equation (4) is nontrivial because the targetvariable ^^ is bounded, i.e., ^^^ = ^^^^^ ∈ [0,1], and the SCAD penalty is not convex. A re-parametrization of ^^ థ^^, ^^^ = log^ିథ^ ,  ^^ = 1, … ,  ^^, was employed to mitigate the boxconstraint, which yields

[0087] Note that ^^'s are monotonic with respect to ^^'s; ^^^ = ^^^ implies ^^^ = ^^^ and viceversa, thus the homogeneity pursuit of ^^ can be achieved by the homogeneity pursuit of ^^. A transformation was performed on the loss function from Equation (4), with an updated penalty function that identified the homogeneity structure in ^^. One may alternativelyAtty. Dkt. No.: 642631-0101 (pmda24-027)enforce ^^ = when< ^^ for some error term ^^. Therefore, the loss function^^ was formulated as follows:

[0088] In order to reduce Equation (7), the alternating direction method of multipliers (ADMM), a popular algorithm for large-scale problems, was employed. Note that SCAD penalty possesses the unbiasedness property, ensuring that it does not shrink large estimated parameters throughout iterations. This property is useful in ADMM algorithms, as biases during the iterations may significantly impact the search for subgroups.

[0089] It is useful for the regularized likelihood-based approach to select a ^^ that balances between over- and under-fitting. Here, the choice of ^^ determined the final number of clusters, with higher values yielding fewer clusters. An ad hoc selection approach was implemented, which focused on the interpretability of the outcome by ensuring that clonal mutations had an estimated CCF around 1. The ^^ selection pipeline operated as follows: for each sample, CliPP was run on the data with 11 different ^^′^^ spanning from 0 to 0.25, specifically: 0.01, 0.03, 0.05, 0.075, 0.1, 0.125, 0.15, 0.175, 0.2, 0.225, 0.25. For eachsample, a score ^^ = ୫ୟ^(^^)ି^௨^^௧௬^௨^^௧௬ was computed. If there were one or more results thatsatisfy ^^ < 0.05, the largest ^^ associated with those results was chosen. If all scores ^^ weregreater than 0.01, the ^^ associated with the smallest ^^ was chosen. Note with applying the CliPP software to any new datasets, the choice of ^^ can be decided by users.

[0090] CliPP has a post-processing pipeline for better biological interpretability. It also allows for down-sampling pre-processing procedure when processing samples with huge amounts of SNVs.

[0091] Several post-filtering steps were implemented to curate the following scenarios: (1) The presence of superclusters, which are clusters with an estimated CCF>>1. This occurrence often correlates with errors in CNA estimates taken as input by CliPP. (2) The presence of insignificant resulting clones, including (a) when the sample has > 2 clusters and the current proportion of clonal mutation ≤ 0.15; (b) small gaps between CP values of any two clusters, i.e., when CP values between any two neighboring clusters < 0.1; and (c) when the number of mutations within a subclone is less than 5% of the total number of mutations in a given sample. CliPP responds to each of these scenarios by merging the affected clusterAtty. Dkt. No.: 642631-0101 (pmda24-027) with its nearest neighboring cluster. The current choice of cutoffs in these filters was trained with a sensitivity analysis using the PCAWG WGS data. These steps can be further modified by users when applying CliPP to a new dataset. CliPP Input Data

[0092] In certain embodiments, CliPP inputs are determined via Next Generation Sequencing techniques or massively parallel sequencing. In some embodiments, high throughput, massively parallel sequencing employs sequencing-by-synthesis with reversible dye terminators. In other embodiments, sequencing is performed via sequencing-by-ligation. In yet other embodiments, sequencing is single molecule sequencing. Examples of Next Generation Sequencing techniques include, but are not limited to pyrosequencing, Reversible dye-terminator sequencing, SOLiD sequencing, Ion semiconductor sequencing, Helioscope single molecule sequencing, Oxford nanopore sequencing, etc. Sequencing may be whole genome sequencing (WGS), whole exome sequencing (WES), or targeted sequencing, which focuses on sequencing specific regions of interest in the genome.

[0093] Nanopore sequencing uses biological or solid-state membranes with nanopores in electrolyte solution. The membrane splits the electrolyte solution into two chambers. A bias voltage is applied for electrophoresis. At high enough concentrations, the electrolyte solution is well distributed, and the voltage drop concentrates near and inside the nanopore, defining a capture region. Inside the capture region, ions have a directed motion that can be recorded as a steady ionic current by placing electrodes near the membrane. DNA or RNA inside the nanopore translocates through via a combination of electro-phoretic, electro-osmotic and / or thermo-phoretic forces. Inside the pore the DNA or RNA occupies a volume that partially restricts the flow of ions, observed as an ionic current drop. Based on this ionic current drop, DNA and RNA can be identified.

[0094] The Ion TorrentTM(Life Technologies, Carlsbad, CA) amplicon sequencing system employs a flow-based approach that detects pH changes caused by the release of hydrogen ions during incorporation of unmodified nucleotides in DNA replication. For use with this system, a sequencing library is initially produced by generating DNA fragments flanked by sequencing adapters. In some embodiments, these fragments can be clonally amplified on particles by emulsion PCR. The particles with the amplified template are then placed in a silicon semiconductor sequencing chip. During replication, the chip is flooded with one nucleotide after another, and if a nucleotide complements the DNA molecule in a particularAtty. Dkt. No.: 642631-0101 (pmda24-027) microwell of the chip, then it will be incorporated. A proton is naturally released when a nucleotide is incorporated by the polymerase in the DNA molecule, resulting in a detectable local change of pH. The pH of the solution then changes in that well and is detected by the ion sensor. If homopolymer repeats are present in the template sequence, multiple nucleotides will be incorporated in a single cycle. This leads to a corresponding number of released hydrogens and a proportionally higher electronic signal.

[0095] The 454TM GS FLXTMsequencing system (Roche, Germany) employs a light- based detection methodology in a large-scale parallel pyrosequencing system. Pyrosequencing uses DNA polymerization, adding one nucleotide species at a time and detecting and quantifying the number of nucleotides added to a given location through the light emitted by the release of attached pyrophosphates. For use with the 454TMsystem, adapter-ligated DNA fragments are fixed to small DNA-capture beads in a water-in-oil emulsion and amplified by PCR (emulsion PCR). Each DNA-bound bead is placed into a well on a picotiter plate and sequencing reagents are delivered across the wells of the plate. The four DNA nucleotides are added sequentially in a fixed order across the picotiter plate device during a sequencing run. During the nucleotide flow, millions of copies of DNA bound to each of the beads are sequenced in parallel. When a nucleotide complementary to the template strand is added to a well, the nucleotide is incorporated onto the existing DNA strand, generating a light signal that is recorded by a CCD camera in the instrument.

[0096] Sequencing technology based on reversible dye-terminators: DNA molecules are first attached to primers on a slide and amplified so that local clonal colonies are formed. Four types of reversible terminator bases (RT-bases) are added, and non-incorporated nucleotides are washed away. Unlike pyrosequencing, the DNA can only be extended one nucleotide at a time. A camera takes images of the fluorescently labeled nucleotides, then the dye along with the terminal 3' blocker is chemically removed from the DNA, allowing the next cycle.

[0097] Helicos's single-molecule sequencing uses DNA fragments with added polyA tail adapters, which are attached to the flow cell surface. At each cycle, DNA polymerase and a single species of fluorescently labeled nucleotide are added, resulting in template-dependent extension of the surface-immobilized primer-template duplexes. The reads are performed by the Helioscope sequencer. After acquisition of images tiling the full array, chemical cleavage and release of the fluorescent label permits the subsequent cycle of extension and imaging.Atty. Dkt. No.: 642631-0101 (pmda24-027)

[0098] Sequencing by synthesis (SBS), like the "old style" dye-termination electrophoretic sequencing, relies on incorporation of nucleotides by a DNA polymerase to determine the base sequence. A DNA library with affixed adapters is denatured into single strands and grafted to a flow cell, followed by bridge amplification to form a high-density array of spots onto a glass chip. Reversible terminator methods use reversible versions of dye-terminators, adding one nucleotide at a time, detecting fluorescence at each position by repeated removal of the blocking group to allow polymerization of another nucleotide. The signal of nucleotide incorporation can vary with fluorescently labeled nucleotides, phosphate-driven light reactions and hydrogen ion sensing having all been used. Examples of SBS platforms include Illumina GA and HiSeq 2000. The MiSeq®personal sequencing system (Illumina, Inc.) also employs sequencing by synthesis with reversible terminator chemistry.

[0099] In contrast to the sequencing by synthesis method, the sequencing by ligation method uses a DNA ligase to determine the target sequence. This sequencing method relies on enzymatic ligation of oligonucleotides that are adjacent through local complementarity on a template DNA strand. This technology employs a partition of all possible oligonucleotides of a fixed length, labeled according to the sequenced position. Oligonucleotides are annealed and ligated and the preferential ligation by DNA ligase for matching sequences results in a dinucleotide encoded color space signal at that position (through the release of a fluorescently labeled probe that corresponds to a known nucleotide at a known position along the oligo). This method is primarily used by Life Technologies’ SOLiDTMsequencers. Before sequencing, the DNA is amplified by emulsion PCR. The resulting beads, each containing only copies of the same DNA molecule, are deposited on a solid planar substrate.

[0100] SMRTTMsequencing is based on the sequencing by synthesis approach. The DNA is synthesized in zero-mode waveguides (ZMWs)-small well-like containers with the capturing tools located at the bottom of the well. The sequencing is performed with use of unmodified polymerase (attached to the ZMW bottom) and fluorescently labeled nucleotides flowing freely in the solution. The wells are constructed in a way that only the fluorescence occurring at the bottom of the well is detected. The fluorescent label is detached from the nucleotide at its incorporation into the DNA strand, leaving an unmodified DNA strand. Methods for Selecting Treatments in Cancer Patients Based on sML

[0101] In one aspect, the present disclosure provides a method for selecting cancer patients for treatment with an immune checkpoint inhibitory therapy comprising: (a) detecting levelsAtty. Dkt. No.: 642631-0101 (pmda24-027) of tumor sML exceeding or meeting a predetermined threshold in a biological sample obtained from the cancer patient; and (b) administering to the cancer patient an effective amount of the immune checkpoint inhibitor therapy. In another aspect, the present disclosure provides a method for prolonging survival of a cancer patient comprising administering to the cancer patient an effective amount of an immune checkpoint inhibitor, wherein the tumor sML level in a biological sample obtained from the cancer patient are at or above a predetermined threshold determined using CliPP comparing survival outcome to tumor sML levels in a cohort of cancer patients that have the same type of cancer as the cancer patient.

[0102] Cancer patients having a large variety of different types of cancer may benefit from the selection and treatment methods disclosed herein. In any embodiment, the cancer type may be one that manifests as a solid tumor. Solid tumors may be heterotypic aggregates of many cell types, including cancer cells, cancer stem cells, connective-tissue cells, and / or immune cells, forming a solid tissue mass. In some embodiments, the cancer patient has a solid tumor with a low to moderate TMB, defined as less than 10 somatic mutations found in the coding region of DNA per million bases in the coding region of DNA of tumor cells in a tumor sample or biopsy. In this cohort of patients, conventional treatment methods would not indicate immunotherapy as an effective treatment, but as described herein, detecting levels of tumor sML may be used to determine which patients within this cohort would benefit from immunotherapy.

[0103] Examples of type of cancer include, but are not limited to, prostate cancer (e.g., metastatic castration resistant prostate cancer), a pheochromocytoma tumor, a thymoma tumor, chromophobe renal cell carcinoma, IDH1-mutant low grade glioma, prostate adenocarcinoma, adrenocortical adenocarcinoma, pancreatic adenocarcinoma, sarcoma, renal cell carcinoma, renal papillary cell carcinoma, ovarian cancer, liver hepatocellular carcinoma, endometrioid uterine cancer, endometrioid cervical cancer, colorectal cancer, head and neck squamous cell carcinoma, lung adenocarcinoma, lung squamous cell carcinoma, acute myeloid leukemia, thyroid carcinoma, esophageal cancer (e.g., esophageal adenocarcinoma), and melanoma. The immune checkpoint inhibitor may include a programmed cell death protein 1 (PD-1) inhibitor, a programmed death-ligand 1 (PD-L1) inhibitor, a cytotoxic T- lymphocyte associated protein 4 (CTLA-4) inhibitor, or any combination thereof. Examples of immune checkpoint inhibitors include ipilimumab, tremelimumab, cadonilimab, zalifrelimab, pembrolizumab, nivolumab, cemiplimab, sintilimab, tislelizumab, toripalimab), camrelizumab, geptanolimab, toripalimab, zimberelimab, penpulimab, serplulimab,Atty. Dkt. No.: 642631-0101 (pmda24-027) prolgolimab, balstilimab, retifanlimab, cadonilimab, pucotenlimab, sasanlimab, cetrelimab, tebotelimab, pidilizumab, dostarlimab, atezolizumab, durvalumab, avelumab, sugemalimab, envafolimab, or combinations thereof.

[0104] The predetermined threshold for determining sML levels with CliPP may be determined for a cancer type using a recursive partitioning survival tree model analyzing at least ten biological samples obtained from different patients with the same cancer type.

[0105] FIG.7 provides a general scheme for selecting treatments in cancer patients based on sML. In step 701, a group of patients 710 are recruited that are diagnosed with one of the following types of cancer: prostate cancer, a pheochromocytoma tumor, a thymoma tumor, chromophobe renal cell carcinoma, IDH1-mutant low grade glioma, prostate adenocarcinoma, adrenocortical adenocarcinoma, pancreatic adenocarcinoma, sarcoma, renal cell carcinoma, renal papillary cell carcinoma, ovarian cancer, liver hepatocellular carcinoma, endometrioid uterine cancer, endometrioid cervical cancer, colorectal cancer, HPV-negative head and neck squamous cell carcinoma, lung adenocarcinoma, lung squamous cell carcinoma, acute myeloid leukemia, thyroid carcinoma, esophageal adenocarcinoma, or melanoma. Biological samples obtained from this group of patients are analyzed to determine sML levels according to the analysis methods disclosed herein.

[0106] In step 703, patients are split into two groups based on their detected sML levels. Patients are stratified into high sML level or low sML level based on the predetermined threshold for each cancer type, as described herein. Patients for which sML predicts favorable immune response are assigned to group 712. Patients for which sML does not predict favorable immune response are assigned to group 714.

[0107] In step 705, patients in group 714, for which sML does not predict favorable immune response are treated with the standard of care therapy. Patients in group 712, for which sML predicts favorable immune response, may be further split into two groups. In group 716, patients with sML levels predicting favorable immune response are treated with immune checkpoint inhibitory therapy along with standard of care therapy. In group 718, patients with sML levels predicting favorable immune response are treated with standard of care therapy, acting as a control group. However, in some embodiments, group 718 may not be included and all patients with sML levels predicting favorable immune response are treated with immune checkpoint inhibitory therapy along with standard of care therapy.Atty. Dkt. No.: 642631-0101 (pmda24-027)

[0108] In any of the methods disclosed herein, the cancer patient may or may not have received a prior anti-cancer therapy, optionally wherein the anti-cancer therapy comprises one or more of chemotherapy, targeted therapy, immunotherapy, radiation therapy, or surgery. In some embodiments, the cancer patient has previously received a prior anti-cancer therapy, optionally wherein the anti-cancer therapy comprises one or more of chemotherapy or radiation therapy. In some embodiments, the cancer patient has previously received radiation therapy. Additionally or alternatively, in some embodiments, the cancer patient is non-responsive or resistant to chemotherapy.

[0109] Radiation therapy may include exposing the cancer patient to high-energy radiation (i.e., having wavelengths of 0.001 millimicrons to 13.6 millimicrons). The radiation therapy may be external beam radiation therapy (EBRT), where radiation is delivered using machines outside the body, or internal radiation, where a small catheter or applicator is positioned in the treatment area to deliver radiation to the treatment area. The EBRT may be 3- dimensional conformal radiation therapy (3D-CRT), where images of the tumor are created with computed tomography or magnetic resonance imaging in order to direct radiation beams at different angles to match the shape of the tumor. The EBRT may be intensity-modulated radiation therapy (IMRT), where the intensity of each radiation beam is individually modulated for better avoidance of nearby normal cells to reduce potential side effects. The EBRT may be stereotactic body radiation therapy (SBRT), where five or fewer treatment sessions are given spaced apart in time (e.g., about 2 weeks), and each session includes precisely delivering a high radiation dose (e.g., 48 Gy to 60 Gy).

[0110] The chemotherapy may comprise one or more of alkylating agents, topoisomerase inhibitors, endoplasmic reticulum stress inducing agents, antimetabolites, mitotic inhibitors, nitrogen mustards, nitrosoureas, alkyl sulfonates, platinum agents, taxanes, vinca agents, anti- estrogen drugs, aromatase inhibitors, ovarian suppression agents, cytostatic alkaloids, cytotoxic antibiotics, antimetabolites, endocrine / hormonal agents, and bisphosphonate therapy agents.

[0111] Specific chemotherapeutic agents include, but are not limited to, cyclophosphamide, fluorouracil (or 5-fluorouracil or 5-FU), methotrexate, edatrexate (10-ethyl-10-deaza- aminopterin), thiotepa, carboplatin, cisplatin, taxanes, paclitaxel, protein-bound paclitaxel, docetaxel, vinorelbine, tamoxifen, raloxifene, toremifene, fulvestrant, gemcitabine, irinotecan, ixabepilone, temozolomide, topotecan, vincristine, vinblastine, eribulin, mutamycin, capecitabine, anastrozole, exemestane, letrozole, leuprolide, abarelix, buserlin,Atty. Dkt. No.: 642631-0101 (pmda24-027) goserelin, megestrol acetate, risedronate, pamidronate, ibandronate, alendronate, denosumab, zoledronate, trastuzumab, tykerb, anthracyclines (e.g., daunorubicin and doxorubicin), cladribine, midostaurin, bevacizumab, oxaliplatin, melphalan, etoposide, mechlorethamine, bleomycin, microtubule poisons, annonaceous acetogenins, chlorambucil, ifosfamide, streptozocin, carmustine, lomustine, busulfan, dacarbazine, temozolomide, altretamine, 6- mercaptopurine (6-MP), cytarabine, floxuridine, fludarabine, hydroxyurea, pemetrexed, epirubicin, idarubicin, SN-38, ARC, NPC, campothecin, 9-nitrocamptothecin, 9- aminocamptothecin, rubifen, gimatecan, diflomotecan, BN80927, DX-8951f, MAG-CPT, amsacrine, etoposide phosphate, teniposide, azacitidine (Vidaza), decitabine, accatin III, 10- deacetyltaxol, 7-xylosyl-10-deacetyltaxol, cephalomannine, 10-deacetyl-7-epitaxol, 7- epitaxol, 10-deacetylbaccatin III, 10-deacetyl cephalomannine, streptozotocin, nimustine, ranimustine, bendamustine, uramustine, estramustine, mannosulfan, camptothecin, exatecan, lurtotecan, lamellarin D9-aminocamptothecin, amsacrine, ellipticines, aurintricarboxylic acid, HU-331, or combinations thereof.

[0112] Examples of antimetabolites include 5-fluorouracil (5-FU), 6-mercaptopurine (6- MP), capecitabine, cytarabine, floxuridine, fludarabine, gemcitabine, hydroxyurea, methotrexate, pemetrexed, and mixtures thereof.

[0113] Examples of taxanes include accatin III, 10-deacetyltaxol, 7-xylosyl-10- deacetyltaxol, cephalomannine, 10-deacetyl-7-epitaxol, 7-epitaxol, 10-deacetylbaccatin III, 10-deacetyl cephalomannine, and mixtures thereof.

[0114] Examples of DNA alkylating agents include cyclophosphamide, chlorambucil, melphalan, bendamustine, uramustine, estramustine, carmustine, lomustine, nimustine, ranimustine, streptozotocin; busulfan, mannosulfan, and mixtures thereof.

[0115] Examples of topoisomerase I inhibitor include SN-38, ARC, NPC, camptothecin, topotecan, 9-nitrocamptothecin, exatecan, lurtotecan, lamellarin D9-aminocamptothecin, rubifen, gimatecan, diflomotecan, BN80927, DX-8951f, MAG-CPT, and mixtures thereof. Examples of topoisomerase II inhibitors include amsacrine, etoposide, etoposide phosphate, teniposide, daunorubicin, mitoxantrone, amsacrine, ellipticines, aurintricarboxylic acid, doxorubicin, and HU-331 and combinations thereof.

[0116] In any and all embodiments of the methods disclosed herein, the methods may further comprise sequentially, simultaneously, or separately administering to the cancer patient an additional anti-cancer therapy, optionally wherein the anti-cancer therapyAtty. Dkt. No.: 642631-0101 (pmda24-027) comprises one or more of chemotherapy, targeted therapy, immunotherapy, radiation therapy, or surgery, optionally wherein the targeted therapy comprises a VEGF / VEGFR inhibitor, EGF / EGFR inhibitor, PARP inhibitor, or a combination thereof or optionally wherein the immunotherapy comprises an immune checkpoint inhibitor therapy.

[0117] In any and all embodiments of the methods disclosed herein, the biological sample comprises plasma, blood, serum, or biopsied tissue. Where the cancer patient has a localized tumor, the biological sample may comprise solid tumor tissue. Where the cancer patient has metastatic tumors, the biological sample may additionally comprise solid tumor tissue from a metastatic tumor site and / or blood. In any of the foregoing embodiments of the methods disclosed herein where the cancer patient has metastatic tumors, the biological sample may comprise circulating tumor RNA (ctRNA). Metrics For Benchmarking Subclonal Reconstruction Accuracy

[0118] The inventors evaluated how accurately CliPP recovers subclonal architecture, as compared to other methods, using three metrics that measure bias in estimated number of clusters, fraction of clonal mutations (clonal fraction), and cellular prevalence (CP) across all variants, respectively.

[0119] Measuring error in estimated number of clusters: The inventors calculated therelative difference in number of clusters ^^^^^^^^ = |ே^ିே^|ே^ , where ^^^ is the estimated numberof clusters and ^^௧the true number of clusters.

[0120] Measuring error in clonal fraction estimates: The inventors calculated the relativedifference in clonal fraction ^^^^^^^^ = |^^ି^^|^^ , where ^^^ and ^^௧ represent the CliPP estimatedclonal fraction and the truth clonal fraction, respectively.

[0121] Measuring error in CP estimates: The inventors further calculated the root mean the total number of SNVs, ^^^ప is theestimated CP, ^^^is the true CP for SNV ^^ , and ^^ is the true purity of the sample which was used to standardize the dynamic ranges of different samples.

[0122] Measuring overall error: The inventors also introduced the total error score =represent the overall performance.Atty. Dkt. No.: 642631-0101 (pmda24-027)

[0123] In all these metrics, smaller values indicated better performance, and 0 indicates a correct reconstruction. To facilitate cross method comparison, The inventors used min-max normalization to force the values to a [0,1] scale. The difference in normalized scores is calculated asଶ൫^^ି^^൯^^ା^^, where ^^^and ^^^are scaled scores for CliPP and PhyloWGS, respectively. A negative normalized difference score suggested better performance in CliPP. EXAMPLES

[0124] The present technology is further illustrated by the following Examples, which should not be construed as limiting in any way. The examples herein are provided to illustrate advantages of the present technology and to further assist a person of ordinary skill in the art with preparing or using the methods of the present technology. The examples should in no way be construed as limiting the scope of the present technology, as defined by the appended claims. The examples can include or incorporate any of the variations, aspects, or embodiments of the present technology described above. The variations, aspects, or embodiments described above may also further each include or incorporate the variations of any or all other variations, aspects, or embodiments of the present technology. Example 1: Pan-cancer subclonal mutation analysis of 7,827 tumors predicts clinical outcome

[0125] In this study, the inventors 1) developed a statistical method called Clonal structure identification through Pairwise Penalization, or CliPP, in order to improve the computational speed per sample by 100-1,000 fold, as compared to existing methods, while maintaining equal accuracy; and 2) generated pan-cancer subclonal structures of pre-treatment tumors with available clinical follow-up data of at least 5 years. The inventors evaluated the clinical impact of subclonal mutational load from 7,708 patient samples in TCGA, 98 esophageal adenocarcinoma samples from the International Cancer Genome Consortium Pan-cancer Analysis of Whole Genomes (ICGC-PCAWG) study, and 21 metastatic castration resistant prostate cancer (mCRPC) samples from an immunotherapy clinical trial. Subclonal mutational load, a surrogate measure for the latency time between the most recent common ancestor and the latest subclonal event was determined to serve as a robust prognostic and therapeutic biomarker, particularly in low- to moderate-TMB cancers.

[0126] Methods. Parallel computing. The Eigen (v3.449) C++ linear algebra library was used, providing efficient multi-threading matrix operations. This approach addressedAtty. Dkt. No.: 642631-0101 (pmda24-027) substantial computational demands, removing a bottleneck. Moreover, by integrating the OpenMP library, CliPP parallelized multiple runs, using resource utilization across modern multi-core systems while exploring different tuning parameters denoted by ^^.

[0127] Benchmarking CliPP performance in simulated datasets. The ability of CliPP to correctly reconstruct the subclonal organization of a tumor was assessed on three simulated datasets. An in-house simulation dataset CliPPSim4k was generated to benchmark the performance in samples with fewer copy number aberrations (CNAs) and higher read depth to cover the scenario of both whole-genome and whole-exome sequencing (WGS, WES) data. CliPP was further assessed using PhylogicSim500, a simulation dataset containing 500 samples where the copy number profiles were sampled from the PCAWG WGS data, with other parameters independently sampled from fixed distributions, and the SimClone1000 dataset, a dataset containing 965 samples that were simulated based on a grid design to cover the spectrum of scenarios encountered in the PCAWG WGS data. CliPP was run on all datasets using the default setting, and PhyloWGS v1.0-rc2 was run on CliPPSim4k using the recommended setting. Available PhyloWGS results were used on the other two simulation datasets. The accuracy of each method was evaluated using three metrics that measure bias in estimated number of clusters (rdNC), fraction of clonal mutations (rdCF), and CP across all variants (RMSE), respectively, and a total error score that averages over the three metrics to represent an overall performance.

[0128] Subclonal reconstruction in PCAWG. The Pan-Cancer Analysis of Whole Genomes (PCAWG) study dataset includes whole-genome sequencing (WGS) data obtained from a cohort of tumor samples spanning 38 cancer types. PCAWG related data were downloaded from the ICGC data portal, using Bionimbus Protected Data Cloud (PDC).

[0129] The PCAWG dataset contains WGS data obtained from 2,778 tumor samples (2,658 distinct donors), including 2,605 primary tumors and 173 metastases or recurrences (minimum average coverage at 30× in the tumor). As input to CliPP, the consensus mutation calls derived from 4 distinct callers and the consensus CNA calls from 6 distinct callers were used.The metric of reads per tumor chromosomal copy (nrpcc) was employed to filter outsamples with insufficient read coverage for subclonal reconstruction.wascalculated, where ^^ is the depth of sequencing, ^^ is the purity and ^^ is the average tumorsample ploidy, i.e., ^^ = ^^^^் + (1is tumor cell ploidy and ^^ே = 2 isnormal cell ploidy. A stringent threshold of ^^^^^^^^^^ ≥ 10 was applied to obtain a high-Atty. Dkt. No.: 642631-0101 (pmda24-027) confidence cohort of 1,993 samples. For 149 samples with >50,000 SNVs, a down-sampling strategy was deployed. The down-sampling strategy included randomly sampling 30,000 SNVs to construct matrices that are within reasonable computer memory limits. In order to cover as many SNVs as possible, the random sampling was repeated 10 times and the CliPP clustering results were merged through kernel smoothing.

[0130] To benchmark the accuracy of CliPP in real data, subclonal reconstruction results from 9 individual methods (Bayclone, Ccube, cloneHD, CTPsingle, DPClust, PhylogicNDT, PhyloWGS, PyClone, Sclust) and the consensus calls, which were available for 1,548 samples, were used. CliPP and PyClone-VI were run on these samples. The comparison was performed based on the proportion of subclonal mutations estimated by each method and summarized by calculating the concordance correlation coefficient (CCC) for each method to each other method and to the consensus subclonal mutation proportion.

[0131] Subclonal reconstruction in TCGA. The Cancer Genome Atlas (TCGA) contains whole-exome sequencing (WES) data from a cohort of tumor samples spanning 32 TCGA cancer types. The TCGA dataset consists of 9,654 WES samples with consensus mutation calls from 5 variant callers and 2 indel callers with matched copy number segments and tumor purity estimates from ASCAT. The CliPP analysis pipeline for TCGA samples included additional steps for quality control and filtering (FIG.6A). In summary, samples were excluded if they had either too few SNVs (resulting in insufficient statistical power) or too many SNVs (indicating potential artifacts). Samples with low read coverage (nrpcc < 10) were excluded. Filters and additional CliPP runs were implemented using its alternative mode for samples with negligible CNA events hence no reliable purity estimates using copy number-based methods. These steps aimed to reduce the potentially large negative impact from inaccurate CNA profiling in WES data. The resulting cohort size in TCGA was n = 7,708 from 32 cancer types. FIG.6B provides acronyms for 38 cancer types analyzed herein.

[0132] Computational speed benchmark. The elapsed real time for each method from initiation to completion was measured on the same machine equipped with Intel(R) Xeon(R) Gold 6,132 CPU @ 2.60GHz, using 28 CPU cores for consistency. In order to compute the speed for PhyloWGS, PyClone, and PhylogicNDT, which are time-consuming to study, a grid-based sampling approach was employed on the PCAWG dataset by selecting five samples across seven predefined SNV grids (100, 500, 1,000, 2,000, 3,000, 4,000 and 5,000 SNVs). In instances of sample scarcity at these exact SNV counts, the nearest equivalentAtty. Dkt. No.: 642631-0101 (pmda24-027) (within a ± 5% window) was selected. For the computational speed benchmark in TCGA WES data a strategy was employed to match the distribution of relatively low SNV numbers to already obtained PCAWG data without further running these methods on more samples.

[0133] Analysis of matched WGS and WES data from the same tumor samples. A set of tumor samples (n=510) was identified that contributed to the generation of both PCAWG WGS and TCGA WES data with neighboring slides through matching study IDs. To provide a fair comparison of the two sequencing platforms, samples with outlier TMBs, i.e., the 3rdquartile+1.5x interquartile range (IQR) of all samples for each cancer type were removed. The cohort comprised 510 samples with 124,901 SNVs before filtering, and after filtering, it consisted of 465 samples with a total of 49,597 SNVs. The CliPP outputs were compared on these samples using Bangdiwala's B statistic, to quantify the agreement of clone / subclonal mutation assignment from WGS and WES data. This statistic can address zero counts in the contingency table better than the commonly used Kappa statistic, which makes it a better fit for this comparison. Bias in disagreement was assessed in the substantial to perfectagreement samples (B ≥ 0.49) and the fair to moderate agreement samples (0.09 ≤ B ≤ 0.49)using McNemar’s test.

[0134] Subclonal mutational load (sML). Subclonal mutational load (sML) wasintroduced as the percentage of subclonal SNVs out of all SNVs. ^^^^^^ = # ^^ ^௨^^^^^^^ ௌே^^.Tumor mutation burden (TMB) was calculated as the total number of SNVs per mega-base (Mb). With WES data, the TMB and sML calculations were restricted to the coding sequence (CDS). CDS was annotated as defined by GENCODE Release 19, and can be updated as new releases become available 38Mb was used as the exome size.

[0135] Association of sML with survival outcomes. To assess clinical relevance of sML through its association with clinical outcomes, the relationship between sML and the patient survival data was characterized (overall survival (OS) and progression-free interval (PFI) on 28 cancer types that presented more than 50 samples). For well-defined clinical cancer subtypes and features, the effect of sML within these patient subgroups was further evaluated, including hormone receptor status in breast cancer, CMS subtype 2 (most frequently occurring) and MSI status in colorectal cancer, adenocarcinoma and squamous cell carcinoma in esophageal cancer, HPV status in head and neck cancer, IDH1 mutation status in low-grade glioma, smoking status in lung cancer, Gleason score in prostate cancer, and endometroid status in uterine and endometrial cancer.Atty. Dkt. No.: 642631-0101 (pmda24-027)

[0136] Appropriate survival measures were identified. sML was associated with clinical outcomes by splitting samples into high and low sML groups using a recursive partitioning survival tree model (rpart) for each cancer type, with the maximum tree depth constrained to 1. To ensure well-balanced grouping, each sML group (“high” versus “low”) for each cancer type had at least 10 samples and ≥ 10% representation. To analyze survival outcomes, Kaplan-Meier (KM) analysis was applied, generating KM curves through the survminer R package. These curves illustrated survival probability across time, stratified by high vs. low sML. Patient outcome stratification in cancer type or subgroups with log-rank test P-values ≤ 0.1 was deemed significant or marginally significant.

[0137] The analysis was furthered on how sML impacts clinical outcomes along with other patient characteristics such as age and sex by utilizing multivariate Cox proportional hazard (PH) models. The PH assumption was deemed appropriate by testing the Schoenfeld residuals for survival data across cancer types. Multivariate CoxPH models were employed in three iterations for all 28 cancer types: 1) using sML as a binary variable, high versus low; 2) using sML as a continuous variable; and 3) using both sML and TMB as continuous variables, plus a model selection step, e.g., with Akaike information criterion (AIC), to determine whether to include their interaction term. Cancer (sub)types were reported where the grouping was balanced as described above for high / low sML, and the estimated hazard ratio for sML and / or its interaction with TMB was statistically significantly different from 1 (P-value < 0.1), excluding cancer types with wide confidence intervals on the hazard ratio. When both OS and PFI showed significant effects, or both cancer type and cancer (sub)type showed significant effects of sML, only one outcome was included in the figures.

[0138] Analysis of WES data from NCT02113657. Raw WES data generated from 30 tumor samples of patients with metastatic castration-resistant prostate cancer (mCRPC) before they received anti-CTLA-4 therapy (ipilimumab) as part of the clinical trial NCT02113657 was obtained. The paired-end FASTQ files were generated on the Illumina HiSeq 2000 after adapter trimming, underwent sequence quality assessment using FASTQC v0.11.8. Subsequently, the reads were aligned to the GRCh38 / hg38 build through BWA mem 0.7.15-r1140. The resulting BAM files were processed according to the Genome Analysis Toolkit (GATK) Best Practices, including steps such as marking duplicates, joint realignment of paired tumor-normal BAMs, and recalibration of base quality scores. Consensus mutation calls were derived as those from 2 out of 3 callers: MuSE 2.0 calls and PASS-filtered results from MuTect2 and Strelka2. Copy number aberrations (CNAs), tumorAtty. Dkt. No.: 642631-0101 (pmda24-027) purity and ploidy were called using ASCAT (v3.1.2) configured for WES data processing. Patients’ response to immunotherapy outcome data as well as immunohistochemistry (IHC) staining for CD8 T cell densities were obtained from a previous trial report. Following the pre-processing pipeline used for TCGA (FIG.6A), samples with nrpcc<10 and TMB>100 were removed, as were samples with less than 10 coding SNVs (i.e., TMB > 0.25) which would otherwise inflate the subclonal fraction estimates. See more details in FIG.5D.

[0139] The clinical trial NCT02113657, conducted at the M.D. Anderson Cancer Center, aimed to assess the impact of the checkpoint inhibitor Ipilimumab on the immune system in men with metastatic castration-resistant prostate cancer receiving hormone therapy. The study involved administering Ipilimumab intravenously at a dose of 3 mg / kg every 3 weeks for a total of 4 doses. Participants continued their standard hormone therapy alongside. The trial sought to enroll up to 30 male patients, aged 18 and above, with various eligibility criteria including histologically or cytologically confirmed carcinoma of the prostate and evidence of metastatic disease. The primary outcome measured was the T-cell response to prostate cancer neoantigens post-treatment.

[0140] The objectives of clinical trial NCT02113657 including determining the impact of Ipilimumab on T cell responses to prostate cancer neoantigens in both primary tumor and metastatic sites in men with metastatic CRPC over 10 weeks. The objectives also included determining T-cell response to neoantigens defined as at least a two-fold increase compared to the response seen against irrelevant control targets, and activating at least 0.1% of cells tested, or at least a two-fold increase compared to baseline. Eligible participants received ipilimumab by vein over 90 minutes at weeks 1, 4, 7, and 10, in addition to the therapy they were already receiving.

[0141] Inclusion criteria for clinical trial NCT02113657 included subjects having histologically or cytologically confirmed carcinoma of the prostate; metastatic prostate cancer mass tissue collection within 3 months of study entry; evidence of metastatic disease on previous bone scan, computed tomography (CT) scan and / or magnetic resonance imaging (MRI); asymptomatic or minimally symptomatic; tumor progression while on hormone therapy at castrate levels serum testosterone (≤ 1.7 nmol / L or 50 ng / dL) defined as biopsy- proven, PSA (prostate-specific antigen) and / or radiographic criteria according to the Prostate Cancer Working Group 2 (PCWG2); castrate levels of testosterone were maintained by surgical or medical means throughout the conduct of the study. ECOG performance status ≤ 1; and normal organ and marrow function as defined below: a) WBC ≥ 2500 / µL, b) ANCAtty. Dkt. No.: 642631-0101 (pmda24-027) (absolute neutrophil count) ≥ 1000 / µL, c) platelets ≥ 75 x 103 / µL, d) hemoglobin ≥ 9 g / dL, e) creatinine ≤ 2.5 × ULN, f) ALT (alanine transaminase) ≤ 2.5 × ULN (upper limit normal) for patients without liver metastases (or patients with liver metastasis ALT ≤ 5 × ULN was allowed), g) bilirubin ≤ 2.5 × ULN (except for patients with Gilbert's Syndrome, who had a total bilirubin ≤ 3mg / dL).

[0142] Exclusion criteria for clinical trial NCT02113657 included treatment with any of the following medications or interventions concomitantly or within 28 days of starting ipilimumab: systemic corticosteroids; external beam radiation therapy or major surgery requiring general anesthetic; any systemic therapy for prostate cancer (with the exception of bisphosphonates and RANK-ligand inhibitors for bone metastases which are allowed) including chemotherapy, secondary hormonal therapies (such as megestrol acetate, diethylstilbestrol, ketoconazole, abiraterone, enzalutamide) and non-steroidal anti-androgens (such as bicalutamide, flutamide or nilutamide); immune modulators, cytokines or vaccines for the management of cancer or non-cancer-related illnesses; any non-oncology vaccine therapy used for prevention of infectious diseases (for up to one month before any dose of ipilimumab); any other investigational product; use of controlled schedule III controlled substances for cancer-related pain control; autoimmune disease; any underlying medical or psychiatric condition, which in the opinion of the Investigator, will make the administration of study drug hazardous or obscure the interpretation of AEs; patients with known brain metastases; uncontrolled intercurrent illness including, but not limited to, ongoing or active infection, history of congestive heart failure, unstable angina pectoris, cardiac arrhythmia, or psychiatric illness / social situations that would limit compliance with study requirements; known HIV, Hepatitis B, or Hepatitis C; untreated symptomatic spinal cord compressions; other malignancies requiring active therapy or known to be associated with altered immune response; and patients who have had a history of acute diverticulitis, intra-abdominal abscess, GI obstruction and abdominal carcinomatosis which are known risk factors for bowel perforation.

[0143] Results. Overview of subclonal reconstruction by CliPP. For a given tumor sample, the variant allele frequency (VAF) spectrum (FIG.1B) exhibited a mixture of multiple peaks corresponding to varying cellular prevalence (CP) of single-nucleotide variants (SNVs). Subclonal reconstruction refers to a procedure to identify cancer cell clones through clustering of variant read counts, adjusting for allele copies and tumor purity, to identify groups of variants with similar CPs. Available conventional methods employ aAtty. Dkt. No.: 642631-0101 (pmda24-027) Bayesian modeling framework, which comes with a very high computational cost, particularly for running WGS data. For example, it can take 1-2 days to analyze one sample. The accuracy of subclonal reconstruction may be better achieved by a consensus approach across callers with distinct modeling strategies to account for different errors while sharing the true signals. The present technology used an orthogonal approach to conventional methods called CliPP, which is rooted in the domain of regularized regression in machine learning. Imposing a pairwise penalty in the form of smoothly clipped absolute deviation (SCAD) to the parameter (e.g., CP) estimation per data point (e.g., mutation) is, in general terms, a modeling advancement over conventional methods. Consequently, in the application of the penalized likelihood statistical framework to subclonal reconstruction, the resulting sparse and homogeneous pattern of CPs across mutations corresponds to a clustering procedure (FIG.1C). FIG.1D illustrates how homogeneity increases as the penalty term (^^) increases, resulting in fewer clusters. CliPP produces the clustered mutations along with their associated homogenized CP values as the output. CliPP offers exceptional computational efficiency owing to three factors: 1) the subclonal reconstruction problem is reformulated as a convex function problem, benefiting from the existence of efficient solvers; 2) conventional Markov chain Monte Carlo methods are generally time-consuming, while CliPP circumvents the need for such procedures; and 3) CliPP utilizes parallel computing of matrix multiplications, which further expedites the process.

[0144] FIG.8 depicts an example of a system 800. The system 800 can be used to implement at least a portion of various techniques described herein, including but not limited to at least a portion of various CliPP techniques to detect subclonal mutations, such as to detect subclonal mutation loads and / or generate outputs regarding selections of subjects for cancer treatment and / or indications of efficacy of cancer treatments.

[0145] One or more components of the system 800 (e.g., clonal structure detector 816; subject selector 820) can include or by implemented by one or more processors 808 and memory 812. The processor 808 may be a general purpose or specific purpose processor, an application specific integrated circuit (ASIC), one or more field programmable gate arrays (FPGAs), a group of processing components, or other suitable processing components. The processor 808 may be configured to execute computer code or instructions stored in memory (e.g., fuzzy logic, etc.) or received from other computer readable media (e.g., CDROM, network storage, a remote server, etc.) to perform one or more of the processes described herein. The memory 812 may include one or more data storage devices (e.g., memory units,Atty. Dkt. No.: 642631-0101 (pmda24-027) memory devices, computer-readable storage media, etc.) configured to store data, computer code, executable instructions, or other forms of computer-readable information. The memory 812 may include random access memory (RAM), read-only memory (ROM), hard drive storage, temporary storage, non-volatile memory, flash memory, optical memory, or any other suitable memory for storing software objects and / or computer instructions. The system 800 or one or more components thereof can be implemented as a hardware processor including a Central Processing Unit (CPU), an Application-Specific Integrated Circuit (ASIC), an Application-Specific Instruction-Set Processor (ASIP), a Graphics Processing Unit (GPU), a Physics Processing Unit (PPU), a Digital Signal Processor (DSP), a Field Programmable Gate Array (FPGA), a Programmable Logic Device (PLD), a Controller, a Microcontroller unit, a Processor, a Microprocessor, an ARM, or the like, or any combination thereof. The memory 812 may include database components, object code components, script components, or any other type of information structure for supporting the various activities and information structures described in the present disclosure. The memory 812 may be communicably connected to the processor 808 and may include computer code for executing (e.g., by processor) one or more of the processes described herein. The memory 812 can include various modules (e.g., circuits, engines) for completing processes described herein. The system 800 can include one or more input and / or output ports for electronic communication with various components of the system 800. The system 800 can be implemented using one or more computing devices that can include various communications and / or user interfaces for communicating data and presenting outputs and / or receiving inputs (e.g., user inputs). The memory 812 can include any one or more functions, routines, code, logic, machine learning models, rules, heuristics, algorithms, or various combinations thereof to implement components such as the clonal structure detector 816 and / or subject selector 820. In some implementations, the one or more processors 808 and memory 812 are implemented as a distributed computing system and / or parallel computing system. In some implementations, at least a subset of the one or more processors 808 include parallel processing units, graphics processing units (GPUs), or various combinations thereof.

[0146] The system 800 can include or be coupled with one or more data sources 804. The data sources 804 can include data regarding one or more tumors, such as sequence data from sequencing of nucleic acids from tumors, tumor samples, and / or cells of tumors. The data sources 804 can include, for example, SNV data, and can include CP data corresponding toAtty. Dkt. No.: 642631-0101 (pmda24-027) the SNV data. The data sources 804 can include, for example, The data sources 804 can be updated to include data from processing by and / or outputs of the system 800.

[0147] The system 800 can include a clonal structure detector 816. The clonal structure detector 816 can retrieve data from the one or more data sources 804 to perform various operations described herein, such as to generate clusters from data of the data sources 804. For example, the clonal structure detector 816 can process sequence data from the data sources 804 to detect one or more clusters of subclonal mutations from the sequence data. A cluster can include various forms of clustered data described herein, such as a group of mutations having similar CPs (e.g., based on processing of variant read counts, adjusting for allele copies and tumor purity, to identify groups of variants with similar CPs).

[0148] The clonal structure detector 816 can include one or more machine learning models to detect the one or more clusters. The machine learning model can be a regression model. The machine learning model can be a regularized regression model. The machine learning model can include or be coupled with an objective function based on a difference (e.g., penalty) between an indication of CP and an indication of an expected proportion of a variant allele (e.g., as described above with respect to Equations 1-7) to be reduced in order to identify the one or more clusters.

[0149] The clonal structure detector 816 can generate an output representative of the one or more detected clusters. For example, the clonal structure detector 816 can generate the output to include one or more characteristics of an inferred subclonal structure corresponding to the one or more detected clusters, including one or more of the total number of clusters, the total number of SNVs in each cluster, the estimated CP for each cluster, and the mutation assignment (e.g., cluster ID) for each mutation.

[0150] In some implementations, the clonal structure detector 816 can determine at least one of an intra-tumor heterogeneity (ITH) or a subclonal mutation load (sML) based on the one or more detected clusters. sML can represent latency time between the most recent common ancestor and the latest subclonal event.

[0151] The system 800 can include at least one subject selector 820. The subject selector 820 can select, according to evaluation of the one or more detected clusters and / or the sML, a corresponding subject for which to provide treatment, e.g., with an immune checkpoint inhibitor. In some implementations, the subject selector 820 compares the sML for a givenAtty. Dkt. No.: 642631-0101 (pmda24-027) subject to a threshold, and selects the given subject for providing the treatment responsive to the sML being greater than the threshold.

[0152] Using three simulation datasets (total n=5,515) covering a wide range of values for features that are known to influence accuracy including tumor purity, percent of genome with copy number alterations (CNA), sequencing read depth, and the number of mutation clusters, CliPP performs similarly or slightly better than PhyloWGS, a conventional method in subclonal reconstruction (FIGS.2A and 2B). Using WGS data from 1,548 patient tumor samples in the PCAWG study, results were further evaluated using 10 existing methods. When compared with these 10 individual methods and their consensus calls in terms of the estimated fraction of clonal mutations, CliPP showed the second highest concordance to the consensus (concordance correlation coefficient = 0.96, FIG.2C), indicating its high accuracy.

[0153] The computational speed of CliPP was compared with those of conventional methods PhyloWGS, PhylogicNDT, PyClone, and PyClone-VI, on TCGA and PCAWG samples. Overall, CliPP finished analyzing 2,778 PCAWG samples within 16 hours, and 9,654 TCGA samples in one hour, presenting at least a 2,000-fold improvement in speed as compared to PyClone and PhyloWGS (FIGS.2D and 2E). While closer to CliPP than PyClone and PhyloWGS, PyClone-VI’s speed remains largely lower than CliPP in samples with more than 5,000 mutations in WGS data, with CliPP showing a median of 4-fold speedup (median absolute deviance [MAD]=2.7 fold). WES data covered only 1% of the genome, and as such, most samples (75%) in TCGA present with less than 200 mutations. It is in samples with lower numbers of SNVs that CliPP had the largest speed advantage (FIG. 2E). Using the processing speeds in PCAWG on samples with a similar distribution of mutation numbers, the speed advantage of CliPP versus PyClone-VI and PhylogicNDT for TCGA was projected to be about 660 times (FIG.2F) and about 1,200 times (by extension using Fig.2E), respectively.

[0154] In summary, the comparable accuracy and substantial speedup achieved by CliPP allowed processing of the large pan-cancer patient cohorts of WES data in TCGA, which had otherwise been prohibited by the very long computing times of conventional methods.

[0155] The subclonal landscape of 7,708 tumors. An updated portrait of the extent of genetic intra-tumor heterogeneity (ITH) across and within tumor types was provided with a larger sample size in TCGA, as compared to the previous two pan-cancer studies with ~1,200Atty. Dkt. No.: 642631-0101 (pmda24-027) tumors and ~2,700 tumors. One bottleneck issue with using TCGA data is that WES data surveys only the coding portion of the cancer genome (FIG.3A) and whether WES can substitute for WGS to provide useful information for cancer evolution remained unclear for researchers designing new studies. With 465 samples from 21 cancer types analyzed by CliPP on both WES and WGS data (profiled by TCGA and PCAWG, FIG.6A), CliPP-based subclonal reconstruction results were highly consistent between the two platforms, with a median inter-rater agreement (B-statistic) of 0.81, indicating close to perfect agreement (FIG. 3B). More mutations changed identity from being clonal in WGS to subclonal in WES, as compared to the opposite direction (McNemar’s test P-values < 0.001). Mutations shifting from WGS-clonal to WES-subclonal were expected, since the higher read coverage provided by WES is more powered to detect deviations from clonality (median ratio of read depth = 1.5 across 49,597 overlapping mutations). In conclusion, WES provided adequate data for subclonal reconstruction, therefore it was appropriate to evaluate subclonality in TCGA, which has a wealth of clinical annotations as compared to PCAWG.

[0156] Across 7,708 tumors from 32 cancer types in TCGA, 93% (n=7,188) had between 1 and 4 clusters. As TMB increased, samples tended to have fewer mutation clusters, indicating less potential for a diverse subclonal structure (FIG.3C). ITH was quantified via subclonal mutational load (sML), the proportion of mutations which are subclonal, merging mutations from multiple subclonal clusters. Consistent with the reported PCAWG results, in TCGA across cancer types, the median sML was moderately negatively correlated with the median TMB (Pearson correlation = -0.55, P-value = 0.001, FIG.3C), but not correlated with the median number of reads per chromosome copy (nrpcc) (Pearson correlation = 0.28, P-value = 0.12). Across samples in each cancer type however, sML was not correlated with TMB (median Pearson correlation = –0.14) but showed weak to moderate correlations with nrpcc (median Pearson correlation = 0.29). Thyroid and thymus cancers, which have low median TMB, exhibited the highest levels of sML. Melanoma, lung squamous, bladder and B-cell lymphoma, which have high TMB, exhibited the lowest levels of sML.

[0157] Subclonal mutation load (sML) is predictive of cancer prognosis. Subclonal mutation load (sML) was shown to be a surrogate measure for the latent time between the most recent common ancestor and the latest subclonal event, and is a biologically and clinically relevant feature for cancer.

[0158] Across TCGA, sML was significantly associated with clinical or pathological features / subtypes, with phenotypes associated with poorer prognosis consistently exhibitingAtty. Dkt. No.: 642631-0101 (pmda24-027) lower sML, including in triple-negative breast cancer, high Gleason score in prostate cancer, smoking status or low PD-L1 expression in lung cancer, and HPV status in head and neck cancer. In contrast, no significant associations were observed between sML and age, sex, TMB, clinical stage, or technical factors such as the size of the tumor biopsy or nrpcc. Together, these observations suggest a unique role of sML as a clinically relevant feature, independent of TMB, stage, or other possible confounders.

[0159] Using clinical outcomes (OS and PFI) in TCGA across 28 cancer types and their subtypes, high sML was associated with better OS / PFI at a pan-cancer level and within 16 (sub)types, with an additional 3 cancer (sub)types demonstrating the opposite trend (FIG. 4A). The three cancer (sub)types where high sML is instead associated with poorer survival are ovarian, lung squamous-smokers and melanoma, which presented the highest levels of genome instability or the highest median TMB among all cancers (FIG.4A, top panel). These effects remained unchanged after adjusting for potential technical confounders. The significant predictive effect of sML persisted for 9 cancer (sub)types as analysis moved from dichotomizing sML to using it as the continuous variable in a Cox model (FIG.4B). Among these cancer types, AML presented an opposite trend for the continuous sML, which may be explained by a different biological underpinning for hematological cancers. As expected, adding TMB as a covariate did not replace the impact of sML (FIG.4C). Using variable selection, significant main effects of sML were found in 7 cancer (sub)types (FIG.4C) and a significant interaction term of sML and TMB in one cancer type, thyroid carcinoma (FIG. 4D). By stratifying the thyroid cancer patients using both TMB and sML, the prognostic effect of sML was prominent within samples with higher TMBs (FIG.4E). This was consistent with the idea that low TMB in thyroid cancers was associated with indolent behavior, thus separation of outcomes was more identifiable in higher TMB tumors, where outcomes were more variable. Replacing sML-based patient classification with TMB-based classification did not provide the same statistical significance, supporting the unique contribution of sML.

[0160] sML patient stratification was also evaluated with the addition of Shannon Index (SI), a diversity measure that increases when there are more mutation clusters and when mutations are more equally divided in each cluster. Given the collinear relationship between sML and SI (FIG.4F), SI only contributed significant additional information in high sML tumors, which represented the minority of TCGA (FIG.3C). For samples with a high subclonal mutation load, a high SI indicated that subclonal mutations belonged to multipleAtty. Dkt. No.: 642631-0101 (pmda24-027) mutation clusters with distinct CPs, whereas a low SI indicated all mutations belonged to the same or a small number of clusters. As these behaviors represented different evolutionary trajectories, the addition of SI partitioned high sML tumors into further distinct survival groups for 11 cancer types. In summary, with low- and moderate-TMB cancers, SI can sometimes be used to complement sML, but overall sML remains an interpretable and representative feature that encodes cancer evolution.

[0161] Validating sML in low to moderate-TMB cancers using independent patient cohorts. Finally, to further validate the clinical observations above, sequencing data was obtained from patient cohorts outside of TCGA, hence not limited to pitfalls in the clinical follow-up data from TCGA. These data were generated from patient samples that presented a range of TMB not meeting the criterion of >10 (high-TMB) for immunotherapy (FIG.5A).

[0162] CliPP was applied to calculate sML in 98 whole-genome sequenced esophageal adenocarcinoma samples from the International Cancer Genome Consortium PCAWG study (FIG.5B). Both TCGA and PCAWG esophageal adenocarcinoma datasets shared similar fractions of high versus low sML samples, with the latter achieving a higher level of statistical significance in distinguishing patients’ OS outcomes (log-rank test P-value = 0.004, FIG.5B), likely due to the larger sample size. Both the continuous and dichotomized sML yielded a significant hazard ratio, even after adjusting for TMB (FIG.5C), and not associated with nrpcc. The association of high sML with better OS in esophageal cancer was shown in an independent, non-WES dataset.

[0163] WES data was obtained from 21 men with mCRPC participating in a clinical trial of ipilimumab (NCT02113657, FIG.5D). mCRPC is a tumor type with low TMB for which benefit from immune checkpoint blockade is restricted to patient subgroups from whom identifying biomarkers remain to be defined. By comparing a subset of patients previously categorized as favorable and unfavorable outcomes (n=7 vs.8), we observed a trend in sML to be higher in the favorable group, although not statistically significant (Wilcoxon test P- value=0.19, FIG.5E). The distribution of sML was not correlated with TMB (Pearson correlation = –0.19, FIG.5F). Using radiographic / clinical PFS (rcPFS) as an endpoint for all patients, including the previously indeterminate group (n=6), a natural cutoff of 0.5 on sML was found that can separate patients into distinct rcPFS outcome groups (n=9 vs.12, log rank P-value = 0.005, FIG.5G), who continued to demonstrate distinct survival outcomes with up to 8 years of follow-up (log rank P-value = 0.03, FIG.5H). These clinical associations were independent of nrpcc, but could be confounded with tissue sites of sample origin, e.g.,Atty. Dkt. No.: 642631-0101 (pmda24-027) prostate versus lymph node or liver. A prostate sample-specific analysis was followed up and still observed significant associations of sML with rcPFS and OS (n = 6 vs.5, log rank P- values = 0.009 and 0.01, respectively). The high versus low sML tumor samples presented significantly different levels of CD8 T-cell densities, as measured by immunohistochemistry (Wilcoxon test P-value=0.008, FIG.5I). CIBERSORTx based immune cell subtype profiling further suggested differences in the myeloid cell populations. In summary, these data indicated that beyond its prognostic value, sML may be a useful marker to identify patients that benefit from immunotherapy.

[0164] Discussion. Accurate characterization of the clonality of single-nucleotide variants across large-scale patient cohorts (7,827 tumors) using CliPP generated useful clinical and biological insights. Among the majority of cancers in TCGA (16 out of 19 cancer types, FIG.5J), unfavorable clinical outcomes were associated with a low subclonal mutational load (sML), found in tumors that have presumably undergone recent clonal sweeps, i.e., positive selection. Favorable clinical outcomes were associated with a high sML, found in tumors that may be slow-growing and have a longer lag time, i.e., long latency, since the last clonal sweep event. Within the three remaining TCGA cancer types, i.e., melanoma, lung squamous cell carcinoma (smoker) and ovarian cancer, those who shared the highest TMB or the highest genome instability (as measured by percent copy number aberrations), had a low sML, corresponding to a high number of clonal mutations, and associated with favorable clinical outcomes. Therefore, this comprehensive evaluation provides evidence supporting distinct evolutionary paths between cancers with high versus low mutational load, and how their divergent evolutionary dynamics may impact clinical outcomes.

[0165] An mCRPC clinical trial study was used to directly evaluate the relationship between sML and immune activities in tumors. Signal in sML was found to identify mCRPC patients with improved survival and progression times after receiving ipilimumab. Therefore, as compared to patients with high-TMB tumors that are conventionally expected to benefit from immunotherapy, this study points to a new target population: patients with low to moderate-TMB tumors (about 90% of all cancers) that may also benefit from immunotherapy, when they present a high subclonal mutational load.

[0166] A high-level summary measure of cancer evolution such as the clonal / subclonal status of single-nucleotide variants, when computed across a large number of patient samples, sheds new light on the relationship between genetic intra-tumor heterogeneity and therapeuticAtty. Dkt. No.: 642631-0101 (pmda24-027) resistance. Highly efficient and accurate methods like CliPP represent a new generation of techniques that bridge the gap between studying tumor biology and developing precision treatment strategies.

[0167] Subclonal mutation load (sML) was identified as a biologically and clinically relevant evolutionary marker in low-to-moderate TMB cancers. Using CliPP to analyze sequencing data from 9,972 tumors across ICGC, TCGA, and two clinical trial cohorts, sML was found to be prognostic of survival (PFI or OS) across 18 cancers. The prognostic effects were in opposite directions between a few high-TMB cancers (n=4) and the remaining majority cancer types (n=14). High sML was also prognostic of immune checkpoint therapy response in two metastatic prostate cancers (low-TMB cancers), where neoantigens alone cannot explain the favorable responses in low TMB tumors. These findings highlight that distinct evolutionary processes and tumor-immune evasion mechanisms may persist in high versus low to moderate TMB cancers.

[0168] Accurate characterization of the clonality of single-nucleotide variants across large-scale patient cohorts (7,827 tumors) using CliPP generated useful clinical and biological insights. Among the majority of cancers in TCGA (16 out of 19 cancer types, FIG.5J), unfavorable clinical outcomes were associated with a low subclonal mutational load (sML), found in tumors that have presumably undergone recent clonal sweeps, i.e., positive selection. Favorable clinical outcomes were associated with a high sML, found in tumors that may be slow-growing and have a longer lag time, i.e., long latency, since the last clonal sweep event. Within the three remaining TCGA cancer types, i.e., melanoma, lung squamous cell carcinoma (smoker) and ovarian cancer, those who shared the highest TMB or the highest genome instability (as measured by percent copy number aberrations), had a low sML, corresponding to a high number of clonal mutations, and associated with favorable clinical outcomes. Therefore, this comprehensive evaluation provides evidence supporting distinct evolutionary paths between cancers with high versus low mutational load, and how their divergent evolutionary dynamics may impact clinical outcomes.

[0169] An mCRPC clinical trial study was used to directly evaluate the relationship between sML and immune activities in tumors. Signal in sML was found to identify mCRPC patients with improved survival and progression times after receiving ipilimumab. Therefore, as compared to patients with high-TMB tumors that are conventionally expected to benefit from immunotherapy, this study points to a new target population: patients with low toAtty. Dkt. No.: 642631-0101 (pmda24-027) moderate-TMB tumors (about 90% of all cancers) that may also benefit from immunotherapy, when they present a high subclonal mutational load.

[0170] A high-level summary measure of cancer evolution such as the clonal / subclonal status of single-nucleotide variants, when computed across a large number of patient samples, sheds new light on the relationship between genetic intra-tumor heterogeneity and therapeutic resistance. Highly efficient and accurate methods like CliPP represent a new generation of techniques that bridge the gap between studying tumor biology and developing precision treatment strategies. Example 2: Expansive Study Validating CliPP to Classify Patients with Cancer for Alternative Target Therapies as Compared to Standard of Care Treatment

[0171] This clinical trial assesses whether CliPP can predict a patient’s tumor sensitivity to standard of care treatment as compared to being placed on a personally designed treatment trial including immune checkpoint therapy. The clinical trial indicates improved responses in patients with high sML tumors using immune checkpoint therapy.

[0172] The trial includes nineteen cohorts with different cancer types. These cancer types include pheochromocytoma (PCPG), thymoma (THYM), chromophobe renal cell carcinoma (KICH), low grade glioma (LGG), prostate adenocarcinoma (PRAD), adrenocortical adenocarcinoma (ACC), pancreatic adenocarcinoma (PAAD), sarcoma (SARC), renal cell carcinoma (KIRC), renal papillary cell carcinoma (KIRP), ovarian cancer (OV), liver hepatocellular carcinoma (LIHC), uterine and cervical cancer (UCEC), colorectal cancer (CRC), HPV-negative head and neck squamous cell carcinoma (HPV-HNSC), lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LUSC), skin cutaneous melanoma (SKCM), and acute myeloid leukemia (AML). PRAD may be split into the cohorts of the subtypes prostate gleason 7, prostate gleason 8, and all.

[0173] The clinical trial follows the general scheme show in FIG.7, with patients with sML levels that predict favorable immune response are assigned to Arm A or Arm B and patients with sML levels that do not predict favorable immune response are assigned to Arm C. In Arm A patients receive immune checkpoint inhibition therapy in addition to standard of care treatment alone and in Arm B patients receive standard of care therapy. In Arm C, patients with sML levels that do not predict favorable immune response are given standard of care treatment.Atty. Dkt. No.: 642631-0101 (pmda24-027)

[0174] Patients in the second arm receive nivolumab intravenously over 30 minutes on days 1, 15 and 29. Treatment repeats every 42 days for up to 17 cycles (2 years) in the absence of disease progression or unacceptable toxicity. After 17 cycles (2 years) of therapy, patients may receive nivolumab once every 14 or 28 days (2 weeks or 4 weeks) in the absence of disease progression or unacceptable toxicity. Standard of care treatment will vary with the type of cancer.

[0175] After completion of study treatment, patients are followed up for up to 5 years.

[0176] Patient eligibility criteria include the following. For all cancer types, patients are 18 years of age or older and have histologically and / or cytologically confirmed cancer of one of the nineteen cancer types listed above. All patients have advanced or metastatic disease for which standard therapy is indicated but not expected to result in a cure. Patients do not have received prior treatment. Alternatively, there is an adequate washout period from previous therapies, including chemotherapy, radiation therapy, or other investigational therapies, to reduce interference with the study treatment's efficacy assessment. Patients have an ECOG (Eastern Cooperative Oncology Group) performance status of 0 or 1, indicating that the patient is fully active or restricted in physically strenuous activity but ambulatory and able to carry out work of a light or sedentary nature, except for GLL patients who are considered ambulatory if using a wheelchair. Patients have no prior adverse reaction to immune checkpoint inhibitors that would preclude further treatment with such therapies. Patients have adequate baseline organ function, as evidenced by specific laboratory criteria (e.g., liver function tests, renal function tests) within acceptable limits to reduce the risk of treatment- related complications. Patients have a life expectancy of at least 12 weeks, to ensure that participants can potentially benefit from the treatment and contribute to the assessment of efficacy and safety. Patients of childbearing potential agree to use effective contraception during the study and for a specified period after the last dose of study treatment, to prevent pregnancy due to the unknown risks of the therapy on a developing fetus. Patients have an ability to understand and willingness to sign a written informed consent document. Patients must have adequate bone marrow, liver, and renal function (hemoglobin ≥ 90 g / L, platelets ≥ 100 x 109 / L, neutrophil count ≥ 1.5 x 109 / L, bilirubin ≤ 1.5 x upper limit of normal, AST or ALT ≤ 2.0 x ULN, creatinine ≤ 1.5x ULN).

[0177] Inclusion criteria for the study differs by cancer type. THYM patients are less than 75 years old, complete resection at pathological examination of the surgical specimen after surgery conducted through standard, recommended approach ensuring accurate assessment ofAtty. Dkt. No.: 642631-0101 (pmda24-027) resection status, and Stage IIb or III disease according to the Masaoka-Koga staging system; this corresponds to stage pT1a with capsule invasion, until stage pT3 N0 M0 in the 8th TNM staging system TNM UICC / AJCC. KICH and OV patients have at least 1 target lesion according to RECIST v1.1. KIRP patients must have at least 1 lesion with measurable disease. LIHC patients do not have extrahepatic metastasis or hepatic encephalopathy, or known history of active Bacillus Tuberculosis (TB). GLL patients must have a Karnofsky score of greater than or equal to 50%. Example 3: Validation of CliPP Clinical Impact in Classifying Patients with Metastatic Castration-Resistant Prostate Cancer (mCRPC) or Oesophageal Cancer for Alternative Target Therapies as Compared to Standard of Care Treatment

[0178] Methods. The CliPP model in Example 1 was used in Example 3.

[0179] An updated method of parameter estimation using a regularized likelihood wasused. The CPs, ^^  =  (^^^,  … ,  ^^ௌ)் may be estimated by maximizing the corresponding log-likelihood. However, a goal was to identify a homogeneous structure, i.e., clusters, of the CPs across all SNVs. Penalized estimation is a canonical tool to achieve homogeneity detection and parameter estimation simultaneously. Therefore, a pairwise penalty was introduced to seek such homogeneity in ^^. To facilitate computation, a normalapproximation of the binomial random variable was employed,∼^^൫0,  ^^^which lead to the following approximated log-likelihoodTo promote homogeneity in ^^, the negative log-likelihood −ℓ^( ^^) was equipped with a pairwise shrinkage penalty, thereby solving the following optimization problem  where ^^ > 0 was a tuning parameter that controlled the degree of the penalization, and ^^ఒ(⋅)denoted a sparsity-inducing penalty function to identify the homogeneity structure in ^^, such as LASSO, SCAD, and MCP to name a few. Here, the SCAD penalty was defined asAtty. Dkt. No.: 642631-0101 (pmda24-027)where ^^ was a hyper-parameter that controlled the concavity of the SCAD penalty. ^^ was set to 3.7. These concave penalties offered sparsity similar to the LASSO penalty, allowing them to automatically yield sparse estimates. One can also choose other penalties, such as LASSO, and the formulation is straightforward.

[0180] Optimizing the objective function in Equation (9) is nontrivial because the targetvariable ^^ is bounded, i.e., ^^^ = ^^^^^ ∈ [0,1], and the SCAD penalty is not convex. A re-parametrization of ^^ was employed by defining థ^= log^ିథ^ ,  ^^ = 1, … ,  ^^, to removebox constraint, which yieldedNote that ^^ 's were monotonic with respect to= ^^^ implied ^^^ = ^^^ and vicethus the homogeneity pursuit of ^^ can be achieved by the homogeneity pursuit of ^^. A transformation on the loss function from Eq. (5) was performed, with an updated penalty function that identified the homogeneity structure in ^^. Consequently, the optimization problem was reformulated as follows:

[0181] This is a non-convex optimization problem with respect to ^^, and its computation is non-trivial due to the complex log-likelihood. To address this challenge, the alternating direction method of multipliers (ADMM) was employed, where the negative log-likelihood was approximated at each iteration by a quadratic function, thereby simplifying the computation. Note that the SCAD penalty possessed the unbiasedness property, ensuring that it does not shrink large estimated parameters throughout iterations.

[0182] Finally, the regularized likelihood-based approach selected a proper ^^, in order to balance between over- and under-fitting. Here, the choice of ^^ determined the final number of clusters, with higher values yielding fewer clusters. Traditional approaches for selecting ^^, including cross-validation, bootstrapping, AIC, and BIC, suffer from respective disadvantages. Cross-validation and bootstrapping require down-sampling and multipleAtty. Dkt. No.: 642631-0101 (pmda24-027) rounds of estimation, causing a heavy computational burden, which conflicts with the major motivation of this work. AIC, BIC, and their extensions select ^^ by minimizing the negative log-likelihood equipped with a penalty term to control the model’s complexity. They are generally applicable to tuning parameter problems but lead to improper clustering results without incorporating the biological requirements in the study. To address this issue, an ad hoc selection approach was implemented, which focused on the interpretability of theoutcome by ensuring that clonal mutations had an estimated CCF of around 1. An automated^^ selection pipeline operated as follows: for each sample, CliPP ran on the data with 11different ^^′^^ spanning from 0 to 0.25, specifically: 0.01, 0.03, 0.05, 0.075, 0.1, 0.125, 0.15,0.175, 0.2, 0.225, 0.25. For each sample, a score ^^ = ୫ୟ^(^^)ି^௨^^௧௬^௨^^௧௬ was computed. A lowerscore indicated a clonal CCF closer to 1, which was more consistent with the model assumption and unlikely to present a superclonal cluster. If there were one or more resultsthat satisfied ^^ < 0.05, the largest ^^ associated with those results was chosen. If all scores^^ were greater than 0.01, the ^^ associated with the smallest ^^ was chosen.

[0183] Analysis of WES data from NCT02113657 and NCT02703623. WES data from two clinical trials involving patients with metastatic castration-resistant prostate cancer (mCRPC) was obtained. Clinical trial NCT02113657 was described in Example 1. In NCT02113657, WES data were collected from 30 tumor samples prior to treatment with anti- CTLA-4 therapy (ipilimumab). In the second trial, NCT02703623, WES data were collected from tumor samples of patients before treatment with abiraterone acetate, prednisone, and apalutamide (ARPi) for 8 weeks. Patients who showed a favorable response (≥50% PSA decline and <5 circulating tumor cells / 7.5 mL) were then assigned to receive either ARPi alone (n=18), ARPi combined with ipilimumab (n = 21), or ARPi combined with carboplatin and cabazitaxel (CC) (n=27). Both datasets were processed using the same pipeline as follows. The paired-end FASTQ files were generated on the Illumina HiSeq 2500 for trial NCT02703623 and HiSeq 2000 for trial NCT02113657. After adapter trimming with Trim_Galore v0.6.10, sequence quality was assessed using FASTQC v0.11.8. Subsequently, the reads were aligned to the GRCh38 build through BWA-mem 0.7.15-r1140. The resulting BAM files were processed according to the Genome Analysis Toolkit (GATK) Best Practices, including steps such as marking duplicates, joint realignment of paired tumor- normal BAMs, and recalibration of base quality scores. Consensus mutation calls were derived as those from 2 out of 3 callers: MuSE 2.0 (v2.1.2) calls and PASS-filtered results from MuTect2 (v4.2.4.0) and Strelka2 (v2.9.10). Copy number aberrations (CNAs), tumorAtty. Dkt. No.: 642631-0101 (pmda24-027) purity and ploidy were called using ASCAT (v3.1.2) configured for WES data processing. Patients’ response to immunotherapy outcome data as well as immunohistochemistry (IHC) staining for CD8 T cell densities were obtained from the previous report of NCT02113657, as was PD-L1 expression, which was assessed using an automated immunohistochemical assay. Following the pre-processing pipeline used for TCGA, samples with nrpcc<10 and TMB>100 and samples with less than 10 coding SNVs (i.e., TMB > 0.25) which would otherwise inflate the subclonal fraction estimates were removed.

[0184] For NCT02113657 samples, since some patients had multiple metastases, only the sample with the lowest sML for each patient were retained to better align with survival outcomes. The sML cutoff was set 0.5 and patients with sML ≥ 0.5 were classified as high sML. To analyze survival outcomes, Kaplan-Meier (KM) analysis was applied, generating KM curves through the survminer R package. These curves illustrated survival probability across time, stratified by high vs. low sML. Survival P-values were obtained via log-rank test. P-value for comparison of CD8 T-cell density and for PD-L1 density between high and low sML samples was calculated using a two-sided Wilcoxon rank-sum test. A multivariate Cox proportional hazards model was employed, implementing bi-directional stepwise variable selection with model comparison via Bayesian information criterion (BIC) to identify the optimal set of variables. The base model used sML as a binary variable (high versus low) along with continuous TMB, and age. Additional candidate variables included PD-L1 density, purity, ploidy, coverage, and the interaction between sML and TMB.

[0185] For NCT02703623 samples, within each treatment arm, samples were split into high and low sML using a recursive partitioning survival tree model (rpart), with the maximum tree depth constrained to 1. Survival P-values were computed via log-rank test. A multivariate Cox proportional hazards model was employed, implementing bi-directional stepwise variable selection with model comparison via Bayesian information criterion (BIC) to identify the optimal set of variables. The base model used sML as a binary variable (high versus low) along with continuous TMB, and age. Additional candidate variables included PD-L1 density, purity, ploidy, coverage, and the interaction between sML and TMB. Age information was only available for 17 of the 21 patients.

[0186] Results and Discussion. Subclonal mutation load is associated with immunotherapy response. mCRPC is a tumor type with low TMB despite which some patient subgroups benefit from immune checkpoint blockade. Biomarkers that can identify them remain to be defined. The effect of sML was evaluated in two cohorts of mCRPCAtty. Dkt. No.: 642631-0101 (pmda24-027) patients exposed to ipilimumab (IPI): monotherapy NCT02113657 and DynAMo trial NCT02703623 (FIG.9A). In the IPI monotherapy trial, patients with high sML experienced improved radiographic / clinical PFS (rcPFS) in this trial (n=8 vs.13, log rank P-value=0.005, FIG.9B) and overall survival, with up to 8 years of follow-up (log rank P-value=0.03) compared to those with low sML. The distribution of sML was not correlated with TMB, metastatic sites, or nrpcc, and further subset-specific analysis for prostate tissues only replicated the significant clinical outcome association and the direction of sML. In the DynAMo trial, patients with high sML also experienced longer failure-free survival (n=6 vs. 15, log rank P-value= 0.004, FIG.9C), after receiving ARPi plus ipilimumab compared to those with low sML. For both datasets, using Cox regression with variable selection, the significant effect of sML remained after adjusting for TMB, age, PD-L1 density, and other technical confounders (FIG.9D). In the DynAMo trial where additional mCRPC patients (n=45) received ARPi alone or combined ARPi and chemotherapy (ARPi+CC), high sML was not significantly associated with outcomes, although still presenting a similar but weaker trend in the ARPi+CC arm. This contrast supported an immune-specific activation that underlies tumors with high sML.

[0187] Additional profiling data of these patients was used to determine whether sML corresponded to changes in the tumor immune microenvironment. Paired immunohistochemistry slides from a subset of patients in the IPI monotherapy trial were allocated and observed high versus low sML tumor samples presented significantly different levels of CD8+T cell densities (Wilcoxon test P-value=0.008, FIG.9E). This distinction was independent of PD-L1 protein expression levels (Wilcoxon P = 0.3, FIG.9F). CIBERSORTx was obtained to deconvolve immune cell type proportions for the IPI trial. The deconvolution-based CD8+T cell proportions showed a similar trend, although not statistically significant, while the myeloid cell population showed a significantly higher proportion of monocytes and a significantly lower proportion of macrophages in the high sML tumors. Together the data suggested that a high subclonal mutation load corresponded with a more favorable tumor immune microenvironment, providing a plausible biological explanation for the prediction of responsiveness to ICB by high sML. EQUIVALENTS

[0188] Having now described some illustrative implementations, it is apparent that the foregoing is illustrative and not limiting, having been presented by way of example. In particular, although many of the examples presented herein involve specific combinations ofAtty. Dkt. No.: 642631-0101 (pmda24-027) method acts or system elements, those acts and those elements may be combined in other ways to accomplish the same objectives. Acts, elements, and features discussed in connection with one implementation are not intended to be excluded from a similar role in other implementations or implementations.

[0189] The phraseology and terminology used herein is for the purpose of description and should not be regarded as limiting. The use of “including” “comprising” “having” “containing” “involving” “characterized by” “characterized in that” and variations thereof herein, is meant to encompass the items listed thereafter, equivalents thereof, and additional items, as well as alternate implementations consisting of the items listed thereafter exclusively. In one implementation, the systems and methods described herein consist of one, each combination of more than one, or all of the described elements, acts, or components.

[0190] The present technology is not to be limited in terms of the particular embodiments described in this application, which are intended as single illustrations of individual aspects of the present technology. Many modifications and variations of this present technology can be made without departing from its spirit and scope, as will be apparent to those skilled in the art. Functionally equivalent methods and apparatuses within the scope of the present technology, in addition to those enumerated herein, will be apparent to those skilled in the art from the foregoing descriptions. Such modifications and variations are intended to fall within the scope of the present technology. It is to be understood that this present technology is not limited to particular methods, reagents, compounds compositions or biological systems, which can, of course, vary. It is also to be understood that the terminology used herein is for the purpose of describing particular embodiments only, and is not intended to be limiting.

[0191] In addition, where features or aspects of the disclosure are described in terms of Markush groups, those skilled in the art will recognize that the disclosure is also thereby described in terms of any individual member or subgroup of members of the Markush group.

[0192] As will be understood by one skilled in the art, for any and all purposes, particularly in terms of providing a written description, all ranges disclosed herein also encompass any and all possible subranges and combinations of subranges thereof. Any listed range can be recognized as sufficiently describing and enabling the same range being broken down into at least equal halves, thirds, quarters, fifths, tenths, etc. As a non-limiting example, each range discussed herein can be readily broken down into a lower third, middle third and upper third,Atty. Dkt. No.: 642631-0101 (pmda24-027) etc. As will also be understood by one skilled in the art all language such as “up to,” “at least,” “greater than,” “less than,” and the like, include the number recited and refer to ranges which can be subsequently broken down into subranges as discussed above. Finally, as will be understood by one skilled in the art, a range includes each individual member. Thus, for example, a group having 1-3 cells refers to groups having 1, 2, or 3 cells. Similarly, a group having 1-5 cells refers to groups having 1, 2, 3, 4, or 5 cells, and so forth.

[0193] Embodiments of the subject matter and the operations described in this specification can be implemented in digital electronic circuitry, or in computer software, firmware, or hardware, including the structures disclosed in this specification and their structural equivalents, or in combinations of one or more of them. The subject matter described in this specification can be implemented as one or more computer programs, e.g., one or more circuits of computer program instructions, encoded on one or more computer storage media for execution by, or to control the operation of, data processing apparatus. Alternatively or in addition, the program instructions can be encoded on an artificially generated propagated signal, e.g., a machine-generated electrical, optical, or electromagnetic signal that is generated to encode information for transmission to suitable receiver apparatus for execution by a data processing apparatus. A computer storage medium can be, or be included in, a computer-readable storage device, a computer-readable storage substrate, a random or serial access memory array or device, or a combination of one or more of them. Moreover, while a computer storage medium is not a propagated signal, a computer storage medium can be a source or destination of computer program instructions encoded in an artificially generated propagated signal. The computer storage medium can also be, or be included in, one or more separate components or media (e.g., multiple CDs, disks, or other storage devices).

[0194] The operations described in this specification can be performed by a data processing apparatus on data stored on one or more computer-readable storage devices or received from other sources. The term “data processing apparatus” or “computing device” encompasses various apparatuses, devices, and machines for processing data, including by way of example a programmable processor, a computer, a system on a chip, or multiple ones, or combinations of the foregoing. The apparatus can include special purpose logic circuitry, e.g., an FPGA (field programmable gate array) or an ASIC (application specific integrated circuit). The apparatus can also include, in addition to hardware, code that creates an execution environment for the computer program in question, e.g., code that constitutes processor firmware, a protocol stack, a database management system, an operating system, a cross-Atty. Dkt. No.: 642631-0101 (pmda24-027) platform runtime environment, a virtual machine, or a combination of one or more of them. The apparatus and execution environment can realize various different computing model infrastructures, such as web services, distributed computing, and grid computing infrastructures.

[0195] A computer program (also known as a program, software, software application, script, or code) can be written in any form of programming language, including compiled or interpreted languages, declarative or procedural languages, and it can be deployed in any form, including as a stand-alone program or as a circuit, component, subroutine, object, or other unit suitable for use in a computing environment. A computer program may, but need not, correspond to a file in a file system. A program can be stored in a portion of a file that holds other programs or data (e.g., one or more scripts stored in a markup language document), in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files that store one or more circuits, subprograms, or portions of code). A computer program can be deployed to be executed on one computer or on multiple computers that are located at one site or distributed across multiple sites and interconnected by a communication network.

[0196] Processors suitable for the execution of a computer program include, by way of example, microprocessors, and any one or more processors of a digital computer. A processor can receive instructions and data from a read only memory or a random-access memory or both. The elements of a computer are a processor for performing actions in accordance with instructions and one or more memory devices for storing instructions and data. A computer can include, or be operatively coupled to receive data from or transfer data to, or both, one or more mass storage devices for storing data, e.g., magnetic, magneto optical disks, or optical disks. A computer need not have such devices. Moreover, a computer can be embedded in another device, e.g., a personal digital assistant (PDA), a Global Positioning System (GPS) receiver, or a portable storage device (e.g., a universal serial bus (USB) flash drive), to name just a few. Devices suitable for storing computer program instructions and data include all forms of non-volatile memory, media, and memory devices, including by way of example semiconductor memory devices, e.g., EPROM, EEPROM, and flash memory devices; magnetic disks, e.g., internal hard disks or removable disks; magneto optical disks; and CD ROM and DVD-ROM disks. The processor and the memory can be supplemented by, or incorporated in, special purpose logic circuitry.Atty. Dkt. No.: 642631-0101 (pmda24-027)

[0197] The implementations described herein can be implemented in any of numerous ways including, for example, using hardware, software, or a combination thereof. When implemented in software, the software code can be executed on any suitable processor or collection of processors, whether provided in a single computer or distributed among multiple computers.

[0198] A computer employed to implement at least a portion of the functionality described herein may comprise a memory, one or more processing units (also referred to herein simply as “processors”), one or more communication interfaces, one or more display units, and one or more user input devices. The memory may comprise any computer-readable media, and may store computer instructions (also referred to herein as “processor-executable instructions”) for implementing the various functionalities described herein. The processing unit(s) may be used to execute the instructions. The communication interface(s) may be coupled to a wired or wireless network, bus, or other communication means and may therefore allow the computer to transmit communications to or receive communications from other devices. The display unit(s) may be provided, for example, to allow a user to view various information in connection with execution of the instructions. The user input device(s) may be provided, for example, to allow the user to make manual adjustments, make selections, enter data or various other information, or interact in any of a variety of manners with the processor during execution of the instructions.

[0199] The various methods or processes outlined herein may be coded as software that is executable on one or more processors that employ any one of a variety of operating systems or platforms. Additionally, such software may be written using any of a number of suitable programming languages or programming or scripting tools, and also may be compiled as executable machine language code or intermediate code that is executed on a framework or virtual machine.

[0200] In this respect, various inventive concepts may be embodied as a computer readable storage medium (or multiple computer readable storage media) (e.g., a computer memory, one or more floppy discs, compact discs, optical discs, magnetic tapes, flash memories, circuit configurations in Field Programmable Gate Arrays or other semiconductor devices, or other non-transitory medium or tangible computer storage medium) encoded with one or more programs that, when executed on one or more computers or other processors, perform methods that implement the various embodiments of the solution discussed above. The computer readable medium or media can be transportable, such that the program or programsAtty. Dkt. No.: 642631-0101 (pmda24-027) stored thereon can be loaded onto one or more different computers or other processors to implement various aspects of the present solution as discussed above.

[0201] The terms “program” or “software” are used herein to refer to any type of computer code or set of computer-executable instructions that can be employed to program a computer or other processor to implement various aspects of embodiments as discussed above. One or more computer programs that when executed perform methods of the present solution need not reside on a single computer or processor, but may be distributed in a modular fashion amongst a number of different computers or processors to implement various aspects of the present solution.

[0202] Computer-executable instructions may be in many forms, such as program modules, executed by one or more computers or other devices. Program modules can include routines, programs, objects, components, data structures, or other components that perform particular tasks or implement particular abstract data types. The functionality of the program modules can be combined or distributed as desired in various embodiments.

[0203] Also, data structures may be stored in computer-readable media in any suitable form. For simplicity of illustration, data structures may be shown to have fields that are related through location in the data structure. Such relationships may likewise be achieved by assigning storage for the fields with locations in a computer-readable medium that convey relationship between the fields. However, any suitable mechanism may be used to establish a relationship between information in fields of a data structure, including through the use of pointers, tags or other mechanisms that establish relationship between data elements.

[0204] Any references to implementations or elements or acts of the systems and methods herein referred to in the singular can include implementations including a plurality of these elements, and any references in plural to any implementation or element or act herein can include implementations including only a single element. References in the singular or plural form are not intended to limit the presently disclosed systems or methods, their components, acts, or elements to single or plural configurations. References to any act or element being based on any information, act or element may include implementations where the act or element is based at least in part on any information, act, or element.

[0205] Any implementation disclosed herein may be combined with any other implementation, and references to “an implementation,” “some implementations,” “an alternate implementation,” “various implementations,” “one implementation” or the like areAtty. Dkt. No.: 642631-0101 (pmda24-027) not necessarily mutually exclusive and are intended to indicate that a particular feature, structure, or characteristic described in connection with the implementation may be included in at least one implementation. Such terms as used herein are not necessarily all referring to the same implementation. Any implementation may be combined with any other implementation, inclusively or exclusively, in any manner consistent with the aspects and implementations disclosed herein.

[0206] References to “or” may be construed as inclusive so that any terms described using “or” may indicate any of a single, more than one, and all of the described terms. References to at least one of a conjunctive list of terms may be construed as an inclusive OR to indicate any of a single, more than one, and all of the described terms. For example, a reference to “at least one of ‘A’ and ‘B’” can include only ‘A,’ only ‘B’, as well as both ‘A’ and ‘B’. Elements other than ‘A’ and ‘B’ can also be included.

[0207] The systems and methods described herein may be embodied in other specific forms without departing from the characteristics thereof. The foregoing implementations are illustrative rather than limiting of the described systems and methods.

[0208] Where technical features in the drawings, detailed description or any claim are followed by reference signs, the reference signs have been included to increase the intelligibility of the drawings, detailed description, and claims. Accordingly, neither the reference signs nor their absence have any limiting effect on the scope of any claim elements.

[0209] The systems and methods described herein may be embodied in other specific forms without departing from the characteristics thereof. The foregoing implementations are illustrative rather than limiting of the described systems and methods. Scope of the systems and methods described herein is thus indicated by the appended claims, rather than the foregoing description, and changes that come within the meaning and range of equivalency of the claims are embraced therein.

[0210] All patents, patent applications, provisional applications, and publications referred to or cited herein are incorporated by reference in their entirety, including all figures and tables, to the extent they are not inconsistent with the explicit teachings of this specification.

Claims

Atty. Dkt. No.: 642631-0101 (pmda24-027) CLAIMS 1. A method for selecting a cancer patient for treatment with an immune checkpoint inhibitor comprising: (a) detecting a level of tumor subclonal mutations at or above a predetermined threshold in a biological sample obtained from the cancer patient; and (b) administering to the cancer patient the immune checkpoint inhibitor.

2. The method of claim 1, wherein the cancer patient has a solid tumor with a tumor mutation burden (TMB) of less than 10.

3. The method of claim 1 or 2, wherein the cancer patient has one or more of prostate cancer, a pheochromocytoma tumor, a thymoma tumor, chromophobe renal cell carcinoma, IDH1-mutant low grade glioma, prostate adenocarcinoma, adrenocortical adenocarcinoma, pancreatic adenocarcinoma, sarcoma, renal cell carcinoma, renal papillary cell carcinoma, ovarian cancer, liver hepatocellular carcinoma, endometrioid uterine cancer, endometrioid cervical cancer, colorectal cancer, head and neck squamous cell carcinoma, lung adenocarcinoma, lung squamous cell carcinoma, acute myeloid leukemia, thyroid carcinoma, esophageal adenocarcinoma, and melanoma.

4. The method of any one of claims 1-3, wherein the immune checkpoint inhibitor comprises a programmed cell death protein 1 (PD-1) inhibitor, a programmed death-ligand 1 (PD-L1) inhibitor, a cytotoxic T-lymphocyte associated protein 4 (CTLA-4) inhibitor, or any combination thereof.

5. The method of any of claims 1-4, wherein the immune checkpoint inhibitor comprises one or more of ipilimumab, tremelimumab, cadonilimab, zalifrelimab, pembrolizumab, nivolumab, cemiplimab, sintilimab, tislelizumab, toripalimab), camrelizumab, geptanolimab, toripalimab, zimberelimab, penpulimab, serplulimab, prolgolimab, balstilimab, retifanlimab, cadonilimab, pucotenlimab, sasanlimab, cetrelimab, tebotelimab, pidilizumab, dostarlimab, atezolizumab, durvalumab, avelumab, sugemalimab, or envafolimab.

6. The method of any of claims 1-5, wherein the level of tumor subclonal mutations are determined via next-generation sequencing or microarray.

7. The method of claim 6, wherein the level of tumor subclonal mutations is determined via targeted next-generation sequencing.Atty. Dkt. No.: 642631-0101 (pmda24-027) 8. The method of any one of claims 1-7, wherein detecting the level of tumor subclonal mutations comprises determining the level of subclonal mutations using clonal structure identification through pairwise penalization.

9. The method of claim 8, wherein the predetermined threshold is determined for a cancer type using a recursive partitioning survival tree model analyzing at least ten biological samples obtained from different patients with the cancer type.

10. The method of claim 8 or 9, wherein detecting the level of tumor subclonal mutations comprises applying, as input to an objective function, at least a metric of cellular prevalence of a given nucleotide variation in the biological sample and a corresponding proportion of cancer cells amongst all cells associated with the biological sample.

11. The method of any one of claims 8-10, wherein detecting the level of tumor subclonal mutations comprises detecting one or more clusters of cellular prevalence data relating to the biological sample, across a plurality of nucleotide variations of the biological sample, that satisfy a criterion of homogeneity.

12. A method for prolonging survival of a cancer patient comprising: administering to the cancer patient an immune checkpoint inhibitor, wherein the tumor subclonal mutations level in a biological sample obtained from the cancer patient are at or above a predetermined threshold determined by comparing survival outcome to tumor subclonal mutations level in a cohort of cancer patients that have the same type of cancer as the cancer patient.

13. The method of claim 12, wherein the cancer patient has a solid tumor with a tumor mutation burden (TMB) of less than 10.

14. The method of claim 12 or 13, wherein the cancer patient has one or more of prostate cancer, a pheochromocytoma tumor, a thymoma tumor, chromophobe renal cell carcinoma, IDH1-mutant low grade glioma, prostate adenocarcinoma, adrenocortical adenocarcinoma, pancreatic adenocarcinoma, sarcoma, renal cell carcinoma, renal papillary cell carcinoma, ovarian cancer, liver hepatocellular carcinoma, endometrioid uterine cancer, endometrioid cervical cancer, colorectal cancer, head and neck squamous cell carcinoma, lung adenocarcinoma, lung squamous cell carcinoma, acute myeloid leukemia, thyroid carcinoma, esophageal adenocarcinoma, and melanoma.Atty. Dkt. No.: 642631-0101 (pmda24-027) 15. The method of any one of claims 12-14, wherein the tumor subclonal mutations level is determined via next-generation sequencing, or microarray.

16. The method of claim 15, wherein the tumor subclonal mutations level is determined via targeted next-generation sequencing.

17. The method of any of claims 12-16, wherein the immune checkpoint inhibitor comprises a programmed cell death protein 1 (PD-1) inhibitor, a programmed death-ligand 1 (PD-L1) inhibitor, a cytotoxic T-lymphocyte associated protein 4 (CTLA-4) inhibitor, or any combination thereof.

18. The method of any one of claims 12-17, wherein the immune checkpoint inhibitor comprises one or more of ipilimumab, tremelimumab, cadonilimab, zalifrelimab, pembrolizumab, nivolumab, cemiplimab, sintilimab, tislelizumab, toripalimab), camrelizumab, geptanolimab, toripalimab, zimberelimab, penpulimab, serplulimab, prolgolimab, balstilimab, retifanlimab, cadonilimab, pucotenlimab, sasanlimab, cetrelimab, tebotelimab, pidilizumab, dostarlimab, atezolizumab, durvalumab, avelumab, sugemalimab, or envafolimab.

19. The method of any one of claims 12-18, wherein the predetermined threshold is determined using clonal structure identification through pairwise penalization.

20. The method of claim 19, wherein the predetermined threshold is determined for a cancer type using a recursive partitioning survival tree model analyzing at least ten biological samples obtained from different patients with the cancer type.

21. The method of any one of claims 1-20, further comprising sequentially, simultaneously, or separately administering to the cancer patient an effective amount of a radiation therapy.

22. The method of any of claims 1-21, further comprising sequentially, simultaneously, or separately administering to the cancer patient an effective amount of a chemotherapeutic agent.

23. The method of claim 22, wherein the chemotherapeutic agent comprises one or more of alkylating agents, topoisomerase inhibitors, endoplasmic reticulum stress inducing agents, antimetabolites, mitotic inhibitors, nitrogen mustards, nitrosoureas, alkyl sulfonates, platinum agents, taxanes, vinca agents, anti-estrogen drugs, aromatase inhibitors, ovarian suppressionAtty. Dkt. No.: 642631-0101 (pmda24-027) agents, cytostatic alkaloids, cytotoxic antibiotics, antimetabolites, endocrine / hormonal agents, and bisphosphonate therapy agents.

24. The method of claim 22 or 23, wherein the chemotherapeutic agent comprises one or more of cyclophosphamide, fluorouracil (or 5-fluorouracil or 5-FU), methotrexate, edatrexate (10-ethyl-10-deaza-aminopterin), thiotepa, carboplatin, cisplatin, taxanes, paclitaxel, protein- bound paclitaxel, docetaxel, vinorelbine, tamoxifen, raloxifene, toremifene, fulvestrant, gemcitabine, irinotecan, ixabepilone, temozolomide, topotecan, vincristine, vinblastine, eribulin, mutamycin, capecitabine, anastrozole, exemestane, letrozole, leuprolide, abarelix, buserlin, goserelin, megestrol acetate, risedronate, pamidronate, ibandronate, alendronate, denosumab, zoledronate, trastuzumab, tykerb, anthracyclines (e.g., daunorubicin and doxorubicin), cladribine, midostaurin, bevacizumab, oxaliplatin, melphalan, etoposide, mechlorethamine, bleomycin, microtubule poisons, annonaceous acetogenins, chlorambucil, ifosfamide, streptozocin, carmustine, lomustine, busulfan, dacarbazine, temozolomide, altretamine, 6-mercaptopurine (6-MP), cytarabine, floxuridine, fludarabine, hydroxyurea, pemetrexed, epirubicin, idarubicin, SN-38, ARC, NPC, campothecin, 9-nitrocamptothecin, 9- aminocamptothecin, rubifen, gimatecan, diflomotecan, BN80927, DX-8951f, MAG-CPT, amsacrine, etoposide phosphate, teniposide, azacitidine (Vidaza), decitabine, accatin III, 10- deacetyltaxol, 7-xylosyl-10-deacetyltaxol, cephalomannine, 10-deacetyl-7-epitaxol, 7- epitaxol, 10-deacetylbaccatin III, 10-deacetyl cephalomannine, streptozotocin, nimustine, ranimustine, bendamustine, uramustine, estramustine, mannosulfan, camptothecin, exatecan, lurtotecan, lamellarin D9-aminocamptothecin, amsacrine, ellipticines, aurintricarboxylic acid, HU-331, or combinations thereof.

25. The method of any one of claims 1-24, wherein the biological sample comprises plasma, blood, serum, or biopsied tissue.

26. The method of any one of claims 1-25, wherein the cancer patient has castration- resistant prostate cancer.

27. The method of any one of claims 1-24, further comprising sequentially, simultaneously, or separately administering to the cancer patient an additional anti-cancer therapy, optionally wherein the additional anti-cancer therapy comprises one or more of chemotherapy, targeted therapy, immunotherapy, radiation therapy, or surgery, optionally wherein the targeted therapy comprises a VEGF / VEGFR inhibitor, EGF / EGFR inhibitor,Atty. Dkt. No.: 642631-0101 (pmda24-027) PARP inhibitor, or a combination thereof, or optionally wherein the immunotherapy comprises an additional immune checkpoint inhibitor therapy.

28. A system, apparatus, or device, comprising one or more processors to perform one or more operations corresponding to any of claims 1-27.

29. A non-transitory computer-readable medium comprising processor-executable instructions that when executed by one or more processors, cause the one or more processors to perform one or more operations corresponding to any of claims 1-27.

Images