SLC2a1 / glut1 as a biomarker for high-risk epithelial dysplasia

Abstract

Provided herein are methods of stratifying oral lesions based upon risk of progressing to malignancy, and methods of treating a subject involving selectively removing oral lesions identified to be high risk. In some aspects, the methods provided herein comprise assessing risk based upon expression of SLC2A1 on the lesion.

Description

[0001]Atty. Docket No. UM-41943.601 SLC2A1 / GLUT1 AS A BIOMARKER FOR HIGH-RISK EPITHELIAL DYSPLASIA PRIORITY STATEMENT This application claims priority to U.S. Provisional Application No.63 / 527,223, filed July 17, 2023, the entire contents of which are incorporated herein by reference for all purposes. STATEMENT OF GOVERNMENT SUPPORT This invention was made with government support under DE029255 awarded by the National Institutes of Health. The government has certain rights in the invention. SEQUENCE LISTING The text of the computer readable sequence listing filed herewith, titled “UM-41943- 601_SQL”, created July 17, 2024, having a file size of 1,960 bytes, is hereby incorporated by reference in its entirety. FIELD Provided herein are methods of stratifying oral lesions based upon the risk of progressing to malignancy and methods of treating a subject involving removing oral lesions identified as high risk. In some aspects, the methods provided herein comprise assessing risk based upon the expression of SLC2A1 on the lesion. BACKGROUND Human papillomavirus (HPV)- head and neck squamous cell carcinomas (HNC) have a dismal 5-year overall survival rate. However, HNC is curable if diagnosed early. Oral epithelial dysplasias (OED) and oral lichen planus (OLP) are potential precursor lesions to HNC that may offer a window for disease eradication. The current standard of care for these lesions involve histological examinations. However, histologic grading has no prognostic value in predicting malignant transformation. Moreover, current standard requires high- frequency follow ups for every patient with OED / OLP, which is very challenging and impractical to maintain. As such, a major need exists in identifying OED / OLP with high risk of progressing to HNC. SUMMARY Atty. Docket No. UM-41943.601 In some aspects, provided herein are methods of stratifying oral lesions based upon risk of progressing to malignancy. In some embodiments, provided herein is a method of stratifying an oral lesion based upon risk of progressing to malignancy comprising measuring expression of solute carrier family 2 member 1 (SLC2A1) in an oral lesion, and identifying the oral lesion as high risk when the expression level of SLC2A1 is equal to or above a threshold value, or identifying the oral lesion as low risk when the expression level of SLC2A1 is below a threshold value. In some embodiments, expression of SLC2A1 in the oral lesion is measured by obtaining a biopsy sample from the oral lesion and measuring expression of SLC2A1 in the biopsy. In some embodiments, expression of SLC2A1 is measured by tissue staining. In some embodiments, expression of SLC2A1 in the oral lesion is measured directly on the oral lesion within the oral cavity of a subject. In some embodiments, expression of SLC2A1 on the oral lesion is measured by contacting the oral lesion with an agent that binds to SLC2A1, wherein the agent comprises a label, and subsequently detecting the label. In some embodiments, detecting the label comprises applying a light source to the oral lesion, thereby activating the label, and detecting the label visually. In some embodiments, the method further comprises selectively abolishing the oral lesion when the oral lesion is identified as high risk, or monitoring the oral lesion when the oral lesion is identified as low risk. In some embodiments, provided herein is a method of stratifying oral lesions based upon risk of progressing to malignancy, the method comprising measuring expression of solute carrier family 2 member 1 (SLC2A1) and expression of an additional risk marker in a plurality of oral lesions within a region of interest of an oral cavity of a subject; and identifying the oral lesion as high risk based upon an interaction score between SLC2A1 and the additional biomarker and / or the Wassterstein distance between the oral lesion and a neighboring oral lesion. For example, in some embodiments the method comprises identifying an oral lesion as high risk when an SLC2A1-additional marker interaction score for the oral lesion is above a threshold interaction score. In some embodiments the method comprises identifying an oral lesion as high risk when the oral lesion expresses SLC2A1 and a Wasserstein distance between the oral lesion and a neighboring oral lesion that expresses the additional risk marker is equal to or less than a threshold Wasserstein distance. In some embodiments, the SLC2A1-additional marker interaction score for an oral lesion is calculated by multiplying the expression of SLC2A1 in the oral lesion by the expression of the additional risk marker in the oral lesion, and dividing the result by the number of cells within the region of interest expressing both SLC2A1 and the additional risk Atty. Docket No. UM-41943.601 marker. In some embodiments, expression of SLC2A1 and expression of the additional risk marker are measured directly on the plurality of oral lesions within the oral cavity of a subject. In some embodiments, expression of SLC2A1 and expression of the additional risk marker are measured by contacting the plurality of oral lesions with an agent that binds to SLC2A1 and an agent that binds to the additional risk marker, wherein each agent comprises a label, and subsequently detecting the labels. In some embodiments, detecting the labels comprises applying a light source to the plurality of oral lesions, thereby activating the labels, and detecting the labels visually. In some embodiments, the method further comprises selectively abolishing an oral lesion when the oral lesion is identified as high risk. In some embodiments, the additional risk marker is epidermal growth factor receptor (EGFR), SRY- Box Transcription Factor 2 (SOX-2), or nuclear expression of Yes1 Associated Transcriptional Regulator (YAP-1). In some aspects, provided herein are methods of treating a subject at risk of having head and neck cancer. In some embodiments, provided herein is a method of treating a subject at risk of having head and neck cancer (HNC), the method comprising measuring expression of solute carrier family 2 member 1 (SLC2A1) in an oral lesion of a subject; identifying the oral lesion as high risk or low risk based upon the expression level of SLC2A1, and selectively abolishing high risk oral lesions or monitoring low risk oral lesions. In some embodiments, the method comprises identifying the oral lesion as high risk when the expression level of SLC2A1 is equal to or above a threshold level, or identifying the oral lesion as low risk when the expression level of SLC2A1 is below a threshold value. In some embodiments, the method comprises selectively abolishing the oral lesion when the oral lesion is identified as high risk, or monitoring the oral lesion when the oral lesion is identified as low risk. In some embodiments, expression of SLC2A1 in the oral lesion is measured by obtaining a biopsy from the oral lesion and measuring expression of SLC2A1 in the biopsy. In some embodiments, expression of SLC2A1 is measured by tissue staining. In some embodiments, expression of SLC2A1 in the oral lesion is measured directly on the oral lesion within the oral cavity of a subject. For example, in some embodiments expression of SLC2A1 on the oral lesion is measured by contacting the oral lesion with an agent that binds to SLC2A1, wherein the agent comprises a label, and subsequently detecting the label. In some embodiments, detecting the label comprises applying a light source to the oral lesion, thereby activating the label, and detecting the label visually. In some embodiments, provided herein is a method of treating a subject at risk of having head and neck cancer (HNC), the method comprising measuring expression of solute Atty. Docket No. UM-41943.601 carrier family 2 member 1 (SLC2A1) in each of a plurality of oral lesions directly within the oral cavity of a subject, identifying an oral lesion as high risk when the expression level of SLC2A1 in the oral lesion is equal to or above a threshold level and / or identifying an oral lesion as low risk when the expression level of SLC2A1 in the oral lesion is below a threshold level, and selectively abolishing high risk oral lesions from the oral cavity of the subject. In some embodiments, the method further comprises monitoring low risk oral lesions. In some embodiments, expression of SLC2A1 is measured by contacting the plurality of oral lesions with an agent that binds to SLC2A1, wherein the agent comprises a label, and subsequently detecting the label. In some embodiments, detecting the label comprises applying a light source to the oral lesion, thereby activating the label, and detecting the label visually. In some embodiments, provided herein is a method of treating a subject at risk of having head and neck cancer (HNC), the method comprising measuring expression of solute carrier family 2 member 1 (SLC2A1) and expression of an additional risk marker in a plurality of oral lesions within a region of interest of an oral cavity of a subject, and identifying an oral lesion as high risk based upon an interaction score between SLC2A1 and the additional biomarker and / or the Wassterstein distance between the oral lesion and a neighboring oral lesion, and selectively abolishing high risk oral lesions from the oral cavity of the subject. In some embodiments, the oral lesion is identified as high risk when an SLC2A1-additional marker interaction score for the oral lesion is above a threshold interaction score; and / or the oral lesion expresses SLC2A1 and a Wasserstein distance between the oral lesion and a neighboring oral lesion that expresses the additional risk marker is equal to or less than a threshold Wasserstein distance. In some embodiments, the SLC2A1-additional marker interaction score for an oral lesion is calculated by multiplying the expression of SLC2A1 in the oral lesion by the expression of the additional risk marker in the oral lesion, and dividing the result by the number of cells within the region of interest expressing both SLC2A1 and the additional risk marker. In some embodiments, the expression of SLC2A1 and expression of the additional risk marker are measured directly on the plurality of oral lesions within the oral cavity of a subject. In some embodiments, the expression of SLC2A1 and expression of the additional risk marker are measured by contacting the plurality of oral lesions with an agent that binds to SLC2A1 and an agent that binds to the additional risk marker, wherein each agent comprises a label, and subsequently detecting the labels. In some embodiments, detecting the Atty. Docket No. UM-41943.601 labels comprises applying a light source to the plurality of oral lesions, thereby activating the labels, and detecting the labels visually. In some aspects, provided herein is a method comprising measuring expression of solute carrier family 2 member 1 (SLC2A1) in at least one oral lesion directly within the oral cavity of a subject, wherein expression of SLC2A1 is measured by contacting the at least one oral lesion with an agent that binds to SLC2A1, wherein the agent comprises a label, and subsequently detecting the label on the at least one oral lesion. In some embodiments, detecting the label comprises applying a light source to the oral lesion, thereby activating the label, and detecting the label visually. BRIEF DESCRIPTION OF THE DRAWINGS FIGS.1A-1C show immune cell markers shift significantly upon the malignant transformation of OED. A longitudinal series of matched OED / HNC specimens were stained with the indicated MSI panels. Each slide was reviewed by a pathologist and 6-13 regions of interest were manually selected for high-resolution scanning. The frequencies of selected immune cell subsets were quantitated using Perkin Elmer’s inFORM. To address the variance heterogeneity between patients, nonparametric ANOVA was conducted using the Aligned Rank Transform to test the significance of changes between OED and HNC. This was implemented in R using the package ARtool. FIGS.2A-2B show IFN-I signatures underpin the immunogenicity of HNC. (FIG.2A) UMAP analysis of 38,862 TIL transcriptomes from HPV- HNC and 20,703 TIL transcriptomes from HPV+ HNC was performed. Significantly changed clusters between the two diseases are highlighted in red. (FIG.2B) The TILs from 520 HNC specimens in the TCGA data set were deconvolved and the relationship between the levels of STING and the frequencies of major subsets was assessed. FIGS.3A-3B show hypoxia induces glycolytic gene signatures in HNC. HNC cells were exposed to normoxic and hypoxic (1% pO2) conditions for 24 hours. The mRNA levels of glycolysis genes were compared by two-tailed t-test. FIGS.4A-4B show malignant transformation upregulates glycolysis markers. (FIG. 4A) Glycolysis signature gene expression levels were compared between 44 normal mucosa and 520 HNC specimens. (FIG.4B) Utilizing FARDEEP, TIL composition for 520 HNC Atty. Docket No. UM-41943.601 patients was deconvolved. Marginal correlation between major TIL subsets and the levels of SLC2A1 was performed. FIGS.5A-5K show histology of the immuno-competent Sox2-driven GEMM for OED / HNC. (FIG.5A) After topical tamoxifen induction, the occurrence rates of OED and HNC in K5-CreER;Sox2+ / +, K5-CreER;Sox2+ / - or K5- CreER control mice treated with or without DMBA are shown. Ten K5-CreER;Sox2+ / + mice were treated with DMBA (n=5) or vehicle only (n=5). K5-CreER;Sox2+ / - mice were treated with DMBA (n=8) or vehicle only (n=8). Ten K5-CreER control mice were treated with DMBA (n=5) or vehicle only (n=5). Tongue and buccal mucosa were harvested and subjected to histologic examination. (FIG. 5B) A representative tongue tumor is shown using wide-field endoscopy with white light illumination. (FIG.5C) GFP-expressing tumor cells (arrow) are visualized in live animal using the first-generation front-view single-axis confocal laser endomicroscope. (FIGS.5D- 5E) Representative H.&E. pictures of two human primary oral squamous cell carcinoma are shown. MEER (FIG.5F) and MOC2-E6 / E7 (FIG.5G) tumor cells were implanted s.c. to C57BL / 6 mice. (FIG.5H-5I) Oral mucosa of tamoxifen-induced K5-CreER;Rosa26Sox2+ / + mice were harvested 4 weeks post-induction. (FIGS.5J-5K) Similar histologic features of OED between human and mouse buccal mucosa (near commissure) are shown. FIGS.6A-6C show OED transformation upregulates Slc2a1. (FIG.6A) Buccal mucosa of induced Sox2-GEMM and littermate control mice were harvested 4 weeks post- induction and subjected to IHC staining of the indicated markers. (FIG.6B) Buccal mucosa from control and Sox2-induced mice were harvested for qPCR analysis. (FIG.6C) Control, OED, and HNC from the Sox2-GEMM was stained for Slc2a1. FIGS.7A-7H show the high-risk immune signatures appear before HNC histology emerges. The buccal mucosa of the K5-CreER;Sox2+ / + GEMM was harvested across the time span of malignant transformation. CD45+ live intra-lesional immune cells were purified by FACS prior to scRNA-Seq. (FIGS.7A-7C) A total of 8,960 transcriptomes across the time span were analyzed. The UMAP showing divergent immune lineage segregation is determined by the differential expression of an immune marker matrix gene set and unique gene signatures. The functional annotation is informed by the highest expressed and most differentially expressed marker genes. (FIG.7D) The myeloid populations revealed by UMAP were subjected to PHATE trajectory analysis. Myeloid cell distribution patterns at Atty. Docket No. UM-41943.601 different time points are shown. (FIGS.7E-H) The most differentially expressed genes associated with OED transformation were mapped to feature plots (****p<.0001). FIG.8 shows YAP1 activation drives the initiation of HNC. A constitutively active form of Yap1 (Yap5sa) was induced by oral mucosa tamoxifen painting for five days. Buccal mucosa from K5- CreER- ;Yap5sa+ / - control and K5- CreER+;Yap5sa+ / - were harvested 6 weeks post-induction and stained for the indicated IHC markers. FIG.9A-9B shows YAP1-driven transformation remodels the local immune landscape. K5-CreER- ;Yap5sa+ / - control and K5-CreER+;Yap5sa+ / - mice were subjected to oral tamoxifen painting (n=4 per group). Six weeks after induction, buccal mucosa were harvested for FACS sorting of CD45+EpCAM-cells, which were subjected to single-cell RNA-Seq. (FIG.9A) Immune subsets were clustered using the highest expressed genes and a panel of anchor genes for lineage separation. (FIG.9B) The composition of each subset is shown. FIGS 10A-10C show confocal laser endomicroscopy provides an optical biopsy tool to capture high-risk cellular features of initiating HNC. (FIG.10A) A confocal laser endomicroscope is placed in contact with oral mucosa for imaging. The mucosa was excited at λ=488 nm. (FIGS.10B-C) C57BL / 6J mice were fed with 50 μg / ml 4-NQO drinking water for 16 weeks and then switched back to regular water for 8 weeks. Normal tongue (FIG.10B) and squamous cell carcinoma (FIG.10C) were treated with topical application of FITC- labeled, EGFR-targeted QRHKPRE (SEQ ID NO: 1) peptide for 5 minutes to generate contrast. The excess peptides were rinsed off with water. Confocal images were collected in vivo at 10 Hz. FIG.11A-11B show that IL-R1 blockade improves outcome for high risk oral lesions. FIG.11A shows improved survival in induced K5-CreER;Sox2+ / +mice treated with 200 µg anti-Il1r1 i.p. injections, starting one-week post-induction. FIG.11B shows an evaluation of tumors in K5-CreER;Sox2+ / +GEMM crossed with Il1r1- / -mice to generate the K5- CreER;Sox2+ / +;Il1r1- / -compound strain. K5-CreER;Sox2+ / +;Il1r1- / -mice showed significantly reduced areas of tumor littermate controls. FIGs.12A-12S demonstrate that blockade of IL1 signaling improves outcomes for high risk oral lesions. FIG.12A shows tumor growth following injection with MOC2 tumor cells Moc2-gW or MOC2-Sox2 cells positive for Sox2. FIG.12B shows results of flow cytometry analysis showing the percentages of CD8+ CD3+ T-cells in gW and Sox2 groups. Atty. Docket No. UM-41943.601 FIG.12C shows the frequency of CD45+ cells within the population of CD8+CD3+ T-cells. FIG.12D shows results of flow cytometry analysis showing the percentages of CD8+ CD11c+ dendritic cells in MOC2-gW and MOC2-Sox2 groups. FIG.12E shows the frequency of CD45+ cells within the population of CD8+ Cd11c+ cells in MOC2-Gw and MOC2-Sox groups. FIG.12F shows flow cytometry analysis of FSCA+ and Gr1+ cells in MOC2-gW and MOC2-Sox2 groups. FIG.12G shows the frequency of CD11b+ cells in the Gr1+ MDSCs. FIG.12H shows heat map analysis of Sox2, Il1a, and Csf2 expression in EV and Sox2 mice. FIG.12I shows mRNA fold change in IL1a in following administration of gW or Sox2 using the squamous cell line UMSC22b. FIG.12J shows probability of survival in K5-Cre / Sox2+ / + mice crossed with IL1r1+ / + or IL1R1- / -. FIG.12K shows cell expression of TCRy / g and CD11b. FIG.12L shows the frequency of CD45+ cells within the population of Cd11b- TCRγ / δ+ cells. FIG.12M shows expression of Glut1 and CD206 cells. FIG.12N shows the frequency of Cd11b+ cells within the population of Glut1+ CD206+ cells. FIG.12O shows tissue staining and FIG.12P quantifies tumor surface area in K5- Cre;Sox2+ / +;IL1r+ / + and K5-Cre;Sox2+ / +;IL1r- / - mice. FIG.12Q shows probability of survival in mice treated with anti-IL1r or PBS (control) following tumor induction. FIG. 12R shows results of flow cytometry performed to assess cell populations following treatment with PBS or anti-IL1r1. FIG.12S shows the frequency of PDL1+ GR1+ CD11B+ MDSCs in the different treatment groups. DEFINITIONS Although any methods and materials similar or equivalent to those described herein can be used in the practice or testing of embodiments described herein, some preferred methods, compositions, devices, and materials are described herein. However, before the present materials and methods are described, it is to be understood that this invention is not limited to the particular molecules, compositions, methodologies or protocols herein described, as these may vary in accordance with routine experimentation and optimization. It is also to be understood that the terminology used in the description is for the purpose of describing the particular versions or embodiments only, and is not intended to limit the scope of the embodiments described herein. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. However, in case of conflict, the present specification, including definitions, will Atty. Docket No. UM-41943.601 control. Accordingly, in the context of the embodiments described herein, the following definitions apply. As used herein and in the appended claims, the singular forms “a”, “an” and “the” include plural reference unless the context clearly dictates otherwise. Thus, for example, reference to “a compound” is a reference to one or more compounds and equivalents thereof known to those skilled in the art, and so forth. As used herein, the term “and / or” includes any and all combinations of listed items, including any of the listed items individually. For example, “A, B, and / or C” encompasses A, B, C, AB, AC, BC, and ABC, each of which is to be considered separately described by the statement “A, B, and / or C.” As used herein, the term “comprise” and linguistic variations thereof denote the presence of recited feature(s), element(s), method step(s), etc. without the exclusion of the presence of additional feature(s), element(s), method step(s), etc. Conversely, the term “consisting of” and linguistic variations thereof, denotes the presence of recited feature(s), element(s), method step(s), etc. and excludes any unrecited feature(s), element(s), method step(s), etc., except for ordinarily-associated impurities. The phrase “consisting essentially of” denotes the recited feature(s), element(s), method step(s), etc. and any additional feature(s), element(s), method step(s), etc. that do not materially affect the basic nature of the composition, system, or method. Many embodiments herein are described using open “comprising” language. Such embodiments encompass multiple closed “consisting of” and / or “consisting essentially of” embodiments, which may alternatively be claimed or described using such language. As used herein, the terms “administration” and “administering” refer to the act of introducing a substance, such as a drug, prodrug, or other agent, or therapeutic treatment to a subject or in vivo, in vitro, or ex vivo cells, tissues, and organs. In general, any route of administration may be utilized including, for example, parenteral (e.g., intravenous), oral, topical, subcutaneous, peritoneal, intra-arterial, inhalation, vaginal, rectal, nasal, introduction into the cerebrospinal fluid, infusion, or instillation into body compartments. Exemplary routes of administration to the human body can be by parenteral administration (e.g., intravenously, subcutaneously, etc.), orally, etc. As used herein, the term “approximately” and “about” is intended to encompass normal statistical variation as would be understood by those of ordinary skill in the art as appropriate to the relevant context. In certain embodiments, the term “approximately” or “about” refers to a range of values that fall within 25%, 20%, 19%, 18%, 17%, 16%, 15%, Atty. Docket No. UM-41943.601 14%, 13%, 12%, 11%, 10%, 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2%, 1%, or less in either direction (greater than or less than) of the stated reference value unless otherwise stated or otherwise evident from the context (except where such number would exceed 100% of a possible value). As used herein, the terms “co-administration” and “co-administering” refer to the administration of at least two agent(s) or therapies to a subject. In some embodiments, the co- administration of two or more agents or therapies is concurrent (e.g., in a single formulation / composition or in separate formulations / compositions). In other embodiments, a first agent / therapy is administered prior to a second agent / therapy. Those of skill in the art understand that the formulations and / or routes of administration of the various agents or therapies used may vary. The appropriate dosage for co-administration can be readily determined by one skilled in the art. In some embodiments, when agents or therapies are co- administered, the respective agents or therapies are administered at lower dosages than appropriate for their administration alone. Thus, co-administration is especially desirable in embodiments where the co-administration of the agents or therapies lowers the requisite dosage of a potentially harmful (e.g., toxic) agent(s), and / or when co-administration of two or more agents results in sensitization of a subject to beneficial effects of one of the agents via co-administration of the other agent. As used herein, the term “pharmaceutical composition” refers to the combination of an active agent with a pharmaceutically acceptable carrier, inert or active, making the composition especially suitable for diagnostic or therapeutic use in vitro, in vivo or ex vivo. As used herein, the term “pharmaceutically acceptable carrier” refers to any of the standard pharmaceutical carriers including, but not limited to, phosphate buffered saline solution, water, emulsions (e.g., such as an oil / water or water / oil emulsions), and various types of wetting agents, any and all solvents, dispersion media, coatings, sodium lauryl sulfate, isotonic and absorption delaying agents, disintegrants (e.g., potato starch or sodium starch glycolate), and the like. The compositions also can include stabilizers and preservatives. For examples of carriers, stabilizers and adjuvants, see, e.g., Martin, Remington's Pharmaceutical Sciences, 15th Ed., Mack Publ. Co., Easton, Pa. (1975), incorporated herein by reference in its entirety. The term “pharmaceutically acceptable” as used herein, refers to compositions that do not substantially produce adverse reactions, e.g., toxic, allergic, or immunological reactions, when administered to a subject. Atty. Docket No. UM-41943.601 The term “selectively” as used herein in reference to selectively removing and / or selectively abolishing a high-risk oral lesion indicates that the high risk oral lesion is targeted for removal from a subject. The term “selectively” does not necessarily indicate that the high risk oral lesion is exclusively removed from the subject. In some embodiments, portions of non-high risk oral lesions may also be removed during selective removal of a high risk oral lesion. As used herein, the term “subject” broadly refers to any animal, including human and non-human animals. In some embodiments, the subject is a vertebrate. In some embodiments, the subject is a mammal. In some embodiments, the subject is a human. As used herein, the term “treatment” (also “treat” or “treating”) refer to obtaining a desired pharmacologic and / or physiologic effect against a particular disease, disorder, or condition. Preferably, the effect is therapeutic, i.e., the effect partially or completely alleviates, ameliorates, relives, inhibits, delays onset of, reduces severity of, and / or reduces frequency, incidence or severity of one or more symptoms, features, and / or causes of a particular disease, disorder, and / or condition. Such treatment may be of a subject who does not exhibit signs of the relevant disease, disorder and / or condition and / or of a subject who exhibits only early signs of the disease, disorder, and / or condition. Alternatively or additionally, such treatment may be of a subject who exhibits one or more established signs of the relevant disease, disorder and / or condition. In some embodiments, treatment may be of a subject who has been diagnosed as suffering from the relevant disease, disorder, and / or condition. In some embodiments, treatment may be of a subject known to have one or more susceptibility factors that are statistically correlated with increased risk of development of the relevant disease, disorder, and / or condition. In some embodiments, “treating” indicates preventing a benign lesion (e.g. OEP, OLP) from progressing to malignancy in a subject (e.g. head and neck cancer (HNC). For example, “treating” encompasses selectively removing a high-risk lesion from a subject, thereby preventing the lesion from otherwise transforming to malignancy in the subject. As used herein, the term “preventing” refers to prophylactic steps taken to reduce the likelihood of a subject (e.g., an at-risk subject) from developing or suffering from a particular disease, disorder, or condition. The likelihood of the disease, disorder, or condition occurring in the subject need not be reduced to zero for the preventing to occur; rather, if the steps reduce the risk of a disease, disorder or condition across a population, then the steps prevent the disease, disorder, or condition for an individual subject within the scope and meaning herein. For example, in some embodiments “treating” Atty. Docket No. UM-41943.601 indicates taking steps to reduce the likelihood of a subject developing HNC by selectively removing high-risk lesions from the subject, thereby effectively “preventing” cancer in a subject. DETAILED DESCRIPTION Diagnosis of oral epithelial dysplasia is based on a histopathological grading system that has no prognostic value in determining which lesions have a high risk of transforming to malignancy. The methods provided herein are predicated at least in part on the discovery that expression of solute carrier family 2 member 1 (SLC2A1, also known as GLUT1) provides a basis for effective stratification of oral lesions based upon risk of progressing to head and neck cancer. Accordingly, the methods provided herein address a long-felt need for early and accurate assessment of dangerous pre-cancerous lesions to enable rapid treatment prior to progressing to head and neck cancer, which has only a 5 year survival rate. Moreover, the methods provided herein facilitate selective removal of high-risk lesions and thus prevent sweeping and invasive surgical treatments from being performed on a subject to remove lesions that actually have minimal risk of transforming, and induce significant damage to qualify of life in the process. In some aspects, provided herein are methods of evaluating one or more oral lesions. The term “oral lesion” is used herein in the broadest sense and is inclusive of any lesion found in the oral cavity of the subject. In some embodiments, the lesion is found in the lining of the oral cavity. This lining is called the mucosa. The mucosa covers the inside of thecheeks, the inside of the lips, the gums, the tongue, and the roof and floor of the mouth. Orallesions include, for example, oral epithelial dysplasia (OED), oral lichen planus (OLP), oral leukoplakia, oral submucous fibrosis (OSMF), proliferative verrucous leukoplakia (PVL), and oral erythroplakia. In some embodiments, a lesion is an oral epithelial dysplasia (OED). In some embodiments, a lesion is an oral lichen planus (OLP). In some embodiments, provided herein are methods of stratifying one or more oral lesions based upon risk of progressing to malignancy. The term “malignant”, “malignancy”, “cancer”, and “cancerous” are used interchangeably herein. The term “progressing” or “transforming” are used interchangeably herein and refer to the process of a benign lesion becoming malignant (e.g. cancerous). In some embodiments, provided herein is a method of stratifying an oral lesion based upon risk of progressing to malignancy, comprising measuring expression of SLC2A1 in an oral lesion and identifying the risk of progressing to Atty. Docket No. UM-41943.601 malignancy based upon expression of SLC2A1 in the lesion. In some embodiments, the methods comprise identifying the oral lesion on a spectrum of high risk to low risk, with high risk indicating a high probability of progressing to malignancy and low risk indicating a low probability of progressing to malignancy (e.g. a high probability of remaining benign). For example, in some embodiments, the methods comprise identifying the oral lesion as high risk, moderate-to-high risk, moderate risk, moderate-to-low risk, or low risk based upon expression of SLC2A1 on the oral lesion. In some embodiments, stratifying the oral lesion based upon risk comprises assigning a risk score to the lesion, as described in further detail below. In some embodiments, other factors are considered in determining how the lesion is classified. For example, in some embodiments expression of one or more markers other than SLC2A1 are measured in the lesion and used to assess risk of transformation. These one or more markers are referred to herein as “risk markers” or “additional risk markers”. Alternatively, in some embodiments one or more additional risk markers are measured in the lesion, and the expression of SLC2A1 and expression of the one or more additional risk markers are used in the identifying the risk of progression to malignancy for the lesion. Suitable markers (e.g. additional risk markers) that may be measured include, for example, epidermal growth factor receptor (EGFR), SRY-Box Transcription Factor 2 (SOX-2), nuclear expression of Yes1 Associated Transcriptional Regulator (YAP-1), expression of phosphorylated YAP-1, and / or type-I interferon (IFN-I) target genes, including C-X-C motif chemokine ligand 10 (CXCL10), Interferon Induced Transmembrane Protein 3 (IFITM3), ISG15 Ubiquitin Like Modifier (ISG15), Major Histocompatibility Complex, Class II, DR Alpha (HLA-DRA), and Major Histocompatibility Complex, Class II, DR Beta 1 (HLA- DRB1). In some embodiments, one or more cell types are measured / quantified in the oral lesion and used to assess risk, alone or in combination with expression level of SCL2A1 in the lesion. For example, in some embodiments myeloid cells are measured in the oral lesion. In some embodiments, increased expression of the aforementioned markers on myeloid cells in the lesion is indicative of a high risk of progressing to malignancy, which can occur prior to the emergence of invasive HNC histology. In some embodiments, lymphocytes (e.g. cytotoxic T-lymphocytes) are measured / quantified in the lesion. In some embodiments, macrophages are measured / quantified in the lesion. In some embodiments, regulatory T-cells (Tregs), B-cells, or regulatory B-cells (Bregs) are measured / quantified in the lesion. In some Atty. Docket No. UM-41943.601 embodiments, decreased lymphocytes (e.g. decreased cytotoxic T-lymphocytes, including CD8+ T cells, TBET+CD8+ T cells, CXCR3+CD8+ T cells, natural killer (NK) cells, and γδ‐ T cells, and / or increased expression of CD206+ macrophages and CD33+PD-L1+ myeloid cells in the lesion are indicative of high risks. In some embodiments, intralesional myeloid cells that assume an IL1(high) IFN-1(low) signature are indicative of a high risk of progressing to malignancy. For example, in some embodiments decreased CD8+ CD3+ T cells or activation thereof (e.g. decreased CD45+ CD8+ CD3+ T cells), decreased CD8+ CD11c+ cells (e.g. decreased type1 conventional dendritic cells, cDC1) or activation thereof (e.g. decreased CD45+ CD8+ CD11c+ cells), increased FSC-A+ GR1+ cells, increased myeloid derived suppressor cells (e.g. increased CD11b+ GR1+ cells), decreased Cd11b- TCRγ / δ+ cells or activation thereof (e.g. decreased Cd11b-TCRγ / δ+ CD45+ cells), decreased Cd11b+ TCRγ / δ+ cells, or decreased Glut1+ CD206+ macrophages or activation thereof (e.g. decreased Glut1+ CD206+ Cd11b+ cells) are indicative of a high risk of progressing to malignancy. In some embodiments, the lesion is selectively abolished when a high risk of progressing to malignancy is identified. In some embodiments, cell types including the above cell types are measured and used to assess risk of progressing to malignancy as described above, without measuring SLC2A1. In some embodiments, cell types including the above cell types are measured and used to assess risk of progressing to malignancy in combination with SLC2A1. In some embodiments, increased expression of PDL1+ GR1+ CD11b+ myeloid depressor cells is indicative of a high risk of progressing to malignancy. In some embodiments, expression of a marker on a specific cell type is assessed in the lesion and used to stratify risk. For example, in some embodiments expression of markers on epithelial cells, expression of markers on immune cells, and / or expression of markers on myeloid cells is measured in the lesion and used to assess risk of transformation. In some embodiments, methods of stratifying one or more oral lesions based upon risk of progressing to malignancy comprise evaluating expression of one or more cell types in a the oral lesion or a sample obtained from the oral lesion. In some embodiments, decreased CD8+ CD3+ T cells or activation thereof (e.g. decreased CD45+ CD8+ CD3+ T cells), decreased CD8+ CD11c+ cells (e.g. decreased type1 conventional dendritic cells, cDC1) or activation thereof (e.g. decreased CD45+ CD8+ CD11c+ cells), increased FSC-A+ GR1+ cells, increased myeloid derived suppressor cells (e.g. increased CD11b+ GR1+ cells), decreased Cd11b- TCRγ / δ+ cells or activation thereof (e.g. decreased Cd11b-TCRγ / δ+ CD45+ cells), decreased Cd11b+ TCRγ / δ+ cells, or decreased Glut1+ CD206+ macrophages Atty. Docket No. UM-41943.601 or activation thereof (e.g. decreased Glut1+ CD206+ Cd11b+ cells) are indicative of a high risk of progressing to malignancy. In some embodiments, the lesion is selectively abolished when a high risk of progressing to malignancy is identified. In some embodiments, the methods comprise identifying the oral lesion as high risk when the expression level of SLC2A1 is equal to or above a threshold value, or identifying the oral lesion as low risk when the expression level of SLC2A1 is below a threshold value. In some embodiments, the methods comprise identifying the oral lesion as high risk when the amount of percentage of a given cell type, including those described above, is equal to or above a threshold value, or identifying the oral lesion as low risk when the amount or percentage of the given cell type is below a threshold value. In some embodiments, the methods comprise identifying the oral lesion as high risk when the amount of percentage of a given cell type, including those described above, is below a threshold value, or identifying the oral lesion as low risk when the amount or percentage of the given cell type is equal to or above a threshold value. The terms “threshold”, “cutoff”, and “control” are used interchangeably herein to refer to a value at or above which a certain risk (e.g. high risk) is determined for a lesion and below which a different risk (e.g. low risk) is determined. In some embodiments, the methods comprise assigning a risk score to the lesion based upon expression of one or more markers and / or cell types in the lesion. In some embodiments, the risk score is assigned based upon SLC2A1 expression, a combination of SLC2A1 expression and expression of additional potential biomarkers, or a combination of SLC2A1 expression and the amount of one or more cell types in the lesion, or the amount or percentage of one or more cell types in the lesion. In some embodiments, expression of SLC2A1 and the additional markers or cell types is given equal value (e.g. equal weight) in determining the risk score for the lesion. In some embodiments, one or more factors (e.g. markers, cell types) receives more weight than another factor in determining the risk score for the lesion. In other words, in some embodiments each different factor does not receive equal weight in determining the overall risk score for the lesion. In some embodiments, expression of solute carrier family 2 member 1 (SLC2A1) and expression of an additional risk marker is measured in a plurality of oral lesions. In some embodiments, an oral lesion is identified as high risk when the expression level of SLC2A1 in the oral lesion is equal to or above a threshold value for SLC2A1 and / or the expression level of the additional risk marker in the lesion is equal to or above a threshold value for the additional risk marker. In some embodiments, an oral lesion is identified as low risk when Atty. Docket No. UM-41943.601 the expression level of SLC2A1 in the oral lesion is less than a threshold value and the expression of the additional risk marker in the oral lesion is less than a threshold value. In some embodiments, an oral lesion is identified as high risk when the SLC2A1- additional marker interaction score for the oral lesion is above a threshold interaction score; and / or the oral lesion expresses SLC2A1 (or the expression level of SLC2A1 in the oral lesion is equal to or above a threshold value for SLC2A1) and a Wasserstein distance between the oral lesion and a neighboring oral lesion that expresses the additional risk marker is equal to or less than a threshold Wasserstein distance. In some embodiments, the interaction score (e.g. SLC2A1-additional marker interaction score) is calculated by multiplying the expression of SLC2A1 in the oral lesion by the expression of the additional risk marker in the oral lesion, and dividing the result by the number of cells within the region of interest expressing both SLC2A1 and the additional risk marker. In some embodiments, an SLC2A1-additional marker interaction score above a threshold interaction score indicates that a lesion is high risk. In some embodiments, an oral lesion is identified as high risk when the lesion expresses SLC2A1 (or expression level of SLC2A1 in the oral lesion is equal to or above a threshold value for SLC2A1) and a Wasserstein distance between the oral lesion and a neighboring oral lesion that expresses the additional risk marker is equal to or less than a threshold Wasserstein distance. The equation to calculate the Wasserstein distance is as follows: where P and Q are the empirical spatial distribution of two cell types and ^(^,^) denotes all joint distributions ^ for (X, Y) with marginal P and Q. In some embodiments, a decreasing Wasserstein distance between a cell that expresses SLC2A1 and a cell that expresses an additional risk marker indicates increasing risk of malignancy for the cell expressing SLC2A1 and / or the cell expressing the additional risk marker (e.g. EGFR). In some embodiments, the additional risk marker is EGFR, YAP-1, or SOX-2. In some embodiments, expression of SLC2A1 and / or expression of the additional risk markers in the oral lesion is measured by obtaining a biopsy sample from the oral lesion and measuring expression in the biopsy. In some embodiments, expression is measured at the protein level. In some embodiments, expression is measured by tissue staining. For example, Atty. Docket No. UM-41943.601 in some embodiments expression is measured by immunohistochemistry. In some embodiments, expression is measured by other immunoassays, including immunofluorescence, FISH, multiplexed assays, and the like. In some embodiments, expression is measured at the nucleic acid level (e.g. mRNA). For example, in some embodiments expression is measured at the nucleic acid level by PCR-based techniques (e.g. qPCR, FISH, and spatial transcriptomics etc.) In some embodiments, expression of SLC2A1 and / or expression of additional risk markers is measured on a cell type, and the expression on the cell type is used in the determination of risk of the lesion progression to malignancy. For example, in some embodiments expression of SLC2A1 is measured on myeloid cells within the lesion. In some embodiments, the amount of Glut1+CD206+CD11b+ macrophages is measured within the lesion. In some embodiments, myeloid cells are identified in a biopsy by immunostaining, and SLC2A1 positive myeloid cells are measured. In some embodiments, myeloid cells are identified by expression of one or more cell surface markers associated with myeloid cell lineage, such as CD33, CD68, or both, and SLC2A1 positive myeloid cells are measured. In some embodiments, the presence of SLC2A1 positive myeloid cells in a lesion indicates that the lesion is high risk. In some embodiments, an amount of Glut1+CD206+CD11b+ macrophages below a threshold value is indicates that the lesion is high risk. In some embodiments, expression of SLC2A1 and / or expression of additional risk markers in the oral lesion is measured directly on the oral lesion within the oral cavity of a subject. In other words, in some embodiments a biopsy is not obtained, and rather the expression of SLC2A1 (and / or expression of an additional risk marker) is measured on the oral lesion within the oral cavity of a subject. In some embodiments, expression of SLC2A1 on the oral lesion is measured by contacting the oral lesion with an agent binds to SLC2A1. In some embodiments, expression of an additional risk marker on the oral lesion is measured by contacting the oral lesion with an agent that binds to the additional risk marker. In some embodiments, the agent comprises a label (also referred to as a “detectable label”), such that the agent will bind to lesions expressing a given marker (e.g. SLC2A1, YAP-1, SOX-2, EGFR, etc.) which lesions can subsequently be identified by detecting the label. In some embodiments, the label comprises a fluorescent label. Any suitable fluorescent label can be used. Exemplary suitable fluorescent labels include fluoresceins, coumarins, rhodamines, cyanines, and derivatives thereof. In some embodiments, the label comprises a near infrared dye. In some embodiments, the label comprises Fluorescein isothiocyanate (FITC). Atty. Docket No. UM-41943.601 In some embodiments, the agent that binds to SLC2A1 is a peptide. In some embodiments, the peptide binds to a portion of the extracellular domain of SLC2A1. In some embodiments, the peptide comprises 5-50 amino acids. In some embodiments, the peptide comprises 5-40 amino acids, 5-35 amino acids, 5-30 amino acids, 5-25 amino acids, 5-20 amino acids, or 5-10 amino acids. In some embodiments, the peptide is conjugated to the detectable label by a linker. The linker may be any suitable length and combination of amino acids to facilitate attachment of the linker to the peptide without disrupting the ability of the peptide to bind to SLC2A1. In some embodiments, detecting the label comprises applying a light source to the oral lesion, thereby exciting the label. For example, in some embodiments detecting the label comprises applying a light source (e.g. a laser) having the appropriate excitation wavelength to the oral lesion, thereby exciting the label and allowing for detection of the label. In some embodiments, the label is detected visually. In some embodiments, the label is detected by the naked eye (e.g. without the need for detection equipment such as a microscope). Accordingly, the methods provided herein facilitate simple and inexpensive determination of high risk oral lesions without necessitating expensive medical equipment or a biopsy, and thus provide for rapid and accessible point-of-care diagnosis and treatment of high-risk lesions in a subject. In some embodiments, the methods provided herein further comprise selectively abolishing the oral lesion when the oral lesion is identified as high risk. In some embodiments, the methods provided herein further comprise providing an inhibitor of IL-1α signaling to the subject. In some embodiments, the continued progression of high risk lesions can be delayed (e.g. arrested, inhibited, slowed) by an inhibitor of IL-1α signaling. In some embodiments, the inhibitor of IL-1α signaling comprises an anti-IL-1R antibody. In some embodiments, the inhibitor of IL-1α signaling (e.g. the anti-IL-1R antibody) is administered to the subject systemically. In some embodiments, the inhibitor of IL-1α signaling (e.g. the anti-IL-1R antibody) is administered to the high risk oral lesion, such as by injection into the lesion. In some embodiments, the methods provided herein further comprise monitoring the oral lesion when the oral lesion is not identified to be high risk. For example, the methods may comprise monitoring the oral lesion when the oral lesion is identified as low risk (or low-to moderate risk, or moderate risk). Atty. Docket No. UM-41943.601 In some aspects, provided herein are methods of treating a subject at risk of having head and neck cancer (HNC). In some embodiments, the subject is identified to be at risk of having HNC due to the presence of one or more oral lesions in the oral cavity of the subject. In some embodiments, provided herein is a method of treating a subject at risk of having head and neck cancer (HNC) comprising stratifying an oral lesion based upon risk of progressing to malignancy as described above, and selectively abolishing the oral lesion when the oral lesion is identified as high risk. In some embodiments, the methods provided herein further comprise monitoring the oral lesion when the oral lesion is not identified to be high risk. For example, the methods may comprise monitoring the oral lesion when the oral lesion is identified as low risk (or low-to moderate risk, or moderate risk). In some embodiments, selectively abolishing the oral lesion comprises removing the oral lesion from the subject. For example, lesions (e.g. high risk lesions) can be removed by incision or other surgical removal technique. In some embodiments, selectively abolishing the oral lesion comprises damaging / destroying the lesion tissue, such as by ablation, heat, radiation, chemotherapy, cryotherapy, etc. In some embodiments, “monitoring” the oral lesion comprises repeating the methods described herein after a suitable duration of time has passed to re-assess the risk of the lesion transforming to malignancy. In some embodiments, “monitoring” the oral lesion comprises repeating the methods described herein after 1 month, 2 months, 3 months, 4 months, 5 months, 6 months, 7 months, 8 months, 9 months, 10 months, 11 months, 12 months, or more than 1 year has passed to re-assess the risk of the lesion transforming to malignancy. In some embodiments, a method of treating a subject at risk of having head and neck cancer (HNC) comprises assessing risk of transforming to malignancy for a single lesion, and selectively abolishing or monitoring the lesion based upon the risk assessment. In some embodiments, a method of treating a subject at risk of having head and neck cancer comprises simultaneously assessing risk of transforming to malignancy for a plurality of oral lesions, and selectively abolishing or monitoring each of the plurality of oral lesions assessed. In some aspects, provided herein are methods of measuring expression of SLC2A1 in at least one oral lesion directly within the oral cavity of a subject (as opposed to measuring expression of SLC2A1 in a biopsy). In some embodiments, expression of SLC2A1 is measured by contacting the at least one oral lesion with an agent that binds to SLC2A1. In some embodiments, the agent comprises a label (e.g. a detectable label). In some Atty. Docket No. UM-41943.601 embodiments, expression of SLC2A1 on the oral lesion is measured by contacting the oral lesion with an agent binds to SLC2A1. In some embodiments, the agent comprises a label (also referred to as a “detectable label”), such that the agent will bind to lesions expression SLC2A1 which lesions can subsequently be identified by detecting the label. In some embodiments, the label comprises a fluorescent label. Any suitable fluorescent label can be used. Exemplary suitable fluorescent labels include fluoresceins, coumarins, rhodamines, cyanines, and derivatives thereof. In some embodiments, the label comprises Fluorescein isothiocyanate (FITC). In some embodiments, the agent that binds to SLC2A1 is a peptide. In some embodiments, the peptide comprises 5-50 amino acids. In some embodiments, the peptide comprises 5-40 amino acids, 5-35 amino acids, 5-30 amino acids, 5-25 amino acids, 5-20 amino acids, or 5-10 amino acids. In some embodiments, the peptide is conjugated to the detectable label by a linker. The linker may be any suitable length and combination of amino acids to facilitate attachment of the linker to the peptide without disrupting the ability of the peptide to bind to SLC2A1. In some embodiments, detecting the label comprises applying a light source to the oral lesion, thereby exciting the label. For example, in some embodiments detecting the label comprises applying a light source (e.g. a laser) having the appropriate excitation wavelength to the oral lesion, thereby exciting the label and allowing for detection of the label. In some embodiments, the label is detected visually. In some embodiments, the label is detected by the naked eye (e.g. without the need for detection equipment such as a microscope). EXAMPLES EXAMPLE 1 Oral leukoplakia and lichen planus (OLP) are the most common oral mucosal lesions, and ~284 million people require long-term follow-ups for head and neck squamous cell carcinoma (HNC) screens globally. Current methods, such as autofluorescence imaging, tissue reflectance, cytology, and circulating tumor DNA, may be helpful in separating a pre- malignant lesion, oral epithelial dysplasia (OED), from HNC. However, this separation has never been a diagnostic challenge histologically. Rather, the main challenge for the care of these patients is that the majority of OED and OLP will remain benign, and surgery to Atty. Docket No. UM-41943.601 remove benign lesions is impractical due to the degradation of overall quality of life should invasive and aggressive surgery be performed. Moreover, it is impossible to maintain high- frequency follow-ups for every patient to determine whether a previously benign OED / OLP has progressed to malignancy, and to date no consistent features have been established to separate high-risk OED from low-risk lesions. Indeed, standard histologic grading cannot predict transformation risks. Although most lesions remain benign, some mild OED can progress to HNC within a year. Once diagnosed, HNC has a dismal 5-year overall survival rate. Thus, clinical examination is integral for HNC risk evaluations. During a standard clinical examination of OED / OLP, erythematous color change and induration warrant a biopsy. These features indicate early inflammatory and mechanical changes in the microenvironment of initiating HNC. Herein, models were engineered to recapitulate these immune and mechanical alterations over the course of HNC initiation. Twi driver oncogenes commonly associated with HNC were selected. The first is SOX2, which is located at the 3q26.3 hotspot and amplified in ~21% of HNC. SOX2 excludes cytotoxic T- lymphocytes (CTL) from the tumor microenvironment (TME) by inhibiting type-I interferon (IFN-I)-mediated immune detection of cancer. The second gene is YAP1, which is a signaling hub for sensing extracellular stiffness. YAP1 frequently activated through genomic amplification at the 11q22 locus, the loss-of-function mutations of the YAP1-inhibitory FAT1 gene, and the environmental mechanical cues. Over 2 / 3 of HNC shows evidence of YAP1 pathway activation. Thus, two genetically engineered mouse models (GEMMs) driven by oral epithelial cell-specific activation of Sox2 and Yap1, respectively, were established. Both GEMMs demonstrated the high-fidelity histological transformation of normal oral mucosa to OED and then to HNC. In addition, matched longitudinal specimens were used to show that the high-risk OED had reproducible immune landscape shifts upon transformation. Taken together, the results presented herein demonstrate that the induction of the SLC2A1 (a.k.a. GLUT1)- glycolysis pathway and the loss of IFN-I signatures establish a powerful tumorigenic niche, which occurs prior to the emergence of HNC histology. To complement on-slide signature discovery, a confocal laser endomicroscopy (CLE) technology was developed. CLE which has been approved by the Food and Drug Administration (FDA) to monitor molecular targets below mucosal surfaces at microscopic resolution and assist the detection of epithelial premalignancy. The detection of OED with a bona fide high-risk can be accomplished using a combination of tissue and clinical evaluation. In sum, the results presented herein demonstrate that immune, metabolic, and Atty. Docket No. UM-41943.601 mechanical signatures inform on-slide and optical biopsy technologies for the precise detection of high-risk OED. RESULTS Driven by distinct etiologies, there are two distinct diseases of HNC defined by the HPV status. HPV+ HNC responds better to chemoradiation therapy. Thus, HPV+ HNC provides a model to identify mechanisms inhibiting immunogenicity. Expansion of suppressive myeloid cells and contraction of CTL are signatures of OED transformation. A multispectral imaging (MSI) analysis platform was used to stain a longitudinal cohort of transformed OED.6-13 areas of interest were selected for high-quality scanning and subsequent analyses for each slide. The malignant transformation of OED was associated with a massive influx of myeloid cells marked as CD33+PD-L1+ (Figure 1A) and CD68+PD- L1+ (data not shown). The dysplastic epithelium and tumor islands were surrounded by variably dense effector CD8+ CTL. Notably, the frequencies of two functional IFN-γ- producing Th1 subsets, T-BET+CD8+ and CXCR3+CD8+ CTL were significantly reduced upon malignant transformation (Figure 1B-C). Loss of IFN-I signature correlates with CTL contraction in HNC Utilizing an enhanced pipeline that stabilizes immune lineage segregation, a UMAP analysis of an HNC single-cell resource was performed. HPV+ HNC showed significantly fewer cDC1 (cluster 25), CXCL10 high macrophages (clusters 10, 19, and 20), and γδ T-cells (cluster 4) (Figure 2A). Clusters 10, 19, and 20 expressed high levels of IFN-I target genes, such as CXCL10, IFITM3, ISG15, HLA-DRA, and HLA-DRB1. IFN-I and its target genes play a role in antigen-presenting cell (APC) maturation and CTL chemotaxis. The induction of IFN-I signatures depends on the proper activation of Pattern Recognition Receptors (PRRs), which are central for the host immune system to detect “non-self”. Unstable cancer genomes and treatments lead to DNA damage, which serves as a danger signal to alert PRRs. STING is a central cytoplasmic DNA-sensing PRR that promotes anti-tumor immunity. STING translates cytoplasmic DNA signals to IFN-I induction and facilitates APC maturation. Utilizing a deconvolution method, the TILs landscape of 520 HNC was resolved. The levels of STING were positively correlated with M1-like macrophages, CD8+ CTL, and Atty. Docket No. UM-41943.601 γδ T-cells (Figure 2B). The STING-CTL correlation was also validated using human HNC tissue microarrays consisting of cores from 264 specimens. The malignant transformation of oral mucosa upregulates glycolysis signatures To recognize transforming cells, APCs migrate to the lesion, process antigens, synthesize high levels of Th1 cytokines de novo, and cross-prime CTL. CD8+ CTL replicate genomes rapidly, engage synapses and deliver cytotoxic cargo into transformed cells. These processes are fueled by extracellular glucose and a lack of metabolic support results in CTL exhaustion. Hypoxia is a key feature of HNC, which shows a median partial pressure of oxygen (pO2) of 9 mmHg, in contrast with a pO2 of 40-60 mmHg in the paired normal mucosa. Hypoxia switches the dominant metabolism from oxidative phosphorylation to glycolysis in HNC, the latter of which depletes the glucose reservoir in the TME. Indeed, hypoxia treatment upregulated glycolytic signatures in HNC cells, including SLC2A1, Hexokinase 1 (HK1), BCL2 interacting protein 3 like (BNIP3L), and pyruvate dehydrogenase kinase 1 (PDK1) (Figure 3A-B). Then, 520 HNC was compared with 44 normal mucosae and it was found that transformation increased glycolysis signatures, including SLC2A1, PDK1, and PKM (Figure 4A). The TILs were deconvolved, and an inverse correlation was found between SLC2A1 levels and anti-tumor TIL subsets such as CTL, memory T-cells, and M1-like macrophages (Figure 4B). SOX2-driven OED progression shows features of IFN-I signature loss and upregulation of glycolysis To identify signatures of high-risk OED, better models to rigorously decipher the epithelial and immune signature changes as a function of time are needed. SOX2 is frequently amplified in squamous cell carcinomas, including HNC. To longitudinally monitor SOX2-driven OED transformation, K5-CreER was crossed with Rosa26- Sox2- Enhanced Green Fluorescence Protein (EGFP) and the oral mucosa was topically painted with tamoxifen to produce the Sox2-GEMM, in which oral epithelial cells expressed Sox2- EGFP. Sox2- GEMM or K5-CreER control mice were treated with a carcinogen, DMBA, or vehicle. Buccal mucosa and tongue were subjected to histologic examination for OED / HNC penetrance rates. The K5-CreER+;Sox2+ / - strain without DMBA showed about 60-80% HNC penetrance 160 days post-induction, and the K5- CreER+;Sox2+ / + strain showed 100% spontaneous penetrance about 4 weeks post-induction (Figure 5A). Some tumors can be seen grossly (Figure 5B), and most of the tumors from the K5-CreER+;Sox2+ / + strain were Atty. Docket No. UM-41943.601 microscopic four weeks post-induction. An endomicroscope was engineered which can access the mouse oral cavity for direct laser confocal imaging and reveal the EGFP- expressing Sox2+ transformed epithelial cells in live animals (Figure 5C). Although valuable for the characterization of HNC TME, the implantable HNC models cannot recapitulate the stromal complexity, cytologic and architectural details, and, importantly, the temporal progression from OED to HNC (Figure 5F-G). Sox2 drove the spontaneous development of invasive HNC. The histology of Sox2-driven cancer bore a high-fidelity resemblance to human disease, including high-grade cytologic features, the formation of keratin pearls, and the establishment of stroma and immune-cell infiltration (Figure 5H-I). Notably, this model showed all histologic features of OED prior to HNC (Figure 5J-K). These tumors were positive for Sox2 with a high proliferation index (Ki-67). The staining patterns of p53 and p63 were similar to those of human HNC (Figure 6A). The Sox2-GEMM showed elevated levels of Slc2a1 mRNA upon transformation (Figure 6B). Normal mucosa was also stained, transforming OED 2 weeks post-induction, and HNC 5 weeks post-induction with a Slc2a1-specific antibody. In normal mucosa, the Slc2a1 stain was either negative or showed a weak cytoplasmic staining pattern. However, in transforming OED, Slc2a1 assumed a conspicuous membranous staining pattern in the lesional areas, with a clear demarcation from the adjacent relatively normal epithelium, which lacked the membrane staining pattern. HNC showed a diffuse strong membranous staining pattern (Figure 6C). Immune signatures emerge before the development of HNC histology To test whether early immune alterations identify high-risk OED, single-cell (sc)RNA-Seq of intra-lesional immune cells was performed as benign mucosa transformed. Normal oral mucosa contained Th1, Th2, Th17 cells as well as balanced IFN-I signature-high M1-like macrophages and suppressive myeloid cells. Sox2- induced transformation triggered expansion of myeloid cells and concomitant CTL loss (Figure 7A-C), also seen in human specimens (Figure 1). The transformation-driven myeloid clusters show distinct compensatory suppressive features. For example, a subset of them expresses Pd-l1. Functional suppression also involves Galectin-9+, Cox2+, or Tgfb1+ clusters that are negative for Pd-l1. The suppressive features become more diverse as the OED progressed (Figure 7B-C). Notably, the two-week time point did not show evidence of an invasive disease but only OED. However, key features of local immune suppression were already established at this early stage. Thus, the high-risk immune signatures appeared before HNC histology, which makes them ideal marker candidates for early detection. Atty. Docket No. UM-41943.601 Myeloid cell clusters were separated their differentiation trajectory was mapped using potential of heat diffusion for affinity-based transition embedding (PHATE) (92). Myeloid expansion was heavily attributed to de novo recruitment (discontinuous clusters), in addition to some in situ proliferation (continuous clusters) (Figure 7D). Then, the differentially expressed genes in myeloid cells from normal and transforming oral mucosa as characterized. The resting myeloid cells in normal mucosa were distributed in the southwest and northeast corners of the PHATE map (Figure 7D, left panel). Transformation resulted in the proliferation and recruitment of the myeloid cells, and these new myeloid cells exhibited a distinct signature, which included high levels of Slc2a1 and Pd-l1, with a concurrent loss of IFN-I signatures such as Stat1 and Isg15 (Figure 7E-H). Yap1-driven GEMM shows key histologic, IHC, and immune features of HPV- HNC Induration is an early clinical sign of OED transformation, which suggests alterations of the lesional mechanical properties. To model this event, another GEMM was engineered by crossing K5-CreER with Rosa26-Yap5sa to generate K5-CreER+;Yap5sa+ / - compound strain. The phosphorylation of Yap1 retains the protein in the cytosol and promotes its degradation. The Yap5sa mutant contains amino acid changes that render Yap1 unphosphorylatable. Expression of the constitutively active Yap5sa was induced in oral epithelium by topical tamoxifen painting in the oral mucosa. Active Yap1 was a potent driver of OED transformation and 100% of K5-CreER+;Yap5sa+ / - mice (n=13) developed microscopic keratinizing HNC by 6-week post-induction. Yap1- GEMM revealed high- fidelity histologic recapitulation of human HNC. These lesions were composed of sheets and nests of squamous tumor cells with abundant eosinophilic cytoplasm, a high-degree of cellular and nuclear pleomorphism, keratin pearls formation, and a high mitotic figures count. Importantly, the Yap1-GEMM exhibited stromal complexity, immune infiltration, and architectural similarities to human HNC. Upon IHC staining, about 30%-60% of the tumor cells showed conspicuous nuclear staining of Yap1, suggestive of Yap1 signaling activation. These tumors were negative for Sox2, in contrast to the Sox2-GEMM. The p63 staining pattern was similar to that in human HNC. Notably, there was a loss of p16 expression in the tumor islands, in comparison with adjacent relatively normal epithelium (Figure 8). This key feature is distinctive for HPV- HNC, in contrast to HPV+ tumors where p16 expression levels are elevated. To characterize the immune signatures associated with Yap1-driven OED transformation, scRNA-Seq was performed of buccal mucosa where most the lesions arose Atty. Docket No. UM-41943.601 in. After filtering, 6,375 high-quality transcriptomes were obtained. Main clusters were identified using the most differentially expressed genes and a panel of ~70 anchor genes to mark main immune lineages (Figure 9A) (58, 59). Similar to human OED transformation, a dramatic reduction of myeloid cells expressing high levels of Tnf and IFN-I signature genes (cluster 1) with concomitant expansion of Tox+ exhausted T-cells (cluster 2) and CD206+ M2-like macrophages (clusters 5 and 9) was found (Figure 9B). Targeted confocal laser endomicroscopy (CLE) is a real-time optical biopsy tool that reveals mucosal epithelial cell details in initiating HNC The current management of OED / OLP depends on on-slide diagnosis and clinical examination. General tissue autofluorescence and reflectance cannot yield molecular information enriched in high-risk OED / OLP. The ADA has not recommended the use of tissue autofluorescence or reflectance as a diagnostic adjunct. To improve clinical capability to monitor high-risk molecular markers, a robust CLE method was developed herein, which utilizes miniature scanner and actuators to interrogate tissue at microscopic resolution by deflecting a laser beam in the lateral and axial directions. These instruments are fiber-coupled and scaled down in size to millimeter dimensions to access the usually small oral cavity lesions with minimal invasiveness (Figure 10A). Front and side-view imaging geometries are available to maintain comprehensive contact of the optics with the tissue surface for high- quality images. Fast scan speeds allow for images to be collected in real time with minimal motion artifact. Overexpression of Epidermal Growth Factor Receptor (EGFR) is observed in >90% of HNC and shows potential in the detection of early carcinogenesis. The peptide QRHKPRE shows high-affinity binding to the extracellular domain of EGFR. C57BL / 6 wildtype mice were fed with 50 μg / ml drinking water for 16 weeks and then switched to regular water for 8 weeks. Normal mucosa of mice that did not develop tongue lesions was sprayed with FITC- QRHKPRE peptides and subjected to CLE. The normal mucosa showed an overall low fluorescence signal-to-background ratio, with only scattered positive cellular boundaries (Figure 10B). In contrast, histology-verified HNC showed a fishnet imaging pattern for the cellular binding of FITC-labeled, EGFR-targeted peptides. Keratin did not bind to EGFR- targeted imaging peptides and its signal was suppressed in the background (Figure 10C). Thus, CLE offers an optical biopsy tool that complements on-slide markers to longitudinally track high-risk features. Atty. Docket No. UM-41943.601 Example 2 Sox2- and Yap1-driven genetically engineered mouse models (as used in Example 1) can be used to establish the optical biopsy features for high-risk OED due to the 100% penetrance rate at the buccal sites. First, the baseline reading for the different fluorescence channels can be established. For example, Cy5-labeled, EGFR-targeted, and IRDye800- labeled, SLC2A1-targeted peptides can be sprayed onto the oral mucosa of the two GEMMs. The excess peptides can be rinsed off after 5-minute incubation. Then, the K5- CreER+;EGFP-Sox2+ / + and K5-CreER+;Yap5sa+ / - GEMMs can be induced with the oral mucosal painting of tamoxifen. The bilateral buccal mucosa can be surveyed weekly until six weeks post-induction, when all mice will be euthanized. Both EGFR and SLC2A1 are plasma membrane proteins which provide signals to identify cellular boundaries in imaging analysis. The number of EGFR+ and SLC2A1+ cells can be quantified, and their average intensity determined. For each region of interest (ROI), the EGFR – SLC2A1 interaction score can be calculated. For example, the EGFR – SLC2A1 interaction score can be calculated using the equation: and ISLC2A1 are the fluorescent intensity of EGFR and SLC2A1, respectively. It can be determined whether the individual marker intensity or the interaction scores increase over time and thus these scores can be used to evaluate transformation risk for a given lesion. Moreover, the spatial distance between positive features for EGFR and SLC2A1 may reduce as mucosa transforms. Therefore, the spatial distance between positive features for EGFR and SLC2A1 may also be a factor for determining the risk of transformation for a given lesion. In some embodiments, Wasserstein distance (W-value) between EGFR+ and SLC2A1+ cells can be used to assess their spatial relationship. The equation to calculate the Wasserstein distance is as follows: denotes all joint distributions ^ for (X, Y) with marginal P and Q. It can be determined whether W-values decrease as a function of time during OED transformation. Example 3 Atty. Docket No. UM-41943.601 Blockade of early IL-1 signaling delays the onset of Sox2-driven HNSCC. To better assess the role of this IL-1α amplification signaling loop in Sox2-driven transformation, induced K5-CreER;Sox2+ / +mice were treated with 200 µg anti-Il1r1 i.p. injections, starting one-week post-induction. Il1r1 blockade significantly extended the survival of these mice (Fig.11A). Buccal mucosa from both groups and RNA-Seq was performed. The most significantly upregulated pathways in the anti-Il1r1 treatment group were IFN-γ response, allograft rejection, and IFN-α response, suggestive of enhanced Th1 immunity. Then, the K5-CreER;Sox2+ / +GEMM were crossed with Il1r1- / -mice to generate the K5-CreER;Sox2+ / +;Il1r1- / -compound strain. Because of the rapid tumor onset in this model, most buccal cancers were microscopic. The areas where invasive tumor islands were investigated, and K5-CreER;Sox2+ / +;Il1r1- / -mice showed significantly reduced areas of tumor than K5-CreER;Sox2+ / +;Il1r1+ / +littermate controls (Fig.11B). Flow cytometric analysis showed that deletion of Il1r1 significantly reduced the infiltrating MDSC in Sox2- driven HNSCC. Additional experiments were conducted to elucidate the role of IL1 signaling in oral lesion malignancy. The results presented in FIG.12 demonstrate that blockade of IL1 signaling improves outcomes for high risk oral lesions. A novel signature of intralesional myeloid cells that assume an IL1(high)IFN-I(low) signature marks elevated risks for malignant transformation. Using genetic and pharmacological approaches significantly reduced the levels of Glut1+CD206+ intralesional macrophages and delayed the onset of cancer. FIG.12A shows tumor growth following injection with MOC2 tumor cells Moc2- gW or MOC2-Sox2 cells positive for Sox2. Tumor volume is shown. FIG.12B shows results of flow cytometry analysis showing the percentages of CD8+ CD3+ T-cells in gW and Sox2 groups. FIG.12C shows the frequency of CD45+ cells within the population of CD8+CD3+ T-cells. As shown in FIG.12C, the frequency of CD45+ cells in the population of Cd8+ CD3+ T cells is lower in the MOC2-Sox2 group compared to the MOC2-gW group. FIG.12D shows results of flow cytometry analysis showing the percentages of CD8+ CD11c+ dendritic cells in MOC2-gW and MOC2-Sox2 groups. FIG.12E shows the frequency of CD45+ cells within the population of CD8+ Cd11c+ cells in MOC2-Gw and MOC2-Sox groups. As shown in FIG.12E, the percentage of CD45+ cells in the population of Cd8+ Cd11c+ cells is lower in the MOC2-Sox2 group compared to the MOC2-gW group. FIG.12F shows flow cytometry analysis of FSCA+ and Gr1+ cells in MOC2-gW and MOC2-Sox2 groups. FIG.12G shows the frequency of CD11b+ cells in the Gr1+ MDSCs. Atty. Docket No. UM-41943.601 As shown in FIG.12G, the frequency of CD11b+Gr1+ MDSCs is significantly higher in the MOC2-Sox2 group compared to the MOC2-gW group. FIG.12H shows heat map analysis of Sox2, Il1a, and Csf2 expression in EV and Sox2 mice. FIG.12I shows mRNA fold change in IL1a in following administration of gW or Sox2 using the squamous cell line UMSC22b. Sox2 mice have a significantly higher fold change in IL1a mRNA compared to gW mice. To further investigate the role of ILR1 signaling, K5-Cre / Sox2+ / + mice were crossed with IL1r1+ / + or IL1R1- / - mice. Probability of survival over 75 days is shown in FIG.12J. K5- Cre; Sox2+ / +; IL1r- / - mice have increased probability of survival for a longer duration of time compared to K5-Cre; Sox2+ / +; IL1r+ / + mice, indicating that IL1r signaling plays a role in tumor growth and development / malignant transformation. To further assess cell types within these groups, flow cytometry was performed on samples obtained from K5-Cre; Sox2+ / +; IL1r+ / + and K5-Cre; Sox2+ / +; IL1r- / - groups. FIG.12K shows cell expression of TCRγ / δ and CD11b. As shown in FIG.12K, a distinct population of Cd11b- TCRγ / δ+ cells was identified in the IL1r+ / + group that was not significantly present in the IL1r- / - group. Both groups had a distinct population of Cd11b+ TCRγ / δ- cells. FIG.12L shows the frequency of CD45+ cells within the population of Cd11b- TCRγ / δ+ cells. The frequency of CD45+ cells was significantly higher in the IL1+ / + group compared to the IL1-- / group. FIG. 12M shows expression of Glut1 and CD206 cells. FIG.12N shows the frequency of Cd11b+ cells within the population of Glut1+ CD206+ cells. As shown, IL1- / - group had less Glut1+ CD206+ Cd11b + macrophages compared to the IL1r+ / + group. Tissue staining was performed to evaluate tissue from K5-Cre; Sox2+ / +; IL1r+ / + and K5-Cre; Sox2+ / +; IL1r- / - mice. The average surface area of tumors is quantified in FIG.12P. As shown in FIG.12P, IL1r- / - groups had significantly smaller tumor surface area compared to IL1r+ / + groups. To further investigate the role of IL1 signaling in oral lesion malignancy, tumor growth was induced in mice and followed by administration of PBS (control) or anti-IL1r1 antibody. As shown in FIG.12Q, the probability of survival was higher for a longer duration of time in the anti-IL1r1 treatment group compared to the control. Flow cytometry was performed to assess cell populations following treatment with PBS or anti-IL1r1. Results are shown in FIG.12R. The anti-IL1r1 group had a significant reduction in the number of PDL1+ GR1+ cells compared to the group treated with PBS. FIG.12S shows the frequency of PDL1+ GR1+ CD11B+ MDSCs in the different treatment groups. The anti-Il1r1 mice had less PDL1+ GR1+ CD11B+ MDSCs compared to control.

Claims

Atty. Docket No. UM-41943.601 CLAIMS 1. A method of stratifying an oral lesion based upon risk of progressing to malignancy, the method comprising: a) measuring expression of solute carrier family 2 member 1 (SLC2A1) in an oral lesion; and b) identifying the oral lesion as high risk when the expression level of SLC2A1 is equal to or above a threshold value, or identifying the oral lesion as low risk when the expression level of SLC2A1 is below a threshold value.

2. The method of claim 1, wherein expression of SLC2A1 in the oral lesion is measured by obtaining a biopsy sample from the oral lesion and measuring expression of SLC2A1 in the biopsy.

3. The method of claim 2, wherein expression of SLC2A1 is measured by tissue staining.

4. The method of claim 2, wherein expression of SLC2A1 in the oral lesion is measured directly on the oral lesion within the oral cavity of a subject.

5. The method of claim 4, wherein expression of SLC2A1 on the oral lesion is measured by contacting the oral lesion with an agent that binds to SLC2A1, wherein the agent comprises a label, and subsequently detecting the label.

6. The method of claim 5, wherein detecting the label comprises applying a light source to the oral lesion, thereby activating the label, and detecting the label visually.

7. The method of any one of claims 1-6, further comprising selectively abolishing the oral lesion when the oral lesion is identified as high risk, or monitoring the oral lesion when the oral lesion is identified as low risk.Atty. Docket No. UM-41943.601 8. A method of stratifying oral lesions based upon risk of progressing to malignancy, the method comprising: a) measuring expression of solute carrier family 2 member 1 (SLC2A1) and expression of an additional risk marker in a plurality of oral lesions within a region of interest of an oral cavity of a subject; and b) identifying an oral lesion as high risk when: i) an SLC2A1-additional marker interaction score for the oral lesion is above a threshold interaction score; and / or ii) the oral lesion expresses SLC2A1 and a Wasserstein distance between the oral lesion and a neighboring oral lesion that expresses the additional risk marker is equal to or less than a threshold Wasserstein distance.

9. The method of claim 8, wherein the SLC2A1-additional marker interaction score for an oral lesion is calculated by multiplying the expression of SLC2A1 in the oral lesion by the expression of the additional risk marker in the oral lesion, and dividing the result by the number of cells within the region of interest expressing both SLC2A1 and the additional risk marker.

10. The method of claim 8 or claim 9, wherein expression of SLC2A1 and expression of the additional risk marker are measured directly on the plurality of oral lesions within the oral cavity of a subject.

11. The method of claim 10, wherein expression of SLC2A1 and expression of the additional risk marker are measured by contacting the plurality of oral lesions with an agent that binds to SLC2A1 and an agent that binds to the additional risk marker, wherein each agent comprises a label, and subsequently detecting the labels.

12. The method of claim 11, wherein detecting the labels comprises applying a light source to the plurality of oral lesions, thereby activating the labels, and detecting the labels visually.Atty. Docket No. UM-41943.601 13. The method of any one of claims 8-12, further comprising selectively abolishing an oral lesion when the oral lesion is identified as high risk.

14. The method of any one of claims 8-13, wherein the additional risk marker is epidermal growth factor receptor (EGFR), SRY-Box Transcription Factor 2 (SOX-2), or nuclear expression of Yes1 Associated Transcriptional Regulator (YAP-1).

15. A method of treating a subject at risk of having head and neck cancer (HNC), the method comprising: a) measuring expression of solute carrier family 2 member 1 (SLC2A1) in an oral lesion of a subject; b) identifying the oral lesion as high risk when the expression level of SLC2A1 is equal to or above a threshold level, or identifying the oral lesion as low risk when the expression level of SLC2A1 is below a threshold value; and c) selectively abolishing the oral lesion when the oral lesion is identified as high risk, or monitoring the oral lesion when the oral lesion is identified as low risk.

16. The method of claim 15, wherein expression of SLC2A1 in the oral lesion is measured by obtaining a biopsy from the oral lesion and measuring expression of SLC2A1 in the biopsy.

17. The method of claim 16, wherein expression of SLC2A1 is measured by tissue staining.

18. The method of claim 15, wherein expression of SLC2A1 in the oral lesion is measured directly on the oral lesion within the oral cavity of a subject.

19. The method of claim 18, wherein expression of SLC2A1 on the oral lesion is measured by contacting the oral lesion with an agent that binds to SLC2A1, wherein the agent comprises a label, and subsequently detecting the label.Atty. Docket No. UM-41943.601 20. The method of claim 19, wherein detecting the label comprises applying a light source to the oral lesion, thereby activating the label, and detecting the label visually.

21. A method of treating a subject at risk of having head and neck cancer (HNC), the method comprising: a) measuring expression of solute carrier family 2 member 1 (SLC2A1) in each of a plurality of oral lesions directly within the oral cavity of a subject; b) identifying an oral lesion as high risk when the expression level of SLC2A1 in the oral lesion is equal to or above a threshold level and / or identifying an oral lesion as low risk when the expression level of SLC2A1 in the oral lesion is below a threshold level; and c) selectively abolishing high risk oral lesions from the oral cavity of the subject.

22. The method of claim 21, further comprising monitoring low risk oral lesions.

23. The method of claim 21 or claim 22, wherein expression of SLC2A1 is measured by contacting the plurality of oral lesions with an agent that binds to SLC2A1, wherein the agent comprises a label, and subsequently detecting the label.

24. The method of claim 23, wherein detecting the label comprises applying a light source to the oral lesion, thereby activating the label, and detecting the label visually.

25. A method of treating a subject at risk of having head and neck cancer (HNC), the method comprising: a) measuring expression of solute carrier family 2 member 1 (SLC2A1) and expression of an additional risk marker in a plurality of oral lesions within a region of interest of an oral cavity of a subject; b) identifying an oral lesion as high risk when: i) an SLC2A1-additional marker interaction score for the oral lesion is above a threshold interaction score; and / orAtty. Docket No. UM-41943.601 ii) the oral lesion expresses SLC2A1 and a Wasserstein distance between the oral lesion and a neighboring oral lesion that expresses the additional risk marker is equal to or less than a threshold Wasserstein distance; and c) selectively abolishing high risk oral lesions from the oral cavity of the subject.

26. The method of claim 25, wherein the SLC2A1-additional marker interaction score for an oral lesion is calculated by multiplying the expression of SLC2A1 in the oral lesion by the expression of the additional risk marker in the oral lesion, and dividing the result by the number of cells within the region of interest expressing both SLC2A1 and the additional risk marker.

27. The method of claim 25 or claim 26, wherein expression of SLC2A1 and expression of the additional risk marker are measured directly on the plurality of oral lesions within the oral cavity of a subject.

28. The method of claim 27, wherein expression of SLC2A1 and expression of the additional risk marker are measured by contacting the plurality of oral lesions with an agent that binds to SLC2A1 and an agent that binds to the additional risk marker, wherein each agent comprises a label, and subsequently detecting the labels.

29. The method of claim 28, wherein detecting the labels comprises applying a light source to the plurality of oral lesions, thereby activating the labels, and detecting the labels visually.

30. A method comprising measuring expression of solute carrier family 2 member 1 (SLC2A1) in at least one oral lesion directly within the oral cavity of a subject, wherein expression of SLC2A1 is measured by contacting the at least one oral lesion with an agent that binds to SLC2A1, wherein the agent comprises a label, and subsequently detecting the label on the at least one oral lesion.Atty. Docket No. UM-41943.601 31. The method of claim 30, wherein detecting the label comprises applying a light source to the oral lesion, thereby activating the label, and detecting the label visually.

Images