BR112015022085B1 - METHODS FOR ASSESSING THE PROGNOSIS OF COLORECTAL CANCER - Google Patents

Abstract

METHODS FOR ASSESSING THE PROGNOSIS OF COLORECTAL CANCER The present invention provides methods for assessing the risk that an individual diagnosed with colorectal cancer has of developing a recurrence of colorectal cancer and methods for predicting the clinical outcome of an individual diagnosed with colorectal cancer by a) determining the expression level for each marker of a panel of markers in a panel of tumor compartments in a tumor tissue sample from the subject, wherein the panel of markers comprises at least two of TEM1, HIF2alpha, CAIX, PDGFRbeta, fibronectin, collagen I , collagen IV, and CD31 and wherein the panel of tumor compartments comprises at least three tumor compartments among the pure stroma, tumor, stromal vessel, and tumor vessel; b) determining the TAPPS score for said individual; and c) comparing the individual's TAPPS score to the TAPPS score of a population of individuals diagnosed with colorectal cancer. Related computer-implemented methods and systems, kits, and tumor microarrays are also provided.

Description

References to related deposit requests

[0001] This application claims the benefit of U.S. provisional patent application 61/793,565, filed on March 15, 2013, which is fully incorporated herein by reference.

Technical field

[0002] The present invention provides a set of biomarkers, the expression levels of which are useful for determining the risk that an individual diagnosed with colorectal cancer will develop recurrence of colorectal cancer and for predicting the clinical outcome of an individual diagnosed with colorectal cancer.

Background

[0003] In 2013, approximately 143,000 people in the United States will be diagnosed with colorectal cancer. In the same year, it is estimated that there will be 50,830 deaths caused by colorectal cancer in the United States. Colorectal cancer is the fourth most common cause of cancer in men, after skin, prostate and lung cancer. It is also the fourth most common cause of cancer in women, after skin, breast and lung cancer. Methods for assessing recurrence risk or prognosis in patients with colorectal cancer would be of great benefit in guiding treatment decisions for these patients. The generation of an independently significant multimarker prognostic method for colorectal cancer is described here. The data demonstrate the potential of multimarker tests to improve the assessment of colorectal cancer prognosis. The methods described here allow the assessment of an individual at the time of colorectal cancer diagnosis in a group with a high risk of recurrence or a low risk of recurrence. The methods will help patients and clinicians determine the need for adjuvant intervention or at least aggressive follow-up surveillance. The aim of the described methods is to optimize the overall survival of the high-risk group of patients without exposing the remaining patients to the risks and costs associated with adjuvant therapy or monitoring for disease recurrence.

summary

[0004] Described here are methods for determining the risk that an individual diagnosed with colorectal cancer will develop a recurrence of this disease based on the gene expression profile of a panel of biomarkers in particular histological compartments of a tumor tissue sample obtained from the individual. Also described are methods for predicting the clinical outcome of an individual diagnosed with colorectal cancer based on the gene expression profile of a panel of biomarkers in particular tumor compartments of a tumor tissue sample obtained from the individual. Additionally provided herein are differential diagnostic methods that allow an individual diagnosed with colorectal cancer to be assigned to a high-risk or low-risk category of TEM-1-expressing colorectal cancer. Also described are materials and methods for producing tissue sample arrays identified with detection agents that allow analysis of the tissue sample to predict various risks associated with the occurrence of colorectal cancer. The described tissue sample arrays can also be used to analyze tissue samples to help determine a beneficial course of treatment for an individual diagnosed with colorectal cancer. Kits that have reagents suitable for identifying the tissue samples or tissue matrices discussed in the present invention are also described herein.

Brief description of the drawings

[0005] Figure 1 illustrates the masking of the biological compartment and a representative biomarker panel. (A) Representative examples of images from each fluorescence channel used to generate the biological compartment masks during AQUA® analysis. DAPI was used to generate a mask of nuclei, fluorescein isothiocyanate (FITC) was used to label cytokeratin to identify the tumor cytoplasm, Cy3 was used to label vimentin to identify the stroma, Cy7 was used to label CD31 to identify the vasculature and in conjunction with cytokeratin to determine tumor vasculature and with vimentin to identify stromal vasculature. (B) Representative examples of images from the seven-biomarker panel (Cy5).

[0006] Figure 2 shows that the TAPPS score is highly prognostic in colorectal cancer. (A) Histogram plot of TAPPS score distribution in YTMA8 (n = 265). Dotted lines represent tertile cutoffs used for survival analysis. (B) Kaplan-Meier survival analysis of all cases with 5-year disease-free survival. Low TAPPS (dark gray line on top at 60-month time) n = 88, events = 15, Intermediate TAPPS (light gray line in the middle at 60-month time) n = 99, events = 38, high TAPPS (black line on bottom at time 60 months) n = 73, events = 51. (C) Kaplan-Meier survival analysis of node-negative cases only with 5-year disease-free survival. Low TAPPS (dark gray line on top at 60-month time) n = 44, events = 4, Intermediate TAPPS (light gray line in the middle at 60-month time) n = 50, events = 9, high TAPPS (black line on bottom at time of 60 months) n = 26, events = 15. (D) Kaplan-Meier survival analysis of node-positive cases only with disease-free survival for 5 years. Low TAPPS (light gray line in the middle at 60-month time) n = 27, events = 6, Intermediate TAPPS (dark gray line at the top at 60-month time) n = 37, events = 24, high TAPPS (black line at the bottom in 60 months) n = 41, events = 32.

[0007] Figure 3 shows a schematic representation of two main groupings of patients identified by the unsupervised hierarchical clustering of the indicated biomarkers.

[0008] Figure 4(A-G) provides a graphical representation of marker expression by histological compartment (tumor, stromal vessel, tumor vessel or stroma). Figure 4A shows the expression of TEM-1, Figure 4B shows the expression of HIF2a, Figure 4C shows the expression of PDGFR-ß, Figure 4D shows the expression of CAIX, Figure 4E shows the expression of fibronectin, the Figure 4F shows collagen I expression, and Figure 4G shows collagen IV expression.

[0009] Figure 5(A-D) shows the Kaplan-Meier survival analysis of the validation set of endosialin/TEM-1 expression in all compartments.

[0010] Figure 6 shows the Kaplan-Meier survival analysis of (A) censored low, intermediate and high risk groups; (B) censored low and high risk groups, (C) patients with stage II colorectal cancer, and (D) patients with stage III/IV colorectal cancer.

[0011] Figure 7 shows the Kaplan-Meier survival analysis for a model that includes only 3 variable parameters for (A) all cases and (B) patients with stage II colorectal cancer.

Detailed description of illustrative embodiments

[0012] Except where otherwise defined, all technical and scientific terms used in the present invention have the same meaning as commonly understood by those skilled in the art to which this invention belongs. Singleton et al., Dictionary of Microbiology and Molecular Biology 3rd ed., J. Wiley & Sons (New York, N.Y. 2006), and March, Advanced Organic Chemistry Reactions, Mechanisms and Structure 6th ed., John Wiley & Sons (New York , N.Y. 2007), provide one skilled in the art with a general guide to many terms used in the present application.

[0013] Those skilled in the art will recognize many methods and materials similar or equivalent to those described herein, which could be used in the practice of the present invention. Indeed, the present invention is in no way limited to the described methods and materials. For the purposes of the present invention, the following terms are defined below.

[0014] As used in this specification and the attached claims, the singular forms “a”, “an” and “the” include plural references unless the content clearly determines otherwise. Thus, for example, reference to “a cell” includes a combination of two or more cells and the like.

[0015] The term “about” (also denoted by the symbol “~”) as used herein when referring to a measurable value, such as a quantity, a duration of time, and the like, is intended to encompass variations of ± 10% of the specified value , as such variations are suitable to perform the presented methods. Except where otherwise indicated, all numbers expressing amounts of ingredients, properties such as molecular weight, reaction conditions, and so on used in the specification and claims are to be understood as being modified in all cases by the term “about ”. Consequently, unless otherwise indicated, the numerical parameters presented in the following specification and in the attached claims are approximations that may vary depending on the desired properties sought to be obtained with the present invention. At a minimum, and not as an attempt to limit the application of the doctrine of equivalents to the scope of the claims, each numerical parameter should be considered in light of the number of significant figures reported and through the application of common rounding techniques. Although the numerical ranges and parameters defining the broad scope of the invention are approximations, the numerical values stated in the specific examples are recorded as precisely as possible. Any numerical value, however, inherently contains certain errors that necessarily result from the standard deviation found in their respective test measurements.

[0016] The terms “express” and “produce” are used synonymously in the present invention, and refer to the biosynthesis of a gene product. These terms cover the transcription of a gene into RNA. These terms also cover the translation of RNA into one or more polypeptides and further encompass all naturally occurring post-transcriptional and post-translational modifications.

[0017] As used herein, the term “antibody” refers to all immunoglobulin isotypes (IgG, IgA, IgE, IgM, IgD and IgY) including the various monomeric and polymeric forms of each isotype, and should be understood to encompass antigen-binding fragments, diabodies and single-chain molecules as well as Fab, F(ab')2, Fc, Fabc and Fv molecules, single-chain antibodies (Sc), individual antibody light chains, individual antibody heavy chains, chimeric fusions between antibody chains or CDRs and other proteins, protein scaffolds, heavy chain monomers or dimers, light chain monomers or dimers, dimers consisting of a heavy and a light chain, and the like.

[0018] Antigen-binding fragments are any proteinaceous structure that can exhibit binding affinity for a specific antigen. Some antigen-binding fragments are composed of intact antibody portions that retain antigen-binding specificity of the original antibody molecule. For example, antigen-binding fragments may comprise at least one variable region (a heavy chain or light chain variable region) or one or more CDRs of an antibody known to bind a specific antigen. Examples of suitable antigen-binding fragments include, but are not limited to, single-chain diabodies and molecules as well as Fab, F(ab')2, Fc, Fabc, and Fv molecules, single-chain (Sc) antibodies, light chains of individual antibodies, individual antibody heavy chains, chimeric fusions between antibody chains or CDRs and other proteins, protein scaffolds, heavy chain monomers or dimers, light chain monomers or dimers, dimers consisting of a heavy and a light chain, and similar. All antibody isotypes can be used to produce antigen-binding fragments. Additionally, the antigen-binding fragments can include non-antibody proteinaceous structures that can successfully incorporate the polypeptide segments in an orientation that confers affinity for a given antigen of interest, such as protein scaffolds. Antigen-binding fragments can be produced recombinantly or produced by enzymatic or chemical cleavage of intact antibodies. The phrase “an antibody or antigen-binding fragment thereof” may be used to denote that a given antigen-binding fragment incorporates one or more amino acid segments of the antibody mentioned in the phrase.

[0019] The term “specific binding”, when used in the context of antibodies, or antibody fragments, represents binding through domains encoded by immunoglobulin genes or fragments of immunoglobulin genes to one or more epitopes of a protein of interest, without binding, preferably, to other molecules in a sample containing a mixed population of molecules. Typically, an antibody binds to a cognate antigen with a Kd less than about 1x10-8 M, as measured by a surface plasmon resonance assay or a cell binding assay. Expressions such as “antibody [antigen]-specific” (e.g. TEM-1 antibody-specific) are intended to convey that the mentioned antibody specifically binds to the mentioned antigen.

[0020] The term “tumor”, as used herein, refers to all neoplastic cell growth and proliferation and all cancerous and precancerous cells and tissues.

[0021] The terms “cancer” and “cancerous” refer to or describe the physiological condition in mammals that is typically characterized by unregulated cell growth. An example of cancer is colorectal cancer.

[0022] The “pathology” of cancer includes all phenomena that compromise the patient's well-being. This includes, without limitation, abnormal or uncontrollable cell growth, metastasis, interference with the normal functioning of neighboring cells, release of cytokines or other secretion products at abnormal levels, suppression or worsening of the inflammatory or immunological response, neoplasia, premalignancy , malignancy, invasion of surrounding or distant tissues or organs such as lymph nodes etc.

[0023] The term “colorectal cancer” is used in its broadest sense and refers to (1) all stages and all forms of cancer that arise from epithelial cells of the large intestine and/or rectum and/or (2) all stages and all forms of cancer that affect the lining of the large intestine and/or rectum. In the staging systems used to classify colorectal cancer, the colon and rectum are treated as one organ.

[0024] According to the American Joint Committee on Cancer (AJCC) tumor-node-metastasis (TNM) staging system (Edge et al. (eds.), AJCC Cancer Staging Manual. 7th Ed. New York, N.Y. : Springer; 2009), the various stages of colorectal cancer are defined as follows: Tumor: T0: no evidence of primary tumor; T1: tumor invades the submucosa; T2: tumor invades the muscularis propria layer; T3: tumor invades the muscularis propria layer in the subserosa, or in the pericolic or perirectal tissues; T4: tumor directly invades other organs or structures, and/or perforations; Node: N0: no regional metastasis in lymph nodes; N1: metastasis in 1 to 3 regional lymph nodes; N2: metastasis in 4 or more regional lymph nodes; Metastasis: M0: no distant metastasis; M1: distant metastasis present; Stage groupings: Stage I: T1 N0 M0; T2 N0 M0; Stage II: T3 N0 M0; T4 N0 M0; Stage III: any T, N1-2; M0; Stage 1V: any T, any N, M1.

[0025] According to the modified Duke staging system, the various stages of colorectal cancer are defined as follows: Stage A: the tumor penetrates the mucosa of the intestinal wall, but not beyond; Stage B: the tumor penetrates and passes through the muscular layer of the intestinal wall; Stage C: the tumor penetrates but does not cross the muscularis propria layer of the intestinal wall; there is pathological evidence of colorectal cancer in the lymph nodes; or the tumor penetrates and passes through the muscularis propria layer of the intestinal wall, there is pathological evidence of cancer in the lymph nodes; Stage D: the tumor has spread beyond the limits of the lymph nodes, into other organs, such as the liver, lungs or bones.

[0026] Prognostic factors are those variables related to the natural history of colorectal cancer, which influence recurrence rates and patient outcomes once they have developed colorectal cancer. Clinical parameters that have been associated with a worse prognosis include, for example, lymph node involvement and high-grade tumors. Prognostic factors are often used to categorize patients into subgroups with different baseline recurrence risks.

[0027] The term “prognosis” is used in the present invention to refer to the prediction of the probability of death or progression attributed to cancer, including recurrence, metastatic spread, and drug resistance, of a neoplastic disease such as colorectal cancer.

[0028] The term “prediction” is used in the present invention to refer to the probability that a patient will have a specific clinical outcome, positive or negative. The anticipated methods of the present invention can be used clinically to make treatment decisions when choosing the most appropriate treatment modalities for any specific patient. The anticipated methods of the present invention are valuable tools for predicting whether a patient is likely to respond favorably to a treatment regimen, such as a surgical intervention. The prediction may include prognostic factors.

[0029] The term “positive clinical outcome” means an improvement in any measure of the patient's condition, including those measures commonly used in the art, such as an increase in the duration of the recurrence-free interval (RFI), an increase in overall survival time (OS), an increase in disease-free survival time (DFS), an increase in the duration of the distant recurrence-free interval (DRFI), and the like.

[0030] The term “risk classification” means the level of risk or the prediction that an individual will experience a specific clinical outcome. An individual can be classified into a risk group or classified into a risk level based on the predictive methods of the present invention. A “risk group” is a group of individuals or subjects with a similar level of risk for a particular clinical outcome.

[0031] The term “individual” or “patient” refers to human and non-human animals, including all vertebrates, e.g., mammals and non-mammals, such as non-human primates, mice, rabbits, sheep, dogs, cats, horses, cows, chickens, amphibians and reptiles. In many embodiments of the described methods, the subject is a human being.

[0032] The terms “treat” or “treatment” refer to any success or sign of success in alleviating or improving an injury, pathology or condition, including any objective or subjective parameter, such as abatement, remission, reduction of symptoms or making the condition more tolerable to the patient, reducing the rate of degeneration or decline, making the end point of degeneration less debilitating, improving an individual's physical or mental well-being, or prolonging survival time. Treatment can be evaluated by objective or subjective parameters; including the results of a physical exam, neurological exam, or psychiatric evaluations.

[0033] The term “long-term” survival is used in the present invention to refer to survival for at least 3 years, more preferably, for at least 5 years.

[0034] The term “recurrence-free interval (RFI)” is used in the present invention to refer to the time until the first censoring of recurrence of colorectal cancer for the second primary cancer as a first event or death without evidence of recurrence.

[0035] The term “progression-free survival (PFS)” is used in the present invention to refer to the time from first-line treatment of colorectal cancer to disease progression or death, during the time the disease is present. but it doesn't get worse.

[0036] The term “overall survival (OS)” is used in the present invention to refer to the time from surgery to death from any cause.

[0037] The term “disease-free survival (DFS)” is used in the present invention to refer to the time to recurrence of colorectal cancer or death from any cause.

[0038] The term “distant recurrence free interval (DRFI)” is used in the present invention to refer to the time from surgery to the first recurrence of anatomically distant cancer.

[0039] In the context of the present invention, reference to “at least one”, “at least two”, “at least five” etc. of the genes mentioned in any particular gene set means any or any and all combinations of the genes mentioned.

[0040] The term “node-negative” cancer, such as “node-negative” colorectal cancer, is used in the present invention to refer to cancer that has not been detected in the lymph nodes and/or has not spread to them. The term “node-positive” cancer, such as “node-positive” colorectal cancer, is used in the present invention to refer to cancer that has spread to the lymph nodes and/or has been detected there.

Methods

[0041] Methods for determining the risk that an individual diagnosed with colorectal cancer will develop a recurrence of this disease based on the gene expression profile of a panel of biomarkers in particular tumor compartments of a tumor tissue sample obtained from the patient are described here. Also described are methods for predicting the clinical outcome of an individual diagnosed with colorectal cancer based on the gene expression profile of a panel of biomarkers in the tumor-specific compartments of a tumor tissue sample obtained from the patient. Also provided herein are differential diagnostic methods that allow an individual diagnosed with colorectal cancer to be assigned to a high-risk or low-risk category for TEM-1-expressing colorectal cancer.

[0042] The individual may be lymph node-positive or lymph node-negative for colorectal cancer. A patient's colorectal cancer can be at any stage, although most often it is stage I, stage II, or stage III cancer.

[0043] The tumor tissue sample can be fixed, fixed and embedded in paraffin or fresh. In some aspects, the tumor tissue sample is obtained from a tissue biopsy, a fine needle aspiration sample, surgically removed tumor tissue, or histological preparations of a biological sample obtained from the patient. In some embodiments, the tumor tissue sample is from the primary colorectal cancer tumor. In other embodiments, the tumor tissue sample is from a metastatic colorectal cancer tumor.

[0044] The panel of markers includes at least two of the endothelial tumor marker-1 (TEM1; also known as endosialin or CD248), hypoxia-inducible factor 2 alpha (HIF2a), carbonic anhydrase 9 (CAIX), platelet-derived growth beta (PDGFRß), fibronectin (FN), collagen I (COLI), collagen IV (COLIV) and CD31. In preferred embodiments, the panel of markers includes three, four, five, six, seven or all eight of TEM1, HIF2a, CAIX, PDGFRß, fibronectin, collagen I, collagen IV and CD31. In some embodiments, the marker panel includes at least TEM1 and CAIX. In some aspects, the panel of markers includes at least TEM1, HIF2a, CAIX and PDGFRß; more preferably, TEM1, HIF2a, CAIX, fibronectin and PDGFRß; and in some embodiments, TEM1, HIF2a, CAIX, PDGFRß, fibronectin, collagen I, collagen IV and CD31.

[0045] The markers are located in a panel of tumor compartments including at least three of pure stroma (stroma that is substantially or completely free of vasculature), tumor, stromal vessel and tumor vessel. Localization may involve identifying and/or labeling at least three compartments in the individual's tumor tissue sample. In preferred embodiments, the panel of tumor compartments in which the biomarkers are located includes pure stroma, tumor, stromal vessel, and tumor vessel.

[0046] In some aspects, the methods of the present invention described herein involve determining the expression level of at least two, three, four, or five of the following biomarkers and tumor compartment combinations: the expression level of TEM1 in the tumor stroma , the expression level of TEM1 in the tumor vessel, the expression level of HIF2a in the stromal vessel, the expression level of HIF2a in the tumor stroma, the expression level of CAIX in the tumor vessel, the expression level of PDGFRß in the tumor stroma and the expression level of fibronectin (e.g., fibronectin-1) in the tumor stroma. In preferred embodiments of the described methods, the level of expression of TEM1 in the tumor stroma, the level of expression of TEM1 in the tumor vessel, the level of expression of HIF2a in the stromal vessel, the level of expression of HIF2a in the tumor stroma, the expression level of CAIX in the tumor vessel and the expression level of PDGFRß in the tumor stroma are determined. In some preferred embodiments of the described methods, the level of expression of TEM1 in the tumor stroma, the level of expression of TEM1 in the tumor vessel, the level of expression of HIF2a in the stromal vessel, the level of expression of HIF2a in the tumor stroma , the expression level of CAIX in the tumor vessel, the expression level of PDGFRß in the tumor stroma and the expression level of fibronectin in the tumor stroma are determined. In some embodiments of the methods described, the level of expression of TEM1 in the stroma, the level of expression of TEM1 in the tumor vessel, the level of expression of HIF2a in the stromal vessel, the level of expression of COLIV in the tumor, and the level of expression of of fibronectin in the stroma are determined. In some embodiments of the described methods, the level of expression of TEM1 in the stroma, the level of expression of COLIV in the tumor, and the level of expression of fibronectin in the stroma are determined. In the methods described herein, at least one sub-cellular compartment (e.g., cell nucleus, a cytoplasm, a nuclear membrane, a cell membrane, a mitochondria, an endoplasmic reticulum, a peroxisome, and a lysosome) can also be identified or labeled. .

[0047] Biomarker expression levels can be determined by any method known in the art. Gene expression profiling methods include methods based on polynucleotide hybridization analysis, methods based on polynucleotide sequencing, and proteomic methods. The most commonly used methods known in the art for quantifying mRNA expression in a sample include Northern blotting and in situ hybridization (Parker & Barnes, Methods in Molecular Biology 106:247-283 (1999)); RNAse protection assays (Hod, Biotechniques 13:852-854 (1992)); and PCR-based methods such as reverse transcription polymerase chain reaction (RT-PCR) (Weis et al., Trends in Genetics 8:263-264 (1992)). Alternatively, antibodies may be employed that can recognize sequence-specific duplexes, including DNA duplexes, RNA duplexes, and DNA-RNA hybrid duplexes or DNA-protein duplexes. Representative methods for sequencing-based gene expression analysis include Serial Analysis of Gene Expression (SAGE), and massively parallel signature sequencing (MPSS) gene expression analysis. Biomarker expression levels can also be determined by immunohistochemical methods and automated tissue microarray analyzes (e.g., see, U.S. Pub Nos. 20120093387 and 20110311123, incorporated herein by reference.)

[0048] The methods provided here can be automated in whole or in part. In some embodiments, the methods involve quantitative immunofluorescence (QIF) signal from the marker panel in a sample of the patient's colorectal cancer tumor. Various quantitative image analysis procedures are known in the art. An example of a quantitative image analysis procedure that can be used to determine expression level is automated quantitative analysis technology (AQUA®), as described in granted US Patent No. 7,219,016, and in the patent application publication US Patent No. 2009/0034823, each of which is incorporated herein by reference in this application in its entirety. AQUA® technology allows quantification not only of the fluorescence signal for a given marker within the tissue sample under analysis, but also allows compartmentalization within molecularly defined tumor and subcellular compartments.

[0049] The individual's expression levels of biomarkers within the panel of tumor compartments allows determination of a TEM1-associated pathways prognostic signature (TAPPS) score for the patient (as used herein, the term “TAPPS” may also , be used in place of the term “R-TAPPS” which is a refined version of the originally defined TAPPS score, as explained in the examples section, below). The individual's TAPPS score is generated by incorporating the individual's expression levels of biomarkers within the panel of tumor compartments into a TAPPS equation. The TAPPS equation is generated by examining the expression profiles of the panel of biomarkers in the panel of tumor compartments of a population of individuals diagnosed with colorectal cancer in a multivariate Cox proportional hazards model using backward elimination employing univariate cutoffs. to provide the prognostic model that satisfies the desired p-value level. From this model, a coefficient is derived for each marker, and, with these coefficients, a TAPPS equation is developed, which provides a total risk score.

[0050] In some embodiments, the TAPPS equation is established by determining the coefficients A, B, C, D, etc. for each of the markers in each of the tumor compartments selected for inclusion in the colorectal cancer patient population to provide a prognostic model that has an acceptable p-value. In some embodiments, the TAPPS equation is wherein the coefficients A, B, C, and D are the coefficients derived from each respective marker in each respective tumor compartment in the multivariate Cox proportional hazards model using backward elimination and employing univariate cutoffs to provide the prognostic model for the given population of patients with colorectal cancer. In other embodiments, the TAPPS equation is: where the coefficients A, B, C, D, E, F, G and H are the coefficients derived from each respective marker in each respective tumor compartment in the multivariate Cox proportional hazards model using backward elimination and employing points univariate cutoffs to provide the optimal prognostic model for the given colorectal cancer patient population. In some embodiments, the TAPPS equation is as follows: wherein the coefficients A, B, C, D, and E are the coefficients derived from each respective marker in each respective tumor compartment in the multivariate Cox proportional hazards model using backward elimination and employing univariate cutoffs to provide the prognostic model for the given population of patients with colorectal cancer. In some embodiments, the TAPPS equation is wherein the coefficients A, B, C, D, and E are the coefficients derived from each respective marker in each respective tumor compartment in the multivariate Cox proportional hazards model using backward elimination and employing univariate cutoffs to provide the prognostic model for the given population of patients with colorectal cancer. In some embodiments, the TAPPS equation is wherein the coefficients A, B, and C are the coefficients derived from each respective marker in each respective tumor compartment in the multivariate Cox proportional hazards model using backward elimination and employing univariate cutoffs to provide the prognostic model for the given population of patients with colorectal cancer. In some embodiments, the TAPPS equation is wherein the coefficients A, B, and C are the coefficients derived from each respective marker in each respective tumor compartment in the multivariate Cox proportional hazards model using backward elimination and employing univariate cutoffs to provide the prognostic model for the given population of patients with colorectal cancer. It is understood by those skilled in the art that the coefficients of the TAPPS equation are subject to some variability depending on the population of individuals diagnosed with colorectal cancer.

[0051] The individual's TAPPS score is then compared to the TAPPS score of a population of individuals diagnosed with colorectal cancer. The individual can be partitioned into a subgroup at any specific value(s) of the TAPPS score, where all patients with values in a given range can be classified as belonging to a particular risk group or group associated with a specific clinical outcome. The results of the analysis can be summarized in a report. This information is useful for the patient and the doctor to evaluate the risk-benefit of observation in relation to adjuvant therapy for that patient. For example, if it is determined that an individual is at high risk for colorectal cancer recurrence, additional therapy may be chosen or administered. This therapy may include TEM-1-targeted therapy (e.g., MORAb-004), chemotherapy, radiotherapy, and/or monitoring for disease recurrence or progression.

[0052] In some aspects, methods for assessing the risk that an individual diagnosed with colorectal cancer has of developing colorectal cancer recurrence involve a) determining the level of expression for each marker in a panel of markers in a panel of tumor compartments in a tumor tissue sample from the individual, wherein the panel of markers comprises at least two of TEM1, HIF2a, CAIX, PDGFRß, fibronectin, collagen I, collagen IV, and CD31 and wherein the panel of tumor compartments comprises at least three tumor compartments of pure stroma, tumor, stromal vessel, and tumor vessel; b) determine the individual's TAPPS score; and c) compare the individual's TAPPS score to the TAPPS score of a population of individuals diagnosed with colorectal cancer. The individual is assigned to the low-risk group for colorectal cancer recurrence if the individual's TAPPS score is low relative to the population's TAPPS score. The individual is assigned to the high-risk group for colorectal cancer recurrence if the individual's TAPPS score is high relative to the population's TAPPS score. In some respects, if the individual's colorectal cancer is node-negative and the individual's TAPPS score is intermediate to the population's TAPPS score, the individual is assigned to the low-risk group for colorectal cancer recurrence. In some respects, if the individual's colorectal cancer is lymph node-positive and the individual's TAPPS score is intermediate to the population's TAPPS score, the individual is assigned to the high-risk group for colorectal cancer recurrence.

[0053] In some aspects, methods of predicting the clinical outcome of an individual diagnosed with colorectal cancer involve a) determining the level of expression for each marker from a panel of markers in a panel of tumor compartments in a tumor tissue sample from the individual, wherein the panel of markers comprises at least two of TEM1, HIF2a, CAIX, PDGFRß, fibronectin, collagen I, collagen IV, and CD31 and wherein the panel of tumor compartments comprises at least three tumor compartments of pure stroma, tumor , stromal vessel, and tumor vessel; b) determine the individual's TAPPS score; and c) compare the individual's TAPPS score to the TAPPS score of a population of individuals diagnosed with colorectal cancer. A low TAPPS score for the individual relative to the population TAPPS score is predictive of a positive clinical outcome. A high TAPPS score for the individual relative to the population TAPPS score is predictive of a poor clinical outcome. If the individual's colorectal cancer is lymph node-negative, an intermediate TAPPS score for the individual relative to the TAPPS score for the population is predictive of a positive clinical outcome for the individual. If the individual's colorectal cancer is node-positive, an intermediate TAPPS score for the individual relative to the population TAPPS score is predictive of a poor clinical outcome for the individual. Clinical outcome can be expressed in terms of progression-free survival (PFS), relapse-free interval (RFI), overall survival (OS), disease-free survival (DFS), or distant relapse-free interval (DRFI).

[0054] If a poor clinical outcome is anticipated for the individual, the individual may choose or may undergo additional therapy, e.g., but not limited to TEM-1 targeted therapy (e.g., MORAb-004), chemotherapy , radiotherapy and/or monitoring for recurrence or progression of the disease.

[0055] Also provided herein are methods of treating colorectal cancer, in which an individual determined by the methods presented to be at high risk for recurrence of colorectal cancer or for whom a poor clinical outcome is predicted, is administered a cancer therapy. colorectal, for example, TEM-1 targeted therapy, chemotherapy and/or radiotherapy. Clinical application of the methods described here would provide an objective assessment of the likelihood of a patient's disease recurrence, which would be complementary to existing criteria. Based on this information, patients who are likely to benefit most from adjuvant therapy or more aggressive monitoring for disease recurrence can be identified. Conversely, patients who may be candidates for adjuvant therapy based on current prognostic criteria, but who are identified as being at low risk for recurrence based on this assay, may avoid the unnecessary risks associated with existing adjuvant therapy.

[0056] The methods described herein allow for the designation of an individual diagnosed with colorectal cancer into a high-recurrence risk or low-recurrence-risk group or into a group in which a positive clinical outcome is predicted or a group in which a positive clinical outcome is predicted. bad is predicted. The methods will help patients and clinicians determine the need for adjuvant intervention or at least aggressive follow-up surveillance. The aim of the described methods is to optimize the overall survival of the high-risk group of patients without exposing the remaining patients to the risks and costs associated with adjuvant therapy or monitoring for disease recurrence. For example, a node-negative patient identified by the methods of the present invention as having a high-risk test result could instruct the patient to choose adjuvant therapy.

[0057] Also described herein are computer-implemented methods for locating a panel of markers within a sample of colorectal cancer tumor tissue. According to these methods, the tumor tissue sample (e.g., one or more of its sections) is incubated with a panel of markers that specifically identify at least three tumor compartments of pure stroma, tumor, stromal vessel and tumor vessel. The tumor tissue sample (e.g., one or more of its sections) is also incubated with a label panel that labels a panel of markers comprising at least two of TEM1, HIF2a, CAIX, PDGFRß, fibronectin, collagen I, collagen IV and CD31. In some preferred embodiments, the panel of markers comprises four, five, six, seven or all eight of TEM1, HIF2a, CAIX, PDGFRß, fibronectin, collagen I, collagen IV and CD31. In a preferred embodiment, the panel of markers comprises TEM1, HIF2a, CAIX and PDGFRß. In another preferred embodiment, the panel of markers comprises TEM1, HIF2a, CAIX, fibronectin and PDGFRß. In one embodiment, the panel of markers includes TEM1, HIF2a, collagen IV, and fibronectin. A high-resolution image of each of the markers in the tumor tissue sample is obtained using an optical imaging device. An image of each of the tumor compartments and an image of each of the markers is generated. Pixel locations are then assigned to each of the tumor compartments based on a marker intensity value that specifically identifies that tumor compartment at that pixel location. Images of each of the markers are then analyzed at the pixel locations assigned to each of the tumor compartments to identify pixel locations that have an intensity value indicative of the presence of the label for the marker; in order to thereby localize each marker from the marker panel in the tumor tissue sample.

[0058] In some embodiments of computer-implemented methods for localizing a panel of labels within a colorectal cancer tumor tissue sample, the tumor tissue sample is additionally incubated with a panel of labels that specifically label two or more subcellular compartments of a cell nucleus, a cytoplasm, a nuclear membrane, a cell membrane, a mitochondria, an endoplasmic reticulum, a peroxisome and a lysosome. A high-resolution image of each of the markers that specifically identifies a subcellular compartment is then obtained using an optical imaging device to image each of the subcellular compartments. Pixel locations are assigned to each of the subcellular compartments based on a label intensity value that specifically identifies that subcellular compartment at that pixel location.

[0059] In some aspects of the methods described herein (including computer-implemented methods for locating a panel of markers within a sample of colorectal cancer tumor tissue, methods of predicting the clinical outcome for an individual diagnosed with colorectal cancer, the methods for determining the risk that an individual diagnosed with colorectal cancer will develop a recurrence of colorectal cancer, methods for differential diagnosis, and methods of treatment), the tissue sample is fixed, fixed, and embedded in paraffin, fresh and/or obtained from a biopsy. In some aspects, the colorectal cancer tumor tissue sample is a tissue microarray. In some embodiments, the tumor tissue sample has a thickness of about 5 microns.

[0060] According to some embodiments of the methods presented, colorectal cancer is lymph node-negative or lymph node-positive. In some respects, colorectal cancer is stage I, stage II, or stage III cancer. The described methods can be performed upon diagnosis of colorectal cancer, following surgical resection of colorectal cancer, or following treatment for colorectal cancer, such as treatment with MORAb-004.

[0061] The labels employed by the methods described herein (including computer-implemented methods for locating a panel of markers within a sample of colorectal cancer tumor tissue, methods of predicting clinical outcome for an individual diagnosed with colorectal cancer, and the methods for determining the risk that an individual diagnosed with colorectal cancer will develop a recurrence of colorectal cancer) are readily known to those skilled in the art. For example, suitable tags include, but should not be considered limited to, antibodies (e.g., detectably identified antibodies), radiolabels, fluorophores, fluorescent tags, epitope tags, biotin, chromophoric or chromogenic tags (e.g., 3, 3-Diaminobenzidine), ECL markers or enzymes. More specifically, the described labels include ruthenium, 111In-DOTA, 111In-diethylenetriaminepenta acetic acid (DTPA), horseradish peroxidase, alkaline phosphatase and beta-galactosidase, polyhistidine (HIS tag), acridine dyes, cyanine dyes , fluorone dyes, oxazine dyes, phenanthridine dyes, rhodamine dyes, Alexafluor® dyes, and the like. Examples of a fluorophore include 4',6-diamidino-2-phenyl indole (DAPI), fluorescein isothiocyanate (FITC), or a cyanine pigment (e.g., Cy2, Cy3, Cy3B, Cy3.5, Cy5, Cy5. 5, Cy7 and Cy7.5). In some embodiments, the tag comprises an antibody labeled with a radiolabel, a fluorescent tag, an epitope tag, biotin, a chromophore tag, an ECL tag, an enzyme, ruthenium, 111In-DOTA, 111In-diethylenetriaminepentaacetic acid (DTPA), horseradish peroxidase, alkaline phosphatase and beta-galactosidase, or polyhistidine, or similar markers known in the art. A signal amplification system (e.g., tyramide signal amplification) can be used with the marker. Labels defining tumor compartments can react with markers including, but not limited to, CD31, CD34, cytokeratin, beta catenin, alpha catenin, and vimentin. In preferred embodiments, the label that reacts with any given tumor compartment marker or biomarker comprises an antibody.

Tissue Sample Arrays

[0062] Described herein are arrays of tissue samples produced from cancerous tissue from one or more individuals, wherein the tissue samples are identified with one or more antibodies that specifically bind to a cellular or extracellular protein that may have varying expression. in a cancer cell. The tissue samples depicted in the matrix may be from a single individual or multiple individuals and may be from a single tumor or more than one tumor. For example, the described tissue matrix could be produced from multiple histological sections of a single tumor sample obtained from a single individual. Alternatively, the cancerous tissue of the described array may be obtained from more than one tumor from an individual. In some embodiments where more than one tumor provides the histological samples for the array, the tumors may be of the same type, stage, or grade as the tumor. Matrix tissue samples can be identified with one or more antibodies in order to detect more or more cellular protein or an extracellular protein that has differential expression in cancer cells or certain types of cancer cells relative to non-cancerous cells. In some embodiments, one or more tissue samples from the array may be identified with more than one antibody, allowing detection of more than one protein, where at least one of the identified proteins does not have differential expression in a cancer cell.

[0063] The described tissue sample matrices can exist in a wide variety of modalities. In some embodiments, the tissue sample arrays may be composed of multiple histological tumor samples, each of which is identified with at least one antibody that specifically binds to a protein that is known to be associated with a histological structure, of so that the associated structure is identified, wherein at least some tissue samples in the array are also identified with at least one antibody that specifically binds to a different protein, which may or may not be a structural protein. In some embodiments, the matrix samples can be identified for structural proteins with one or more antibodies that specifically bind to any of cytokeratin, vimentin, and CD31, while the matrix samples can also be identified with antibodies that specifically bind to any of TEM1, fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a or CAIX. In some embodiments, the matrix samples can be identified for structural proteins with one or more antibodies that specifically bind to cytokeratin, vimentin, and CD31, while the matrix samples can also be identified with antibodies that specifically bind to TEM1. , fibronectin, collagen IV and HIF2a.

[0064] In some embodiments, the described arrays may be prepared such that each tissue sample in the array is identified with a different combination of antibodies to allow a variety of individual histological structures to be identified across the array, while these same tissue samples tissue can be simultaneously identified with one or more antibodies specific for different proteins, so that multiple protein expression profiles can be assessed simultaneously across the array. Using the array in this way can allow a combination of parameters related to protein expression of interest to be obtained for a tissue or tumor of interest by making use of multiple search parameters in the array. For example, in one aspect, a single tissue sample on the array could be identified to detect a protein of interest (such as TEM-1) and also be identified to detect cytokeratin, vimentin, and CD31 to denote certain cellular or histological structures of interest; Furthermore, other elements of the cell, such as the nucleus, could be detected using a marker capable of binding to DNA. On the same matrix, other histological samples could be identified in a similar manner, except that the marker for TEM-1 could be replaced with a marker for a different protein of interest, such as fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a or CAIX. It would then be possible to evaluate the expression levels of the marked proteins of interest in relation to each other and in relation to the structural elements marked in each sample. In this way, it would be possible to determine, for example, the expression levels of any combination of TEM1, fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a and CAIX in a single tumor, evaluate the expression levels of these proteins in relation to each other. to the others and also determine the location of their expression in relation to the structural elements identified in each sample of the matrix. Detection of identified proteins can be achieved by immunohistochemical means. For example, fluorescence-based identification using an identified primary antibody or an identified secondary antibody, and detection using a fluorescence detector.

[0065] In view of the aforementioned description, the embodiments described below can provide additional knowledge of the tissue sample matrices disclosed herein. In one embodiment, the described tissue sample matrix may include multiple histological slides of a tumor, where the majority of the histological slides are identified to allow: detection of nuclei in the cells of the tissue samples, detection of one or more of cytokeratin , vimentin and CD31, and detection of one or more of TEM1, fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a and CAIX, where the identified proteins are labeled to allow their localization and quantification in the cells of the histological slides. In one embodiment, the described tissue sample matrix may include multiple histological slides of a tumor, where the majority of the histological slides are identified to allow: detection of cell nuclei in tissue samples, detection of one or more of cytokeratin , vimentin and CD31, and at least one cell identified with any one or more of cytokeratin, vimentin and CD31 is also identified with one or more antibodies specific for TEM1, fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a or CAIX, where the identified proteins are marked to allow their localization and quantification in the cells of the histological slides.

[0066] In one embodiment, the described tissue sample matrix may include multiple histological slides of a tumor, where the majority of the histological slides are identified to allow: the detection of cell nuclei in the tissue samples, the detection of a or more of cytokeratin, vimentin and CD31, and at least one histological slide identified with an antibody specific for one or more of cytokeratin, vimentin and CD31 is also identified with an antibody specific for TEM1, fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a or CAIX, such that the array as a whole includes at least one histological slide identified with any one or more of cytokeratin, vimentin and CD31 that is also identified with an antibody specific for TEM1, at least one histological slide identified with any one or more of cytokeratin, vimentin and CD31 is also identified with an antibody specific for fibronectin-1, at least one histological slide identified with any one or more of cytokeratin, vimentin and CD31 is also identified with an antibody specific for PDGFRß, at least at least one histological slide identified with any one or more of cytokeratin, vimentin and CD31 is also identified with an antibody specific for collagen I, at least one histological slide identified with any one or more of cytokeratin, vimentin and CD31 is also labeled with a antibody specific for collagen IV, at least one histological slide labeled with any one or more of cytokeratin, vimentin and CD31 is also labeled with an antibody specific for HIF2a, and at least one histological slide labeled with any one or more of cytokeratin, vimentin and CD31 is also identified with a CAIX-specific antibody, where the identified proteins are labeled to allow their localization and quantification in the cells of the histological slides.

[0067] In one embodiment, the described tissue sample matrix may include multiple histological slides of a tumor, where the majority of the histological slides are identified to allow: detection of cell nuclei in the tissue samples, detection of one or more of cytokeratin, vimentin and CD31, and at least one histological slide identified with any one or more of cytokeratin, vimentin and CD31 is also identified with an antibody specific for TEM1, fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a or CAIX, such that the array as a whole includes at least one histological slide identified with: an antibody specific for vimentin and is also identified with an antibody specific for TEM1; an antibody specific for cytokeratin and is also identified with an antibody specific for TEM1; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for TEM1; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for TEM1; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for HIF2a; an antibody specific for vimentin, and is also identified with an antibody specific for HIF2a; an antibody specific for cytokeratin and is also identified with an antibody specific for CAIX; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for CAIX; an antibody specific for vimentin and is also labeled with an antibody specific for PDGFRß; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for PDGFRß; an antibody specific for vimentin and is also identified with an antibody specific for fibronectin-1; an antibody specific for vimentin and is also identified with an antibody specific for collagen I; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for collagen I; an antibody specific for vimentin and is also identified with an antibody specific for collagen IV; and an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for collagen IV.

[0068] In one embodiment, the described tissue sample matrix may include multiple histological slides of a tumor, where the majority of the histological slides are identified to allow: the detection of cell nuclei in the tissue samples, the detection of a or more of cytokeratin, vimentin and CD31, and at least one histological slide identified with any one or more of cytokeratin, vimentin and CD31 is also identified with an antibody specific for TEM1, fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a or CAIX, such that the array as a whole includes any one or more of the following identification combinations: an antibody specific for vimentin and is also identified with an antibody specific for TEM1; an antibody specific for cytokeratin and is also identified with an antibody specific for TEM1; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for TEM1; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for TEM1; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for HIF2a; an antibody specific for vimentin, and is also identified with an antibody specific for HIF2a; an antibody specific for cytokeratin and is also identified with an antibody specific for CAIX; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for CAIX; an antibody specific for vimentin and is also identified with an antibody specific for PDGFRß; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for PDGFRß; an antibody specific for vimentin and is also identified with an antibody specific for fibronectin-1; an antibody specific for vimentin and is also identified with an antibody specific for collagen I; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for collagen I; an antibody specific for vimentin and is also identified with an antibody specific for collagen IV; or an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for collagen IV.

[0069] In one embodiment, the described tissue sample matrix may include multiple histological slides of a tumor, where the majority of the histological slides are identified to allow: the detection of cell nuclei in the tissue samples, the detection of a or more of cytokeratin, vimentin and CD31, and at least one histological slide identified with any one or more of cytokeratin, vimentin and CD31 is also identified with an antibody specific for TEM1, fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a or CAIX, such that the array as a whole includes at least one histological slide identified with: an antibody specific for vimentin and is also identified with an antibody specific for TEM1; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for TEM1; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for HIF2a; an antibody specific for vimentin, and is also identified with an antibody specific for HIF2a; an antibody specific for cytokeratin and is also identified with an antibody specific for CAIX; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for CAIX; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for PDGFRß; and an antibody specific for vimentin and is also identified with an antibody specific for fibronectin-1.

[0070] In one embodiment, the described tissue sample matrix may include multiple histological slides of a tumor, where the majority of the histological slides are identified to allow: the detection of cell nuclei in the tissue samples, the detection of a or more of cytokeratin, vimentin and CD31, and at least one histological slide identified with any one or more of cytokeratin, vimentin and CD31 is also identified with an antibody specific for TEM1, fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a or CAIX, such that the array as a whole includes at least one histological slide identified with: an antibody specific for vimentin and is also identified with an antibody specific for TEM1; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for TEM1; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for HIF2a; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for CAIX; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for PDGFRß; and an antibody specific for vimentin and is also identified with an antibody specific for fibronectin-1.

[0071] In one embodiment, the described tissue sample matrix may include multiple histological slides of a tumor, where the majority of the histological slides are identified to allow: the detection of cell nuclei in the tissue samples, the detection of a or more of cytokeratin, vimentin and CD31, and at least one histological slide identified with any one or more of cytokeratin, vimentin and CD31 is also identified with an antibody specific for TEM1, fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a or CAIX, so that the array as a whole includes 2 to 5 histological slides identified with: an antibody specific for vimentin and is also identified with an antibody specific for TEM1; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for TEM1; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for HIF2a; an antibody specific for vimentin, and is also identified with an antibody specific for HIF2a; an antibody specific for cytokeratin and is also identified with an antibody specific for CAIX; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for CAIX; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for PDGFRß; and an antibody specific for vimentin and is also identified with an antibody specific for fibronectin-1.

[0072] In one embodiment, the described tissue sample matrix may include multiple histological slides of a tumor, where the majority of the histological slides are identified to allow: the detection of cell nuclei in the tissue samples, the detection of a or more of cytokeratin, vimentin and CD31, and at least one histological slide identified with any one or more of cytokeratin, vimentin and CD31 is also identified with an antibody specific for TEM1, fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a or CAIX, so that the array as a whole includes 2 to 5 histological slides identified with: an antibody specific for vimentin and is also identified with an antibody specific for TEM1; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for TEM1; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for HIF2a; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for CAIX; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for PDGFRß; and an antibody specific for vimentin and is also identified with an antibody specific for fibronectin-1.

[0073] In one embodiment, the described tissue sample matrix may include multiple histological slides of a tumor, where the majority of the histological slides are identified to allow: the detection of cell nuclei in the tissue samples, the detection of a or more of cytokeratin, vimentin and CD31, and at least one histological slide identified with any one or more of cytokeratin, vimentin and CD31 is also identified with an antibody specific for TEM1, fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a or CAIX, so that the array as a whole includes only histological slides identified with: an antibody specific for vimentin and is also identified with an antibody specific for TEM1; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for TEM1; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for HIF2a; an antibody specific for vimentin, and is also identified with an antibody specific for HIF2a; an antibody specific for cytokeratin and is also identified with an antibody specific for CAIX; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for CAIX; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for PDGFRß; and an antibody specific for vimentin and is also identified with an antibody specific for fibronectin-1.

[0074] In one embodiment, the described tissue sample matrix may include multiple histological slides of a tumor, where the majority of the histological slides are identified to allow: the detection of cell nuclei in the tissue samples, the detection of a or more of cytokeratin, vimentin and CD31, and at least one histological slide identified with any one or more of cytokeratin, vimentin and CD31 is also identified with an antibody specific for TEM1, fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a or CAIX, so that the array as a whole includes only histological slides identified with: an antibody specific for vimentin and is also identified with an antibody specific for TEM1; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for TEM1; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for HIF2a; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for CAIX; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for PDGFRß; and an antibody specific for vimentin and is also identified with an antibody specific for fibronectin-1.

[0075] In one embodiment, the described tissue sample matrix may include multiple histological slides of a tumor, where the majority of the histological slides are identified to allow: the detection of cell nuclei in the tissue samples, the detection of a or more of cytokeratin, vimentin and CD31, and at least one histological slide identified with any one or more of cytokeratin, vimentin and CD31 is also identified with an antibody specific for TEM1, fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a or CAIX, such that the array as a whole includes histological slides labeled with only: an antibody specific for vimentin and is also labeled with an antibody specific for TEM1; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for TEM1; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for HIF2a; an antibody specific for vimentin, and is also identified with an antibody specific for HIF2a; an antibody specific for cytokeratin and is also identified with an antibody specific for CAIX; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for CAIX; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for PDGFRß; and an antibody specific for vimentin and is also identified with an antibody specific for fibronectin-1.

[0076] In one embodiment, the described tissue sample matrix may include multiple histological slides of a tumor, where the majority of the histological slides are identified to allow: the detection of cell nuclei in the tissue samples, the detection of a or more of cytokeratin, vimentin and CD31, and at least one histological slide identified with any one or more of cytokeratin, vimentin and CD31 is also identified with an antibody specific for TEM1, fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a or CAIX, such that the array as a whole includes histological slides labeled with only: an antibody specific for vimentin and is also labeled with an antibody specific for TEM1; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for TEM1; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for HIF2a; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for CAIX; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for PDGFRß; and an antibody specific for vimentin and is also identified with an antibody specific for fibronectin-1.

[0077] In one embodiment, the described tissue sample matrix may include multiple histological slides of a tumor, where the majority of the histological slides are identified to allow: the detection of cell nuclei in the tissue samples, the detection of a or more of cytokeratin, vimentin and CD31, and at least one histological slide identified with any one or more of cytokeratin, vimentin and CD31 is also identified with an antibody specific for TEM1, fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a or CAIX, such that the array as a whole includes histological slides identified with only: an antibody specific for vimentin and is also identified with an antibody specific for TEM1; an antibody specific for cytokeratin, an antibody specific for CD31, and is also identified with an antibody specific for TEM1; an antibody specific for vimentin, an antibody specific for CD31, and is also identified with an antibody specific for HIF2a; an antibody specific for cytokeratin, and is also identified with an antibody specific for collagen IV; and an antibody specific for vimentin and is also identified with an antibody specific for fibronectin-1.

[0078] As will be understood by those skilled in the art, other modalities of the antigen detection parameters defined herein could also be employed to produce results substantially similar to those provided by the methods or tissue matrices described. For example, it would be possible to simply combine certain identification conditions separately described above, but use distinct detection conditions to avoid interference that would otherwise be caused by adding markers. Other such modifications known to those skilled in the art could also be used in an attempt to avoid the described detection methods; however, these alternatives should be considered within the knowledge of one skilled in the art when considered in view of the present disclosure.

Kits

[0079] The materials for use in the methods of the present invention are suitable for the preparation of kits produced according to well-known procedures. The invention thus provides kits comprising agents for detecting and/or quantifying expression of the disclosed panel of markers to predict prognostic outcome, which may include labels specific to the panel of markers, e.g., at least two markers from one label. specific for TEM1, a specific label for PDGFRß, a specific label for HIF2a, a specific label for CAIX, a specific label for fibronectin, a specific label for collagen I, a specific label for collagen IV and a specific label for CD31. The TEM1-specific marker may include an antibody directed against TEM1, for example, the 9G5 antibody (antibody-producing cells that produce the 9G5 antibody have been deposited with the American Type Culture Collection, 10801 University Blvd., Manassas, Virginia, USA 20110- 2209, on February 16, 2012 and were assigned ATCC accession number PTA- 12547). The specific marker for PDGFRß may include an antibody directed against PDGFRß. The specific marker for HIF2a may include an antibody directed against HIF2a. The CAIX-specific marker may include an antibody directed against CAIX. The fibronectin-specific marker may include an antibody directed against fibronectin. The collagen I-specific marker may include an antibody directed against collagen I. The collagen IV-specific marker may include an antibody directed against collagen IV. Labels defining tumor compartments can react with markers including, but not limited to, CD31, CD34, cytokeratin, beta catenin, alpha catenin, and vimentin. Labels for biomarkers and tumor compartment markers are readily known to those skilled in the art. For example, suitable labels include, but should not be considered limited to, antibodies (e.g., detectably identified antibody), radiolabels, fluorophores, fluorescent labels, epitope tags, biotin, chromophoric or chromogenic labels (e.g., 3, 3- Diaminobenzidine), ECL markers or enzymes. More specifically, the described labels include ruthenium, 111In-DOTA, 111In-diethylenetriaminepenta acetic acid (DTPA), horseradish peroxidase, alkaline phosphatase and beta-galactosidase, polyhistidine (HIS tag), acridine dyes, cyanine dyes , fluorone dyes, oxazine dyes, phenanthridine dyes, rhodamine dyes, Alexafluor® dyes, and the like. Examples of a fluorophore include 4',6-diamidino-2-phenyl indole (DAPI), fluorescein isothiocyanate (FITC), or a cyanine pigment (e.g., Cy2, Cy3, Cy3B, Cy3.5, Cy5, Cy5. 5, Cy7 and Cy7.5). In some embodiments, the tag comprises an antibody labeled with a radiolabel, a fluorescent tag, an epitope tag, biotin, a chromophore tag, an ECL tag, an enzyme, ruthenium, 111In-DOTA, 111In-diethylenetriaminepentaacetic acid (DTPA), horseradish peroxidase, alkaline phosphatase and beta-galactosidase, or polyhistidine, or similar markers known in the art. A signal amplification system (e.g., tyramide signal amplification) can be used with the marker and thus included in the described kits.

[0080] Furthermore, the kits may optionally comprise the reagent(s) with an identification or marker description or instructions relating to their use in the methods described herein. Kits may comprise containers, each containing one or more of the various reagents (typically in concentrated form) used in the methods. Mathematical algorithms used to estimate or quantify prognostic or predictive information are also appropriately potential components of kits.

[0081] In some aspects, a kit is provided for localizing or quantifying a panel of markers in colorectal cancer, including a TEM1-specific label, a PDGFRß-specific label, a HIF2a-specific label, and a CAIX-specific label. In some embodiments, instructions for using the kit are included.

[0082] In some aspects, a kit is provided for localizing or quantifying a panel of markers in colorectal cancer, including a label specific for TEM1, a label specific for collagen IV, a label specific for HIF2a, and a label specific for fibronectin-1. . In some embodiments, instructions for using the kit are included.

[0083] The kits described herein may also include one or more of a marker specific for cell nuclei, a marker specific for the tumor cytoplasm, a marker specific for the tumor stroma, a marker specific for the vasculature, a marker for the tumor vasculature and a marker for tumor stromal vasculature. For example, the marker specific for the tumor cytoplasm may react with cytokeratin, the marker specific for the tumor stroma may react with vimentin, and the marker specific for the vasculature may react with CD31.

[0084] In some embodiments of the described kits, the TEM1-specific marker comprises an antibody directed against TEM1, the PDGFRß-specific marker comprises an antibody directed against PDGFRß, the HIF2a-specific marker comprises an antibody directed against HIF2a, the specific marker for CAIX comprises an antibody directed against CAIX, the fibronectin-specific marker comprises an antibody directed against fibronectin, the collagen I-specific marker comprises an antibody directed against collagen I, and the collagen IV-specific marker comprises an antibody directed against collagen IV.

Reports

[0085] The methods described herein can produce a report or summary of the probability of recurrence of colorectal cancer, the prediction of clinical outcome or the location of markers. The methods and reports described may also include storing the report in a database. Alternatively, the methods may additionally create a record in a database for the individual and populate the record with the data. In one embodiment, the report is a paper report, in another embodiment, the report is an audit report, in another embodiment, the report is an electronic record. It is contemplated that the report is provided to a doctor and/or the patient. Receiving the report may further include establishing a network connection to a server computer that includes the data and report and requesting the data and report from the server computer.

Selected aspects

[0086] Selected aspects of the subject matter described herein are provided to exemplify how the subject matter described may be used or applied. These selected aspects are provided by way of illustration and are not intended to limit the present disclosure in any way.

[0087] The following aspects are provided: 1. Method for assessing the risk that an individual diagnosed with colorectal cancer has of developing a recurrence of colorectal cancer comprising: a) determining the expression level for each marker of a panel of markers in a panel of tumor compartments in a tumor tissue sample from the subject, wherein the panel of markers comprises at least two of TEM1, HIF2a, CAIX, PDGFRß, fibronectin, collagen I, collagen IV and CD31, and wherein the panel of tumor compartments comprises at least three tumor compartments of pure stroma, tumor, stromal vessel, and tumor vessel; b) determine the TAPPS score for said individual; and c) compare the individual's TAPPS score to the TAPPS score of a population of individuals diagnosed with colorectal cancer; where the individual is at low risk for colorectal cancer recurrence if the individual's TAPPS score is low relative to the population's TAPPS score and where the individual is at high risk for colorectal cancer recurrence if the individual's TAPPS score is high in relation to the population's TAPPS score. 2. Method according to aspect 1, wherein the expression level for each marker of the marker panel is determined using an automated pathology system. 3. Method according to aspect 1 or 2, wherein the expression level for each marker of the marker panel is determined using a quantitative image analysis procedure. 4. Method according to any of aspects 1 to 3, where the colorectal cancer is lymph node-negative. 5. Method according to aspect 4, in which the individual is at low risk for colorectal cancer recurrence if the individual's TAPPS score is intermediate in relation to the population's TAPPS score. 6. Method according to any of aspects 1 to 3, where the colorectal cancer is lymph node-positive. 7. Method according to aspect 6, in which the individual is at high risk for colorectal cancer recurrence if the individual's TAPPS score is intermediate in relation to the population's TAPPS score. 8. Method according to the previous aspect, where colorectal cancer is stage I, stage II or stage III cancer. 9. Method, according to any of the previous aspects, further comprising identifying at least three compartments in the individual's tumor tissue sample. 10. Method according to any of the preceding aspects which further comprises identifying at least one of the tumor cell nuclei and the tumor cell cytoplasm. 11. Method according to any of the previous aspects wherein said panel of markers comprises three of TEM1, HIF2a, CAIX, PDGFRß, fibronectin, collagen I, collagen IV and CD31. 12. Method according to any of the previous aspects wherein said panel of markers comprises four of TEM1, HIF2a, CAIX, PDGFRß, fibronectin, collagen I, collagen IV and CD31. 13. Method according to any one of aspects 1 to 10, wherein said panel of markers comprises five of TEM1, HIF2a, CA9, PDGFRß, fibronectin, collagen I, collagen IV and CD31. 14. Method according to any one of aspects 1 to 10, wherein said panel of markers comprises six of TEM1, HIF2a, CAIX, PDGFRß, fibronectin, collagen I, collagen IV and CD31. 15. Method according to any one of aspects 1 to 10, wherein said marker panel comprises TEM1 and CAIX. 16. The method of aspect 12, wherein said panel of markers comprises TEM1, HIF2a, CAIX and PDGFRß. 17. The method of aspect 13, wherein said panel of markers comprises TEM1, HIF2a, CAIX, fibronectin and PDGFRß. 18. Method according to any one of aspects 1 to 10, wherein said panel of markers comprises TEM1, HIF2a, fibronectin and collagen IV. 19. Method according to any of the preceding aspects, wherein the panel of tumor compartments comprises one or more of stroma and tumor. 20. Method according to any one of aspects 1 to 19, wherein the panel of tumor compartments comprises one or more of stroma, tumor, stromal vessel and tumor vessel. 21. Method according to any previous aspect, wherein step a) comprises determining the expression level of TEM1 in the tumor stroma, determining the level of TEM1 in the tumor vessel, determining the level of CAIX in the tumor vessel, and determining the level of CAIX in the tumor. 22. Method according to any one of aspects 1 to 20, wherein step a) comprises determining the level of expression of TEM1 in the tumor stroma, determining the level of TEM1 in the tumor vessel, determining the level of expression of HIF2a in the stromal vessel, determine the level of HIF2a in the tumor stroma, determine the level of CAIX in the tumor, determine the level of CAIX in the tumor vessel, determine the level of PDGFRß in the tumor stroma. 23. The method of aspect 22, wherein step a) further comprises determining the level of fibronectin expression in the tumor stroma. 24. Method, in accordance with any of the previous aspects, which additionally comprises the step of creating a report summarizing said risk. 25. Method, in accordance with any of the foregoing aspects, wherein, if it is determined that said individual is at high risk for recurrence of colorectal cancer, said individual is subjected to additional therapy. 26. Method according to aspect 25, wherein said additional therapy comprises TEM-1 targeted therapy, chemotherapy and/or radiotherapy. 27. The method of aspect 26, wherein said TEM-1 targeted therapy comprises MORAb-004. 28. Method, in accordance with any of the foregoing aspects, wherein, if it is determined that said individual is at high risk for recurrence of colorectal cancer, said individual is subjected to monitoring for recurrence or progression of the disease. 29. Method, according to any of the previous aspects, in which the tumor tissue sample is fixed and embedded in paraffin. 30. Method, according to any of the preceding aspects, in which the tumor tissue sample is fresh. 31. Method according to any of the preceding aspects, wherein the tumor tissue sample is obtained from a tissue biopsy. 32. Method, according to any of the previous aspects, wherein said individual is a human being. 33. Method according to aspect 1, wherein the TAPPS equation is established by determining the coefficients A, B, C, D etc. for each of the markers in each of the tumor compartments selected for inclusion in the population of patients with colorectal cancer. 34. Method according to aspect 1, wherein the TAPPS equation is established by determining the coefficients A, B, C and D for the colorectal cancer patient population to provide any of the following TAPPS equations: by determining the individual's TAPPS score using the said TAPP equation. 35. Method according to aspect 1, wherein the TAPPS equation is established by determining the coefficients A, B, C, D, E, F, G and H for the colorectal cancer patient population to provide the TAPPS equation: and by determining the individual's TAPPS score using said TAPPS equation. 36. Method according to aspect 1, wherein the individual's TAPPS score is determined using any of the following TAPPS equations: 37. Method of predicting the clinical outcome of an individual diagnosed with colorectal cancer, comprising: a) determining the level of expression for each marker from a panel of markers in a panel of tumor compartments in a tumor tissue sample from the individual, wherein the panel of markers comprises at least two of TEM1, HIF2a, CAIX, PDGFRß, fibronectin, collagen I, collagen IV and CD31 and wherein the panel of tumor compartments comprises at least three tumor compartments of pure stroma, tumor, stromal vessel, and tumor vessel; b) determine the TAPPS score for said individual; and c) compare the individual's TAPPS score to the TAPPS score of a population of individuals diagnosed with colorectal cancer; wherein a low TAPPS score for said individual relative to the population TAPPS score is predictive of a positive clinical outcome and wherein a high TAPPS score for said individual relative to the population TAPPS score is predictive of a poor clinical outcome. 38. The method of aspect 37, wherein the expression level for each marker of the marker panel is determined using an automated pathology system. 39. Method according to aspect 37 or 38, wherein the expression level for each marker of the marker panel is determined using a quantitative image analysis procedure. 40. Method according to any of aspects 37 to 39, in which the colorectal cancer is lymph node-negative. 41. Method according to aspect 34, wherein an intermediate TAPPS score for said individual relative to the population TAPPS score is predictive of a positive clinical outcome for said individual. 42. Method according to any of aspects 37 to 39, in which the colorectal cancer is lymph node-positive. 43. Method according to aspect 42, wherein an intermediate TAPPS score for said individual relative to the population TAPPS score is predictive of a poor clinical outcome for said individual. 44. Method according to any one of aspects 37 to 43, wherein the colorectal cancer is stage I, stage II or stage III cancer. 45. The method of any one of aspects 37 to 44, further comprising identifying at least two compartments in the subject's tumor tissue sample. 46. Method according to any one of aspects 37 to 45, further comprising identifying at least one of the tumor cell nuclei and tumor cell cytoplasm. 47. Method according to any one of aspects 37 to 46, wherein said panel of markers comprises three of TEM1, HIF2a, CAIX, PDGFRß, fibronectin, collagen I, collagen IV and CD31. 48. Method according to any one of aspects 37 to 46, wherein said panel of markers comprises four of TEM1, HIF2a, CAIX, PDGFRß, fibronectin, collagen I, collagen IV and CD31. 49. Method according to any one of aspects 37 to 46, wherein said panel of markers comprises five of TEM1, HIF2a, CAIX, PDGFRß, fibronectin, collagen I, collagen IV and CD31. 50. Method according to any one of aspects 37 to 46, wherein said panel of markers comprises six of TEM1, HIF2a, CAIX, PDGFRß, fibronectin, collagen I, collagen IV and CD31. 51. The method of aspect 48 wherein said panel of markers comprises TEM1, HIF2a, collagen IV and fibronectin. 52. The method of aspect 51, wherein said panel of tumor compartments comprises any one of a) stroma, tumor vessel, stromal vessel and tumor or b) stroma and tumor. 53. The method of aspect 48, wherein said panel of markers comprises TEM1, HIF2a, CAIX and PDGFRß. 54. The method of aspect 49, wherein said panel of markers comprises TEM1, HIF2a, CAIX, fibronectin and PDGFRß. 55. Method according to any one of aspects 37 to 46, wherein said panel of markers comprises TEM1, HIF2a, CAIX, PDGFRß, fibronectin, collagen I, collagen IV and CD31. 56. The method of any one of aspects 37 to 55, wherein the panel of tumor compartments comprises pure stroma, tumor, stromal vessel and tumor vessel. 57. Method according to any one of aspects 37 to 56, wherein step a) comprises determining the level of expression of TEM1 in the tumor stroma, determining the level of TEM1 in the tumor vessel, determining the level of expression of HIF2a in the stromal vessel, determine the level of HIF2a in the tumor stroma, determine the level of CAIX in the tumor, determine the level of CAIX in the tumor vessel and determine the level of PDGFRß in the tumor stroma. 58. The method of aspect 57, wherein step a) further comprises determining the level of fibronectin expression in the tumor stroma. 59. Method according to any one of aspects 37 to 58, which further comprises the step of creating a report summarizing said prediction. 60. Method according to any one of aspects 37 to 59, wherein, if a poor clinical outcome is predicted for said individual, said individual is subjected to additional therapy. 61. Method according to aspect 60, wherein said additional therapy comprises TEM-1 targeted therapy, chemotherapy and/or radiotherapy. 62. The method of aspect 61, wherein said TEM-1 targeted therapy comprises MORAb-004. 63. Method according to any one of aspects 37 to 62, wherein, if a poor clinical outcome is predicted for said individual, said individual is subjected to monitoring for recurrence or progression of the disease. 64. Method according to any one of aspects 37 to 63, wherein the tumor tissue sample is fixed and embedded in paraffin. 65. Method according to any one of aspects 37 to 64, wherein the tumor tissue sample is fresh. 66. The method of any one of aspects 37 to 65, wherein the tumor tissue sample is obtained from a tissue biopsy. 67. Method according to any one of aspects 37 to 66, wherein said individual is a human being. 68. Method according to any one of aspects 37 to 67, wherein said clinical result is expressed in terms of progression-free survival (PFS), recurrence-free interval (RFI), overall survival (OS), survival of disease (DFS) or distant recurrence-free interval (DRFI). 69. Method according to aspect 37, wherein the TAPPS equation is established by determining the coefficients A, B, C, D etc. for each of the markers in each of the tumor compartments selected for inclusion in the population of patients with colorectal cancer. 70. Method according to aspect 37 wherein the TAPPS equation is established by determining the coefficients A, B, C and for the colorectal cancer patient population to provide the TAPPS equation: and by determining the individual's TAPPS score using the said TAPP equation. 71. Method according to aspect 37, wherein the TAPPS equation is established by determining the coefficients A, B, C, D, E, F, G and H for the colorectal cancer patient population to provide the TAPPS equation: and by determining the individual's TAPPS score using said TAPPS equation. 72. Method in accordance with aspect 37, wherein the individual's TAPPS score is determined using the following TAPPS equation: 73. Kit comprising at least two markers selected from the group consisting of a TEM1 specific marker, a PDGFRß specific marker, a HIF2a specific marker, a CAIX specific marker, a fibronectin specific marker, a collagen specific marker I, a specific marker for collagen IV and a specific marker for CD31 and instructions on how to use the kit. 74. Kit according to aspect 73, comprising a TEM1-specific marker, a PDGFRß-specific marker, a HIF2a-specific marker, a CAIX-specific marker, and instructions on how to use the kit. 75. Kit, according to the aspect of 73 or 74, further comprising one or more of a marker specific for cell nuclei, a marker specific for the tumor cytoplasm, a marker specific for the tumor stroma, a specific marker for the vasculature, a marker for the tumor vasculature and a marker for the tumor stromal vasculature. 76. Kit according to aspect 75, wherein the tumor cytoplasm-specific marker reacts with cytokeratin, the tumor stroma-specific marker reacts with vimentin, and the vasculature-specific marker reacts with CD31. 77. Kit according to aspect 73, wherein said TEM1-specific marker comprises an antibody that specifically binds to TEM1, said PDGFRß-specific marker comprises an antibody that specifically binds to PDGFRß, said specific marker for HIF2a comprises an antibody that specifically binds to HIF2a, said CAIX-specific marker comprises an antibody that specifically binds to CAIX, said fibronectin-specific marker comprises an antibody that specifically binds to fibronectin, said collagen I-specific marker comprises an antibody that specifically binds to collagen I, and said collagen IV-specific marker comprises an antibody that specifically binds to collagen IV. 78. Kit according to aspect 77, wherein said antibody directed against TEM1 is produced by the hybridoma deposited with the ATCC having accession number PTA-12547. 79. Kit according to aspect 73, wherein each of said labels comprises a radiolabel, a fluorophore, an epitope tag, biotin, a chromophore tag, an ECL tag, an enzyme, ruthenium, 111In-DOTA , DAB (3, 3'-diaminobenzidine), 111In-diethylenetriaminepentaacetic acid (DTPA), horseradish peroxidase, alkaline phosphatase and beta-galactosidase or polyhistidine. 80. Kit according to aspect 79, wherein the fluorophore is selected from the group consisting of 4',6-diamidino-2-phenylindole (DAPI), fluorescein isothiocyanate (FITC) and a cyanine dye. 81. Tissue sample matrix, comprising tumor tissue samples identified with one or more antibodies that specifically bind to cytokeratin, vimentin, CD31, TEM1, fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a or CAIX, wherein each matrix comprises at least one tissue sample identified with at least one antibody that specifically binds to cytokeratin, vimentin, CD31, TEM1, fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a or CAIX. 82. Tissue sample array according to aspect 80, wherein one or more of the tissue samples is identified with a nuclear marker. 83. Tissue sample matrix according to aspect 82, wherein the nuclear marker is DAPI. 84. Tissue sample matrix according to any one of aspects 80 to 83, wherein the matrix comprises at least one tissue sample identified with an antibody that specifically binds to any one of cytokeratin, vimentin or CD31, and an antibody that specifically binds to any of TEM1, fibronectin-1, PDGFRß, collagen I, collagen IV, HIF2a, or CAIX. 85. A tissue sample array according to any one of aspects 80 to 84, wherein the array comprises a plurality of tissue samples identified with an antibody that specifically binds to cytokeratin, wherein: at least one tissue sample identified with an antibody that specifically binds to cytokeratin, further comprises an antibody that specifically binds to TEM1; at least one tissue sample identified with an antibody that specifically binds to cytokeratin further comprises an antibody that specifically binds to fibronectin-1; at least one tissue sample identified with an antibody that specifically binds to cytokeratin further comprises an antibody that specifically binds to PDGFRß; at least one tissue sample identified with an antibody that specifically binds to cytokeratin further comprises an antibody that specifically binds to collagen I; at least one tissue sample identified with an antibody that specifically binds to cytokeratin further comprises an antibody that specifically binds to collagen IV; at least one tissue sample identified with an antibody that specifically binds to cytokeratin further comprises an antibody that specifically binds to HIF2a; and at least one tissue sample identified with an antibody that specifically binds to cytokeratin further comprises an antibody that specifically binds to CAIX; 86. A tissue sample array according to any one of aspects 80 to 84, wherein the array comprises a plurality of tissue samples identified with an antibody that specifically binds to vimentin, wherein: at least one tissue sample identified with an antibody that specifically binds to vimentin, further comprises an antibody that specifically binds to TEM1; at least one tissue sample identified with an antibody that specifically binds to vimentin further comprises an antibody that specifically binds to fibronectin-1; at least one tissue sample identified with an antibody that specifically binds to vimentin further comprises an antibody that specifically binds to PDGFRß; at least one tissue sample identified with an antibody that specifically binds to vimentin further comprises an antibody that specifically binds to collagen I; at least one tissue sample identified with an antibody that specifically binds to vimentin further comprises an antibody that specifically binds to collagen IV; at least one tissue sample identified with an antibody that specifically binds to vimentin further comprises an antibody that specifically binds to HIF2a; and at least one tissue sample identified with an antibody that specifically binds to vimentin, further comprises an antibody that specifically binds to CAIX. 87. A tissue sample array according to any one of aspects 80 to 84, wherein the array comprises a plurality of tissue samples identified with an antibody that specifically binds to CD31, wherein: at least one sample of tissue identified with an antibody that specifically binds to CD31, further comprises an antibody that specifically binds to TEM1; at least one tissue sample identified with an antibody that specifically binds to CD31 further comprises an antibody that specifically binds to fibronectin-1; at least one tissue sample identified with an antibody that specifically binds to CD31 further comprises an antibody that specifically binds to PDGFRß; at least one tissue sample identified with an antibody that specifically binds to CD31 further comprises an antibody that specifically binds to collagen I; at least one tissue sample identified with an antibody that specifically binds to CD31 further comprises an antibody that specifically binds to collagen IV; at least one tissue sample identified with an antibody that specifically binds to CD31 further comprises an antibody that specifically binds to HIF2a; and at least one tissue sample identified with an antibody that specifically binds to CD31, further comprises an antibody that specifically binds to CAIX. 88. Tissue sample array according to any one of aspects 80 to 84, wherein the array comprises a plurality of tissue samples identified with an antibody that specifically binds to CD31 and an antibody that specifically binds to vimentin , wherein: at least one tissue sample identified with an antibody that specifically binds to CD31 and an antibody that specifically binds to vimentin further comprises an antibody that specifically binds to TEM1; at least one tissue sample identified with an antibody that specifically binds to CD31 and an antibody that specifically binds to vimentin further comprises an antibody that specifically binds to fibronectin-1; at least one tissue sample identified with an antibody that specifically binds to CD31 and an antibody that specifically binds to vimentin further comprises an antibody that specifically binds to PDGFRß; at least one tissue sample identified with an antibody that specifically binds to CD31 and an antibody that specifically binds to vimentin further comprises an antibody that specifically binds to collagen I; at least one tissue sample identified with an antibody that specifically binds to CD31 and an antibody that specifically binds to vimentin further comprises an antibody that specifically binds to collagen IV; at least one tissue sample identified with an antibody that specifically binds to CD31 and an antibody that specifically binds to vimentin further comprises an antibody that specifically binds to HIF2a; and at least one tissue sample identified with an antibody that specifically binds to CD31 and an antibody that specifically binds to vimentin further comprises an antibody that specifically binds to CAIX. 89. Tissue sample array according to any one of aspects 80 to 84, wherein the array comprises a plurality of tissue samples identified with an antibody that specifically binds to CD31 and an antibody that specifically binds to cytokeratin , wherein: at least one tissue sample identified with an antibody that specifically binds to CD31 and an antibody that specifically binds to cytokeratin further comprises an antibody that specifically binds to TEM1; at least one tissue sample identified with an antibody that specifically binds to CD31 and an antibody that specifically binds to cytokeratin further comprises an antibody that specifically binds to fibronectin-1; at least one tissue sample identified with an antibody that specifically binds to CD31 and an antibody that specifically binds to cytokeratin further comprises an antibody that specifically binds to PDGFRß; at least one tissue sample identified with an antibody that specifically binds to CD31 and an antibody that specifically binds to cytokeratin further comprises an antibody that specifically binds to collagen I; at least one tissue sample identified with an antibody that specifically binds to CD31 and an antibody that specifically binds to cytokeratin further comprises an antibody that specifically binds to collagen IV; at least one tissue sample identified with an antibody that specifically binds to CD31 and an antibody that specifically binds to cytokeratin further comprises an antibody that specifically binds to HIF2a; and at least one tissue sample identified with an antibody that specifically binds to CD31 and an antibody that specifically binds to cytokeratin further comprises an antibody that specifically binds to CAIX. 90. Tissue sample array according to any one of aspects 80 to 84, wherein the array comprises a plurality of tissue samples identified with an antibody that specifically binds to cytokeratin and an antibody that specifically binds to vimentin , wherein: at least one tissue sample identified with an antibody that specifically binds to cytokeratin and an antibody that specifically binds to vimentin further comprises an antibody that specifically binds to TEM1; at least one tissue sample identified with an antibody that specifically binds to cytokeratin and an antibody that specifically binds to vimentin further comprises an antibody that specifically binds to fibronectin-1; at least one tissue sample identified with an antibody that specifically binds to cytokeratin and an antibody that specifically binds to vimentin further comprises an antibody that specifically binds to PDGFRß; at least one tissue sample identified with an antibody that specifically binds to cytokeratin and an antibody that specifically binds to vimentin further comprises an antibody that specifically binds to collagen I; at least one tissue sample identified with an antibody that specifically binds to cytokeratin and an antibody that specifically binds to vimentin further comprises an antibody that specifically binds to collagen IV; at least one tissue sample identified with an antibody that specifically binds to cytokeratin and an antibody that specifically binds to vimentin further comprises an antibody that specifically binds to HIF2a; and at least one tissue sample identified with an antibody that specifically binds to cytokeratin and an antibody that specifically binds to vimentin further comprises an antibody that specifically binds to CAIX. 91. Tissue sample array according to aspect 80, wherein the array comprises a plurality of tissue samples, any one or more of which are identified with any of the identification combinations of aspects 85 to 90. 92. A tissue sample matrix according to aspect 80, wherein the matrix comprises: at least one tissue sample identified with an antibody specific for vimentin and is also identified with an antibody specific for TEM1; at least one tissue sample identified with an antibody specific for cytokeratin, an antibody specific for CD31 and is also identified with an antibody specific for TEM1; at least one tissue sample identified with an antibody specific for vimentin, an antibody specific for CD31 and is also identified with an antibody specific for HIF2a; at least one tissue sample identified with an antibody specific for vimentin and is also identified with an antibody specific for HIF2a; at least one tissue sample identified with an antibody specific for cytokeratin and is also identified with an antibody specific for CAIX; at least one tissue sample identified with an antibody specific for cytokeratin, an antibody specific for CD31 and is also identified with an antibody specific for CAIX; at least one tissue sample identified with an antibody specific to vimentin, an antibody specific to CD31 and is also identified with an antibody specific to PDGFRß; and at least one tissue sample identified with an antibody specific for vimentin and is also labeled with an antibody specific for fibronectin-1. 93. Tissue sample array according to aspect 80, wherein the array comprises a plurality of tissue samples labeled with one or more of: an antibody that specifically binds to TEM1, an antibody that specifically binds to HIF2a , an antibody that specifically binds to CAIX, and an antibody that specifically binds to PDGFRß. 94. The tissue sample array according to aspect 93, wherein the array comprises a plurality of tissue samples, wherein: at least one tissue sample in the array is identified with an antibody that specifically binds to TEM1, at least one tissue sample on the array is identified with an antibody that specifically binds to HIF2a, at least one tissue sample on the array is identified with an antibody that specifically binds to CAIX, and at least one tissue sample on the array is identified with an antibody that specifically binds to PDGFRß. 95. Tissue sample matrix according to aspect 94, wherein the tissue samples identified with an antibody that specifically binds to TEM1, HIF2a, CAIX or PDGFRß, further comprises any one or more of: an antibody that specifically binds to vimentin, an antibody that specifically binds to cytokeratin, or an antibody that specifically binds to CD31. 96. Tissue sample array according to aspect 93 or 94, wherein at least one tissue sample in the array is identified with a combination of any two of said antibodies. 97. Tissue sample array according to aspect 93 or 94, wherein at least one tissue sample in the array is identified with a combination of any three of said antibodies. 98. Tissue sample array according to aspect 93 or 94, wherein at least one tissue sample in the array is identified with a combination of any four of said antibodies. 99. Tissue sample array according to aspect 93 or 94, wherein at least one tissue sample in the array is identified with a combination of any five of said antibodies. 100. Tissue sample array according to aspect 93 or 94, wherein at least one tissue sample in the array is identified with a combination of any six of said antibodies. 101. Tissue sample array according to aspect 93 or 94, wherein at least one tissue sample in the array is identified with a combination of all seven said antibodies. 102. Tissue sample array according to aspect 80, wherein the array comprises a plurality of tissue samples, wherein the array is identified with any three of the following: an antibody that specifically binds to TEM1, an antibody that specifically binds to HIF2a, an antibody that specifically binds to CAIX, and an antibody that specifically binds to PDGFRß. 103. Tissue sample matrix according to aspect 102, wherein the matrix further comprises any one or more of: an antibody that specifically binds to vimentin, an antibody that specifically binds to cytokeratin, or an antibody which specifically binds to CD31. 104. Tissue sample array according to aspect 80, wherein the array comprises a plurality of tissue samples, wherein the array is identified with any two of the following: an antibody that specifically binds to TEM1, an antibody that specifically binds to HIF2a, an antibody that specifically binds to CAIX, and an antibody that specifically binds to PDGFRß. 105. Tissue sample matrix according to aspect 104, wherein the matrix further comprises any one or more of: an antibody that specifically binds to vimentin, an antibody that specifically binds to cytokeratin, or an antibody which specifically binds to CD31. 106. Tissue sample matrix, according to any one of aspects 103 to 105, further comprising one or more of: an antibody that specifically binds to fibronectin-1, an antibody that specifically binds to collagen I , and an antibody that specifically binds to collagen IV. 107. Tissue sample array according to any one of aspects 80 to 106, wherein the tumor tissue samples of the array are obtained from the same tumor. 108. Tissue sample array according to any one of aspects 80 to 106, wherein the tumor tissue samples of the array are obtained from different tumors. 109. Tissue sample matrix according to aspect 107 or 108, wherein the tumor tissue samples are histological sections. 110. Tissue sample matrix according to aspect 109, wherein the tumor tissue samples are histological sections obtained from an individual who has colorectal cancer.

[0088] Having described the invention, the same will be more readily understood by reference to the following examples, which are provided by way of illustration and are not intended to limit the invention in any way.

Examples Example 1: Primary Biomarker Panel Location

[0089] The cellular localization and expression profiles of putative biomarkers associated with TEM-1 were evaluated using histological samples. Histological samples for these studies were obtained from the Yale Colorectal Cancer Cohort, YTMA8, which consisted of tumor microarrays (TMAs) constructed at the Yale University Tissue Microarray Facility. Microarray Facility, New Haven, CT). The TMAs contained 599 primary colorectal carcinomas (CRCs) from formalin-fixed paraffin-embedded tumor samples obtained at Yale University-New Haven Hospital (New Haven, CT) from 1970 to 1981. Each tumor sample block was cut and first stained. by hematoxylin and eosin (H&E) so that areas of invasive tumor could be identified and circled. The circled region was then transcribed onto the original block from which 0.6 mm cores were taken. Each core was arrayed in receiver blocks on a grid with 1 mm spacing and 5 micron thick sections were prepared. Clinicopathologic variables available for this cohort include T stage, N stage, histologic grade, organ site, sex, and age. The univariate Cox proportional hazards model based on the five-year disease-specific survival of each clinicopathological variable for the entire cohort (all clinical information available), as well as only for cases with complete data used in the multivariate analysis in this study, can be found in Table 1.

[0090] Each putative biomarker associated with TEM-1 (TEM-1, HIF2a, CAIX, PDGFRß, fibronectin (FN), collagen I (COLI) and collagen IV (COLIV)) was quantified in all cellular and sub-cellular compartments of interest using quantitative immunohistochemistry (IHC) analysis (AQUA® technology, HistoRx, Inc.). Each TEM-1-associated biomarker was analyzed in the presence of DAPI (4',6-diamidino-2-phenylindole) to identify nuclei, fluorescently labeled cytokeratin to identify tumor cytoplasm, fluorescently labeled vimentin to identify stroma, and CD -31 fluorescently labeled to identify vasculature. The labeled histological samples were then analyzed using advanced image analysis algorithms to generate a binary image that identifies each pixel as included in or excluded from the histological compartment of interest. Figure 1A provides examples of each target and the respective compartments used in this study. Figure 1B provides an example of the marker location seen with each of the potential biomarkers tested. The amount of each biomarker of interest was determined for each histological compartment of interest and these values were translated into AQUA® scores for each biomarker. Four histological compartments of interest were evaluated: tumor (membrane/cytoplasm), tumor vasculature, pure stroma (stroma without contributing vasculature), and stromal vasculature. The mean and 95% confidence interval for AQUA® scores for each compartment were plotted to determine the histological compartment(s) with the highest expression. Based on these data and for subsequent statistical analyses, AQUA® scores from all compartments were included for TEM-1, stroma and stromal vasculature for HIF2a, tumor and tumor vasculature for CAIX, stroma and stromal vasculature for PDGFRß, stroma for FN, stroma and stromal vasculature for COLI, and stroma and stromal vasculature for COLIV.

Immunohistochemistry (IHC)

[0091] Immunofluorescence staining for AQUA® analyzes has been described in granted US Patent No. 7,219,016, and in US Patent Application Publication No. 2009/0034823, each of which is incorporated herein by reference in this request in its entirety. Briefly, pre-cut paraffin-coated tissue microarray slides were deparaffinized and antigen retrieval for the slides to be stained for TEM-1 and CAIX was done using a Decloaking chamber with a 10X DIVA buffer, pH 6.2 ( Biocare Medical DV2004MX). The slides were incubated for 15 minutes inside the chamber, where the pressurized incubation reached a maximum of 125°C, 0.10 to 0.14 MPa (15-20 PSI) for 30 seconds followed by cooling for 15 minutes to 95° W. Antigen retrieval for slides to be stained for PDGFRß, fibronectin, HIF2a, collagen I and collagen IV was done in the PT module (Labvision, Fremont, CA) with Tris EDTA buffer, pH 9. Staining was done in a LabVision equipment Autostainer (Labvision, Fremont, CA) according to previously described protocols. The slides were incubated for one hour at room temperature with mouse anti-CD-31 (DAKO, clone JC70A, lot no. 00079267) for TEM-1, CAIX, fibronectin, PDGFRß, collagen I, collagen IV, or anti-CD -31 rabbit (Abcam, lot no. GR-61759-4) for HIF2a. After incubation with CD-31 antibody, slides were rinsed in 1X TBS-Tween and then incubated with mouse Envision™ Plus (DAKO, Carpinteria, CA) or rabbit Envision™ Plus (DAKO, Carpinteria, CA ) for 30 minutes at room temperature. Slides were then rinsed with 1X TBS-Tween and incubated with biotinylated tyramide (TSA plus biotin system, Perkin Elmer) for 10 minutes at room temperature. Slides were rinsed with 1X TBS-Tween® and this was followed by two incubations of benzoic hydrazide (Sigma Aldrich) and hydrogen peroxide first for 8 min and second for 7 min. Slides were rinsed with 1X TBS-Tween® and incubated with Alexa Fluor® 750 streptavidin. This was followed by incubation with the primary mouse antibody for TEM-1 (Morphotek, clone 9G5, lot no. 26089C) or rabbit for fibronectin. -1 (Abcam, lot no. GR16465-1) or from rabbit for PDGFRß (Cell Signaling, clone 28E1, lot no. 3169S) or from rabbit for collagen I (Fitzgerald, lot no. 70R-CR007x) or from rabbit for collagen IV (Fitzgerald, lot no. 70R-CR013x) or mouse for HIF2a (Novus Biologicals, clone ep190b, lot no. J-3) or rat for CAIX (Morphotek, clone 165F3) for one hour at room temperature. The primary antibody was included in a cocktail with chicken anti-vimentin (Millipore, lot no. 2028291, 1:200). After incubation with the primary antibody, the slides were rinsed in 1X TBS-Tween and then incubated with the secondary antibody Immpress anti-mouse (Vector laboratories, TEM-1 and CAIX) or Impress anti-rabbit (Vector Laboratories, fibronectin, PDGFRß, collagen I, collagen IV) or anti-mouse ImmPRESS™ (Vector Laboratories, HIF2a) for thirty minutes at room temperature. The slides were then incubated with an anti-chicken antibody cocktail conjugated to Alexa Fluor® 555 (Invitrogen, A21437, Carlsbad, CA, 1:200) for half an hour at room temperature. After staining with Alexa Fluor®, the slides were incubated with mouse IgG blocking (Biocare Medical, NC494 H) and then with Alexa Fluor® 488 Pan Cytokeratin, clone AE1/AE3 (eBioscience, 53-9003-82) for 30 minutes at room temperature. Slides were then incubated with the Cy5 tyramide amplification system (Perkin Elmer, SAT705A, 1:50 dilution in amplification buffer, Waltham, MA) for 10 minutes, mounted with Prolong antifade solution with DAPI (Invitrogen, P36931, Carlsbad , CA) to identify nuclei within each histological point and left to dry overnight.

Imaging and image analysis

[0092] The relative concentration of protein within subcellular compartments can be measured with a high degree of accuracy using the AQUA® analysis system. Briefly, high-resolution, 12-bit digital images (resulting in 4096 distinct intensity values per pixel of a captured image) of cytokeratin with FITC, vimentin with Cy3, nuclear staining with DAPI, biomarker panel staining with Cy5 and CD-31 with Cy7 were captured and saved for each histological point on the array using the PM2000 epi-fluorescence microscopy system (HistoRx, Inc., New Haven, CT). Before statistical analysis, images were analyzed for quality (e.g., poor tissue, saturation, focus, and other artifacts) and signal intensity of DAPI and Cy3 signals. The pan-cytokeratin signal was used to create an epithelial “mask” to distinguish regions of epithelial tissue from stromal elements within normal and tumor samples. Vimentin was used to create a mask specific to the total stroma. CD-31 was used to create a mask specific to the vasculature. Using the combination of these masks, compartments of tumor vasculature, stromal vasculature, and pure stroma (total stroma without stromal vasculatures) were generated.

Statistical analysis

[0093] Unsupervised hierarchical clustering was done using the Multiple Experiment Viewer of the TM4 Microarray Software Suite (Saeed et al., Biotechniques 34(2):374-8 (2003)). Clustering was done using average linkage clustering by Person's uncentered correlation. All analyzes were conducted using SPSS v17 or later (SPSS Inc., Chicago, IL, USA). AQUA® scores for all biomarkers showed a skewed distribution and therefore log base 2 transformed scores were used for subsequent parametric analyzes (i.e., mean comparisons). However, to provide linear comparisons, AQUA® raw score data is sometimes reported. Differences in mean scores between clinical characteristics were assessed by general linear modeling based on one-way ANOVA. Optimal and Kaplan-Meier cutoff analysis for biomarkers with respect to 5-year disease-specific survival was done using X-Tile™ (Camp REF) correcting for multiple comparisons using Monte-Carlo simulations. For the X-tile™ analysis, two p-values are reported, the uncorrected p-value and the p-value corrected by Montel Carlo simulations. Although the uncorrected p-value may be significant (<0.05), the corrected p-value may not be significant, indicating the need for further validation of the finding. The Cox proportional hazards model was performed in the multivariate setting using backward elimination by Wald statistics to determine the optimal analytical model. Log likelihood chi-square ratios (LR-x2) were used to compare models. All survival analyzes were based on five-year disease-specific survival. Kaplan-Meier survival was compared using the log-rank statistic.

Example 2: Association with clinicopathological variables

[0094] Analyzes were conducted to determine the associations for AQUA® scores in the primary location compartments of each biomarker with the available clinicopathological variables of T stage, N stage, histological grade, sex, tumor site and age (data not shown) . Briefly, the findings indicated that stromal TEM-1 had significantly increased expression in well-differentiated tumors; CAIX expression in tumor vasculature was significantly higher in younger patients; PDGFRß expression in the stroma and stromal vasculature was significantly higher in men; FN expression was significantly higher in node-positive cases; stromal HIF2a expression was significantly higher in poorly differentiated tumors while HIF2a expression in the stromal vasculature was significantly higher in node-negative cases; and collagen I expression in the stromal vasculature was significantly higher in node-positive cases.

Example 3: Univariate survival analysis

[0095] Tumor-associated expression and quantification of each of the evaluated biomarkers were analyzed in view of the clinical results of the evaluated samples. Optimal Kaplan-Meier survival cut-points were defined within continuous AQUA® score data using X-Tile for optimal cut-point analysis. Five-year disease-free survival data associated with TEM-1 expression are summarized in Table 2. Table 2

[0096] The results of these analyzes showed that TEM-1, especially stromal expression, is significantly associated with better (increased) 5-year disease-specific survival in CRC. CAIX expression in the tumor vessel is significantly associated with better (increased) five-year disease-specific survival, whereas stromal expression of FN, collagen I and collagen IV is significantly associated with worse (reduced) five-year disease-specific survival. . No significant association with survival was observed for PDGFRß or HIF2a.

Example 4: TAPPS score model generation and prognostic value

[0097] The expression profiles of the selected biomarkers (TEM1 in the stroma, TEM1 in the tumor, TEM1 in the stromal vessel, TEM1 in the tumor vessel, HIF2a in the stromal vessel, HIF2a in the stroma, CAIX in the tumor vessel, CAIX in the tumor, PDGFRß in the stromal vessel, PDGFRß in the stroma, fibronectin in the stroma, collagen I in the stromal vessel, collagen I in the stroma, collagen IV in the stromal vessel, and collagen IV in the stroma) were examined together in a multivariate Cox proportional hazards model using the univariate cutoff points. In total, these 15 binarized variables (low expression versus high expression) were entered into a Cox proportional hazards model using backward elimination with an a set at 0.10 based on the Wald statistic to provide the model. excellent prognosis. The results are shown in Table 3, where p = 3.92 x 10-10 for the optimal AQUA® score and optimal model cutoff points for each marker/biological compartment included in the model are shown. Table 3. TAPPS score model for optimal prognosis in colorectal cancer.

[0098] From this model, a coefficient was derived for each marker and, therefore, with these coefficients an equation was developed, which provides a total risk score called TAPPS for the Prognostic Signature of the Route Associated with TEM-1 .

[0099] The TAPPS score equation for this cohort is as follows:

[0100] As a proof of concept, TAPPS scores for each patient in this CRC cohort were generated. The distribution of TAPPS scores in Figure 2A, which shows a near-normal distribution of TAPPS scores. The patient population was then divided into tertiles (Figure 2A). The TAPPS score ranges from -1.75 to 2.10. The TAPPS score range of the low-risk group is -1.75 to 0.00; the intermediate TAPPS score range is 0.00 to 0.91, and the high TAPPS score range is 0.91 to 2.10. Kaplan-Meier survival analysis showed, as expected, that patients with elevated TAPPS scores have a significantly reduced 5-year survival compared to patients with intermediate TAPPS scores and patients with intermediate TAPPS scores have a significantly decreased prognosis when compared to patients with low TAPPS scores (p < 0.0001, Figure 2B). Interestingly, when using the same tertile cutoffs of the TAPPS score only in node-negative patients, it was found that the elevated TAPPS score could distinguish a subset of node-negative patients with significantly diminished prognosis when compared to populations of intermediate or low TAPPS score (p < 0.0001, Figure 2C). Additionally, the same analysis only in node-positive patients indicated that patients with a low TAPPS score had significantly increased survival when compared to patients with intermediate or high TAPPS scores (p < 0.0001, Figure 2D). As shown in Table 4, a Cox proportional hazards model was used to determine the independent prognostic value of the TAPPS score as a continuous variable in all cases with complete data (left) or only node-negative patients (right). The continuous TAPPS score was 10 independent and significant in a multivariate Cox proportional hazards model adjusted for nodal status (N stage), T stage and histological grade in all patients (hazards ratio 2.7, p < 0.001) and independent of T stage and histological grade only in the 15 node-negative patients (hazard ratio of 3.7, p < 0.001). Table 4. Multivariate analysis of the TAPPS score with clinicopathological variables.

[0101] An additional model was generated, in which the expression profiles of the same set of selected biomarkers (TEM1 in stroma, TEM1 in tumor, TEM1 in stromal vessel, TEM1 in tumor vessel, HIF2a in stromal vessel, HIF2a in stroma , CAIX in tumor vessel, CAIX in tumor, PDGFRß in stromal vessel, PDGFRß in stroma, fibronectin in stroma, collagen I in stromal vessel, collagen I in stroma, collagen IV in stromal vessel, and collagen IV in stroma) were examined together in a multivariate Cox proportional hazards model using the univariate cutoffs. In total, these 15 binarized variables (low expression versus high expression) were entered into a Cox proportional hazards model using backward elimination with an a set at 0.10 based on the Wald statistic to provide a prognostic model. additional. The results are shown in Table 5. Table 5. Variables

[0102] From this model, a coefficient for each marker was derived and, therefore, with these coefficients, an alternative TAPPS equation was developed.

[0103] The alternative TAPPS equation for this cohort is as follows:

Example 5: Total Expression Analysis

[0104] TMAs originally containing a total of 599 primary colorectal carcinomas (CRCs) from formalin-fixed, paraffin-embedded tumor samples obtained at Yale University-New Haven Hospital (New Haven, CT) from 1970 to 1981 were constructed at the Tissue Microarray from Yale University (New Haven, CT) as described in detail elsewhere (REFS: Rimm TMA reviews). Of the 599 cases, 494 had biomarker data and clinical data. The cohort was divided into training (67%) and validation (33%) groups based on sequential inclusion from the date of diagnosis (2 cases for the training set, and 1 case for the validation set).

[0105] Clinical variables available for this cohort include Duke staging, histological grade, sex and age. Chi-square analysis showed no significant differences in the proportion of clinical variables between the training and validation sets (Table 6). Median disease-specific follow-up was 24 months with a median age of 68 years. The Cox proportional hazards model based on five-year disease-specific survival for each clinical variable shows expected reductions in survival for advanced stage and for men, but no significant differences were observed for histological grade and age (Table 6 ).

[0106] Since all biomarkers associated with TEM-1 (TEM-1, HIF2a, CAIX, PDGFR-ß, fibronectin (FN), collagen I and collagen IV) showed putative expression in the tumor and stromal cellular compartments, experiments were conducted to examine the total tissue expression level in the CRC cohort (Table 6). Unsupervised hierarchical clustering showed two main groups of patients (Figure 3). In cluster “gene” 1, all patients express relatively low levels of all proteins, while in cluster 2 all patient samples express relatively high levels of all proteins except CAIX. For biomarker clustering (“matrix”), endosialin/TEM-1 grouped most closely with PDGFR-ß, collagen I, and collagen IV. This was also corroborated in the bivariate correlation analysis with significant Spearman's Rho correlation coefficients (p < 0.001) of 0.62, 0.43 and 0.31, respectively. A full bivariate correlation analysis is provided in Table 7. CAIX did not cluster with any of the other biomarkers tested and, as such, is the least related. Survival outcomes were examined as a function of total expression, although all markers demonstrated an optimal cutoff in the training set that significantly (p < 0.10) predicted survival, none of these cutoffs were validated in the set. validation.

Example 6: Expression of cellular compartment-specific biomarkers

[0107] Since it is known that biomarkers of interest can have varied cellular expression, compartment-specific “masks” have been developed to specifically identify and isolate the tumor, stroma, stromal-specific vasculature, and tumor-specific vasculature. Briefly, each endosialin/TEM-1 associated biomarker was combined with DAPI to identify nuclei, cytokeratin to identify tumor cytoplasm, vimentin to identify stroma, and CD-31 to identify vasculature. Advanced image analysis algorithms were then used to generate a binary image that identifies each pixel as included or excluded from each compartment of interest. AQUA scores for each biomarker were generated for 4 biological compartments of interest: tumor (membrane/cytoplasm), tumor vessel, stroma (stroma without contributing vasculature), and stromal vessel. Media analysis showed significant differential expression by compartment (Figure 4A-G). All markers except CAIX showed significantly higher expression level in the stroma and/or vasculature compared to the tumor. Conversely, CAIX showed higher expression level in the tumor and tumor vessel compared to stromal components. Expression as a function of available clinical variables showed only minimal association (data not shown). Specifically, PDGFR-ß and collagen IV showed significantly increased expression in men compared to women; and stromal fibronectin and collagen I showed increased expression as a function of advanced stage disease.

Example 7: Univariate survival analysis

[0108] Optimal Kaplan-Meier survival cutpoints were defined within continuous AQUA score data using X-Tile for optimal cutpoint analysis on the training set (Table 6) and subsequently applied to the validation set. A biomarker was considered validated if significance (p < 0.05) was reached in both the training set and the validation set. Five-year disease-free survival data for each individual marker/compartment combination is summarized in Table 8. Although all markers showed a trend toward association with survival (training p < 0.20), TEM-1 expression in the tumor and stromal vessel as well as collagen I and HIF2a from the stromal vessel were the only biomarkers to be validated in the univariate setting. High tumor and stromal vessel expression of TEM-1 as well as stromal vessel HIF2a, associated with increased five-year disease-specific survival (HR = 0.47/p = 0.03; HR = 0.49/p = 0.03; and HR = 0.33/p = 0.02, respectively), while high expression of collagen I in the stromal vessel associated with reduced five-year survival (HR = 2.31/p = 0.008) . Kaplan-Meier survival analysis of the validation set for endosialin/TEM-1 expression in all compartments is shown in Figures 5A-D demonstrating the differential survival benefit and thus the rationale for quantifying biomarker expression. in different cellular compartments.

Example 8: R-TAPPS score generation

[0109] Because of the association of the markers described here (Figure 5) and the supposed impact they may have collectively on survival as well as the potential response to endosialin/TEM-1 targeted therapy through ontuxizumab, we combined the markers in a multivariate Cox proportional hazards model using the univariate cutoffs. The criteria for initial entry into the model were significance at level 10 of 20% in both the training set and the validation set (biomarkers highlighted in bold in Table 8; n = 13). Using backward elimination modeling based on Wald statistics on the training set, the initial 13 biomarker model was refined to 5 biomarkers (Table 9): TEM-1 in stroma, TEM-1 in tumor vessel, HIF2a in the stromal vessel, collagen IV in the tumor and FN in the stroma. This overall model was highly significant with a p-value of 9.0x10-8. Table 8

[0110] Using the model coefficients for each marker (Table 9) an equation called the “refined TAPPS” (TEM-Associated Route Prognostic Signature) score was developed, which would provide a total risk score of 5. The R-TAPPS score is defined as: R-TAPPS = (TEM-1 Stroma(0/1) * -0.89) + (TEM-1 Tumor Vessel(0/1) + 1.19) + (HIF2a Stromal Vessel(0/1) * -0.76) + (Collagen IV in Tumor(0/1) * 0.62) + (FN in Stroma(0/1) * 0.83). Table 9 10

[0111] The resulting R-TAPPS score had a range of -0.89 to 3.88 with a median score of 1.13. As a continuous variable in the training set, as expected, the R-TAPPS score was significantly associated with reduced survival [HR = 1.76 (95% CI: 15 1.44-2.15); p = 4.3x10-9; Table 3], but was then applied to the validation set with significance [HR = 1.38 (95% CI: 1.02 - 1.88); p = 0.04; Table 3] The R-TAPPS score was independent and provided significant additional prognostic value [HR = 1.66 (95% CI: 1.37-2.03); p<0.001; LR—X2 = 67.2 ] when placed in the highly significant model with known clinical variables (LR—x2 = 49.7).

[0112] In order for the R-TAPPS score to be used as a diagnosis in a clinical setting, risk groups (e.g., Oncotype Dx) must be established as a function of the continuous risk score to allow classification of patients. Therefore, the R-TAPPS score was divided into tertiles representing 3 groups, supposedly low risk (R-TAPPS < 0.3), intermediate risk (R-TAPPS from 0.3 to 1.86) and high risk (R- TAPPS > 1.86). However, when survival was examined by Kaplan-Meier, although highly significant (P < 0.0001), there was no substantial difference between the intermediate and high risk groups (Figure 6A). In this way, the intermediate and high risk groups were combined to form a high risk group (Figure 6B).

[0113] An important clinical question in the treatment of CRC is whether chemotherapy should be given to stage II patients; therefore, analyzes were conducted to examine the association between R-TAPPS risk groups in stage II patients. The resulting observations showed a highly significant association (p = 0.006) with survival (Figure 6C). We also observed a significant association in stage III/IV patients (Figure 6D).

[0114] For validation of the assay and to evaluate whether fewer markers could generate useful results, studies were conducted to reduce the number of markers in the model. Removing the vessel markers would eliminate part of the CD31 compartment (vasculature). Therefore, a model with only TEM-1 in the stroma, collagen IV in the tumor, and FN in the stroma was tested (Figure 7). Although this minimal model did not provide as much prognostic value (LR-X2 = 32.25 for the minimal versus LR-X2 = 37.2 for the full model), it was nevertheless highly significant [HR = 2.7 ( 95% CI: 1.9-3.8); p = 2.5x10-8).

Claims (6)

1. Tissue sample matrix comprising colorectal cancer tissue samples, characterized in that the matrix is identified with the antibody that specifically binds pan-cytokeratin, the antibody that specifically binds vimentin, the antibody that binds specifically TEM1, the antibody that specifically binds to fibronectin-1, the antibody that specifically binds to collagen IV, the antibody that specifically binds to CD31, and the antibody that specifically binds to HIF2a, wherein the matrix comprises: a plurality of tissue samples identified with the antibody that specifically binds to vimentin, a plurality of tissue samples identified with the antibody that specifically binds to pan-cytokeratin and the antibody that specifically binds to CD31, a plurality of samples from tissue identified with the antibody that specifically binds to vimentin and the antibody that specifically binds to CD31, and a plurality of tissue samples identified with the antibody that specifically binds to pan-cytokeratin, and wherein at least one sample of tissue identified with the antibody that specifically binds to vimentin is also identified with the antibody that specifically binds to TEM-1, at least one tissue sample identified with the antibody that specifically binds to pan-cytokeratin and the antibody that binds specifically to CD31 is also identified with the antibody that specifically binds to TEM-1; at least one tissue sample identified with the antibody that specifically binds to vimentin and the antibody that specifically binds to CD31 is also identified with the antibody that specifically binds to HIF2a, at least one tissue sample identified with the antibody that binds binds specifically to pan-cytokeratin is also identified with the antibody that specifically binds to collagen IV, and at least one tissue sample identified with the antibody that specifically binds to vimentin is also identified with the antibody that specifically binds to fibronectin- 1. 2. Tissue sample matrix according to claim 1, characterized by the fact that one or more of the tissue samples is identified with a nuclear marker. 3. Tissue sample matrix according to claim 2, characterized by the fact that the nuclear marker is DAPI. 4. Tissue sample matrix according to claim 1 or 2, characterized in that the tumor tissue samples of the matrix are obtained from the same tumor or from different tumors. 5. Tissue sample matrix according to claim 4, characterized by the fact that the tumor tissue samples are histological sections. 6. Method for identifying a tissue sample array, comprising: obtaining a tissue sample array that has a plurality of colorectal cancer tissue samples from an individual, contacting a plurality of the samples of colorectal cancer tissue with an antibody that specifically binds to vimentin, contact a plurality of the colorectal cancer tissue samples with an antibody that specifically binds to pan-cytokeratin and a second antibody that specifically binds to CD31, place contacting a plurality of the colorectal cancer tissue samples with an antibody that specifically binds to vimentin and a second antibody that specifically binds to CD31, contacting a plurality of the tissue samples with an antibody that specifically binds to pan- cytokeratin, wherein at least one colorectal cancer tissue sample placed in contact with an antibody that specifically binds to vimentin is also placed in contact with an antibody that specifically binds to TEM-1, at least one cancer tissue sample colorectal tissue placed in contact with an antibody that specifically binds to pan-cytokeratin and an antibody that specifically binds to CD31 is also placed in contact with an antibody that specifically binds to TEM-1, at least one cancer tissue sample At least one colorectal cancer tissue sample placed in contact with an antibody that specifically binds to vimentin and an antibody that specifically binds to CD31 is also placed in contact with an antibody that specifically binds to HIF2a, at least one colorectal cancer tissue sample placed in contact with an antibody that specifically binds to pan-cytokeratin is also contacted with an antibody that specifically binds to collagen IV, and at least one colorectal cancer tissue sample placed in contact with an antibody that specifically binds to vimentin is also contacted. placed in contact with an antibody that specifically binds to fibronectin-1, where the colorectal cancer tissue sample matrix is then assessed for the degree of labeling.