METHODS AND SYSTEMS UTILIZING COLONIC TISSUE TOPOGRAPHY AS A

DIAGNOSTIC MARKER

STATEMENT REGARDING FEDERALLY SPONSORED DEVELOPMENT

Development for this invention was supported in part by Contract No. 1 RO1

CA1 13975-A2, awarded by the National Institutes of Health . Accordingly, the United States Government may have certain rights in this invention.

CROSS-REFERENCE TO RELATED APPLICATIONS

This application is related to U.S. Provisional application no. 61 /782,419 to which priority is claimed under 35 USC 1 1 9. The entire disclosure of this provisional is incorporated herein by reference.

FIELD OF THE INVENTION

The present invention relates to systems, methods, and computer-readable media for classifying colonic tissue as normal or diseased, and if diseased a state of the disease along a grading scale.

BACKGROUND OF THE INVENTION

Colon cancer remains one of the leading causes of mortality worldwide.

According to American Cancer Society 103,1 70 new cases of colon cancer 40,290 new cases of rectal cancer are expected in the U.S. this year ¹ ' ² . One of the major concerns surrounding cancer today is the lack of early diagnosis. Colon cancer generally is diagnosed during a colonoscopy exam where a physician looks for any abnormalities in the tissue which include polyps and other signs of inflammation within the walls of the colon. A further histological exam is also conducted to confirm the diagnosis. The current diagnostic methods highly depend on extracting tissue samples from patients to confirm cancer diagnosis or to identify any dysplasia associated with colonic disorders. Current guidelines for dysplasia surveillance recommend a minimum of 33 biopsies to get a 90% predictive value and a minimum of 56 biopsy samples to get a predictive value of 95% ³ ' ⁴ . Although histological exams have proved to be effective, the disadvantages of this method far outweigh the benefits. This method of evaluation creates a significant delay in diagnosis, introduces the possibility of sampling error, and adds to the risk and cost of the procedure ^5"8 . Due to the limitations of current endoscopy-histology technique, several new ways of diagnosis have been pursued by researchers; one such way is optical biopsy.

Optical biopsies have an advantage over current biopsy techniques in that these do not require physical biopsy samples and provide an instantaneous way to visualize and evaluate tissue in-vivo (9-1 1 ). Optical biopsies utilize imaging techniques to determine differences in cancerous and surrounding normal tissue based on several different factors such as tissue penetration depth, spatial resolution, and image contrast. However, these factors provide insufficient data to diagnose cancer with a certainty that is usually found in histological examination of ex-vivo tissue. To overcome this limitation, several researchers have shifted their focus onto examining the extra cellular matrix (ECM) or tissue property as an alternative to histological analysis of the cells. Several studies have shown that ECM alone plays an important role in altering cellular behavior such as differentiation, proliferation as well as apoptosis (1 2, 13). Thus, a closer look at colonic tissue may give further insight on an existing medical condition.

One of the major topics of tissue mechanics that currently holds researchers' interest is tissue topography. Several studies have focused on determining the effect of topography on cellular function. Certain in-vivo studies show the significance of tissue specific surface topography in determining the differentiation lineage of stem cells (14- 16) as well as their expansion (17). Tissue specific topography has been highlighted in many different studies as being unique to each organ and/or disease condition (18-20) which may aid in predicting or diagnosing an existing condition. For instance, researchers have studied fat tissue topography extensively and have established that adipose tissue topography provides insight to the pathogenesis of several metabolic diseases (21 ). Similarly, colon tissue in-vivo is altered in the event of colon cancer and other Gl disorders such as Inflammatory Bowel Disease (IBD). These alterations in tissue are what the physician primarily looks for in a colonoscopy procedure.

Optical biopsy, although an attractive alternative to current endoscopic technique, has its own disadvantages. Although a physician may visualize colonic abnormalities instantaneously with this technique, it is essential to categorize and classify these abnormalities. Classification of these abnormalities not only gives the certainty that is required for a diagnosis but may also help in predicting a disorder. As previously mentioned, tissue topography is unique to each organ and/or disease condition. Isolation of specific and unique physical features of colonic tissue seen in normal and tumor state is a key component in making optical biopsy technique a clinically usable modality. In this study, we proposed to isolate specific topographical features, both micron as well as nano-scaled protrusions that are commonly found in normal and tumorigenic milieu. We believe that these topographical features may serve as 'signatures' since each condition, normal and tumor, has a unique set of mechanical features.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic representation of a hyper-plane created from a Support Vector Machine (SVM).

FIG. 2 shows SEM images of normal colonic tissue clearly exhibit crypts and valleys compared to tumorigenic tissue. Scanning Electron Images of ex-vivo tissue were obtained as described in the methods section. Normal colonic tissue (left) had clear crypts and valleys (indicated by arrows), whereas in tumorigenic colonic tissue samples the crypts and valleys were not clearly seen. In tumorigenic colonic tissue an increase in porosity, and a significant increase in the roughness of the ECM was also observed. Magnification = 300x; Scale bar = 100 μηι.

FIG. 3 shows normal colonic tissue exhibits smoother topography compared to tumorigenic colonic tissue. Digitally Elevated Models (DEMs) were obtained as described in methods section. DEM of normal colonic tissue (Left Panel) had a smoother appearance than tumorigenic (Right Panel). Micron and nano-scaled protrusions were more prominent in tumorigenic colonic tissue. Scale bar = 50 μηι.

FIG. 4 shows tumorigenic colonic tissue is more porous than normal colonic tissue. 2D stereopair SEM images were analyzed using MeX and surface area and volume were determined for each tissue, normal and tumorigenic. Approximately 2-fold increase in surface area to volume ratio was noted in tumorigenic colonic tissue compared to normal. % porosity of normal tissue was measured to be 26% and more than 2-fold increase in porosity was observed in tumorigenic colonic tissue (Table 2).

Scale bar = 50 μηι.

FIG. 5 shows that there was a significant difference between normal and tumorigenic colonic tissue in roughness parameters especially, Rz, Rp and Rv.

Roughness parameters were obtained from ex-vivo colonic tissue as described in methods section. The overall height (Rz) and the peak height (Rp) of the protrusions was 2-fold higher in tumorigenic tissue when compared to the normal colonic tissue.

Student t-test was performed to test for statistical significance; p<0.05 was considered significant (indicated by ^* ).

FIG. 6 shows binary classification of normal and tumorigenic tissue using

Support Vector Machine (SVM). Roughness parameters, specifically Rz and Rp, were used to classify normal and tumorigenic tissue samples using SVM. Rz and Rp were linearly separable with a margin width of 10 nm. Misclassifications were noted to be

<10% for both normal and tumorigenic colonic tissue.

FIG. 7 shows a decision tree algorithm for classifying colonic tissue as normal or diseased in accordance with an aspect of the present invention.

FIG. 8 shows the roughness parameters and threshold values used in the decision tree algorithm of FIG. 7. DETAILED DESCRIPTION

The present invention provides novel systems, apparatuses, and methods for classifying colonic tissue as normal or diseased (e.g., carcinogenic). Advantageously, aspects of the present invention substantially improve diagnostic capabilities over known physical biopsy techniques and provide a classification framework for optical biopsy techniques that substantially improves classification accuracy. In this way, aspects of the present invention reduce the occurrence of unnecessary surgical procedures and improve therapies for colonic diseases, namely colon cancer and inflammatory bowel disease, through improved diagnostics.

The two major forms of Inflammatory Bowel Disease (IBD) are ulcerative colitis (UC) and Crohn's disease (CD). IBD is a chronic and remitting disease causing inflammation of the intestinal diseases. UC and CD have symptoms and pathologies in common, but they differ in the severity and location of the inflammation along the intestinal tract. Inflammation in UC patients is limited to the mucosal layer, and involves only the rectum and colon, while inflammation in CD patients penetrates the entire wall of the intestine and can occur anywhere along the intestinal tract. A clear diagnosis of the type of IBD is crucial to treatment decisions.

UC typically is characterized by ulcers in the colon and chronic diarrhea mixed with blood, weight loss, blood on rectal examination, and occasionally abdominal pain. UC patients may also present with a variety of other symptoms and extraintestinal manifestations including but not limited to anemia, weight loss, iritis, seronegative arthritis, ankylosing spondylitis, sacroiliitis, erythema nodosum, and pyoderma gangrenosum. Toxic megacolon is a life threatening complication of UC and requires urgent surgical intervention. UC usually requires treatment to go into remission. UC therapy includes anti-inflammatories, immunosuppressants, steroids, and colectomy (partial or total removal of the large bowel, which is considered curative). There is a significantly increased risk of colorectal cancer in UC patients several years

after diagnosis, if involvement is beyond the splenic flexure, and a significant risk of primary sclerosing cholangitis, a progressiveinflammatory disorder of the bile ducts. Crohn's disease (CD) is also an IBD feat can affect the colon with symptoms similar to UC. Unlike UC, CD may affect any part of the gastrointestinal tract, and the

inflammation penetrates deeper into the layers of the intestinal tact. Patients with CD may have symptoms and intestinal complications including abdominal pain, diarrhea, occult blood, vomiting, weight loss, anemia, fecal incontinence, intestinal obstructions, perianal disease, fistulae, and strictures, and apthous ulcers of the mouth.

Extraintestinal complications include skin rashes, arthritis, uveitis, seronegative arthritis, peripheral neuropathy, episcleritis, fatigue, depression, erythema nodosum, pyoderma gangrenosum, growth failure in children, headache, seizures, and lack of concentration. The risk of small intestine malignancy is increased in CD patients. CD is believed to be an autoimmune disease, while it is uncertain whether there is an autoimmune component to UC. In one aspect, there is provided a prediction model for the classification of colonic tissue with an in vivo imaging modality, such as Laser Confocal Endomicroscopy, to form an in vivo imaging system that can be used during regular colonoscopy examination.

In another aspect, there is provided a method to diagnose the state of the colonic tissue using a set of roughness parameters based upon a three-dimensional (3D) image of the colon. For example, the present invention may provide the necessary parameters, e.g., roughness parameters or topographical features, to form a

mathematical prediction module that may output a 3D view of portions of the colon, as well as classify colonic tissue on a grading continuum between normal and cancerous tissue.

In another aspect, there is provided a method for classifying colonic tissue as normal or diseased. The method comprises obtaining an image of the colonic tissue. The method further comprises processing the image to provide two or more roughness parameters based upon a three-dimensional model of the surface of the colonic tissue. In addition, the method includes classifying the tissue as normal or diseased based upon a comparison of the two or more roughness parameters with a corresponding two or more predetermined threshold values.

In another aspect, there is provided a method for classifying colonic tissue as normal or diseased. The method comprises obtaining three-dimensional image data that describes an image of colonic tissue. The image data comprises two or more roughness parameters characterizing the surface of the colonic tissue. In addition, the method includes classifying the colonic tissue as normal or diseased based upon a comparison of the two or more roughness parameters with a corresponding two or more predetermined threshold values.

In another aspect, there is provided an imaging system for classifying colonic tissue as normal or diseased in a subject. The imaging system comprises an imaging modality configured to obtain an image of the colonic tissue of the subject. An image processor is operably connected to the imaging modality. The imaging processor is programmed to: process the image to provide two or more roughness parameters of the surface of the colonic tissue; and classify the tissue as normal or diseased based upon a comparison of the two or more roughness parameters with a corresponding two or more predetermined threshold values. In another aspect, there is a computer system for classifying colonic tissue as normal or diseased in a subject. The computer system comprises a processor and memory coupled to the processor and having stored therein instructions that, if executed by the computer system, cause said computer system to execute a method comprising: processing an image of the colonic tissue to provide two or more roughness parameters of the surface of the colonic tissue; and classifying the tissue as disease as normal or diseased based upon a comparison of the two or more roughness parameters with a corresponding two or more predetermined threshold values.

In another aspect, a computer-readable medium, in which a computer program for classifying colonic tissue as normal or diseased in a subject is stored which, when executed by a processor, causes the processor to carry out the steps of: processing an image of the colonic tissue to provide two or more roughness parameters of the surface of the colonic tissue; and classifying the tissue as disease as normal or diseased based upon a comparison of the two or more roughness parameters with a corresponding two or more predetermined threshold values.

1 .1 Definitions

As used herein, the term "diseased tissue" refers to any tissue that differs from normal tissue in any physical form, and includes tissue that differs due to a disease, a disorder, a medical condition or other abnormal physical state. In specificembodiments, diseased tissue involves colonic tissue possessing characteristics ofcancer, dysplasia, or IBD (e.g. ulcerative colitis or Crohn's disease). Dysplasia is a term that describes how much colonic tissue (e.g polyp) has characteristics of cancerous tissue. Specific characteristics of colonic diseased tissue include but are not limited to severe crypt architectural distortion, severe widespread decreased crypt density, frankly villous surface, and distorted dialated or branching crypts.

As used herein, term "subject" refers to any animal (e.g., a mammal), including, but not limited to, humans, which may be the recipient of a particular treatment. The term is intended to include living organisms susceptible to conditions or diseases caused or contributed to by unrestrained cell proliferation and/or differentiation.

As used herein, the term "Rv" refers to a Maximum Profile Valley Depth;

As used herein, the term "Rp" refers to a Maximum Profile Peak Height; As used herein, the term "Rz" refers to Average Maximum Height of the Profile; As used herein, the term "S" refers to a Mean spacing of Local Peaks of the

Profile.

As used herein, the term "Sm" refers to a Mean Spacing of Profile Irregularities; As used herein, the term "D" refers to a Peak Profile Density.

As used herein, the term "Pc" refers to a Peak Count.

1 .2 Imaging Modality

In one aspect, there is provided a prediction model for the classification of colonic tissue with an in vivo imaging modality, such as Laser Confocal

Endomicroscopy, to form an in vivo imaging system that can be used during regular colonoscopy examination.

In another aspect, there is provided a method to diagnose the state of the colonic tissue using a set of roughness parameters based upon a three-dimensional (3D) image of the colon. For example, the present invention may provide the necessary parameters, e.g., roughness parameters or topographical features, to form a

mathematical prediction module that may output a 3D view of portions of the colon, as well as classify colonic tissue on a grading continuum between normal and diseased tissue.

In another aspect, there is provided a method for classifying colonic tissue as normal or diseased. The method comprises obtaining an image of the colonic tissue. The method comprises processing the image to provide two or more roughness parameters based upon a three-dimensional model of the surface of the colonic tissue. In addition, the method includes classifying the tissue as disease as normal or diseased based upon a comparison of the two or more roughness parameters with a

corresponding two or more predetermined threshold values.

1 .3 Computer system

Aspects of the present invention may be executed on a processing module, which may include a single processor or a plurality of processors. Such a processor or processing device may be a microprocessor, micro-controller, digital signal processor, microcomputer, central processing unit, field programmable gate array, programmable logic device, state machine, logic circuitry, analog circuitry, digital circuitry, and/or any device that manipulates signals (analog and/or digital) based on operational instructions. The processing module may have operationally coupled thereto, or integrated therewith, a memory device. The memory device may be a single memory device or a plurality of memory devices. Such a memory device may be a read-only memory, random access memory, volatile memory, non-volatile memory, static memory, dynamic memory, flash memory, and/or any device that stores digital information. A computer, as used herein, is a device that comprises at least one processing module.

As will be appreciated by one of skill in the art, embodiments of the present invention may be embodied as a device, method, or system comprising a processing module, and/or computer program product comprising at least one program code module. Accordingly, the present invention may take the form of an entirely hardware embodiment or an embodiment combining software and hardware aspects.

Furthermore, the present invention may include a computer program product on a computer-usable storage medium having computer-usable program code means embodied in the medium. Any suitable computer-readable medium may be utilized including hard disks, CD-ROMs, DVDs, optical storage devices, or magnetic storage devices.

The computer-usable or computer-readable medium may be or include, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, device, or propagation medium. More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, and a portable compact disc read-only memory (CD-ROM), a CD ROM, a DVD (digital video disk), or other electronic storage medium. Note that the computer-usable or computer- readable medium could even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via, for instance, optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner if necessary, and then stored in a computer memory. Computer program code for carrying out operations of certain embodiments of the present invention may be written in an object oriented and/or conventional procedural programming languages including, but not limited to, Java, Smalltalk, Perl, Python, Ruby, Lisp, PHP, "C", FORTRAN, or C++. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer. In the latter scenario, the remote computer may be connected to the user's computer through a local area network (LAN) or a wide area network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet Service Provider).

Certain embodiments of the present invention are described herein with reference to descriptions, flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and/or block diagrams, as well as combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer-readable program code modules. These program code modules may be provided to a processing module of a general purpose computer, special purpose computer, embedded processor or other programmable data processing apparatus to produce a machine, such that the program code modules, which execute via the processing module of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart and/or block diagram block or blocks.

These computer program code modules may also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the program code modules stored in the computer-readable memory produce an article of manufacture. The computer program code modules may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart and/or block diagram block or blocks. 1 .4 Obtaining Image and Image Data

Aspects of the present invention aim to first obtain an image or data representing an image of the colonic tissue, e.g., colonic mucosa of a subject. Once obtained, the image can ultimately be processed by the processor to obtain two or more

topographical parameters characterizing the topography of the colonic tissue. These parameters will be discussed in detail below.

In one embodiment, the image data is obtained from an imaging modality. In one embodiment, the imaging modality is a scanning electron microscope (SEM) as is well- known in the art. An exemplary scanning electron microscope for use in the invention is the Hitachi S-3000N scanning electron microscope available from Hitachi, Ltd. An SEM is utilized for ex vivo tissue analysis, such as with tissue obtained from physical biopsies from a subject. In another embodiment, the imaging modality comprises a confocal microscope. A confocal microscope typically utilizes a laser to provide excitation light to a sample and generates a two dimensional image from the emitted light from the sample on a point by point basis. As explained in Goetz, M., Kiesslich, R., Anticancer Research 28: 353-360 (2008), confocal microscopy is an adaptation of microscopy in which white light passes through a system of consecutive pinholes before an image is detected. Typically, a fluorescent dye in added to the medium being viewed or administered to the subject for in vivo imaging applications in order to obtain an image. The confocal microscope may be a Confocal Laser Scanning Microscope, a Spinning- Disk Confocal Microscope, or a Programmable Array Microscope (PAM) as is known in the art. The image data collected may be utilized for the subsequent reconstruction of a two- or three-dimensional image, and preferably a three-dimensional image.

In a preferred embodiment, an imaging modality, such as a confocal microscope described above, is integrated into a tip of a typical colonoscope as is known in the art so that the image can be captured in real time in vivo. When utilized with a subject in typical colonoscopy procedure, the captured image data can immediately be processed and classified to characterize the colonic tissue from normal to cancerous along the tissue characterization system described herein.

1 .5 Image Processing Once obtained, the image data can be processed into a two- or three- dimensional image as is desired. In certain embodiments, a three-dimensional image is reconstructed from the image data such that the topographical features described below can be obtained from the image. Primarily, the topographical parameters are physical (measureable) parameters that may be automatically determined by the processing module after reconstruction of an image from the image data. In one embodiment, the obtained image data from the imaging modality is reconstructed into a three- dimensional data set from which the topographical parameters can be obtained. For example, in one embodiment, imaging analysis software, such as MeX image analysis software, available from Alicona Imaging, Graz, Austria. MeX software works by automatically retrieving 3D information and presenting a highly accurate, robust and dense 3D dataset, which then may be utilized to perform traceable metrology examination. Briefly, MeX image analysis software converts 2D stereopair images obtained via SEM imaging into 3D Digitally Elevated Models (DEMs) by superimposing images obtained at two different angles, 0° and 10° respectively. DEMs provide the depth information (Z-axis) required to generate a roughness profile for each

superimposed image. A working distance, which describes the distance between the tip of the lens of a microscope and the surface of the tissue sample being observed, of 350 mm with an objective of 300X is utilized to obtain a field of view (FOV) of 300 μηι during SEM imaging. MeX image analysis software deconstructs the superimposed image to generate roughness profile comprising two or more roughness parameters.

1 .6 Roughness parameters

The roughness parameters may be determined from the reconstructed image data by any processing method as is known in the art, including the MeX software described above. In one embodiment, the present inventors have innovatively development a classification system based upon the relationship between selected roughness parameters for each image. In one embodiment, the roughness parameters comprise at least two of the following parameters: Rv (Maximum Profile Valley Depth), Rp (Maximum Profile Peak Height), Rz (Average Maximum Height of the Profile), S (Mean spacing of Local Peaks of the Profile), Sm (Mean Spacing of Profile Irregularities), D (Peak Profile Density), and Pc (Peak Count). In a preferred embodiment, all the parameters are determined for a particular colonic tissue sample.

Once determined for the associated colonic tissue sample, the determined roughness parameters may be compared to corresponding predetermined threshold values. In one embodiment, the threshold parameters may be identified from the analysis of normal tissue samples, and in certain embodiments, from diseased tissue samples as well. For example, a value or range of values that correspond to a normal tissue sample for each rough parameter may be identified by determining the value or range of values that are statistically different from normal and diseased tissue.

As shown in FIG. 7, the present inventors have identified a logic tree developed by determining which parameters are typically different in diseased vs. normal tissue, whether the parameters are greater than or less than a particular threshold value in a diseased state, and which parameters indicate a more significant or advanced disease state along a continuum between a normal and a disease state.

In one embodiment, the programming module is configured to compare roughness parameters captured for a particular colonic tissue to threshold parameters to determine whether the tissue is normal or indicates a disease state. In a particular embodiment, the disease state is characterized by two or more grades with one grade being indicative of more advanced state of disease of the colonic tissue.

Referring to FIG. 1 , there is shown an exemplary logic tree or algorithm for use in classifying the colonic tissue sample to be carried out by the processing module. In this embodiment, the threshold values include values T1 -T7, which correlate to roughness parameters Rv, Rp, Rz, S, Sm, D, and Pc, respectively. Exemplary values for each roughness parameter are shown in FIG. 8, for example (e.g., T1 has a threshold maximum profile valley depth value of 700 nm). The values shown were calibrated for analysis of a system utilizing a SEM microscope and MeX software, but it is understood the present invention is not so limited and that threshold values may be modified for the particular system being used by characterizing a sufficient number of known and diseased samples.

As shown in FIG. 7, tissue may be classified as normal when Rp < a

predetermined threshold value T1 ; Rv < a predetermined threshold value T2; Rz < a predetermined threshold value T3; S > a predetermined threshold value T4; Sm > a predetermined threshold value T5; Pc < a predetermined threshold value T6; and D < a predetermined threshold value T7.

Second, the colonic tissue may be classified as grade one diseased tissue when Rp > T1 and Rv > T2.

Third, the colonic tissue may be classified as grade two diseased tissue when Rp

> T1 and Rz > T3.

Fourth, the colonic tissue may be classified as grade three diseased tissue when Rz > T3 and Pc > T6.

Fifth, the colonic tissue may be classified as grade four diseased tissue when Rz > T3 and D > 17.

Sixth, the colonic tissue may be classified as grade five diseased tissue when Rz > T3; S < T4 and Sm < T5

Seventh, the colonic tissue may be classified as grade six diseased tissue when the colonic tissue has been classified as each of grade one, grade two, and grade three diseased tissue.

Eighth, the colonic tissue may be classified as grade seven diseased when the colonic tissue has been classified as each of grade three, grade two, and grade three diseased tissue.

Ninth, the colonic tissue may be classified as grade eight diseased tissue when the colonic tissue has been classified as each of grade six and grade seven diseased tissue.

By carrying out the described decision tree for a given colonic tissue sample, the tissue can be accurately classified as normal or diseased, and a state of the disease can be provided. It is appreciated that the present invention is not limited to the above parameters nor are all parameters required to be utilized. In certain embodiments, fewer parameters and grades may be utilized. When a processor for carrying out the logic tree is incorporated into an in vivo imaging modality, accurate assessment of the tissue can be provided in real time

1 .7 Examples

The following example(s) are intended for the purpose of illustration of the present invention. However, the scope of the present invention should be defined as the claims appended hereto, and the following example(s) should not be construed in any way limiting the scope of the present invention.

METHODS AND MATERIALS

1 .8 Reagents and Supplies

15 normal colonic tissue samples and 15 tumorigenic colonic tissue samples were obtained from human cooperative tissue network. These tissue samples were kept frozen at -80 °C until ready to be imaged. 1 .9 Scanning Electron Microscopy

Frozen tissue samples were thawed by immersing in sterile phosphate buffered saline (PBS) at 37 °C for 30 min, followed by incubating in Hank's balanced salt solution (HBSS) for 10 min. The epitheiial surface was gently stroked with a 13-mm glass cover- slip. The tissue was then incubated in 10 mM dithiothreitol in PBS at RT for 30 min. This was followed by a second incubation with HBSS followed by gentle scraping with a glass cover-slip. The tissue was then subjected to two rounds of incubation with 1 0 mM EDTA in PBS for 30 min at 37°C followed by HBSS for 30 min. The tissue was finally fixed with Trump's fixative (80 mM sodium monobasic phosphate, 67.5 mM sodium hydroxide, 3.75% formaldehyde, and 1 % giutaraldehyde in water) for at least 1 hour. Samples were then washed with distilled water and prepared for SEM by alcohol dehydration. Samples were finally dried using hexamethyidisiiazane (HMDS). Uncoated specimens were glued onto metal stubs with carbon-coated tabs. SEM images were obtained using a Hitachi S-3000N scanning electron microscope at 20 kV in variable pressure mode at 10 Pa.

1 .1 0 Surface Reconstruction

To three-dimensional!y reconstruct the ECM topography, stereopair images were obtained at 0° and +10° from the horizontal and input into the MeX image analysis software program (Alicona Imaging, Graz, Austria). The tilt angle along with the working distance of 350 mm and the size of image pixels were also input for proper calibration. Reconstruction was then performed in two steps. First, corresponding points were extracted from the stereoscopic images. Secondly, metrically correct 3D points were calculated using the geometric relationships from the SEM identified in the first step as described previously (22).

1 .1 1 Roughness Analysis

There are many different ways to characterize surfaces and to compare them to each other, visual comparison being the most common. In addition, roughness parameters that have previously been identified as important to regulating cell motility, adhesion, and morphology were quantified for each tissue type (23-27). These roughness parameters were directly calculated using the MeX software and were derived from surface profiles extracted from the reconstructed 3-D models.

The extraction of roughness parameters from surface profiles (26, 27) is based on the decomposition of a primary profile or primary curve into a roughness profile that contains the high frequency information and a waviness profile that contains the low frequency information. The surface being analyzed can be broken up into specific surface texture components, such as hills, valleys, or bumps. The surface bumps correspond to finer irregularities of the surface texture, which represent roughness. The hills and valleys represent irregularities that are more spaced out and correspond to waviness. All these components put together make up what is known as the primary profile curve. Each of the parameters used to form a primary curve are listed in Table 1 below.

Table 1 : Roughness Parameters

Parameter Description

RL Ratio of True Profile Length to Projected Profile Length

Ra Roughness Average

Rq Root Mean Square (RMS) Roughness

Rz Average Maximum Height of Profile

Rp Maximum Profile Peak Height

Rv Maximum profile Valley Depth

Rsm Mean Spacing of Profile Irregularities of Roughness Profile

Rsk Skewness of Roughness Profile Kurtosis of Roughness Profile

Root- Mean-Square Slope of Roughness Profile

Filter Wavelength for Roughness Profile

1 .1 2 Porosity Analysis

Porosity was determined by analyzing the DEMs constructed previously using MeX software. Porosity is an important factor in determining extent of tissue

degradation. Since, the SEM images revealed drastic differences in the tissue topography of colonic normal and tumorigenic tissue, it was imperative to test for porosity. A higher degree of porosity has been shown to increase cellular invasion and formation of invadopodia (28). A general method to determine porosity is to take the ratio of the overall volume (V _v ) of the substrate and the volume of the void space (V _T ) within the substrate as shown in equation 1 .

Porosity = (Equation 1 )

v _T

The void space (V _v ) can be determined by subtracting the volume above (V _A ) and volume below (V _B ) the surface of the tissue sample. The quantities, V _A and V _B, were measured from the DEMs using the MeX software.

V _V = V _A - V _B (Equation 2)

1 .1 3 Support Vector Machine (SVM)

Using the roughness parameters listed in Table 1 , SVMs were employed to classify normal and tumorigenic colonic tissue. SVM is a binary classifier algorithm that creates a hyperplane (equation 1 ) to separate positive and negative training samples and maximizes the distance between the samples and the hyperplane.

f(x) = w - x + b (Equation 3)

The positive training sample in our case corresponds to tumorigenic ex-vivo tissue and negative corresponds to normal colonic tissue. Briefly, SVM algorithm takes the roughness profile of each sample and assigns it a value of either +1 (tumor) and -1 (normal) and plots it across a hyperplane as described in equation 3. The distance A and B in Figure 1 correspond to the minimum confidence required to classify the data point.

Bioinformatics toolpack in MATLAB was used to train and test the data set using SVM algorithm. Briefly, training data set contained roughness parameters obtained from 6 tumorigenic and 6 normal colonic tissue samples. 8 normal and 8 tumorigenic tissue samples were then used as test set to classify based on roughness parameters. Each tissue sample was 'scanned' using MeX software for topographical features at least 3 times, to obtain 48 data points representing the roughness parameters.

1 .14 Statistical analysis

Student t-test was performed to determine significant changes within the samples. A p-value of less than 0.05 was considered significant. RESULTS

1 .15 Colonic tissue is significantly compromised in the event of colon cancer

Ex-vivo colonic samples were imaged using Scanning Electron Microscopy (SEM) as described in the methods section. Normal colonic tissue samples showed well defined and structured colonic crypts; whereas, in tumorigenic tissue, colonic crypts were absent. Normal colonic tissue also had a globular appearance, indicative of white blood cells (Figure 2).

SEM images were further analyzed by transforming these into 3D images to obtain depth information as described in methods section. DEMs of normal colonic tissue were more planar (Figure 3, left panel) in appearance when compared to the DEMs of tumorigenic colonic tissue. All three tumorigenic tissue samples (Figure 3, right panel) showed micro-scaled peak like structures. To quantify these structures, depth images were generated for each tissue sample. The depth images of normal and tumorigenic tissue samples clearly show the irregularities. More than 2-fold increase in the overall height of the peaks was observed in tumorigenic colonic tissue when compared to the normal tissue. Tumorigenic colonic tissue had micro-scaled features that ranged from approximately 80 μηι in height to 160 Dm in depth, with an overall average height of approximately 220 μηι. Features observed in normal colonic tissue ranged from 80 μηι in height to a depth of 20 μηι and the average overall height of the features was approximately 100 μηι (Figure 4).

Total surface area (SA) and volume (V) of each tissue sample were evaluated and the ratio, SA/V, was determined. SA/V ratio for tumorigenic colonic tissue, 0.388 μιη ^"1 ' was approximately 2-fold higher than normal colonic tissue. Similarly, a 2-fold increase in % porosity of tumorigenic colonic tissue was also observed (Table 2). The average porosity was noted to be 26% for normal colonic tissue and 60% for tumorigenic colonic tissue. 1 .16 Roughness profile of tumorigenic colonic tissue is twice as normal colonic

tissue

Since normal and tumorigenic colonic tissue topography was drastically different at a micro-scale level, it is important to analyze these structures at a nano-scale.

Roughness profile was generated as described in the methods section to quantify nano- scaled topographical features found in both normal as well as tumorigenic colonic tissue.

There was a 1 .5-fold increase in the average roughness profile (Ra) of tumorigenic tissue when compared to normal colonic tissue sample. A significant increase was noted in Rz, Rp and Rv parameters. Rz, which represents the overall height of the nano-scaled feature, was approximately 2-fold higher in tumorigenic colonic tissue. Normal colonic tissue had nano-scaled protrusions that were

approximately 2000 nm in height. Whereas, tumorigenic colonic tissue samples had nano-scaled structures that were approximately 4400 nm in height. Rp and Rv, which represent the peak and valley depth respectively, were also 2-fold higher in magnitude (Figure 5). Rsk, the skewness of roughness profile, indicates that normal colonic tissue distribution is left skewed, whereas, for tumorigenic tissue, the distribution is right skewed. This directional skewness indicates that normal colonic tissue, with a negative skewness, has fewer values that are high in magnitude. In contrast, tumorigenic colonic tissue has a positive skewness indicating that this distribution has more values that are higher in magnitude than the normal colonic tissue. There was no significant change in the root mean square of the slope of roughness profile (Rdq) for both normal and tumorigenic colonic tissue. 1 .1 7 Roughness parameters are linearly separable

Roughness parameters obtained as described previously were further analyzed using Bioinformatics Tool-pack in MATLAB. Parameters such maximum profile peak height (Rp) and average maximum height of roughness profile (Rz) were considered mainly because these parameters provided maximum change in magnitude between normal and tumorigenic colonic tissue (Figure 5). A Support Vector Machine (SVM) plot was created focusing mainly on Rz and Rp. Rz for the samples tested using SVM ranged from 1000 to 4000 nm. Rp for normal colonic tissue ranged from 500 to 2000 nm. The parameters were drastically different for tumorigenic tissue samples. Rz for tumorigenic tissue samples ranged from 3500 nm to 7000 nm and Rp ranged from 1500 nm to 4000 nm. A clear distinction between normal and tumorigenic tissue samples is seen in the SVM plot (Figure 6). The margin, which measures the minimum distance from the hyper-plane and the data point, was approximately 10 nm. Rp and Rz values were obtained for 48 trials and a hyper-plane was generated using SVM.

Misclassification of data points was observed although this was not very significant (<5% for normal and <10% for tumor). Only one tumorigenic colonic tissue sample was classified in the normal space and three normal samples were classified in the tumor space.

1 .18 Discussion

Tissue topography is an essential parameter in diagnosing and recognizing dysplasia commonly seen in IBD, colon cancer and many other diseases affecting the colon. One of the most common identifiers in colon cancer is tissue inflammation.

Colonoscopy examinations reveal irregularities in tissue topography that are most often caused by persistent inflammation. Fibroblasts have been shown to help define tissue topography and regulate the switch from resolving to persistent inflammation (29).

Since change in tissue topography is an imminent outcome of inflammation, focusing on topographical features may provide physicians with a better diagnostic tool for colon cancer, as well as other diseases affecting colon.

Colonoscopy imaging technique utilizes the differences in tissue topography to determine and diagnose physiological abnormalities. However, colonoscopy lacks the means to capture micron scaled irregularities within colonic tissue; the minimum diameter of a lesion that can be captured during a colonoscopy exam is approximately 1 .5 mm (30). Our data suggests that tumorigenic colonic tissue has micron as well as nano-scaled irregularities. Several imaging techniques, such as confocal

endomicroscopy and optical coherence tomography, have been shown to be effective in visualizing micro-scaled neoplastic lesions (31 , 32). Although optical biopsy technique is a remarkable diagnostic tool, this technique is rendered ineffective without proper testable parameters.

In the event of colon cancer, colonic tissue is inflamed and as a result is severely degraded. Our data indicate that tumorigenic colonic tissue topographical features are drastically different when compared to normal colonic mucosa. Roughness profile generated using ex-vivo colonic tissue of both normal and colonic mucosa suggest that the overall roughness profile, the peak height as well as the valley depth are 2-fold higher in tumorigenic colonic tissue than normal tissue. The roughness parameters, especially Rz and Rp, are linearly separable and a hyper-plane can be established using data mining tools such as SVM. Since the topography can easily be

distinguished and quantified, topographical features alone can be used as diagnostic aids.

The present inventors have shown for the first time that topographical features are unique and drastically different in tumorigenic and normal colonic tissue. These roughness parameters obtained from a roughness profile serve as 'signatures' that can be used to identify and classify tissue as being normal or tumorigenic. Logistic regression and other data mining tools may also be employed to create a prediction model based on the roughness parameters. It would be appreciated by those skilled in the art that the while the above-described topography is described in the event of colon cancer, aspects of the present invention may be applied to study and diagnose other physiological conditions that affect tissue topography. Furthermore, from a clinical perspective, these micro-scaled 'signatures' when combined with optical biopsy techniques may potentially serve as a diagnostic modality that is not only accurate, but also less daunting to patients.

While various embodiments of the present invention have been shown and described herein, it will be obvious that such embodiments are provided by way of example only. Numerous variations, changes and substitutions may be made without departing from the invention herein. Accordingly, it is intended that the invention be limited only by the spirit and scope of the appended claims. REFERENCES

1 . Haggar FA, Boushey RP. Colorectal cancer epidemiology: incidence, mortality, survival, and risk factors. Clin Colon Rectal Surg. 2009 Nov;22(4):1 91 -7.

2. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2012. CA Cancer J Clin. Jan-Feb;62(1 ):10-29.

3. Rubin DT, Turner JR. Surveillance of dysplasia in inflammatory bowel disease: The gastroenterologist-pathologist partnership. Clin Gastroenterol Hepatol. 2006 Nov;4(1 1 ):1309-13.

4. Rubin CE, Haggitt RC, Burmer GC, Brentnall TA, Stevens AC, Levine DS, et al. DNA aneuploidy in colonic biopsies predicts future development of dysplasia in ulcerative colitis. Gastroenterology. 1992 Nov;103(5):161 1 -20.

5. Provenzale D, Onken J. Surveillance issues in inflammatory bowel disease: ulcerative colitis. J Clin Gastroenterol. 2001 Feb;32(2):99-105.

6. Reid BJ, Blount PL, Feng Z, Levine DS. Optimizing endoscopic biopsy detection of early cancers in Barrett's high-grade dysplasia. Am J Gastroenterol. 2000

Nov;95(1 1 ):3089-96.

7. Levine DS, Haggitt RC, Blount PL, Rabinovitch PS, Rusch VW, Reid BJ. An endoscopic biopsy protocol can differentiate high-grade dysplasia from early

adenocarcinoma in Barrett's esophagus. Gastroenterology. 1 993 Jul;105(1 ):40-50.

8. Reid BJ, Weinstein WM, Lewin KJ, Haggitt RC, VanDeventer G, DenBesten L, et al. Endoscopic biopsy can detect high-grade dysplasia or early adenocarcinoma in

Barrett's esophagus without grossly recognizable neoplastic lesions. Gastroenterology. 1988 Jan;94(1 ):81 -90.

9. Wang T, Triadafilopoulos G. S, m, I, xl. Gut. 2003 Jan;52(1 ):5-7.

10. Dacosta RS, Wilson BC, Marcon NE. New optical technologies for earlier endoscopic diagnosis of premalignant gastrointestinal lesions. J Gastroenterol Hepatol. 2002 Feb;17 Suppl:S85-104. 1 1 . Pfau PR, Sivak MV, Jr. Endoscopic diagnostics. Gastroenterology. 2001

Feb;120(3):763-81 .

12. Schernthaner M, Reisinger B, Wolinski H, Kohlwein SD, Trantina-Yates A, Fahrner M, et al. Nanopatterned polymer substrates promote endothelial proliferation by initiation of beta-catenin transcriptional signaling. Acta Biomater. Aug;8(8):2953-62.

13. Zhang L, Webster TJ. Decreased lung carcinoma cell functions on select polymer nanometer surface features. J Biomed Mater Res A. Jan;100(1 ):94-102.

14. Kulangara K, Yang Y, Yang J, Leong KW. Nanotopography as modulator of human mesenchymal stem cell function. Biomaterials. Jul;33(20):4998-5003.

15. Zouani OF, Chanseau C, Brouillaud B, Bareille R, Deliane F, Foulc MP, et al. Altered nanofeature size dictates stem cell differentiation. J Cell Sci. Mar 1 ;125(Pt 5):1 217-24.

16. Dado D, Sagi M, Levenberg S, Zemel A. Mechanical control of stem cell differentiation. Regen Med. Jan;7(1 ):101 -16.

17. Dellatore SM, Garcia AS, Miller WM. Mimicking stem cell niches to increase stem cell expansion. Curr Opin Biotechnol. 2008 Oct;19(5):534-40.

18. Natali AN, Fontanella CG, Carniel EL. A numerical model for investigating the mechanics of calcaneal fat pad region. J Mech Behav Biomed Mater. Jan;5(1 ):216-23.

19. Agosta F, Canu E, Stojkovic T, Pievani M, Tomic A, Sarro L, et al. The topography of brain damage at different stages of Parkinson's disease. Hum Brain Mapp. Apr 24.

20. Horst M, Madduri S, Gobet R, Sulser T, Milleret V, Hall H, et al. Engineering functional bladder tissues. J Tissue Eng Regen Med. Mar 21 .

21 . Machann J, Thamer C, Schnoedt B, Haap M, Haring HU, Claussen CD, et al. Standardized assessment of whole body adipose tissue topography by MRI. J Magn

Reson Imaging. 2005 Apr;21 (4):455-62.

22. Anderson RS, L. Anderson, E. Glover S,. Image Analysis of Extracellular matrix Topography of Colon Cancer Cells. Microscopy & Analysis. 2006;20(3):5-7.

23. Ranucci CS, Moghe PV. Substrate microtopography can enhance cell adhesive and migratory responsiveness to matrix ligand density. J Biomed Mater Res. 2001

Feb;54(2):149-61 . 24. Lampin M, Warocquier C, Legris C, Degrange M, Sigot-Luizard MF. Correlation between substratum roughness and wettability, cell adhesion, and cell migration. J Biomed Mater Res. 1997 Jul;36(1 ):99-108.

25. Chung TW, Liu DZ, Wang SY, Wang SS. Enhancement of the growth of human endothelial cells by surface roughness at nanometer scale. Biomaterials. 2003

Nov;24(25):4655-61 .

26. Fan YW, Cui FZ, Hou SP, Xu QY, Chen LN, Lee IS. Culture of neural cells on silicon wafers with nano-scale surface topograph. J Neurosci Methods. 2002 Oct 15;120(1 ):17-23.

27. Thapa A, Miller DC, Webster TJ, Haberstroh KM. Nano-structured polymers enhance bladder smooth muscle cell function. Biomaterials. 2003 Aug;24(17):291 5-26.

28. Parekh A, Weaver AM. Regulation of cancer invasiveness by the physical extracellular matrix environment. Cell Adh Migr. 2009 Jul-Sep;3(3):288-92.

29. Buckley CD. Why does chronic inflammation persist: An unexpected role for fibroblasts. Immunol Lett. Jul;138(1 ):12-4.

30. Cushman P, Jr. Alcohol and opioids: possible interactions of clinical importance. Adv Alcohol Subst Abuse. 1987 Spring;6(3):33-46.

31 . Lund A, Mjellem-Jolly N, Hole K. Desipramine, administered chronically, influences 5-hydroxytryptamine1 A-receptors, as measured by behavioral tests and receptor binding in rats. Neuropharmacology. 1992 Jan;31 (1 ):25-32.

32. Goetz M, Kiesslich R. Confocal endomicroscopy: in vivo diagnosis of neoplastic lesions of the gastrointestinal tract. Anticancer Res. 2008 Jan-Feb;28(1 B):353-60.

It should be borne in mind that all patents, patent applications, patent publications, technical publications, scientific publications, and other references referenced herein are hereby incorporated by reference in this application in order to more fully describe the state of the art to which the present invention pertains.

Reference to particular buffers, media, reagents, cells, culture conditions and the like, or to some subclass of same, is not intended to be limiting, but should be read to include all such related materials that one of ordinary skill in the art would recognize as being of interest or value in the particular context in which that discussion is presented. For example, it is often possible to substitute one buffer system or culture medium for another, such that a different but known way is used to achieve the same goals as those to which the use of a suggested method, material or composition is directed. It is important to an understanding of the present invention to note that all technical and scientific terms used herein, unless defined herein, are intended to have the same meaning as commonly understood by one of ordinary skill in the art. The techniques employed herein are also those that are known to one of ordinary skill in the art, unless stated otherwise. For purposes of more clearly facilitating an understanding the invention as disclosed and claimed herein, the following definitions are provided.

While a number of embodiments of the present invention have been shown and described herein in the present context, such embodiments are provided by way of example only, and not of limitation. Numerous variations, changes and substitutions will occur to those of skill in the art without materially departing from the invention herein. For example, the present invention need not be limited to best mode disclosed herein, since other applications can equally benefit from the teachings of the present invention. Also, in the claims, means-plus-function and step-plus-function clauses are intended to cover the structures and acts, respectively, described herein as performing the recited function and not only structural equivalents or act equivalents, but also equivalent structures or equivalent acts, respectively. Accordingly, all such modifications are intended to be included within the scope of this invention as defined in the following claims, in accordance with relevant law as to their interpretation.