Diagnosis and localization of disease states using an ensemble of feature-fused quantitative parameter arrays

Abstract

Universal biomarker arrays enable earlier diagnosis and localization of disease states than conventional methods, while also improving the diagnostic accuracy of disease states in general. Universal biomarker arrays consist of a variety of mathematical parameters calculated from digital tissue images that reveal tissue details at the cellular level. A universal biomarker array includes one or more novel parameter arrays (i.e., spatial entropy arrays, bin arrays, and / or quartile arrays) in combination with one or more parameter arrays of different levels of technology (i.e., pattern arrays, distance arrays, and / or morphological arrays) and / or one or more other parameters. Using AI / ML systems, universal biomarker arrays can be analyzed against database-based reference data for early diagnosis and localization of disease processes.

Description

Technical Field

[0001] Cross - reference to Related Applications This patent application claims the benefit of U.S. Provisional Patent Application No. 63 / 464,058, entitled "UNIVERSAL BIOMARKER ARRAY," filed on May 4, 2023, the entire contents of which are incorporated herein by reference.

[0002] Field of the Invention The present invention generally relates to a universal biomarker array for use in clinical and scientific imaging and image analysis for early disease diagnosis, localization of disease processes, and real - time feedback, in combination with AI / machine learning / statistical analysis, and used across various tissues and diseases.

[0003] Background of the Invention In clinical medicine and basic science, an important focus is to improve methods for diagnosing early disease processes (early diagnosis), more accurate diagnosis, and precise localization of early disease stages or diseases in general within tissues, preferably integrated with near - immediate feedback.

[0004] Currently, depending on the type of tissue and disease, there are various levels of specialized diagnostic methods for non-clinical and clinical use, such as laboratory tests (blood and urine tests), functional tests, and genetic tests, clinical imaging (magnetic resonance imaging (MRI), CT scans), endoscopy, and biopsies for histology or RNA-Seq, for example. Non-clinical tests, such as biochemical assays in basic science, can accurately assess and localize the early disease process, but often require tissue biopsy and tissue destruction analysis, which is not possible or ethically justifiable for all tissues. Clinical tests, such as genetic tests, laboratory tests, and functional tests, do not allow for the localization of the disease process; they assess the risk of a particular disease, the systemic presence of disease-specific biomarkers, or the degree of limitation due to the disease. Liquid biopsies are also not site-specific. In contrast, imaging tests are important for localizing disease, but their technical resolution is limited. Several research studies have shown that MRI scanners achieve a spatial resolution of less than 0.5 mm, while clinical MRI scanners achieve a spatial resolution of approximately 1-2 mm in routine imaging examinations, but generally do not reach the resolution necessary to clinically visualize the early disease process, for example, at a microscopic scale. Therefore, MRI does not have sufficient sensitivity to detect small changes in tissue structure in early disease states. Today's endoscopy, such as confocal microendoscopy, can have microscopic resolution and, using the disease-specific markers and downstream methodologies described herein, can help identify early-stage diseases in near real-time, including accurately diagnosing early diseases. Other uses of organ / tissue / tissue section images for medical diagnosis include evaluation by pathologists or other specialists relying on their professional experience, or evaluation by AI / machine learning systems relying on parameters of the level of technology.

[0005] Overview of various embodiments According to one embodiment of the present invention, a method and system for disease diagnosis includes receiving a digital tissue image showing tissue details at the cellular level; performing automated image analysis on the received digital tissue image to identify cells of interest for analysis; calculating a variety of mathematical parameters at multiple levels for collective use as a universal biomarker array, which includes a pattern array, a distance array, a morphological array, a spatial entropy array, a bin array that converts absolute parameter values ​​into relative information by assigning them to specific bin positions, and a quartile array that converts absolute parameter values ​​into relative information by assigning them to specific quartiles; and analyzing the universal biomarker array based on database-based reference data for initial diagnosis and localization of the disease process.

[0006] In various alternative embodiments, digital tissue images may include endoscopy digital tissue images. Performing automated image processing may include identifying cells of interest relative to the image background, constructing at least one region of interest (ROI) in the non-background region, and performing (i) cell segmentation for analysis of segmented cells and (ii) inverted segmentation for analysis of intercellular space based on the ROI. The array of spatial entropy parameters may include Batty (absolute, relative), Contagion, Karlstrom (absolute, relative), O Neill (absolute, relative), and Parredw parameters. Calculating the bin array may include assigning each of several parameters to class and range-specific bin relative positions across the entire range of parameter values, the total number of bins being calculated by multiplying the number of disease states by the number of bins per state, and the conversion of data to class and range-specific bin relative positions may be performed based on parameters of the spatial entropy array, pattern array, distance array, and morphological array. Computing the quartile array may involve assigning each of several parameter values ​​to a disease state-specific quartile position and transforming each of several parameter values ​​to a disease range-specific quartile position, and the transformation of data to class and range-specific quartile relative positions may be performed based on parameters of a spatial entropy array, pattern array, distance array, and morphological array. Analyzing the universal biomarker array based on database-based reference data may involve providing the universal biomarker array to an AI / ML system trained on universal biomarker array data to detect localization of initial diagnosis and disease process, and the AI / ML system may use random forest regression to detect localization of initial diagnosis and disease process.

[0007] Additional embodiments may be disclosed and requested.

[0008] Those skilled in the art will better understand the advantages of various embodiments of the present invention, which will be discussed with reference to the drawings summarized immediately following the "Description of Exemplary Embodiments." [Brief explanation of the drawing]

[0009] [Figure 1] This figure shows a random forest classification using a universal biomarker array to diagnose KIN vs. SCC according to a specific embodiment. [Figure 2] This figure shows a predictive classification using a universal biomarker array to diagnose colon adenoma versus cancer, according to a specific embodiment. [Figure 3] This graph shows that the classification accuracy of disease states in human articular cartilage was significantly improved when a combination of a novel parameter array (n=3) and a novel unique array was used as a separate predictive modeling input dataset (n=57) compared to the standard parameter (n=3). [Figure 4] This graph shows that the classification accuracy of human skin disease states was significantly improved when using a combination of a novel parameter array (n=3) and a novel unique array as separate predictive modeling input datasets (n=57) compared to using the standard level of technology parameters (n=3). [Figure 5] This graph shows that the classification accuracy of disease states in the human colon was significantly improved when a combination of a novel parameter array (n=3) and a novel unique array was used as a separate predictive modeling input dataset (n=57) compared to using a standard parameter (n=3). [Figure 6] This graph shows a significant improvement in classification accuracy for a wide variety of unique array combinations, which function as separate predictive modeling input datasets, within the complex environment generated by pooling all tissue / disease states. [Figure 7]This figure shows a graph illustrating the level of technology and the top unique array combinations that consistently achieved the best classification accuracy across all analyzed tissues / disease states, including novel single arrays. [Figure 8] This figure shows a graph illustrating the feature-fused ensemble of top-performing unique array combinations and their average disease state classification performance across all tissues / disease states, based on technical level parameters. [Figure 9] This is a schematic diagram showing six constituent arrays of a universal biomarker array according to a specific embodiment.

[0010] It should be noted that the aforementioned figures and the elements shown therein are not necessarily drawn to a consistent or arbitrary scale. Unless otherwise indicated in the context, similar elements are numbered similarly. The drawings are for illustrative purposes only and are not intended to limit the scope of the subject matter of the invention as described herein.

[0011] Description of Exemplary Embodiments Definitions. As used herein and in the appended claims, the following terms shall have the meanings shown unless otherwise indicated by the context.

[0012] A "set" includes one or more members, even if the set is described in the plural form (for example, a set of X can contain one or more X).

[0013] The concept of "real-time" can range from instantaneous to a few seconds or minutes, depending on the context. For example, real-time feedback in a medical procedure generally requires that feedback be provided during the course of the procedure, and in some cases, within a required timeframe (e.g., if tissue damage occurs within X minutes of detecting the onset of a disease condition / process such as oxygen deficiency, real-time feedback needs to be provided within X minutes so that it is useful in taking corrective action).

[0014] In this specification, the term “disease” in relation to disease conditions, processes, diagnoses, etc., is used collectively to refer to attributes or combinations of attributes that exhibit any characteristic symptoms, which may be diseases, injuries, defects, allergic reactions or chemical / pharmaceutical reactions, hyperactivity, hypoactivity, healing, regeneration, growth, or other symptoms that can be analyzed using the techniques described herein.

[0015] Given the background described above, there is a clear need to improve the sensitivity, accuracy, localization, and feedback rate of methods for detecting, diagnosing, and localizing early disease processes. Three short examples highlight the advantages of methodologies that can meet these requirements: ● Early detection of cancer A methodology that can detect specific skin tumors or other tumors in their early stages non-invasively and in near real-time would be immediately clinically significant because it would enable early treatment. ● Early detection of tissue damage / degeneration Current clinical techniques cannot accurately identify the early tissue damage / degeneration that constitutes the potentially reversible disease state in the early stages of osteoarthritis (OA). OA remains a leading cause of disability worldwide. As a result, we are missing a golden opportunity by failing to adequately test novel / promising drug candidates in the early disease state where therapeutic effects are likely. A methodology capable of detecting the early stages of OA (or other diseases) would enable the development of novel drugs specific to the disease state, and / or the testing of existing promising drugs that were not useful in clinical trials because they were tested too late in the disease process. ● Real-time feedback Feedback on whether a given tissue area is affected by disease improves local accuracy in diagnostic biopsies (skin, intestines), tumor resections (resection margins), and the precise implantation of regenerative medicine implants during surgical procedures (treatment of cartilage defects).

[0016] In summary, earlier and more accurate diagnosis and precise localization, combined with real-time feedback, will significantly improve the development of treatment options and treatment outcomes, creating opportunities for meaningful improvements in medical treatment pathways.

[0017] The inventors have developed an ensemble of high-performance quantitative parameter arrays using feature fusion to enable earlier diagnosis and localization of disease states and to improve the diagnostic accuracy of disease states in general. Specifically, in certain embodiments, various mathematical parameters (e.g., spatial entropy, cell morphology, distribution and density, and population-based parameters as described / defined below) are calculated from images of organs, tissues, or tissue sections (collectively referred to herein as “tissues”) that show details at the cellular level. These are used collectively to quantitatively describe the healthy state versus a specific disease state of organs / tissues or tissue sections across various tissues and diseases. For convenience, the relevant sets of mathematical parameters can be logically classified into individual arrays (in which case, each array essentially translates distinct aspects of healthy tissue structure and changes in diseased tissue structure into comprehensive quantitative parameters based on images taken with a resolution / methodology system that depicts cells), and multiple arrays can be collectively analyzed as a feature-fused ensemble (referred to herein as a “universal biomarker array”). The term "universal" here reflects that the biomarker array is applicable to a wide range of organs / tissues or tissue sections and / or healthy state versus disease state. It should be noted that various embodiments may use different arrays and / or arrays with different constituent mathematical parameters as universal biomarker arrays for, for example, different organs / tissues and / or for the detection of different healthy state versus disease state. Universal biomarker arrays may be provided as predictive modeling input data to artificial intelligence / machine learning systems, for example, to classify healthy state versus disease state, or for basic statistics.

[0018] Among other things, such mathematical biomarker arrays and related methodological systems can be used, for example, for the following matters: ● Identification of healthy state vs. early disease state (early disease diagnosis) in an organ / tissue or tissue section ● Localization of disease processes in an organ / tissue or tissue section ● Real-time feedback regarding disease state and localization during imaging in an organ / tissue (e.g., a decision support tool during a surgical procedure) ● Monitoring of disease processes and tissue structure (e.g., healing / regeneration, results of drug treatment, results of cosmetic applications, etc.) ● Diagnosis and staging of tumors ● Improvement by enhancing the accuracy in overall diagnosis of disease state in an organ / tissue or tissue section as compared to the state of the art For example, it can be used for statistical and / or artificial intelligence (AI) / machine learning (ML)-assisted examination of an image / data of interest against reference data based on a database specific to a healthy state and / or a disease state. Although this specification frequently refers to the detection of "early stages", it should be noted that the techniques described herein may also provide benefits for other applications. For example, this universal biomarker array may be superior to current methods and thus may be beneficial for late-stage disease diagnosis.

[0019] Certain embodiments innovatively and quantitatively describe tissue structures in a healthy state and at an early disease stage using mathematical parameters (referred to as universal biomarker arrays) that are used as AI / ML inputs for diagnosis or other statistical analysis and can be used as long as imaging / images enable the identification / segmentation of cells (and matrices) within an organ / tissue / tissue section and the imaging of cells / images, which describe the characteristics of cell populations and matrices.

[0020] Generally speaking, the characteristics of a given tissue's cell population are remarkably specific to the tissue type and disease-sensitive. For example, in trauma, cells are lost; in tissue swelling, the spacing between cells widens; and in cancer and other proliferative disorders, the number of cells increases, the arrangement of cells differs significantly from that of healthy tissue, or different types of cells are present. Therefore, mathematical parameters that quantify the diverse characteristics of a tissue's cell population constitute a disease-sensitive digital fingerprint of tissue structure and function as a whole. Similarly, the extracellular matrix—the space between cells—of a given tissue can undergo various changes in the early stages of disease, which can be described using mathematical parameters in both the healthy and early disease states. This data is used as AI input for initial diagnosis and disease localization across tissue types and diseases, constituting a universal biomarker array usable for diagnosis.

[0021] The AI / ML inputs used here differ from other diagnostic AI / ML applications where segmentation does not identify cell populations / matrices, because they use either (i) annotated medical images without further quantification, e.g., annotated medical images for convolutional neural network analysis, or (ii) simpler quantification of segmented images (e.g., halalic texture features), or (iii) images from MRI (or other imaging modalities) that are too low resolution to analyze cells / matrices.

[0022] The following is a description of the workflow according to a specific embodiment.

[0023] The workflow generally begins with generating images of organs / tissues / tissue sections (e.g., using an endoscopy microscope, microscopy, or other imaging techniques) and / or using existing images of organs / tissues / tissue sections (e.g., images from histological sections generated for evaluation by a pathologist or other specialist). Importantly, as discussed above, embodiments of the present invention operate on images with a resolution that depicts cells. A specific method for generating images for the analysis described herein is to use an endoscopy microscope commercially available for clinical use, e.g., an endoscopy microscope from Mauna Kea Technologies (e.g., Cellvizio, a probe-based confocal laser endoscopy system) or Zeiss (e.g., CONVIVO, a confocal endoscopy system), or another system suitable for clinical use. A specific method for using images from histological sections for the analysis described herein is to use a digital pathology slide scanner to generate images or to use existing scans. A specific method for generating non-clinical images from organs / tissues / tissue sections for the analysis described herein is to use any type of microscope, e.g., a fluorescence microscope, a confocal microscope, or other microscopy techniques. When images are generated for clinical use (for example, using an endoscopy microscope), certain embodiments may use autofluorescent drugs, such as antibiotics like tetracycline or ciprofloxacin, or tinine as a fluorescent dye, or fluorescent molecules (i.e., in this case, such molecules are not drugs in a pharmacological sense), because currently clinically available fluorescent cell dyes are not available. Note that the drugs (e.g., antibiotics) are not used for pharmacological activity. This is not necessary for all tissues, but for some tissues that do not readily recognize cells, such as articular cartilage. If the autofluorescent drug does not readily penetrate the tissue, such as cartilage, drug polarization is performed by adjusting the pH / using another charged molecule / polar molecule in the solution.For example, when using tetracycline as a fluorescent dye to visualize tissue cells, such as cartilage, tetracycline penetration into negatively charged tissue can be achieved by adding a positive charge to the solution to interact with the negatively charged functional groups of tetracycline. This can be achieved using a polar solvent such as PEG and, for example, a 70% or 20% glucose solution. Specifically, tetracycline solutions for clinical use (e.g., doxycycline vials for intravenous administration) can be used in combination with approximately 20% to 70% glucose solutions for clinical use. If necessary, tissue penetration of ciprofloxacin and other autofluorescent polar drugs can be achieved by similar means. Since doxycycline can enhance fluorescence by interacting with divalent cations such as magnesium and calcium, divalent cations can be used in combination with doxycycline for imaging by adding, for example, magnesium ions, or by using isotonic solutions containing divalent cations, such as magnesium ions, which are commonly used in arthroscopic surgery. In this context, the fluorescence signal can be excited by, for example, UV-A light in the range of 350-400 nm. Another possibility for using autofluorescent drugs as fluorescent dyes to visualize tissue cells is to synthesize drug derivatives to which a positively charged group, such as an amino group, is attached. It should be noted that in alternative embodiments of the present invention, other imaging techniques, including future imaging techniques or even enhanced / improved versions of current imaging techniques that do not have cell-level resolution (e.g., existing MRI and CT imaging techniques), may also be used.

[0024] The workflow typically involves distinguishing organs / tissues / tissue sections / cells from the image background, constructing a region of interest (ROI) in the non-background area, and then performing automated image analysis based on the ROI, such as (i) cell segmentation for analyzing segmented cells and (ii) inverted segmentation for analyzing intercellular spaces.

[0025] The workflow is typically used collectively as a universal biomarker array, followed by the calculation of various mathematical parameters at multiple levels, as described / defined below.

[0026] The workflow typically involves storing curated (e.g., organ / tissue / tissue section / cell and disease state specific) diagnosed / classified / annotated universal biomarker array data in a database to create database-based reference data.

[0027] The workflow typically proceeds to a statistical and / or AI / ML-assisted examination of database-based reference data using images / data of interest for initial diagnosis and localization of the disease process. One specific AI technique is to use random forest classification for diagnostic purposes, e.g., to examine whether (i) "new" data from images of interest represents a healthy or diseased state, and / or (ii) whether it determines a specific disease classification / score. Here, the "new" data from "new" images of interest is examined against reference data stored in the database or otherwise available to the system. Another specific AI technique is to use random forest regression to calculate continuous numerical values, e.g., continuous scores or other values ​​of interest, where the "new" data from "new" images of interest is examined against database reference data. Yet another specific AI technique is to use LightGBM classification, as described for random forest classification. Another specific AI approach involves using PLS-DA for discriminant analysis (e.g., with max.dist, centroids.dist, and / or mahalanobis.dist) and / or using sPLS-DA for discriminant analysis and prediction / classification of "new" data to select a subset of variables. Generally, a particular AI model is determined using the precision, precision, recall, and F1 score of each model.

[0028] As a result, the workflow can determine whether the image / data of interest represents data from a healthy organ / tissue / tissue section or deviates from healthy data, for example, to diagnose and localize a specific disease state, and to determine the accuracy (classification) of the diagnosis and other quality control-related parameters. Furthermore, since the embodiments enable the analysis of tissues and tissue sections, the universal biomarker array is applicable to cells in tissues and cells visible in tissue sections. Therefore, for example, the universal biomarker array may be particularly applicable to pathology laboratories (tissue sections) and clinical trials / routine practice (patient / live cell imaging). Moreover, since the embodiments enable analysis using live cell imaging as well as analysis using different staining solutions (dyes), the universal biomarker array of the present invention is dye-independent.

[0029] The following is a description of various mathematical parameters currently assumed for a universal biomarker array of a particular embodiment, which may be used as AI / ML inputs or other statistical analyses for initial diagnosis and disease localization across tissue types and diseases: ● Calculations are performed using the "R" language and the "SpatEntropy" library, based on point patterns created from cell coordinates: • Spatial entropy Batty • Spatial entropy (contagion) • Spatial entropy (Karlstrom) • Spatial entropy O'Neill (oneill) • Spatial entropy Parredw(parredw) • Spatial entropy Leibovici(leibovici) • Spatial entropy Altieri (altieri) • Use absolute and relative values ​​when available. ● Using a marked point pattern created from cell coordinates, all parameters calculated for each cell are used as marks, and calculations are performed using the "R" language and the "SpatEntropy" library: • Spatial entropy Shannon (using absolute and relative values) ● Calculations are performed using the R language and the "spatstat" library: ○ About each cell: · Nearest cell distance (nndist) ○ Regarding images created from point patterns derived from each image's ROI / cell coordinates: • arithmetic mean and mean of nndist • The first quartile of nndist • The second quartile of nndist • The third quartile of nndist • Standard deviation of nndist • Standard error of nndist · Spatial intensity • Calculate the Clark-Evans aggregate index (clarkevans) with no edge correction, with "cdf" edge correction, and with "Donnelly" edge correction. • Arithmetic mean and mean of each version of the Clark-Evans aggregate index • Maximum value of the paired correlation function (maximum value of the PCF calculated with "trans" and "iso" corrections) • The r-value of the maximum value of the paired correlation function on the y-axis (x-axis) (calculated using both the Ohser-Stoyan transform and Ripley's isotropy correction) ● Calculations are performed using ImageJ / Fiji or the "R" language: ○ About each cell: · length · width Aspect ratio • Roundness • Circularity · Density · Perimeter · Area • Orientation ● For each image ROI / image / parameter (if applicable): • Mean, median, standard deviation, standard error, quartiles

[0030] The parameters listed above are calculated using cell identification / segmentation, focusing on a given cell population in a tissue. Inverting the segmentation mask selects the extracellular matrix, i.e., the intercellular space within the tissue. From the segmentation / area data, quantitative parameters such as spatial entropy (see above for parameters, library "SpatEntropy") and Haralick texture features can be calculated.

[0031] A specific embodiment introduces three novel quantitative parameter arrays, as follows: (a) A novel quantitative parameter array called spatial entropy (abbreviated as entropy), which calculates various spatial entropy parameters from cell location (details below). (b) A novel quantitative parameter array called a bin, which converts absolute parameter values ​​into relative information by assigning them to specific bin locations. Binning refers to the process of dividing a sequential parameter dataset into sequentially numbered intervals (bins) and assigning each data point to a specific bin number based on its value. Specifically, each parameter value is converted to a disease state-specific bin location, and this conversion is determined individually for each disease state using 20 bins per disease state. Additionally, each parameter value is converted to a disease range-specific bin location. Here, bin locations are assigned across the entire range of parameter values, and the total number of bins is calculated by multiplying the number of disease states by the number of bins per state, for example, 20 bins. The conversion of data to relative class and range-specific bin relative locations has so far been carried out based on novel entropy arrays as well as morphological, pattern, and distance arrays of technological standards, but will not necessarily be limited to these in the future. (c) A novel quantitative parameter array called quartiles, which converts absolute parameter values ​​into relative information by assigning them to specific quartiles. Specifically, each parameter is assigned to a quartile position specific to the disease state, and calculations are performed independently for each state. Additionally, each parameter value is converted to a quartile position specific to the disease range, quartiles are calculated using the entire range of values, and each value is assigned to one of these quartiles. Conversion to class and range-specific quartile relative positions has so far been performed based on entropy, morphology, pattern, and distance arrays, but will not necessarily be limited to these in the future.

[0032] In one exemplary embodiment, the entropy array includes spatial entropy parameters Batty (absolute and relative), Contagion, Karlstrom (absolute and relative), O Neill (absolute and relative), and Parredw, which can be computed using the R package "SpatEntropy". It is important to note that the exact number and / or types of spatial entropy parameters currently assumed for an entropy array should not be understood as an exclusive final list, as the number and / or types of spatial entropy parameters may differ in different embodiments. Furthermore, it should be noted that the exact number and / or types of parameters currently assumed for any array should not be understood as an exclusive final list, as the number and / or types of parameters may vary in different embodiments.

[0033] In certain embodiments, the entropy, bin, and quartile arrays are used in combination with one or more of the following additional arrays (sometimes called level-of-technology arrays) as a feature-fused ensemble that can be used for predictive modeling: (d) In one embodiment, a pattern array in which parameters include spatial intensity, Clark-Evans index without edge correction, Clark-Evans index with cdf edge correction, Clark-Evans index with Donnelly edge correction, the maximum value of the paired correlation function, and the x-value at which the maximum value is observed (using both Ohser-Stoyan transformation and Ripley isotropy correction), which can be computed using the R package "spatstat". (e) In one embodiment, using the individual nearest neighbor distances determined for each cell and its nearest neighbor cells (for example, using the spatstat package in R), a distance array comprising the calculated arithmetic mean, median, standard deviation, and first, second, and third quartiles. (f) A morphological array that, in one embodiment, utilizes the individual cell morphology determined for each cell and includes, in one embodiment, the average values ​​of the calculated cell area, length, width, orientation, perimeter, roundness, and density.

[0034] A particular embodiment may use all six of these arrays, as schematically shown in Figure 9, and other parameters as needed, but it should be noted that alternative embodiments may include one or more novel arrays in combination with one or more arrays of the art, and / or in combination with one or more other parameters. Also, as discussed above, the configuration parameters of each array may differ between different embodiments.

[0035] To generate distinct and unique array combinations and compare the disease state classification accuracy of single arrays versus unique array combinations versus skill level parameters as predictive modeling inputs, array values ​​for morphology, pattern, distance, entropy, bin, and quartiles were calculated for diverse disease / disease states in three human tissues: articular cartilage, skin, and colon (details below). Arrays were used as single arrays (six in total; three of which were skill level arrays: pattern, distance, and morphology) and to generate unique array combinations (two, three, four, five, and six array combinations; a total of 57 unique combinations). This resulted in 63 different arrays / array combinations, which were used as predictive model inputs for separate random forest modeling. The modeling process, performed in R, included 5-fold cross-validation and 5 cross-validations, performed twice using a balanced class distribution to ensure an equal number of data rows per class (disease state) in all runs. This step allowed the resulting accuracy to be analyzed as a function of 63 distinct predictive modeling inputs (Figures 3-5), which was performed to identify single arrays and unique combinations of arrays that had significantly higher classification accuracy than the level of skill parameters. Statistical analysis was performed on three data subsets. Each subset contained all novel arrays, all novel unique array combinations, and one of the three level of skill parameter sets, as detailed in the legend of Figures 3-5.

[0036] In human articular cartilage (Figure 3), statistical testing revealed that the disease status classification accuracy of all novel arrays and all unique array combinations differed significantly from the accuracy obtained using articulate parameter arrays: pattern (p<0.001), distance (p<0.001), and morphology (p<0.001). As shown in Figure 3, using only novel arrays and most unique array combinations as predictive modeling inputs (excluding entropy_pattern, distance_entropy, distance_pattern, and distance_entropy_pattern array combinations) resulted in significantly higher classification accuracy in articular cartilage than using articulate parameters, clearly outperforming the performance of articulate parameters.

[0037] As shown in Figure 3, this graph demonstrates a significant improvement in the classification accuracy of human articular cartilage disease states when novel parameter arrays (n=3) and novel unique array combinations were used as separate predictive modeling input datasets (n=57) compared to the technical level parameters (n=3). The box plots show the overall accuracy obtained by using the separate predictive modeling input datasets, displaying the median, 25th percentile, and 75th percentile, with whiskers displaying the 10th and 90th percentile. The line graphs show the specific accuracy calculated for each disease state (class) using the confusion matrix obtained from the modeling. Each disease state describes how cells are arranged in spatial patterns (e.g., string-like, double-string-like, large and small clusters, and diffuse arrangements without identifiable spatial patterns), because such arrangements correlate with structural and functional pathology (details below). The accuracy specific to these disease states indicates the accuracy of correctly classifying a particular disease state and how well it performs for individual disease states. A Shapiro-Wilk test to assess the normality of the data revealed that all datasets were non-normally distributed. A Wilcoxon rank-sum test for pairwise comparisons, adjusted for p-values ​​using the Bonferroni method, revealed that each of the skill level arrays—pattern, distance, and morphology—differed significantly (p<0.0005 each) from the novel arrays and the intrinsic array combinations. Therefore, no statistically significant differences were observed.

[0038] In human skin (Figure 4), statistical testing revealed that the disease status classification accuracy of all novel arrays and all unique array combinations differed significantly from the accuracy obtained by using level-of-technical parameter arrays such as pattern (p<0.001), distance (p<0.001), and morphology (p<0.001) as predictive modeling inputs. As shown in Figure 4, the separate array combination located to the right of the morphology array on the X-axis of Figure 4 yielded significantly higher classification accuracy in human skin compared to using level-of-technical parameters, thereby clearly exceeding the performance achieved with current level-of-technical parameters.

[0039] As shown in FIG. 4, this is a graph showing that the classification accuracy of human skin disease states was significantly improved as a result of using a new parameter array (n = 3) and a new unique array combination as separate predictive modeling input data sets (n = 57) compared to the prior art parameters (n = 3). The box-and-whisker plots show the overall accuracy obtained by using separate predictive modeling input data sets, which display the median and the 25th percentile and 75th percentile, and the whiskers display the 10th percentile and 90th percentile. The line graph shows the specific accuracy calculated for each disease state (class) using the confusion matrix obtained from the modeling. The individual disease states are actinic keratosis graded into the precancerous lesions KIN (keratinocyte intraepithelial neoplasia) I, II, or III, and moderately differentiated squamous cell carcinoma (SCC G2), which is a malignant cancer (details will be described later). The accuracy specific to these disease states indicates the accuracy of correctly classifying a specific disease state and shows how well it functions for each individual disease state. The Shapiro-Wilk test for testing the normality of the data revealed that all data sets were non-normally distributed. The Wilcoxon rank sum test for pairwise comparison with the p-value adjusted using the Bonferroni method revealed that all new arrays were significantly different from each of the prior art arrays of pattern, distance, and shape (0.000001 < p < 0.007). Furthermore, each unique array combination was significantly different from the prior art arrays and the new single arrays (0.000001 < p < 0.007). Therefore, no significant difference was shown.

[0040] In human colon (FIG. 5), statistical tests revealed multiple significant differences in the prior art vs. new array entropy and the classification accuracy when using various unique array combinations whose details are shown in the figure legend. As shown in FIG. 5, the various unique array combinations (indicated by "§") resulted in a significantly higher classification accuracy in human colon than when using the prior art parameters as predictive modeling input, thereby clearly outperforming the performance of the prior art parameters.

[0041] As shown in FIG. 5, this is a graph showing that the classification accuracy of human colon disease states was significantly improved as a result of using a new parameter array (n = 3) and a new unique array combination as separate predictive modeling input data sets (n = 57) compared to the prior art parameters (n = 3). The box-and-whisker plots show the overall accuracy obtained by using separate predictive modeling input data sets, which display the median and the 25th percentile and 75th percentile, and the whiskers display the 10th percentile and 90th percentile. The line graph shows the specific accuracy calculated for each disease state (class) using the confusion matrix obtained from the modeling. The individual disease states are colon adenoma and colon cancer (details are shown in the "Analyzed Tissues, Diseases, and Disease States" section of the text). The accuracy specific to these disease states indicates the accuracy of correctly classifying a specific disease state and shows how well it functions for individual disease states. The Shapiro-Wilk test for testing the normality of the data revealed that all data sets were non-normally distributed. The Wilcoxon rank-sum test for pairwise comparison with p-values adjusted using the Bonferroni method revealed that the new entropy array was significantly different from each of the prior art arrays of pattern, distance, and shape (each p < 0.0001), while no significant difference was seen in the new bin array and quartile array. Furthermore, the prior art arrays of pattern and distance were significantly different from each unique array combination (0.000001 < p < 0.02). The prior art array form was significantly different from the various unique array combinations denoted by "m" (0.000001 < p < 0.02). No other significant differences were shown. The "§" symbol indicates the unique array combination that brought about a significantly higher classification accuracy than any of the prior art parameters when used as a predictive modeling input.

[0042] Next, the novel single arrays and unique array combinations were examined against the state-of-the-art arrays as inputs for predictive modeling in a complex environment that pooled all tissue / disease states, thereby generating 12 classification options. By using the pooled data (Figure 6), statistical tests showed that all disease state classification accuracies, except for the comparison between the distance array and the entropy array, reached a significance level (0.0000001 < p < 0.03). Importantly, the novel array bins and all distinct array combinations located to the right of the morphological array on the X-axis in Figure 6 yielded significantly higher classification accuracies than the state-of-the-art parameters. This excellent performance was achieved in the context of cross-classification across different tissues and disease states, and it should be noted that it emphasizes the effectiveness of the novel and characteristic array combinations as inputs for predictive modeling for the classification task. This result bears the evidence of the practical utility that these unique array combinations bring to complex real-world scenarios and supports their potential as general-purpose tools for advanced classification applications. Figures 3 - 5 presented the classification accuracies of disease states in specific tissues, and Figure 6 presented the classification accuracies in the pooled complex dataset. Taken together, these figures statistically and convincingly showed that when used as inputs for predictive modeling, diverse unique array combinations clearly outperformed the accuracies obtained from state-of-the-art parameters.

[0043] As shown in FIG. 6, this is a graph showing that the classification accuracy of a variety of unique array combinations, which function as separate predictive modeling input data sets in a complex environment generated by pooling all tissue / disease states, has been significantly improved. The individual tissue / disease states used in FIGS. 3 - 5 were pooled to generate a difficult classification task. The box-and-whisker plots (pink) show the overall accuracy obtained using separate predictive modeling input data sets, which display the median and 25th and 75th percentiles, and the whiskers display the 10th and 90th percentiles. The line graph shows the specific accuracy calculated for each disease state (class) using the confusion matrix obtained from the modeling. The accuracy specific to these disease states indicates the accuracy of correctly classifying a specific disease state and shows how well it functions for an individual disease state. The Shapiro-Wilk test for testing the normality of the data revealed that all data sets were non-normally distributed. The Wilcoxon rank-sum test for pairwise comparison with p-values adjusted using the Bonferroni method showed that the disease state classification accuracy of all new arrays and all unique array combinations was significantly different (0.0000001 < p < 0.03) from the accuracy obtained from the state-of-the-art parameter arrays of pattern, distance, and shape used as predictive modeling inputs. The only exception was the comparison between the distance array and the entropy array, which did not reach a significant difference.

[0044] Next, unique array combinations that consistently demonstrated the highest level of performance across all tissues, disease states, and aggregated complex datasets were identified. This was achieved by sorting the numerical values ​​of each classification precision (Figures 3-6) in ascending order of median precision in box plots, and then sorting the disease states represented by each by ascending order of precision in line graphs. In each sorting iteration, the resulting order of the unique array combinations was recorded, and their (ascending) performance was captured as predictive modeling input for specific disease states (data not shown). Using this method, the top 11 unique array combinations that consistently achieved the highest classification precision across all analyzed tissues / disease states were identified. Figure 7 is a graph showing the skill level and novel single arrays, as well as the top unique array combinations that consistently achieved the highest classification precision across all analyzed tissues / disease states. The classification precision of the three skill level arrays (pattern, distance, morphology) and their combined array (pattern_distance_morphology) for each of those tissues / disease states against the top 11 unique array combinations is shown. Figure 8 is a graph showing the feature-fused ensembles of top-performing unique array combinations and their average disease state classification accuracy performance across all tissue / disease states, relative to the level of technology parameters. The feature-fused ensembles of level of technology parameters, novel arrays, and top-performing quantitative parameter arrays are shown, which "transform" distinct aspects of tissue structure in healthy states and changes in tissue structure in disease states into comprehensive mathematical parameters as predictive modeling input data for classifying disease states, significantly outperforming the accuracy of the level of technology parameters.

[0045] In the above embodiment, the array values ​​were calculated using threshold-segmented images of human articular cartilage tissue that are macroscopically normal and degenerated, depicting the extent of cells beneath the tissue surface, as well as images of tissue sections from human skin and colon biopsies, depicting the extent of cells within the tissue.

[0046] Articular cartilage images depicting cells stained with calcein AM were classified according to publication [1] as to how the cells were arranged in spatial patterns (e.g., string-like, double-string-like, large and small clusters, and diffuse arrangements without identifiable spatial patterns). These spatial patterns indicate structural [2] and functional pathologies [8,10]. It should be noted that the content of these publications partially quantifies distance and / or pattern, which are parameters of the state of the art. This is not inconsistent with the present application, as it does not mention the novel arrays and feature-fused ensembles of higher-performing quantitative parameter arrays presented here.

[0047] Skin images depicting hematoxylin-eosin stained tissue sections were diagnosed by specialist pathologists as actinic keratosis or moderately differentiated squamous cell carcinoma (SCC G2) with histological grade KIN (keratinocyte intraepithelial neoplasia) I, II, or III

[11] . Actinic keratosis is the most common precancerous lesion in humans, while SCC accounts for approximately 20% of non-melanoma skin cancers and is the second most common type after basal cell carcinoma. Other images depict tissue sections diagnosed by specialist pathologists as Morbus Bowen disease, an early-stage squamous cell carcinoma confined to the outermost layer of skin (epidermis).

[0048] Colon imaging of biopsy tissue sections was diagnosed by a specialist pathologist as either a colon adenoma (a benign colon tumor considered a precancerous condition) or colon cancer (a malignant colon tumor). The 5-year survival rate for colon cancer is 91% when diagnosed locally, 72% when the cancer has spread to surrounding tissues / organs / regional lymph nodes, and 13% when it has spread to distant parts of the body. These figures highlight the need for early diagnosis and localization, which is usually achieved by biopsy for histopathological examination using some of the parameters of the technical level described above.

[0049] The following are some specific examples of using the biomarker array described above for classification / diagnosis with AI / ML learning support.

[0050] Specific example I (skin cancer): Actinic keratosis (KIN I-III), also known as sun-damaged skin, is a precancerous area characterized by thickened, scaly, or crusty skin. It is considered a precursor to squamous cell carcinoma (SCC) and may progress to SCC over time. SCC is the second most common form of skin cancer, typically arising from areas chronically exposed to sunlight. Here, we used scans of pathological section images for analysis. A universal biomarker array for distinguishing KIN I-III from SCC yielded an accuracy of 89.64% using a random forest classification model and images diagnosed by specialist pathologists. See Figure 1 – Random Forest Classification using a Universal Biomarker Array for Diagnosing KIN vs. SCC (in this figure, the upper left and lower panels show multiple parameters of model performance, and the right panel shows SHAP (SHapley Additive exPlanations) values, which show the impact of each array parameter on prediction). Examination images were not included in the training dataset. It should be noted that these results were achieved with a relatively small number of images (n<100) in each category, suggesting that accuracy improves as the number of curated and classified images increases.

[0051] Specific example II (colon cancer): Colon adenomas are a type of polyp. Up to 10% of colon adenomas have the potential to become malignant and are therefore precursor lesions to colon adenocarcinoma (colon cancer). Here, we used scans of pathological section images for analysis.

[0052] A universal biomarker array for distinguishing between colon adenoma and colon cancer yielded a 96.53% accuracy using a LightGBM classification model and images diagnosed by specialist pathologists. See Figure 2 – Predictive classification using a universal biomarker array for diagnosing colon adenoma vs. cancer (in this figure, the upper left and lower panels show multiple parameters of model performance, and the right panel shows SHAP (SHapley Additive exPlanations) values, which show the impact of each array parameter on prediction). Examination images were not included in the training dataset. Note that these results were achieved with a relatively small number of images (n<100) in each category, suggesting that accuracy improves with a larger number of curated and classified images.

[0053] Specific example III (cartilage degeneration): Cartilage degeneration in osteoarthritis can be recognized by spatial organization of superficial chondrocytes (SCSO), a neologism that indicates a loss of tissue function, such as loss of nanoscale stiffness. Therefore, by identifying specific stages of SCSO and determining whether a given SCSO is typical of healthy articular cartilage, it is possible to classify healthy versus early disease (cartilage degeneration). Here, fluorescence microscopy was used to generate images of tissue biopsies / explants for analysis.

[0054] A universal biomarker array for distinguishing between SCSO [healthy] and SCSO [disease] yielded a 91.15% accuracy using a random forest classification model and diagnosed images (see Figure 1). Examination images were not included in the training dataset. Note that these results were achieved with a relatively small number of images (n<100) in each category, suggesting that accuracy improves with a larger number of curated and classified images.

[0055] The following is a description of potential future applications envisioned by the inventor: ● Using universal biomarker arrays and related methodologies, clinically applicable and commercially available solutions are available, particularly for early-stage diseases that have previously been undetectable or difficult to detect. A tool to support medical decision-making in diagnosis and treatment during medical procedures. We are developing a tool that provides real-time, user-friendly feedback during medical procedures, intended to enable surgeons, internists, and other specialists to diagnose the onset of early disease earlier than currently clinically possible and to localize the initial disease process within the tissue, for example, in combination with endoscopic microscopy or other appropriate imaging modalities during medical procedures. ● Another exemplary application of this tool is, for example—as described above—to assist in guided biopsy, evaluation / assurance of tumor resection, tissue debridement, and implantation of biomaterials / ATMP / RMAT (FDA, EMA) for regeneration and other purposes, in combination with an endoscopy microscope or other appropriate imaging modality during medical procedures. Localizing early / advanced disease processes within tissues during medical procedures. That is the case. ● Using a universal biomarker array and related methodology, fully clinically applicable and commercially available, A tool to support medical decision-making in diagnosis and staging for pathologists and other specialists who evaluate medical images of organs / tissues / tissue sections. We will develop a system that streams / transmits images through an imaging device and a dedicated software interface, allows us to perform subsequent analysis and examination, and provides AI-assisted diagnostic and staging recommendations to pathologists or other specialists. Examples include, but are not limited to, diseases requiring biopsy for diagnosis, such as those listed herein. ● The universal biomarker array and related methodologies are used in combination with lymph node / bone marrow biopsies, peripheral blood smears, bronchoalveolar lavage, cerebrospinal fluid aspiration, brain biopsies, and other biopsies to perform cellular profiling and the analyses described herein. ● Universal biomarker arrays and related methodologies can be used in in vitro studies, animal experiments, and clinical trials to enable the above applications, including statistical and / or AI / ML-assisted analysis of images / data of interest against database-based reference data. Generate a database containing quantitative information specific to organs, tissues, and / or tissue sections, as well as quantitative information specific to healthy and diseased states, to be used as reference data.In this context, the database is a quantitative atlas / digital fingerprint of organ / tissue / tissue section structures in healthy and diseased tissues. ● Universal biomarker arrays and related methodologies are used in combination with molecular imaging (e.g., by endoscopic microscopy as well as PET, SPECT, MRI, optical imaging, ultrasound imaging, photoacoustic imaging, CT, and other imaging modalities) using markers of interest (e.g., antibodies, antibody fragments, small molecules, and other sensors or substances, including commercially available and / or clinically applicable markers) to correlate molecular imaging with initial diagnosis, staging, monitoring, and disease localization in organs, tissues / tissue sections generated by the methods described herein. ● The Universal Biomarker Array and related methodologies are applied to molecular imaging signals / data, for example, by analyzing markers of interest recorded by molecular imaging (e.g., antibodies, fragment antibodies, small molecules, and other sensors or substances, including commercially available and / or clinically applicable markers) to generate quantitative data usable as AI / ML input / for statistical analysis. Examples of molecular marker targets include metabolism, inflammation, autoimmune processes, and / or anatomical and / or pathological structures. ● Using the database described above, information on disease diagnosis and localization obtained from the Universal Biomarker Array (see above) is linked with other data measured at the molecular, intracellular, cellular, tissue, organ, and patient levels (e.g., patient-reported outcomes) in animal experiments and clinical trials. Examples of applications include, but are not limited to, bioinformatics analyses for testing existing drugs in the early stages of disease or for developing novel drugs that target early stages of disease that were previously not clinically identifiable, and testing of promising / existing / novel / development-oriented drugs that target specific / early disease processes. ● Use the universal biomarker array and related methodologies in combination with “omics analysis,” e.g., genomics, transcriptomics including spatial transcriptomics, proteomics, metabolomics, radiomics, and other fields, to correlate with initial diagnosis, staging, monitoring, and disease localization in organs, tissues / tissue sections generated by the methods described herein, and, where applicable, use the universal biomarker array and related methodologies with these signals / data. ● The universal biomarker array and related methodologies are used in combination with other experimental data of interest, e.g., other specific readouts from a given in vitro or ex vivo / in vivo experiment, to determine their statistical relationships. These relationships are then used to predict / determine experimental data of interest by using parameters or subsets of parameters from the universal biomarker array as surrogate markers. For example, (after the statistical relationship has been established by appropriate analysis) the OARSI score or the degree of surface damage of articular cartilage can be predicted / determined from the parameters or subsets of parameters. ● Universal biomarker arrays and related methodologies are used as quantitative markers and predictors for the performance of cells, organs, and tissue cultures, such as cellular phenotype, differentiation, and other functions, in in vitro cells, organs, and tissue cultures. ● For the purposes listed above, the universal biomarker array and related methodology system will be used in all mammalian species / tissues / cells, such as human volunteers, human patients, and animals such as dogs, horses, and camels. ● Examples of diseases associated with the uses listed above include, but are not limited to, the following: ● Cancer and precancerous lesions, for example: ○ Skin, ○ Gastrointestinal tract, ○ Lungs, ○ Kidney ○ Breasts, ○ Liver, ○ Esophagus, ○ Brain, ○ Prostate, ○ Pancreas, ○ Glioblastoma, ○ Metastasis, ● Orthopedic symptoms / diseases (including, but not limited to, the following) ○ Damage to cartilage and other joint structures, ○ Degeneration of cartilage and other joint structures, ○ Hereditary diseases related to cartilage, ○ Osteochondritis dissecans, ○ Inflammatory joint diseases (rheumatoid arthritis, juvenile idiopathic arthritis, gout, systemic lupus erythematosus, seronegative spondyloarthropathy), ○ Osteoarthritis, ○ Post-traumatic osteoarthritis, ○ Chondrodysplasia, ● Vascular diseases, for example, ○ Atherosclerosis, ○ Stenosis, ○ High blood pressure, ○ Stroke, ● Heart disease, for example, ○ Coronary artery disease / arteriosclerosis ○ Cardiomyopathy, ○ Myocarditis, ○ Heart transplant rejection, ● Gastrointestinal diseases, for example, ○ Polyps, ○ Gastritis / gastric ulcer, ○ Gastrointestinal mass / lesion of unknown cause ○ Cancer, ● Kidney disease, for example, ○ Chronic kidney disease, ○ Glomerulonephritis, ○ Renal fibrosis, ● Liver disease, for example, ○ Liver cirrhosis, ● Inflammatory / autoinflammatory / autoimmune diseases, ● Eye diseases, for example, ○ Glaucoma, ○ Macular degeneration, ● Neurodegenerative diseases, for example, ○ Parkinson's disease, ○ Alzheimer's disease, Huntington's disease, ○ Amyotrophic lateral sclerosis (ALS), ○ Motor neuron disease, ● Chronic inflammatory brain diseases, for example, ○ Multiple sclerosis, ○ Acute demyelination, ● Vascular diseases, for example, ○ Cerebral amyloid angiopathy (CAA), ● Peripheral neuropathy, ● Rare diseases, ● Metabolic disorders / diseases, for example, ○ Diabetes.

[0056] The following is information regarding past publications (reference numbers in parentheses refer to citation numbers in the references listed below). Note that all these studies focused solely on chondrocytes / cartilage / osteoarthritis (OA). In studies from 2008 to 2014, we used the term “spatial organization” and focused solely on chondrocytes / cartilage / osteoarthritis (OA). These analyses used a very limited amount of quantitative data, for example, using one or two markers (Clark-Evans index, nearest neighbor distance) that are part of the universal biomarker array described here. Since 2016, we have used the term “superficial chondrocyte spatial organization (SCSO).” A very limited number of quantitative parameters were used in later studies on articular cartilage. Some terms used, associated measurements, and key disease-related descriptions are highlighted in bold to indicate the very limited and contextual use of small subgroups of the universal biomarker array (cartilage, chondrocytes, degeneration, osteoarthritis only). Note that while specific cell morphological parameters are used to describe the quantitative effects of stimulation, they are not combined with diagnostic tools or other applications.

[0057] ● Clear horizontal patterns in the spatial organization of the outermost chondrocytes of human joints [1]: A better understanding of the unique cellular and functional characteristics of the outermost chondrocytes could contribute to current tissue engineering strategies that attempt to hierarchically design the repair of cartilage lesions to avoid or delay the onset of osteoarthritis. However, data on the cellular tissue structure of the non-denatured outermost chondrocytes are not available for most human joints. In this study, the arrangement of chondrocytes in non-denatured human joints (shoulder, elbow, knee, and ankle) was analyzed using a fluorescence microscope with the superficial layer viewed upward. The resulting horizontal arrangement of chondrocytes was examined for randomness, uniformity, or the presence of significant grouping by point pattern analysis, and a correlation was found with the type of joint in which they were present. This study showed that human outermost chondrocytes exist in four distinct patterns: string-like, cluster-like, pair-like, or single-chondrocyte. These patterns represented significant grouping aligned horizontally (p<0.0001). Each joint surface was occupied by only one of these four patterns (p<0.001). Certain patterns correlated with specific types of movable joints (p<0.001). Further research is needed to determine whether these organizational patterns are attributable to the surrounding environment or can be linked to a functional purpose.

[0058] ● Proliferative remodeling of spatial organization of human outermost chondrocytes located distant from localized early osteoarthritis [2]: Objective: Human superficial chondrocytes exhibit distinct spatial organization, and they generally aggregate near the fissures of osteoarthritis (OA). The aim of this study was to determine whether remodeling or destruction of the spatial organization of chondrocytes located distant from local (early) lesions occurs in OA patients. Methods: Samples of normal cartilage (condyle, patellofemoral groove, and proximal tibia) located distant from local lesions in grade 2 OA joints were compared with location-matched non-degenerative (grade 0-1) cartilage samples. Chondrocyte nuclei were stained with propidium iodide, observed under a fluorescence microscope, and recorded upward. Chondrocyte arrangement was examined for randomness or significant grouping using point pattern analysis (Clark and Evans aggregation index), which correlated with OA grade and surface cell density. Results: In grade 2 cartilage samples, superficial chondrocytes were located in horizontal patterns such as string-like, cluster-like, pair-like, and solitary, similar to the patterns in non-degenerative cartilage. In normal cartilage samples from grade 2 joints, spatial reorganization included a novel pattern of chondrocytes aligning along two parallel lines to form double strings. These double strings were significantly correlated with an increase in the number of chondrocytes per group and a corresponding increase in the density of the outermost layer cells. They were observed in all grade 2 condyles and some grade 2 tibias, but not in grade 0–1 cartilage. Conclusion: This study is the first to identify distinct spatial reorganization of human outermost layer chondrocytes in response to distantly located early OA lesions, suggesting that proliferation occurred at locations distant from the localized early OA lesion. This spatial reorganization may play a role in mobilizing metabolically active units as an attempt to repair localized damage.

[0059] ● Onset of preclinical osteoarthritis: Angular spatial organization enables early diagnosis [3]: Objective: The outermost articular chondrocytes exhibit a distinct spatial remodeling process in response to the development of osteoarthritis (OA) in distant locations. Such a process may be useful for diagnosing early events before manifested OA causes tissue destruction and clinical symptoms. Using a novel method to quantify space by calculating the angles between chondrocytes and their surrounding neighboring cells, we compared the maturation and degeneration processes of cellular tissue in rat and human cartilage specimens. Methods: The nuclei of the outermost chondrocytes obtained from normal rat cartilage and human knee cartilage, as well as cartilage with focal and severe OA, were digitally recorded in an upward view. Using their orthogonal coordinates, the angles between each chondrocyte's nearest neighbor cell and between these two cells and a reference cell were determined. These angles, cell density, nearest neighbor cell distance, and aggregation were analyzed as functions of location and OA severity. Results: Adjacent rat chondrocytes presented a complex angular pattern with four dominant angles maintained throughout the maturation process and the development and progression of OA. Within normal cartilage, human chondrocytes exhibited one dominant angle, and consequently, significantly different angular organization. Upon the onset of early OA, human chondrocytes located within normal cartilage showed increased frequency of four angles; the resulting angular patterns were indistinguishable from those observed in rats. This angular remodeling was associated with location- and OA-severity-dependent changes in cell density and aggregation. Conclusion: This study is the first to reveal the angular characteristics of chondrocyte spatial organization and the existence of species-specific remodeling processes correlated with OA development. The emergence of distinct angular and spatial patterns between adjacent chondrocytes allows for the identification of OA development in distant locations, even before microscopically visible tissue damage and even before clinical onset. Further development could make this novel concept suitable for the diagnosis and monitoring of OA-prone patients.

[0060] ● Modeling of chondrocyte patterns using the elliptical clustering process [4]: The outermost chondrocytes (CH) of human joints are spatially organized in a distinct horizontal pattern. Among other factors, the type of spatial organization of CH within a given joint surface differs depending on whether the cartilage originates from a normal joint or whether the joint is affected by osteoarthritis (OA). Furthermore, specific changes in the type of spatial organization are associated with specific states of OA. This association may prove important for early disease recognition based on the quantitative structural characteristics of CH patterns. Therefore, we present a point process model that describes the distinct morphology of CH patterns within the joint surface of normal human cartilage. This reference model of normal CH organization can be seen as a first step toward a model-based statistical diagnostic tool. Model parameters are fitted to fluorescence microscopy data using novel statistical methods with cluster analysis and principal component analysis tools. This allows the complex morphology of surface CH patterns to be represented by a relatively small number of model parameters. We validate the point process model by comparing biologically important structural characteristics between the fitted model and data obtained from micrographs of human joint surfaces using spatial statistical techniques.

[0061] ● Spatial organization of articular surface chondrocytes: A review of its potential role in tissue function, disease, and early and preclinical diagnosis of osteoarthritis [5]: Chondrocytes within articular cartilage exhibit depth-dependent changes in many of their properties, similar to the depth-dependent changes in the properties of the surrounding extracellular matrix. However, little is known about the spatial organization of chondrocytes throughout the tissue. Recent studies have revealed that human chondrocytes exhibit distinct spatial organization patterns within articular surfaces, with each articular surface typically occupied by one of four basic spatial patterns. The resulting complex spatial organization correlates with specific types of movable joints, suggesting that the organization of chondrocytes within articular surfaces is related to the biomechanical forces at play. In response to localized osteoarthritis (OA), the outermost chondrocytes undergo spatial organization disruption within the OA lesion, while simultaneously undergoing a defined remodeling process on the remaining normal cartilage surface, at locations distant from the OA lesion. One biological insight that can be derived from this spatial remodeling process is that chondrocytes can respond comprehensively and cooperatively to localized OA at distant locations. The spatial characteristics of this process differ significantly from the cell aggregation typical of OA lesions, suggesting a different underlying mechanism. Here, we summarize the available information regarding the spatial organization of chondrocytes and their potential role in cartilage function. This spatial organization may be useful in diagnosing early OA onset before manifested OA causes tissue destruction and clinical symptoms. With further development, this concept may become clinically applicable to the diagnosis of preclinical OA.

[0062] ● (Abstract) Spatial organization and pericellular matrix of human fetal chondrocytes are not innate features but acquired features [6]: Introduction: In adult articular cartilage, the pericellular matrix mediates biomechanical, biophysical, and biomechanical interactions between chondrocytes and the extracellular matrix. The pericellular matrix (PCM) is also associated with the spatial organization of human outermost chondrocytes, which are located in four distinct patterns: string-like, cluster-like, pair-like, or single-cell. However, little is known about the PCM and spatial organization during fetal development. In this study, we investigated whether fetal chondrocytes exhibit spatial organization similar to that of adult chondrocytes, and whether the PCM is present in the early stages of fetal chondrogenesis. Methods: Articular cartilage sections (100 μm thick) were prepared from human fetal knee joint condyles (7-10 weeks of gestation) and macroscopically normal condylar regions (knee arthroplasty). Samples were characterized by immunofluorescence microscopy and multiphoton-induced autofluorescence imaging combined with quantitative SHG signal profiling, which can quantify fibrous collagen content. Spatial organization was analyzed by point pattern analysis, as previously described. For these analyses, the orthogonal coordinates of each nucleus were determined by converting immunofluorescence images to grayscale images and finding local grayscale maximums using ImageJ (NIH). Results: In the outermost adult cartilage, PCMs surrounded single or multiple chondrocyte populations and defined the spatial tissue structure of cell strings. PCMs were characterized by a strong type VI collagen staining signal and high collagen intensity (156.7 ± 12.4) as measured by SHG. In fetal cartilage, condyles were characterized by high cell density, lack of recognizable spatial tissue structure, and a small amount of extracellular matrix. Type VI collagen was not detected in the matrix surrounding fetal chondrocytes. Furthermore, SHG imaging revealed that fibrous collagen intensity was significantly weaker compared to adult tissue (4.7 ± 0.8; p<0.001). Cell density per unit volume significantly decreased during maturation from fetal to adult (p<0.001). The distance from fetal chondrocytes to nearest neighbor cells was 15.70 ± 0.12 μm, which was significantly longer in adult chondrocytes (35.86 ± 1.37 μm, p < 0.001).The level of spatial clustering, measured by the integral of the paired correlation function, was significantly higher in adult cartilage (p<0.001). Overall, cell populations similar to those of adult chondrocytes were not present in fetal cartilage. Taken together, these parameters suggest that the spatial organization typical of adult condylar cartilage was absent in human fetal condylar cartilage. Discussion / Conclusion: Human fetal articular cartilage lacked the spatial organization typical of adult chondrocytes. Instead, chondrocytes were densely packed and located in close proximity to one another. This study is the first to reveal the absence of collagen components typical of adult PCM in fetal cartilage. In conclusion, the spatial organization of PCM and human outermost chondrocytes develops with cartilage maturation and is therefore an acquired rather than congenital feature. Summary: In adult articular cartilage, the pericellular matrix (PCM) mediates chondrocyte-matrix interactions and is associated with spatial cell organization. Immunofluorescence microscopy, multiphoton-induced autofluorescence, and second-harmonic generation (SHG) imaging, as well as point pattern analysis, revealed that both PCM and spatial organization are absent in fetal chondrocytes.

[0063] ● Loss of spatial tissue structure and disruption of the pericellular matrix in early osteoarthritis in vivo and novel in vitro methods [7]: Objective: Current articular cartilage (AC) repair techniques fail to restore the tissue's original morphology and function because the lifelong changes in its structural blueprint and the corresponding biological understanding are not fully achieved. We investigated whether two inherent elements of human cartilage structure—the pericellular matrix (PCM) surrounding chondrocytes and the superficial chondrocyte spatial organization (SCSO) beneath the articular surface (AS)—are congenital, stable throughout life, or dynamic. We hypothesized that in vitro induction of chondrocyte proliferation impairs organization and PCM, leading to a progressive osteoarthritis (OA)-like structural phenotype in human cartilage. Methods: Fetal and adult cartilage explants stained with propidium iodide were recorded, and the stages of organization were arranged chronologically to create a lifelong SCSO model. To reproduce the OA-related dynamics revealed by our model and test our hypothesis, we specifically introduced hFGF-2 into early OA explants to induce proliferation. PCM was examined using immunofluorescence and autofluorescence, multiphoton second harmonic generation (SHG), and scanning electron microscopy (SEM). Results: Spatial organization developed from fetal homogeneity and peaked in adult string-like arrays, but was completely lost in OA. Loss of organization included PCM perforation (decreased intensity of local microfibrous collagen) and disruption [weakening or disappearance of regional type VI collagen (Coll VI) signaling]. Importantly, both loss of organization and PCM disruption were reproduced in FGF-2-transplanted explants. Conclusion: Induced proliferation of spatially characterized early OA chondrocytes within standardized explants reproduced the full range of SCSO loss and PCM disruption, introducing a novel in vitro technique. This technique induces a structural phenotype of human cartilage similar to that of progressive OA and is potentially important and useful.

[0064] ● Proof of concept for the detection of early osteoarthritis lesions using clinically applicable endoscopic microscopy and quantitative AI-assisted optical biopsy [8]: Objective: Clinical trials for osteoarthritis (OA), a leading cause of disability worldwide, cannot accurately identify early, potentially reversible disease using current clinical techniques. Therefore, disease-modifying drug candidates cannot be tested in the early stages of the disease. To overcome this limitation, we investigated whether early OA lesions could be detected using current clinical techniques. Methods: We examined the relationship between two highly sensitive early OA markers—human articular cartilage (AC) surface stiffness measured by atomic force microscopy (AFM) and position-correlated spatial organization of the outermost chondrocytes (SCSO)—to determine whether a significant decrease in surface stiffness could be detected at the SCSO stage of early OA. Next, we tested whether SCSO could be visualized and accurately diagnosed using current clinical techniques with an approved probe-type confocal laser endoscopy and a random forest (RF) model. Results: A correlation was shown between AC surface stiffness and SCSO (r(rm) = -0.91, 95% CI: -0.97, -0.73), with a significant decrease in surface stiffness observed, particularly in ACs with SCSO typical of early OA (95% CI: string SCSO: 269-173 kPa, double string SCSO: 77-46 kPa). This established SCSO as a visualized and functionally important surrogate marker for AC surface lesions in early OA. Furthermore, stiffness discrimination based on SCSO functioned well in each patient's AC. Next, we demonstrated the feasibility of visualizing SCSO using clinical laser endoscopy and, importantly, accurate SCSO diagnosis using RF. Conclusion: We present a proof-of-concept for early OA pathology detection using available clinical technologies, incorporating future-oriented AI-assisted, non-destructive, quantitative optical biopsy for early disease detection. By putting SCSO recognition into practical use, this approach allows us to examine the correlation between local tissue structure and other experimental and clinical readouts, but clinical validation and larger sample sizes are needed to define diagnostic thresholds.

[0065] ● Hybrid fluorescence AFM explores joint surface degeneration in early osteoarthritis across length scales.[9] Atomic force microscopy (AFM) has become a powerful tool for material characterization at the nanoscale. However, its application to hierarchical biological tissues such as cartilage remains limited. One reason for this is that such samples are typically millimeter-sized, while AFM provides more localized information. Here, we present a combination of AFM and fluorescence microscopy, in which feature regions on millimeter-sized tissue samples are selected at the micrometer scale using fluorescence microscopy and then mapped to nanometer precision by AFM under native conditions. This shows that localized changes in the organization of fluorescently stained cells, markers of early osteoarthritis, correlate with a significant local decrease in elastic modulus, local thinning of collagen fibers, and roughening of the joint surface. This approach is important not only for cartilage but also for the characterization of native biological tissues in general, from the macroscale to the nanoscale. Statement of significance: To understand the function and dysfunction of hierarchically organized biomaterials or tissues, it is necessary to study different length scales. Here, we correlate the micro and nanoscale properties of articular cartilage under native conditions in millimeter-sized samples by combining a highly stable AFM with a fluorescence microscope and a precisely motorized moving mechanism. This is necessary to elucidate the relationship between the microscale organization of chondrocytes, the micrometer-scale changes in articular cartilage properties, and the nanoscale organization of collagen (including D-banding). We expect that such research will pave the way for the guided design of hierarchical biomaterials.

[0066] ● The surface of articular cartilage is damaged in early osteoarthritis due to the loss of thick collagen fibers and the formation of type I collagen.

[10] Osteoarthritis (OA) is a joint disease affecting millions of people worldwide. While articular cartilage is destroyed during the onset and progression of OA, the underlying complex mechanisms remain unclear. Here, we elucidate the changes in collagen fiber thickness and composition at the onset of OA. In articular cartilage exgrafts from the knee joints of OA patients, we found the formation of type I collagen-rich fibrocartilage-like tissue in macroscopically normal cartilage distant from the OA lesion. Importantly, the number of thick fibers (>100 nm) decreased in the early stages of the disease, and these thick fibers completely disappeared in progressive OA. We obtained these results by combining high-resolution atomic force microscopy imaging under near-native conditions, immunofluorescence, scanning electron microscopy, and fluorescence-based classification of the spatial organization of the outermost chondrocytes. Taken together, our data suggest that the loss of tissue function in early OA cartilage is caused by a decrease in type II collagen thick fibers, likely due to the formation of type I collagen-rich fibrocartilage, leading to localized defects in the later stages of OA. We expect that such integrated characterization will be highly beneficial for a deeper understanding of other native biological tissues and for the development of sustainable biomaterials. Statement of Significance: In early osteoarthritis (OA), cartilage appears macroscopically normal. However, this study shows that the collagen network is already altered in early OA due to thinning of collagen fibers and formation of fibrocartilage-like tissue. These nanoscale defects already occur in macroscopically normal areas of the human knee joint and are likely related to processes that lead to weakening of the extracellular matrix. This study deepens our understanding of the earliest progressive cartilage degeneration in the absence of external injury. These results suggest that measuring mean collagen fiber thickness is a new target for early OA detection, and that regulating type I collagen synthesis may open up new pathways for OA treatment.

[0067] The following references are used in the above discussion. [1] Rolauffs, B., et al., Distinct horizontal patterns in the spatial organization of superficial zone chondrocytes of human joints. Journal of Structural Biology, 2008. 162(2): p. 335-344. [2] Rolauffs, B., et al., Proliferative remodeling of the spatial organization of human superficial chondrocytes distant from focal early osteoarthritis. Arthritis Rheum, 2010. 62(2): p. 489-98. [3] Rolauffs, B., et al., Onset of preclinical osteoarthritis: the angular spatial organization permits early diagnosis. Arthritis Rheum, 2011. 63(6): p. 1637-47. [4] Meinhardt, M., et al., Modeling chondrocyte patterns by elliptical cluster processes. 2012. 177(2): p. 447-458. [5] Aicher, W.K. and B. Rolauffs, The spatial organisation of joint surface chondrocytes: review of its potential roles in tissue functioning, disease and early, preclinical diagnosis of osteoarthritis. Ann Rheum Dis, 2014. 73(4): p. 645-53. [6] Felka, T., et al., THE SPATIAL ORGANISATION AND THE PERICELLULAR MATRIX OF HUMAN FOETAL CHONDROCYTES ARE NOT INBORN BUT INSTEAD ACQUIRED CHARACTERISTICS. Bone & Joint Journal Orthopaedic Proceedings Supplement, 2014. 96-B(SUPP 11): p. 323. [7] Felka, T., et al., Loss of spatial organization and destruction of the pericellular matrix in early osteoarthritis in vivo and in a novel in vitro methodology. Osteoarthritis Cartilage, 2016. 24(7): p. 1200-9. [8] Tschaikowsky, M., et al., Proof-of-concept for the detection of early osteoarthritis pathology by clinically applicable endo-microscopy and quantitative AI-supported optical biopsy. Osteoarthritis Cartilage, 2021. 29(2): p. 269-279. [9] Tschaikowsky, M., et al., Hybrid fluorescence-AFM explores articular surface degeneration in early osteoarthritis across length scales. Acta Biomater, 2021. 126: p. 315-325.

[10] Tschaikowsky, M., et al., The articular cartilage surface is impaired by a loss of thick collagen fibers and formation of type I collagen in early osteoarthritis. Acta Biomaterialia, 2022. 146: p. 274-283.

[11] Yantsos, V., et al., Incipient intraepidermal cutaneous squamous cell carcinoma: a proposal for reclassifying and grading solar (actinic) keratoses. Semin Cutan Med Surg 1999;18(1):3-14. (In eng). DOI: 10.1016 / s1085-5629(99)80003-0.

[0068] Various embodiments of the present invention may be implemented, at least in part, in any conventional computer programming language. For example, some embodiments may be implemented in a procedural programming language (e.g., "C") or an object-oriented programming language (e.g., "C++"). Other embodiments of the present invention may be implemented as pre-configured standalone hardware elements and / or pre-programmed hardware elements (e.g., application-specific integrated circuits, FPGAs, and digital signal processors) or other related components.

[0069] In alternative embodiments, the disclosed apparatus and methods (e.g., the flowchart or logic flow described above) may be implemented as a computer program product for use in a computer system. Such an implementation may include a set of computer instructions fixed on a tangible, non-temporary medium such as a computer-readable medium (e.g., a diskette, CD-ROM, ROM, or fixed disk). This set of computer instructions can embody all or part of the system functions described herein.

[0070] Those skilled in the art will understand that such computer instructions can be written in numerous programming languages ​​for use in many computer architectures or operating systems. Furthermore, such instructions can be stored in any memory device, such as tangible non-temporary semiconductor memory, magnetic memory, optical memory, or other memory devices, and can be transmitted over any suitable medium, such as wired (e.g., wires, coaxial cables, fiber optic cables, etc.) or wireless (e.g., in the air or in outer space), using any communication technology, such as optical, infrared, RF / microwave, or other transmission technologies.

[0071] Among other methods, such computer program products may be distributed as removable media (e.g., shrink-wrapped software) with attached printed or electronic documents, pre-loaded onto a computer system (e.g., system ROM or fixed disk), or distributed from a server or electronic bulletin board via a network (e.g., the Internet or the World Wide Web). In fact, some embodiments may be implemented in a Software-as-a-Service ("SAAS") model or a cloud computing model. Naturally, some embodiments of the present invention may be implemented as a combination of both software (e.g., computer program products) and hardware. Yet another embodiment of the present invention may be implemented entirely as hardware or entirely as software.

[0072] Computer program logic implementing all or part of the functions described herein may be executed on a single processor at different times (e.g., simultaneously), or on multiple processors at the same or different times, and may run under a single operating system process / thread or under different operating system processes / threads. Therefore, the term “computer process” generally refers to the execution of a set of computer program instructions, regardless of whether different computer processes run on the same or different processors, or whether different computer processes run under the same operating system process / thread or under different operating system processes / threads. Software systems may be implemented using various architectures, such as monolithic architecture or microservices architecture.

[0073] Importantly, embodiments of the present invention may employ conventional components, such as conventional computers (e.g., commercially available PCs, mainframes, microprocessors), conventional programmable logic devices (e.g., commercially available FPGAs or PLDs), or conventional hardware components (e.g., commercially available ASICs or individual hardware components), which, when programmed or configured to perform the unconventional methods described herein, will produce unconventional devices or systems. Therefore, there are no conventional elements in the inventions described herein, because even if embodiments were implemented using conventional components, the resulting devices and systems would inevitably be unconventional, since conventional components do not inherently perform the unconventional functions described unless specially programmed or configured.

[0074] The effects described and claimed herein provide technical solutions to problems that arise directly in the art. These solutions, as a whole, are not well understood, and are not routine or customary, but in any case, they provide practical applications for modifying and improving computer and computer routing systems.

[0075] While various inventive embodiments are described and illustrated herein, those skilled in the art will readily conceive of various other means and / or structures to perform the functions described herein, obtain the results described herein, and / or obtain one or more of the advantages described herein, and each of such variations and / or modifications will be considered to fall within the scope of the inventive embodiments described herein. More generally, those skilled in the art will readily understand that all parameters, dimensions, materials, and configurations described herein are illustrative, and that actual parameters, dimensions, materials, and / or configurations will depend on the particular use or combination of uses in which the inventive teachings are used. Those skilled in the art will recognize, or be able to recognize, many embodiments equivalent to the particular inventive embodiments described herein without requiring more than routine experimentation. Therefore, it should be understood that the embodiments described herein are presented only as examples, and within the scope of the appended claims and their equivalents, the inventive embodiments may be performed in ways different from those specifically described and claimed. The inventive embodiments of this disclosure are directed to the individual features, systems, articles, materials, kits, and / or methods described herein. Furthermore, any combination of these features, systems, articles, materials, kits, and / or methods is included within the inventive scope of this disclosure, provided that they do not conflict with each other.

[0076] Various inventive concepts can be embodied in one or more methods, and examples are provided. The actions performed as part of a method can be ordered in any suitable manner. Thus, embodiments can be constructed in which the actions are performed in a different order than those illustrated, which may include performing several actions simultaneously, even though they are shown as sequential actions in the exemplary embodiments.

[0077] All definitions defined and used herein should be understood to supersede dictionary definitions, document definitions incorporated by reference, and / or the ordinary meanings of the defined terms.

[0078] As used herein and in the claims, the indefinite articles "a" and "an" should be understood to mean "at least one" unless explicitly stated otherwise.

[0079] As used herein and in the claims, the phrase “and / or” should be understood to mean “either or both” of the elements thus combined, that is, that in some cases the elements exist conjunctively and in other cases they exist disjunctly. Multiple elements listed using “and / or” should be interpreted similarly, that is, they should be interpreted to mean “one or more” of the elements thus combined. There may be elements other than those specifically identified in the “and / or” phrase, and these may or may not be related to the identified elements. Therefore, as a non-restrictive example, a reference to “A and / or B” when used in combination with an open-ended expression such as “comprising” may refer to A only in one embodiment (optionally including elements other than B), B only in another embodiment (optionally including elements other than A), and both A and B in yet another embodiment (optionally including other elements).

[0080] As used herein and in the claims, “or” should be understood to have the same meaning as “and / or” as defined above. For example, when separating items in a list, “or” or “and / or” is interpreted as inclusive, meaning that an item includes at least one of several elements or lists of elements, two or more elements, and additional items that are optionally not included in the list. Only terms that explicitly indicate the opposite meaning, such as “only one of” or “exactly one of” or, as used in the claims, “consisting of”, refer to an item including exactly one element of several elements or lists of elements. In general, the term “or” as used herein is interpreted as indicating an exclusive choice (i.e., “one or the other but not both”) only when preceded by an exclusive term such as “either,” “one of,” “only one of,” or “exactly one of.” The phrase “consisting essentially of” as used in the claims has its usual meaning as used in the field of patent law.

[0081] As used herein and in the claims, the phrase “at least one” used in reference to a list of one or more elements should be understood to mean at least one element selected from any one or more elements contained in the list of elements, and not requiring each element in the list to be included at least once, nor excluding any combination of elements in the list. This definition also allows for the optional presence of elements other than those specifically identified in the list of elements to which the phrase “at least one” refers, whether or not they are related to those specifically identified elements. Therefore, as a non-restrictive example, "at least one of A and B" (or, as a synonym, "at least one of A or B" or as a synonym, "at least one of A and / or B") may, in one embodiment, mean that there is at least one (optionally two or more) A and no B (optionally including elements other than B); in another embodiment, mean that there is at least one (optionally two or more) B and no A (optionally including elements other than A); and in yet another embodiment, mean that there is at least one (optionally two or more) A and at least one (optionally two or more) B (optionally including other elements).

[0082] As used herein and in the claims, transitional phrases such as “comprising,” “including,” “carrying,” “having,” “containing,” “involving,” “holding,” and “composed of” are all understood to be open-ended, meaning “including, but not limited to.” Only the transitional phrases “consisting of” and “consisting essentially of” are closed or semi-closed transitional phrases, respectively, as defined in Section 2111.03 of the U.S. Patent and Trademark Examination Manual.

[0083] Various embodiments of the present invention may be characterized by potential claims enumerated in the paragraphs following this paragraph (and before the actual claims listed at the end of this application). These potential claims constitute part of the specification of this application. Accordingly, the subject matter of the following potential claims may be presented as actual claims in later proceedings in this application or any application claiming priority thereunder. Including such potential claims should not be interpreted as meaning that the actual claims do not cover the subject matter of the potential claims. Accordingly, a decision not to present these potential claims in later proceedings should not be interpreted as transferring the subject matter to the public. Nor are these potential claims intended to limit the various claims pursued.

[0084] Potentially patentable subject matter (preceded with a "P" to avoid confusion with the actual claims presented below) includes, but is not limited to, the following: P1. A method for disease diagnosis, comprising: receiving images of organs / tissues / tissue sections; recognizing organs / tissues / tissue sections / cells in the images against the image background; constructing a region of interest (ROI) in the non-background area; performing automated image analysis to analyze (i) cell segmentation for analysis of segmented cells and (ii) inverted segmentation for analysis of intercellular (matrix) spaces based on the ROI; calculating diverse mathematical parameters at multiple levels for collective use as a universal biomarker array; storing curated (e.g., specific to organs / tissues / tissue sections / cells and disease states) diagnosed / classified / annotated universal biomarker array data in a database to create database-based reference data; and using images / data of interest for statistical and / or AI / ML-assisted examination against database-based reference data for initial diagnosis and localization of disease processes. P2. A system comprising at least one processor and at least one memory containing instructions that, when executed by the at least one processor, cause the system to perform a process including receiving images of organs / tissues / tissue sections; recognizing the background of organs / tissues / tissue sections / cells in the images, constructing a region of interest (ROI) in non-background areas, performing automated image analysis to analyze (i) cell segmentation for analysis of segmented cells and (ii) inverted segmentation for analysis of space between cells (matrixes) based on the ROI; calculating diverse mathematical parameters at multiple levels for collective use as a universal biomarker array; storing curated (e.g., specific to organs / tissues / tissue sections / cells and disease states) diagnosed / classified / annotated universal biomarker array data in a database to create database-based reference data; and using images / data of interest against database-based reference data for statistical and / or AI / ML-assisted examination for initial diagnosis and localization of disease processes.

[0085] While the above discussion discloses various exemplary embodiments of the present invention, it should be apparent to those skilled in the art that various modifications can be made to achieve some of the advantages of the present invention without departing from the true scope of the invention. Any reference to “the present invention” is intended to refer to exemplary embodiments of the present invention and should not be construed as referring to all embodiments of the present invention unless the context suggests otherwise. The embodiments described should be considered in all respects to be illustrative only and not limiting.

Claims

1. A method for diagnosing a disease, Receiving digital tissue images that show tissue details at the cellular level, Automated image analysis is performed on the received digital tissue image to identify cells of interest for analysis, To collectively use a universal biomarker array that includes pattern arrays, distance arrays, morphological arrays, spatial entropy arrays, bin arrays that convert absolute parameter values ​​into relative information by assigning them to specific bin positions, and quartile arrays that convert absolute parameter values ​​into relative information by assigning them to specific quartiles, it is necessary to compute diverse mathematical parameters at multiple levels, For initial diagnosis and localization of the disease process, the universal biomarker array is analyzed based on database-based reference data. Methods that include...

2. The method according to claim 1, wherein the digital tissue image includes an endoscopy digital tissue image.

3. The method according to claim 1, wherein the automated image processing comprises: identifying cells of interest relative to the image background; constructing at least one region of interest (ROI) in a non-background region; and performing (i) cell segmentation for analysis of segmented cells and (ii) inverted segmentation for analysis of intercellular spaces based on the ROI.

4. The method according to claim 1, wherein the array of spatial entropy parameters includes Batty (absolute value, relative value), Contagion, Karlstrom (absolute value, relative value), O Neill (absolute value, relative value), and Parredw parameters.

5. The method according to claim 1, wherein calculating the bin array includes assigning each of several parameters to a class and range-specific bin relative position over the entire range of parameter values, and the total number of bins is calculated by multiplying the number of disease states by the number of bins per state.

6. The method according to claim 5, wherein the conversion of data to class and range-specific bin relative positions is performed based on the parameters of the spatial entropy array, the pattern array, the distance array, and the morphology array.

7. Calculating the aforementioned quartile array is Assigning each of the multiple parameter values ​​to a disease-specific quartile position, Converting each of the multiple parameter values ​​to a disease-specific quartile position and The method according to claim 1, including the method described in claim 1.

8. The method according to claim 7, wherein the transformation of data to class and range-specific quartile relative positions is performed based on the parameters of the spatial entropy array, the pattern array, the distance array, and the morphology array.

9. Analyzing the universal biomarker array based on reference data from a database is possible. To provide the universal biomarker array to an AI / ML system trained on universal biomarker array data, and to detect initial diagnosis and localization of the disease process. The method according to claim 1, including the method described in claim 1.

10. The method according to claim 9, wherein the AI / ML system uses random forest regression to detect initial diagnosis and localization of the disease process.

11. A system for disease diagnosis, At least one computer processor, and when executed by the at least one computer processor, Receiving digital tissue images that show tissue details at the cellular level, Automated image analysis is performed on the received digital tissue image to identify cells of interest for analysis, To collectively use a universal biomarker array that includes pattern arrays, distance arrays, morphological arrays, spatial entropy arrays, bin arrays that convert absolute parameter values ​​into relative information by assigning them to specific bin positions, and quartile arrays that convert absolute parameter values ​​into relative information by assigning them to specific quartiles, it is necessary to compute diverse mathematical parameters at multiple levels, For initial diagnosis and localization of the disease process, the universal biomarker array is analyzed based on database-based reference data. A computer system comprising associated memory containing computer program instructions that execute computer processes including a computer system.

12. The system according to claim 11, wherein the digital tissue image includes an endoscopy digital tissue image.

13. The system according to claim 11, wherein the automated image processing includes identifying cells of interest relative to the image background, constructing at least one region of interest (ROI) in a non-background region, and performing (i) cell segmentation for analysis of segmented cells and (ii) inverted segmentation for analysis of intercellular spaces based on the ROI.

14. The system according to claim 11, wherein the array of spatial entropy parameters includes Batty (absolute value, relative value), Contagion, Karlstrom (absolute value, relative value), O Neill (absolute value, relative value), and Parredw parameters.

15. The system according to claim 11, wherein calculating the bin array includes assigning each of several parameters to a class and range-specific bin relative position over the entire range of parameter values, and the total number of bins is calculated by multiplying the number of disease states by the number of bins per state.

16. The system according to claim 15, wherein the conversion of data to class and range-specific bin relative positions is performed based on the parameters of the spatial entropy array, the pattern array, the distance array, and the morphology array.

17. Calculating the aforementioned quartile array is Assigning each of the multiple parameter values ​​to a disease-specific quartile position, Converting each of the multiple parameter values ​​to a disease-specific quartile position and The system according to claim 11, including the above.

18. The system according to claim 17, wherein the conversion of data to class and range-specific quartile relative positions is performed based on the parameters of the spatial entropy array, the pattern array, the distance array, and the morphology array.

19. Analyzing the universal biomarker array based on reference data from a database is possible. To provide the universal biomarker array to an AI / ML system trained on universal biomarker array data, and to detect initial diagnosis and localization of the disease process. The system according to claim 11, including the above.

20. The system according to claim 19, wherein the AI / ML system uses random forest regression to detect initial diagnosis and localization of the disease process.

21. The method according to claim 1, wherein the digital tissue image includes a digital tissue image from a digital pathology slide scanner.

22. The method according to claim 1, wherein the digital tissue image includes a confocal microscope digital tissue image.

23. The method according to claim 1, wherein the digital tissue image includes a fluorescence microscope digital tissue image.

24. The method according to claim 1, wherein the digital tissue image includes a digital tissue image of stained cells.

25. The method according to claim 1, wherein the digital tissue image includes a digital tissue image of cells treated with a fluorescent material.

26. The method according to claim 25, wherein the fluorescent material comprises fluorescent molecules.

27. The method according to claim 25, wherein the fluorescent material contains an autofluorescent drug.

28. The method according to claim 27, wherein the autofluorescent drug comprises tetracycline, ciprofloxacin, or tinine.

29. The method according to claim 27, wherein the autofluorescent drug is combined with an isotonic solution commonly used in surgery and contains a doxycycline having a divalent cation such as magnesium.

30. The method according to claim 25, wherein the digital tissue image includes a digital tissue image of cells treated with the fluorescent material excited, for example, by UV-A light, in the range of 350 to 400 nm.

Images