Systems and methods for guided diagnostic testing and cloud-based analysis

Abstract

A diagnostic testing and analysis system operating on a network, which may be a cloud-based architecture, includes one or more user terminals for obtaining user instructions, providing diagnostic test recommendations based on user input, a central hub for accessing and ordering one or more tests, two or more data pre-processing modules (DPPMs), each DPPM for receiving raw medical test data and converting the data into a data matrix, an analytical engine for interpreting data from each of the data matrices and generating a report including diagnostic test results and recommendations, wherein the user terminals, clinical test recommendation tool, DPPM, and analytical engine each include interfaces for communicating over the network, and wherein a user may order one or more medical tests.

Description

[Technical Field]

[0001] The present disclosure relates generally to medical diagnostic and prognostic systems, and more particularly to such systems with integrated recommendation tools and cloud-based analytical capabilities. [Background technology]

[0002] Typically, disease diagnosis and prognosis are based on the analysis of a constellation of features, such as symptoms, physical signs, laboratory abnormalities, and radiographic data or images. When a specific diagnosis cannot be derived, the constellation of clinical symptoms is called a syndrome. Based on the diagnosis and prognosis, treatment is administered, and responses to treatment can vary widely. One reason for the variable treatment response is that syndromes may actually represent the manifestation of many diverse disease processes with heterogeneous pathogenic mechanisms. For example, arthritis can be caused by multiple etiologies, and each etiology may have several subtypes. While there are therapies that can be effective for almost any form of arthritis (e.g., prednisone), the optimal treatment for each subtype depends on knowledge of the underlying pathogenesis.

[0003] The scientific basis of disease pathogenesis and subclassification has advanced thanks to improved clinical research, developments in multiparametric analytical technologies, and increasingly sophisticated radiographic studies.

[0004] Precision diagnostics emerging from such converging technologies often rely on high-dimensional input data, such as clinical and pathological data, molecular data from various tissues, and digital radiographic data. For example, many prognostic tools require knowledge of extensive clinical and laboratory data; cancer mutation patterns provide insight into potentially actionable therapeutic targets; diseases can be identified based on patterns of circulating metabolites or proteins; and subtle physical and radiographic features of tumor masses can help distinguish benign from malignant lesions. Such complex biomarkers enable more precise subclassification of disease processes. However, they are also more difficult to interpret, and few tools exist that enable rapid interpretation of clinical, molecular, and radiographic data for important diagnostic signatures. Processing complex datasets relevant to precision diagnostics often requires the use of machine learning (ML) methods, and implementations of ML methods can vary significantly in their structure, capabilities, and methods for disease subclassification. These implementations may require significant processing power and computational resources, a problem that may be particularly evident in applications involving large numbers of diverse samples that are processed in parallel.

[0005] Diagnostic providers face several challenges related to commercializing their products. The available diagnostic tests, many based on complex biomarkers, are expanding rapidly. Physicians cannot keep up with the available tests. Academic and commercial entities are constantly discovering and validating new biomarkers, and the most clinically and economically useful biomarkers are commercialized. The accelerating pace of commercial biomarker development makes it difficult for physicians to keep up. Physicians need to know what tests are available to guide their clinical management. Physicians must have access to information about each test's performance and to evidence of the test's clinical utility to enable knowledgeable interpretation of test results and cost-effective test selection.

[0006] Similarly, patients have difficulty navigating the many diagnostic offerings. Often, patients conduct their own research to identify diagnostic tests that may expedite their medical care. Given patients' often-limited medical knowledge, the process of finding and selecting tests that are valid, reliable, and have a strong impact on a patient's clinical care can be overwhelming. Healthcare costs are increasing, in part, as a result of complex diagnostic testing. While common laboratory tests are relatively inexpensive, newer tests based on complex biomarkers are prohibitively expensive. Therefore, there is a need to ensure that any test paid for by a healthcare system or patient has a strong clinical impact. Furthermore, when several of these more expensive tests are needed, bundling related tests (e.g., based on analytical platform and sample type) is cost-effective.

[0007] Diagnostic tests are becoming more fragmented, and most precision medicine tests are performed in centralized laboratories. The proliferation of biomarkers has led to a proliferation of companies focused on biomarker development. While the adoption of multiple parties accelerates progress, it also leads to fragmentation. Thus, for example, in the current environment, a complete molecular characterization of a tumor may require engaging multiple commercial entities to determine the tumor's molecular subtype, its propensity to recur after surgery, its likelihood of metastasizing to various parts of the body, and its sensitivity to various chemotherapy regimens. A physician must be familiar with each of these tests, and ordering all of them is prohibitively expensive and time-consuming. Currently, most precision medicine tests are performed in centralized laboratories, which can require samples, especially from hospitals, to be sent extremely long distances for testing. This becomes particularly problematic when a physician requires multiple tests performed by different companies, as a single, often limited, sample or dataset may need to be submitted to several disparate geographic locations. As a result, there is an increased risk of biohazard, an increased risk of sample or data loss, increased costs, and the time to obtain and consolidate all test results may be prolonged, negatively impacting clinical care. Summary of the Invention [Means for solving the problem]

[0008] Embodiments of the systems and methods disclosed herein guide physicians and / or patients to select diagnostic tests most likely to have a high clinical impact and be cost-effective, expediting and streamlining the decision-making process. The systems and methods may provide a means to access evidence regarding each test's performance characteristics and clinical utility, including informing users about stages of regulatory approval and insurance coverage. Embodiments of the present disclosure may identify tests that can be grouped together to reduce the need to transport samples to multiple central laboratories and for cost efficiency. Embodiments of the present disclosure may include an analytics engine that provides the ability to rapidly process complex data sets for analysis and interpretation and delivers test reports within a short period of time; embodiments of the present disclosure may enable decentralization of some tests, limiting the need to transport samples over significant distances and increasing the efficiency of testing and reporting of test results.

[0009] Some embodiments of the systems and methods disclosed herein guide test selection, provide centralized test ordering and interpretation, generate corresponding test results, or a combination thereof. Embodiments of the present disclosure may include a clinical test recommendation tool, which may include a study and clinical test library and may be linked to a laboratory information system (LIS), a radiology information system (RIS), and / or an electronic medical record (EMR) system.

[0010] In some embodiments of the present disclosure, a system for use by a user over a network includes a user terminal for obtaining one or more instructions from the user; two or more data pre-processing modules (DPPMs), each DPPM for obtaining a sample including raw medical test data and / or receiving the raw medical test data and converting the medical test data into a data matrix; and an analytical engine including a plurality of diagnostic tests, for performing a diagnostic test from the plurality of diagnostic tests on each of the data matrices, providing diagnostic test results, providing one or more recommendations based on the diagnostic test results, and generating a report including the diagnostic test results and the one or more recommendations, wherein the user terminal, the DPPM, and the analytical engine each include an interface for communicating over the network, and wherein the user may order one or more medical tests.

[0011] Some embodiments of the present disclosure relate to a system that operates on a cloud-based computer architecture, where the analytics engine includes allocated, scalable computing power.

[0012] Some embodiments of the present disclosure relate to diagnostic tests that include one or more machine learning (ML) methods.

[0013] Some embodiments of the present disclosure relate to a user terminal that includes a user interface for operation of the system by a user.

[0014] In some embodiments of the present disclosure, the user terminal includes a clinical test recommendation tool.

[0015] Some embodiments of the present disclosure relate to a lab test recommendation tool for communicating with at least one of an LIS, a RIS, and an EMR database or system.

[0016] In some embodiments, the lab test recommendation tool includes clinical data and a lab test library of diagnostic tests.

[0017] In some embodiments of the present disclosure, the lab test recommendation tool includes a survey questionnaire.

[0018] In some embodiments of the present disclosure, the one or more recommendations include a recommendation for additional testing.

[0019] In some embodiments of the present disclosure, the system further includes a network storage device for storing the data matrix and the results of the diagnostic test.

[0020] In some embodiments of the present disclosure, the raw medical test data includes one or more of tissue samples, fluid samples, and imaging data.

[0021] In some embodiments of the present disclosure, the diagnostic test comprises immunohistochemical staining with one or more selected antibodies.

[0022] In some embodiments of the present disclosure, the diagnostic test comprises determining a molecular signature.

[0023] In some embodiments of the present disclosure, the diagnostic examination includes radiographic imaging analysis to characterize features of the radiographic image.

[0024] In some embodiments of the present disclosure, the diagnostic test involves mapping the epigenome to assess one or more epigenetic modifications of the genome.

[0025] In some embodiments of the present disclosure, the diagnostic test comprises assessing chromatin accessibility.

[0026] In some embodiments of the present disclosure, the diagnostic test involves evaluating the proteome.

[0027] In some embodiments of the present disclosure, the diagnostic test comprises assessing the metabolome.

[0028] In some embodiments of the present disclosure, a method includes obtaining a sample including raw medical test data and / or receiving the raw medical test data, converting the raw medical test data into a data matrix, performing a diagnostic classification based on the data matrix to provide diagnostic test results to assist in assessing the patient's health, providing one or more recommendations based on the diagnostic test results, and generating a report based on the diagnostic test results and the one or more recommendations.

[0029] In some embodiments of the present disclosure, the diagnostic test comprises a method of ML.

[0030] In some embodiments of the present disclosure, the step of providing one or more recommendations includes accessing at least one of an LIS, a RIS, and an EMR database or system.

[0031] In some embodiments of the present disclosure, providing the one or more recommendations includes accessing a clinical data and clinical test library.

[0032] In some embodiments of the present disclosure, providing one or more recommendations includes asking a survey question.

[0033] In some embodiments of the present disclosure, providing one or more recommendations includes recommending additional testing.

[0034] In some embodiments of the present disclosure, the method further comprises storing the data matrix.

[0035] In some embodiments of the present disclosure, the method further comprises storing the results of the one or more diagnostic tests.

[0036] In some embodiments of the present disclosure, the raw medical test data includes one or more of data from a tissue sample, data from a fluid sample, and imaging data.

[0037] In some embodiments of the present disclosure, the diagnostic test comprises immunohistochemical staining with one or more selected antibodies.

[0038] In some embodiments of the present disclosure, the diagnostic test comprises determining a molecular signature.

[0039] In some embodiments of the present disclosure, diagnostic testing includes radiological image analysis to characterize features in x-rays, computed tomography scans, magnetic resonance imaging scans, and other radiographic tests.

[0040] In some embodiments of the present disclosure, the diagnostic test involves mapping the epigenome to assess one or more epigenetic modifications of the genome.

[0041] In some embodiments of the present disclosure, the diagnostic test comprises assessing chromatin accessibility.

[0042] In some embodiments of the present disclosure, the diagnostic test involves evaluating the proteome.

[0043] In some embodiments of the present disclosure, the diagnostic test comprises assessing the metabolome.

[0044] For a more complete understanding of the present disclosure, reference is made to the following description and accompanying drawings. [Brief explanation of the drawings]

[0045] [Figure 1] FIG. 1 is a schematic diagram of a computerized guided diagnosis and integrated analysis system according to some embodiments of the present disclosure. [Figure 2]FIG. 2 is a schematic diagram illustrating a simplified hardware structure of a computing device of the guided diagnostic and integrated analysis system shown in FIG. 1. [Figure 3] FIG. 2 is a schematic diagram illustrating a simplified software architecture of the computing device of the guided diagnostic and integrated analysis system shown in FIG. 1. [Figure 4] FIG. 2 is a schematic diagram illustrating the functional structure of the guided diagnostic and integrated analysis system shown in FIG. 1 implemented as a cloud-based system according to an embodiment of the present disclosure. [Figure 5A] 1 is a flow diagram illustrating the initial steps in an embodiment of a survey module. [Figure 5B] 5B is a flow diagram showing additional steps of the research module of FIG. 5A for cancer-related diseases. [Figure 5C] 5B is a flow diagram illustrating additional steps of the research module of FIG. 5A for respiratory-related disorders. [Figure 6] FIG. 1 is a schematic diagram illustrating an exemplary embodiment of a data pre-processing module. [Figure 7] FIG. 1 is a schematic diagram illustrating a cloud-based analytics engine. [Figure 8] 1 is a flow chart illustrating steps of an embodiment of a diagnostic and analytical method. DETAILED DESCRIPTION OF THE INVENTION

[0046] Unless otherwise defined, all technical and scientific terms used herein generally have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains. To facilitate understanding of the subject matter of this disclosure, exemplary terms are defined below.

[0047] The term "a" or "an" refers to one or more of that entity; for example, "a terminal" refers to one or more terminals or at least one terminal. Thus, the terms "a" (or "an"), "one or more," and "at least one" are used interchangeably herein. Furthermore, reference to an element or feature by the indefinite article "a" or "an" does not exclude the possibility that more than one of the element or feature is present, unless the context clearly requires that only one of the element is present. Furthermore, unless clearly intended, reference to a feature in the plural (e.g., systems) does not imply that the systems or methods disclosed herein must include a plurality.

[0048] The expression "and / or" refers to and includes all possible combinations (e.g., one or the other, or both) of one or more of the associated listed items, as well as the lack of combinations when interrupted in the alternative (or).

[0049] Some embodiments of the systems and methods disclosed herein allow users to order tests, provide raw medical test data, and receive results therefrom. In some embodiments of the systems and methods disclosed herein, diagnostic tests are selected, ordered, and analyzed based on one or more multiparametric assays to provide personalized clinical care and guide patients or physicians to select appropriate clinically relevant tests that inform clinical management. The guidance provided may include consideration of the following user inputs: patient characteristics, including demographics; clinical features related to disease categories; clinical questions that may inform clinical management; commercial availability of tests; sample type; and analysis platform. In some embodiments disclosed herein, a clinical test recommendation tool includes a survey that can be completed by the patient or physician and may include a clinical test library. The clinical test recommendation tool may be linked to a laboratory information system (LIS), a radiology information system (RIS), and / or an electronic medical record (EMR) system to further inform guidance related to diagnostic tests. Test recommendations that are personalized and appropriate for a particular patient may then be generated.

[0050] In some embodiments disclosed herein, the systems and methods include collecting and / or acquiring medical test data. In some embodiments disclosed herein, the collecting or acquiring includes streaming the medical test data from an instrument performing sample analysis in real time as the medical test data is created or as a bulk payload after an experiment has completed a batch test or batches of tests. In some embodiments disclosed herein, the systems and methods provide guidance to a user regarding which tests may be appropriate.

[0051] Embodiments of the systems and methods disclosed herein provide an integrated system. In some embodiments disclosed herein, the user interface includes a dashboard for a user to order one or more recommended tests suggested by the clinical test recommendation tool. In some embodiments disclosed herein, the system is cloud-based and includes an analytical engine, a key vault library, and a reporting module that may be connected to machine learning (ML) capabilities, as well as an audit system and associated database. In some embodiments disclosed herein, the analytical engine includes a bank of tests, including test criteria and test objectives, which may be manually curated or manually / electronically curated. In some embodiments disclosed herein, the integrated analytical system further includes one or more data pre-processing modules (DPPMs) that may accept test requests directly from the user interface or through the clinical test recommendation tool.

[0052] In some embodiments of the systems and methods disclosed herein, a user is guided to identify one or more appropriate clinical tests and then order, run, and process the tests. In some embodiments disclosed herein, not all of these steps are required. For example, a user may simply use a recommendation tool to identify appropriate tests. Alternatively, a user may simply order, run, and process the tests without the recommendation tool.

[0053] A user may be a patient or a non-professional. For example, a patient may have a serious cancer with few treatment options and want to identify and request tests that identify actionable therapeutic targets.

[0054] The user may also be a clinician. For example, the clinician may have some available clinical and laboratory information that may be used to predict cerebrovascular accident (stroke) or raise concerns about an underlying autoimmune disease. The raw medical data needed for these inquiries may be provided to the system, and the user selects tests based on the specific question. Alternatively, there may be a specific set of tests required. The raw medical data may be analyzed using the corresponding predictive method, and a test report may be generated.

[0055] The user may be in a laboratory where biological samples are tested. For example, a tumor may be subjected to next generation sequencing (NGS) to identify genomic variants and transcriptomic features. Optionally, blood, plasma, or serum samples may be analyzed by mass spectrometry to characterize proteome or metabolome. The raw molecular data and linked clinical data may be preprocessed and uploaded to the cloud. A test corresponding to a particular predictive method may be selected. The data may then be analyzed in the context of that method, and test results may be generated.

[0056] The user may be a radiologist having difficulty interpreting radiological examinations. For example, a CT scan may show a lesion in the liver that is difficult to determine whether it is benign or malignant. Digital Imaging and Communications in Medicine (DICOM) data or a variant thereof may be provided to the system. A predictive method designed to distinguish between benign and malignant liver lesions may be used to generate an examination result.

[0057] In some embodiments disclosed herein, a user is guided to select clinically appropriate tests from a clinical test library, which may be based on criteria derived from a survey question. In some embodiments disclosed herein, a recommendation tool includes a survey that uses a question-branching method to elicit information that guides the selection of potentially appropriate tests from a clinical test library. Data from the LIS, RIS, and / or EMR system may be used to inform the recommendation tool. Alternatively, the recommendation tool may be primarily informed by the results of the survey question. Data elicited from the survey may include one or more of the following: Demographics: age, gender, or sex. Available samples: tissue biopsy, blood, urine, saliva, sputum, stool, imaging data, routine lab test. Disease categories: for example, cancer, hematology, skin, rheumatology, digestive, respiratory, kidney, and heart. · The type of information derived from the report and / or test: for example, diagnostic, prognostic, or predictive information.

[0058] The clinical test library is a repository of commercially available tests that is updated regularly. In some embodiments disclosed herein, data associated with each test is used to match patient needs (based on recommendations from a recommendation tool survey) with each test's characteristics. In other embodiments, a user and / or physician may select one or more tests without using a recommendation tool. The clinical test library includes the following data elements: test indication (disease type, test purpose), test category (early disease detection and early detection of recurrence, disease diagnosis and subclassification, treatment response detection, identification of actionable treatment targets, prognosis, and treatment prediction), sample type (e.g., blood, tissue, urine, saliva, stool, imaging data, routine clinical test), level of evidence, regulatory stage (e.g., research use only, laboratory developed test, regulatory approval from a specific jurisdiction), insurance coverage determination, test website, and relevant published literature.

[0059] A guidance report is generated that includes a list of tests that may be appropriate for the current clinical problem. The guidance report may include information that can be reviewed by the user and / or physician. The guidance report may include the following information: name of the test, indication for the test, scientific rationale for the test, sample type / data required to perform the test, analysis platform (e.g., whole transcriptome or targeted gene expression panel, mutation panel, single nucleotide polymorphism (SNP) panel, metabolomic assay, proteomic assay, computed tomography (CT) scan, magnetic resonance imaging (MRI) scan, light microscopy imaging), clinical evidence and level of evidence, published literature, regulatory stage, insurance coverage determination, website associated with the test, and instructions for ordering the test.

[0060] In some embodiments disclosed herein, a cloud-based analytics engine includes multiple containers (e.g., Docker® containers) residing in the cloud, each container containing a diagnostic method corresponding to a specific diagnostic test. In some embodiments disclosed herein, the method includes a predictive model with specific clinical applications such as disease classification, identifying events with common attributes (clustering), forecasting, identifying outliers, and tracking perturbations over time. The model can be based on any prediction or classification or regression method. Examples of predictive methods include linear regression, logistic regression, random forests, neural networks, gradient boosted models, K-means clustering, and generative adversarial networks. The method corresponding to the predictive model (and including the diagnostic test) can be stored in an encrypted, frozen format in the cloud-based container, where "frozen" refers to code that cannot be modified once in the container, although updated versions can be installed and tracked. The cloud-based analysis engine may be a central repository of clinically useful methods, guiding users to select a diagnostic test or function, then upload the relevant data and run the selected method.

[0061] In some embodiments disclosed herein, the cloud-based analytical engine is linked to a clinical laboratory or radiology suite equipped with a DPPM that communicates with the LIS, RIS, or EMR and simultaneously processes data generated by the analytical instruments. In some embodiments disclosed herein, the system and method include an analytical instrument and acquires raw data. Examples of analytical instruments include next-generation sequencers, mass spectrometers, magnetic resonance imaging devices, computed tomography devices, and microscopic imaging devices. In some embodiments disclosed herein, the DPPM processes the raw data from the analytical instrument and converts the raw data into a data matrix in a format suitable for uploading to the cloud-based analytical engine. The system architecture disclosed herein allows for scaling, with one or more DPPMs operatively connected to each user. The use of a data matrix allows for synchronous processing of batches containing data from multiple patients.

[0062] The data used for the predictive model may include molecular data such as mutation data, transcriptome data, circular ribonucleic acid (RNA) expression, microRNA (miRNA) expression, long non-coding RNA (lncRNA) expression, epigenomic data, proteome data, metabolomic data, or any other molecular data that may be obtained from a subject sample. Similarly, clinical data, pathology data, and laboratory data may be included. Digital imaging data, such as radiographic data or images from pathology slides, may be used as input. The dashboard allows the user to request diagnostic tests, and the user can select one or more diagnostic tests that process the appropriate input data using relevant methods.

[0063] Embodiments of the systems and methods disclosed herein determine whether uploaded data associated with one or more patients matches known diagnostic signatures and deliver a report to an ordering physician, lab, or other designated individual. Embodiments of the systems and methods disclosed herein can quickly interpret complex, high-dimensional data to generate a test report that is sent to a user and / or interested parties. Some embodiments of the systems and methods disclosed herein include sufficient security features to enable them to handle sensitive and private data. In some embodiments disclosed herein, a clinical report with test results and interpretation is generated, and the test results are sent to an ordering user, EMR, LIS, and / or RIS.

[0064] The development of instruments capable of simultaneously quantifying multiple parameters has increased the complexity of diagnostic tests. Multiple data points can be collected to derive test results, and interpretation of such diagnostic tests requires converting a wide variety of physical characteristics into an appropriate digital format. Diagnostic tests can utilize a wide variety of data types and provide results in a wide variety of data formats.

[0065] For example, during pathology examination of tissue samples, it is common to perform immunohistochemical staining with selected antibodies to make a diagnosis. Staining intensity is an important consideration in diagnosis. Recently, this has been done using image analysis. Images are converted to DICOM, and then staining intensity can be quantified across the entire tissue section or in selected regions. Similarly, immunofluorescence microscopy images can be evaluated and quantified. Applications of this technology include establishing tissue diagnoses, detecting invasive tumors, and identifying and characterizing biologically important features such as inflammatory infiltrates.

[0066] Many transcriptome-based molecular signatures have been described. These include signatures that allow for disease subclassification, signatures associated with specific biological functions, and prognostic signatures. Methods for quantifying the expression levels of mRNA, miRNA, and non-coding RNA use next-generation sequencing (e.g., RNASeq, microarrays, etc.). Sequence data are converted into text file formats such as FASTQ, which must be further processed for downstream bioinformatics data analysis.

[0067] Radiography image analysis is used to precisely characterize features from radiographic images such as X-rays, CT scans, MRI scans, and other radiographic tests. DICOM images are generated, and the images can be classified by the pattern of features observed.

[0068] Some diagnostic tests are based on the ability to map the epigenome. These tests evaluate epigenetic modifications to the genome, including histone modifications, chromatin remodeling, and methylation and / or hydroxymethylation of cytosine bases. Epigenetic modifications to the genome reflect the genetic influence of environmental exposures and disease. The epigenome can be characterized by high-throughput technologies such as methylation arrays, which quantitatively examine methylation sites. Methylation data are sometimes expressed as beta values ​​(the ratio of methylated probe intensity to the total intensity, which is the sum of methylated and unmethylated probe intensity). Other technologies have different data output formats.

[0069] Chromatin accessibility can be assessed using a technology called assay for transposase-accessible chromatin with high-throughput sequencing (ATAC-Seq), which uses next-generation sequencing. Identifying patterns associated with accessible deoxyribonucleic acid (DNA) regions has proven useful for disease subclassification and investigating disease biology. The data generated requires a count matrix of the number of reads per open chromatin region. Another technology may be chromatin immunoprecipitation sequencing (ChIP-sequencing), which combines chromatin immunoprecipitation with parallel DNA sequencing to analyze protein interactions with DNA.

[0070] Proteomics is used for the diagnosis, subclassification, and prognosis of disease. Many technologies are available for targeted or untargeted analysis of the proteome, including the identification of post-translational modifications. Untargeted proteomics uses various types of mass spectrometry to create a "spectrum" of features, such as mass-to-charge ratios, that correspond to different protein fragments. The magnitude of the peaks corresponds to the protein abundance. Targeted proteomics involves antibody-based quantification of proteins.

[0071] Metabolomics has been used for disease diagnosis, subclassification, and prognosis. Depending on the physicochemical properties of the metabolites of interest, various mass spectrometry-based platforms are used to characterize the metabolome. Similar to untargeted proteomics, spectra are generated with features corresponding to metabolites, and peak magnitudes correspond to abundances.

[0072] Any of these radiographic and molecular features can be integrated to create a more comprehensive understanding of disease subclassification and associated biology. To benefit from such integration, users need both an understanding of the different tests and convenient access to those tests. Because tests are performed in disparate locations by various companies and there is limited access to information about the potential value of each test, patients generally have difficulty finding tests that provide useful information about their health problems. Medical professionals generally have difficulty keeping up with the rapidly increasing number of available tests, often do not fully understand the science behind each test, and ordering tests from multiple different suppliers is time-consuming.

[0073] Typically, ML methods are required to interpret these multiparametric tests, customized for the test's specific indications and / or purposes. To implement such methods, the raw medical data generated by each test must be compiled into a simplified format that can be used as input for the method. With the rapid development and expansion of these complex tests, it has become increasingly difficult to coordinate the tests, as they are typically performed in highly specialized facilities. One solution is to centralize data interpretation. This is difficult due to the diverse data formats generated by each test. In some embodiments disclosed herein, a DPPM compiles each data format into a simplified matrix that serves as input for the ML method housed in a central hub.

[0074] example The following examples illustrate how users may benefit from embodiments of the present disclosure.

[0075] Example 1: A patient undergoes a (virtual) CT colonoscopy to screen for colorectal cancer. There is a lesion in the colon that could represent a polyp, cancer, or a stool fragment. The radiologist's uncertainty prompts them to access ML methods present in the engine. Data from the DICOM file is converted into an analyte matrix by a pre-processing module, and then the data is uploaded to the analytics engine. The lesion has the characteristics of cancer. A colonoscopy is performed, which confirms this. The patient undergoes resection, and pathology confirms the diagnosis. RNA is extracted from the tumor and then subjected to whole-transcriptome RNASeq to assess prognosis. Two separate tests are performed based on the whole-transcriptome sequencing data, requiring processing by two separate methods. One test determines that the prognosis is poor, and the second test determines that this molecular subtype has a high incidence of lymph node metastasis. The pathologist did not identify lymph node metastases by visual inspection, so he stained the lymph nodes for cytokeratin and then created images stored in DICOM file format. The data was preprocessed and then uploaded to an engine to access ML methods capable of identifying extremely small metastases. Two micrometastases were identified. The patient was referred to a clinical oncology department for chemotherapy consideration. The oncologist accessed a guide to identify tests that would help determine the best chemotherapy regimen. Five chemosensitivity tests were performed to assess the appropriateness of different chemotherapy regimens: two based on RNA-Seq, two based on methylome assessment, and one based on ATAC-Seq. The whole-transcriptome data from the first RNA-Seq was processed by two separate ML methods to perform the first two chemosensitivity tests. DNA was isolated from the tumor and then subjected to next-generation sequencing to perform methylation sequencing and ATAC-Seq. The diverse data formats generated by the tests are compiled into a simplified analyte matrix and then analyzed in an analytical engine using the appropriate diagnostic method.

[0076] Example 2: A patient's health is deteriorating. Associated signs and symptoms are nonspecific and include progressive generalized weakness, muscle weakness, mild cognitive decline, and joint pain. The physician consults a guide, which suggests blood tests to analyze the proteome, metabolome, and methylome. The diverse and complex data formats are simplified and compiled into an analyte matrix. The data from each of these tests is compiled and uploaded to an analytics engine. The compiled data suggests that the symptoms are due to Lyme disease. There is no evidence of an autoimmune etiology.

[0077] Example 3: A patient has a mammogram showing an equivocal lesion in the right breast. The radiologist recommends an MRI scan, which is also equivocal based on the radiologist's visual inspection. The radiologist accesses the clinical test recommendation tool, which instructs them to upload DICOM data from the MRI to access ML methods to assist in characterizing the lesion. The test results indicate that the lesion is likely malignant. An excision biopsy is performed, which shows ductal carcinoma in situ. The pathologist orders special stains including cytokeratin, estrogen receptor, progesterone receptor, and HER2 receptor. The pathologist wants to confirm the absence of an invasive component and uploads the DICOM images to access ML image analysis, which shows an invasive component. Treatment is then customized for invasive breast cancer. The clinical test recommendation tool provides a means to analyze the transcriptome for biomarkers that aid in disease subclassification and prognosis. Additional tests are recommended for consideration. Blood is analyzed for circulating tumor cells. Serum proteome and metabolome are analyzed to estimate the likelihood of occult metastatic disease. Chemosensitivity is assessed using an ATAC-Seq-based test. Data from all of these tests are converted into a data matrix by a pre-processing module. The matrix data is uploaded to a central hub to access the corresponding classification methods present in the system.

[0078] Example 4: A patient has a malignant lesion in the liver, and routine pathology cannot definitively determine the type of cancer. The physician completes a survey, which determines that the 58-year-old man has cancer of unknown etiology and that local treatments (e.g., surgery, radiation, ablation) and systemic therapy are treatment considerations. The clinical test recommendation tool generates a report recommending tests, including whole-transcriptome analysis using RNASeq. The recommended tests are tests designed to determine cancer type, prognosis, and sensitivity to specific chemotherapy agents. RNASeq data, consisting of sequenced reads stored in FASTQ files, are processed by the relevant DPPM. A gene expression matrix is ​​generated, which is then uploaded to a cloud-based analytics engine for further processing and prediction by applying selected methods to the data matrix (in this case, corresponding to the gene expression matrix). A test report is generated describing the cancer type, prognosis, and sensitivity to relevant chemotherapy agents. These tests are all based on gene expression signatures.

[0079] Example 5: A breast cancer patient has undergone surgery and the doctor wants to know the prognosis. The doctor fills out a survey that includes a recommendation tool. The recommendation tool allows the doctor to enter staging data. Several tests based on gene expression signatures are recommended, as well as a test related to chemotherapy sensitivity. The doctor selects a test for prognosis based on a panel of genes quantified with a cDNA microarray. The microarray data and important clinical features are converted into a data matrix, specifically an analyte matrix, by the associated DPPM. The combination of demographic and staging data improves the prognostic accuracy of genomic risk.

[0080] Example 6: A radiologist identifies a pancreatic mass on a CT scan. It is suspicious of pancreatic adenocarcinoma, but could also represent a benign diagnosis such as autoimmune pancreatitis. The radiologist wants to access an AI (deep learning) method that provides an estimate of the tumor's likelihood of being malignant. The radiologist is guided by a series of questions in the recommendation tool. DICOM data is processed and corresponding tests are recommended for how to distinguish between benign and malignant pancreatic tumors. The DICOM data is processed by DPPM, generating an imaging data matrix, which is then uploaded to a cloud-based analytics engine. Features are compared to diagnostic signatures previously identified by the artificial intelligence, housed in an encrypted, cloud-based container. The DICOM image features are compared to the signatures of various benign and malignant pancreatic masses, and it is determined that the mass likely represents pancreatic adenocarcinoma.

[0081] Example 7: A cardiologist is responsible for a patient who is believed to be at high risk for myocardial infarction. A clinical test recommendation tool recommends a test that involves serially monitoring a set of clinical and laboratory variables over time. The test is based on a clever combination of perturbations that predict myocardial infarction. When a test is ordered, data is acquired from the electronic medical record (EMR) or a wearable activity / fitness tracker (e.g., Apple® watch, Garmin® watch), and the DPPM transforms the data into a data matrix. The data is processed in a cloud-based analytics engine. A report is generated indicating whether the combination of perturbations (signature) raises concern of an impending myocardial infarction.

[0082] Example 8: A patient is diagnosed with a rare cancer type and there is little data related to the efficacy of chemotherapy. The oncologist suggests administering a chemotherapy regimen that has some activity based on case reports. The patient is hesitant because the chemotherapy regimen has some troublesome toxicities. The patient accesses the recommendation tool and completes a survey. The recommendation tool's output is a report listing the specific tests recommended, the test's indications and characteristics, each test's stage of regulatory approval, relevant literature and publications, and insurance coverage. The recommendation tool recommends whole-genome analysis (to look for actionable mutations) and transcriptome-based molecular pathway analysis (to identify upregulated, targetable molecular signatures). The patient shares the report with their oncologist for further discussion.

[0083] Computing Architecture Embodiments of the guided diagnostic testing and integrated analysis systems and methods disclosed herein may be implemented in a variety of computer network architectures. Referring to FIG. 1 , a computer network system for project and program management is shown and generally identified using reference numeral 100. As shown, the computer network system 100 includes one or more server computers or analysis engines 102, a plurality of computing devices or terminals 104 for interfacing with users, and a plurality of computing devices or terminals 106 for acquiring medical raw data and performing data preprocessing, each operatively interconnected by a network 108, such as the Internet, a local area network (LAN), a wide area network (WAN), or a metropolitan area network (MAN), via appropriate wired and wireless network connections. In some embodiments disclosed herein, other devices 103, such as a RIS, LIS, EMR, data warehouse, key vault library, and audit device, may also be connected to the network 108.

[0084] The analysis engines 102 may be computing devices specially designed for use as servers and / or general-purpose computing devices that may be used by a variety of users but also act as server computers. Each analysis engine 102 may execute one or more server programs.

[0085] The terminals 104, 106 may be portable and / or non-portable computing devices, such as laptop computers, tablets, smartphones, personal digital assistants (PDAs), desktop computers, etc. Each terminal 104, 106 may run one or more client application programs, sometimes referred to as "apps."

[0086] Generally, the analysis engine 102 and the terminals 104, 106 have similar hardware structures, such as the hardware structure 120 shown in Figure 2. As shown, the analysis engine 102 and the terminals 104, 106 include a processing structure 122, a control structure 124, one or more non-transitory computer-readable memory or storage devices 126, a network interface 128, an input interface 130, and an output interface 132, operatively interconnected by a system bus 138. The analysis engine 102 and the terminals 104, 106 may also include other components 134 coupled to the system bus 138.

[0087] The processing structure 122 may be one or more single-core or multi-core computing processors (also referred to as "central processing units" (CPUs)), such as an INTEL® microprocessor (INTEL is a registered trademark of Intel Corp., Santa Clara, CA, USA), an AMD® microprocessor (AMD is a registered trademark of Advanced Micro Devices Inc., Sunnyvale, CA, USA), an ARM® microprocessor (ARM is a registered trademark of Arm Ltd., Cambridge, UK) manufactured by various manufacturers such as Qualcomm of San Diego, California, USA based on the ARM® architecture. When the processing structure 122 includes multiple processors, the processors of the processing structure 122 may cooperate via dedicated circuitry, such as a dedicated bus, or via a system bus 138.

[0088] The processing structure 122 may also include one or more real-time processors, programmable logic controllers (PLCs), microcontroller units (MCUs), μcontrollers (UCs), specialized / customized processors and / or controllers using, for example, field programmable gate array (FPGA) or application specific integrated circuit (ASIC) technology, and the like.

[0089] Generally, each processor of processing structure 122 includes the necessary circuitry, implemented using technology such as electrical and / or optical hardware components, to execute one or more processes to perform various tasks depending on the purpose and / or use case of the implementation. In many embodiments, one or more processes may be implemented as firmware and / or software stored in memory 126 and executed by one or more processors of processing structure 122. Those skilled in the art will understand that in these embodiments, one or more processors of processing structure 122 are typically completely useless without meaningful firmware and / or software.

[0090] For example, each processor of the processing structure 122 may include semiconductor-implemented logic gates for performing various operations, calculations, and / or processes. Examples of logic gates include AND gates, OR gates, XOR (exclusive OR) gates, and NOT gates, each of which takes one or more inputs and generates or otherwise provides an output from the inputs based on the logic implemented therein. For example, a NOT gate receives an input (e.g., a high voltage, a current-carrying state, a light-emitting state, etc.), inverts the input (e.g., to form a low voltage, a no-current state, a light-emitting state, etc.), and provides the inverted input as an output.

[0091] Although the inputs and outputs of logic gates are generally physical signals, and the logic or processing of a logic gate is a tangible operation with a physical result (e.g., the output of a physical signal), the inputs and outputs of logic gates are generally described using numbers (e.g., the numbers "0" and "1"), and the operation of logic gates is generally described as "computing" (hence the name "computer" or "computing device"), or "calculation," or more broadly "processing" to generate or bring about an output from the inputs of the logic gate.

[0092] Sophisticated combinations of logic gates in the form of circuits of logic gates, such as one or more processors of processing structure 122, may be formed using multiple AND, OR, XOR, and / or NOT gates. Such combinations of logic gates may be implemented using discrete semiconductors or, more often, are implemented as integrated circuits (ICs).

[0093] A circuit of logic gates may be a "hard-wired" circuit that, once designed, may only perform the task for which it was designed. In this example, the task is "hard-coded" into the circuit.

[0094] As technology advances, circuits of logic gates, such as one or more processors of processing fabric 122, are alternatively implemented as firmware and / or software, often designed in a generic manner such that the circuitry may perform various tasks according to a set of "programmed" instructions stored in memory 126. In this example, circuits of logic gates, such as one or more processors of processing fabric 122, are typically completely useless without meaningful firmware and / or software.

[0095] Of course, those skilled in the art will appreciate that the processor may be implemented using other technologies, such as analog technology.

[0096] The control structure 124 includes one or more control circuits, such as a graphics controller, input / output chipsets, etc., for coordinating the operation of the various hardware components and modules of the analysis engine 102 and terminals 104, 106.

[0097] The memory 126 includes one or more non-transitory computer-readable storage devices or media accessible by the processing structure 122 and the control structure 124 for reading and / or storing instructions for execution by the processing structure 122, and for reading and / or storing data, including input data and data generated by the processing structure 122 and the control structure 124. The memory 126 may be volatile and / or non-volatile non-removable or removable memory, such as RAM, ROM, EEPROM, solid-state memory, hard disk, CD, DVD, flash memory, etc. During use, the memory 126 is generally divided into multiple portions for different purposes. For example, one portion of the memory 126 (referred to herein as storage memory) may be used for long-term data storage, e.g., for storing files or databases. Another portion of the memory 126 may be used as system memory (referred to herein as working memory) for storing data being processed.

[0098] The network interface 128 may be any of a variety of wireless technologies, including Ethernet, WI-FI (registered trademark) (WI-FI is a registered trademark of Wi-Fi Alliance, Austin, TX, USA), BLUETOOTH (registered trademark) (BLUETOOTH is a registered trademark of Bluetooth Sig Inc., Kirkland, WA, USA), Bluetooth Low Energy (BLE), Z-Wave, Long Range (LoRa), ZIGBEE (registered trademark) (ZIGBEE is a registered trademark of ZigBee Alliance Corp., San Ramon, CA, USA), Global System for Mobile Communications (GSM), Code Division Multiple Access (CDMA), Universal Mobile Telecommunications System (UMTS), Worldwide Interoperability for Microwave Access (WiMAX), CDMA2000, Long Term Evolution (LTE), 3GPP, 5G New Radio (5G and one or more network modules for connecting to other computing devices or networks through network 108 by using a suitable wired or wireless communication technology, such as a wireless broadband communication technology, such as 5G networks (NW), 5G (Radio Frequency), and / or 5G (5G Internetworking) networks. While parallel ports, serial ports, USB connections, optical connections, etc. are typically thought of as input / output interfaces for connecting input / output devices, in some embodiments they may also be used to connect other computing devices or networks.

[0099] The input interface 130 includes one or more input modules for one or more users to input data via, for example, a touch screen, a touch whiteboard, a touchpad, a keyboard, a computer mouse, a trackball, a microphone, a scanner, a camera, etc. The input interface 130 may be a physically integrated part of the analysis engine 102 or the terminal 104, 106 (e.g., a touchpad on a laptop computer or a touch screen on a tablet), or it may be a device that is physically separate from but functionally coupled to other components of the analysis engine 102 or the terminal 104 (e.g., a computer mouse). The input interface 130 may, in some implementations, be integrated with a display output to form a touch screen or a touch whiteboard.

[0100] The output interface 132 includes one or more output modules for outputting data to a user. Examples of output modules include a display (such as a monitor, LCD display, LED display, or projector), speakers, a printer, a virtual reality (VR) headset, or an augmented reality (AR) goggle. The output interface 132 may be a physically integrated part of the analysis engine 102 or the terminal 104, 106 (e.g., a laptop computer or tablet display), or may be a device that is physically separate from but functionally coupled to other components of the analysis engine 102 or the terminal 104 (e.g., a desktop computer monitor). The analysis engine 102 or the terminal 104 may also include other components 134, such as one or more positioning modules, temperature sensors, barometers, inertial measurement units (IMUs), etc.

[0101] A system bus 138 interconnects the various components 122 to 134 and allows those components to send and receive data and control signals to and from each other.

[0102] From a computational perspective, the analysis engine 102 and the terminals 104, 106 may include multiple modules. As used herein, a "module" is a descriptive term referring to a hardware structure, such as a circuit, implemented using technologies such as electrical and / or optical technologies (and semiconductors being a more particular example) to perform defined operations or processes. A "module" may alternatively refer to a combination of hardware and software structures, where the hardware structure may be implemented using technologies such as electrical and / or optical technologies (and semiconductors being a more particular example) in a conventional manner to perform defined operations or processes in accordance with a software structure in the form of a set of instructions stored on one or more non-transitory computer-readable storage devices or media.

[0103] As part of a device, apparatus, system, etc., a module may be coupled or integrated with other parts of the device, apparatus, or system such that the combination forms the device, apparatus, or system. Alternatively, a module may be implemented as a stand-alone device or apparatus.

[0104] 3 shows a simplified software architecture 160 of the analysis engine 102 or terminal 104, 106. The software architecture 160 includes an application layer 162, an operating system 166, a logical input / output (I / O) interface 168, and a logical memory 172. The application layer 162, the operating system 166, and the logical I / O interface 168 are generally implemented as computer-executable instructions or code in the form of software or firmware programs stored in the logical memory 172 that may be executed by the processing fabric 122.

[0105] As used herein, a software or firmware program is a set of computer-executable instructions or code stored in one or more non-transitory computer-readable storage devices or media, such as memory 126, that may be read and executed by the processing structure 122 and / or other suitable components of the analysis engine 102 and terminals 104, 106 to perform one or more processes. Those skilled in the art will understand that a program may be implemented as either software or firmware depending on design objectives and requirements. Therefore, for ease of description, the terms "software" and "firmware" may be used interchangeably hereinafter.

[0106] As used herein, process has a broad meaning equivalent to that of method and does not necessarily correspond to the concept of a computing process (which is an instance of an executing computer program). More specifically, a process herein is a defined method implemented as a software or firmware program executable by a hardware component to process data (such as data received from a user, another computing device, other components of the analysis engine 102 or terminals 104, 106, etc.). A process may include or use one or more functions to process data as designed. As used herein, a function is a defined sub-process or sub-method to operate, calculate, or otherwise process input data in a defined manner and generate or otherwise result in output data.

[0107] Alternatively, the processes may be implemented as one or more hardware structures having the necessary electrical and / or optical components, circuits, logic gates, integrated circuit (IC) chips, and the like.

[0108] Referring again to FIG. 3, application layer 162 includes one or more application programs 164 that are executed or performed by processing structure 122 to perform various tasks.

[0109] The operating system 166 manages the various hardware components of the analysis engine 102 or the terminal 104 via a logical I / O interface 168, manages logical memory 172, and manages and supports the application programs 164. The operating system 166 also communicates with other computing devices via the network 108 to enable the application programs 164 to communicate with programs running on other computing devices (not shown). As one skilled in the art would appreciate, the operating system 166 may be any suitable operating system, such as MICROSOFT® WINDOWS® (MICROSOFT and WINDOWS are registered trademarks of Microsoft Corp., Redmond, WA, USA), APPLE® OS X, APPLE® iOS (APPLE is a registered trademark of Apple Inc., Cupertino, CA, USA), Linux, ANDROID® (ANDROID is a registered trademark of Google Inc., Mountain View, CA, USA), etc. The analysis engines 102 and terminals 104 of the computer network system 100 may all have the same operating system, or may have different operating systems.

[0110] Logical I / O interface 168 includes one or more device drivers 170 for communicating with each input interface 130 and output interface 132 to receive data therefrom and transmit data thereto. Received data may be sent to application layer 162 for processing by one or more application programs 164. Data generated by application programs 164 may be sent to logical I / O interface 168 (via output interface 132) for output to various output devices.

[0111] Logical memory 172 is a logical mapping of physical memory 126 for easy access by application program 164. In this embodiment, logical memory 172 includes a storage memory area that may generally be mapped to non-volatile physical memory, such as a hard disk, solid-state disk, or flash drive, for long-term data storage. Logical memory 172 also includes a working memory area that may generally be mapped to high-speed, and in some implementations, volatile, physical memory, such as RAM, for application program 164 to temporarily store data during program execution. For example, application program 164 may load data from the storage memory area into the working memory area and store data generated during its execution in the working memory area. Application program 164 may store some data in the storage memory area as needed or in response to a user command.

[0112] As discussed above, processing fabric 122 is typically completely useless without meaningful firmware and / or software. Similarly, computer network system 100 may have the potential to perform various tasks, but without meaningful firmware and / or software, it cannot perform any tasks and is completely useless. As explained in more detail below, computer network system 100 described herein, as a combination of hardware and software, generally produces tangible results that are tied to the physical world, and tangible results such as those described herein may lead to improvements in the computer and the system itself.

[0113] In some embodiments disclosed herein, the guided diagnostic testing and integrated analysis systems and methods are configured to operate on a cloud-based computer architecture, where the analysis engine 102 is a virtual computing environment that includes allocated, scalable computer power, sometimes referred to as an instance. In some embodiments, each analysis engine 102 may have a variable configuration of CPU, memory, storage, and networking capabilities. In some embodiments, each analysis engine 102 is assigned a unique Internet Protocol (IP) address within the network 108. The processing, memory, storage, and networking capabilities may be provided by a single physical analysis engine 102 or a more complex computer architecture that includes multiple interconnected service, storage, and network components. Cloud-based computer architectures may be provided by several cloud computer service platforms, such as Amazon Web Services® or AWS® (Amazon Web Services and AWS are registered trademarks of Amazon Web Services, Inc., a subsidiary of Amazon, Seattle, Washington, USA), Microsoft Azure™ (Azure is a trademark of Microsoft Corporation, Redmond, Washington, USA), and Google Cloud Platform or GCP.

[0114] In embodiments disclosed herein, the systems and methods for guided diagnostic testing and integrated analysis allow raw data from analytical devices to be preprocessed and then converted or transformed into a data matrix. Examples of analytical devices include thermal cyclers, next-generation sequencers, mass spectrometers, complementary DNA (cDNA) microarrays, protein arrays, magnetic resonance imaging (MRI) devices, and computed tomography (CT) scanners. Data from analytical devices is converted into a format compatible with disease classification using prediction methods included in the analysis engine 102. Data is organized into a matrix by patient identifier. Simultaneous testing of multiple patients is enabled, and data arrays may include data from multiple patients, allowing batches of tests to be performed.

[0115] In some embodiments disclosed herein, the data matrix is ​​sent to a cloud-based analysis engine 102. The cloud-based analysis engine 102 examines the data for important features. Some embodiments disclosed herein are configured to handle data from different analysis platforms, including mutation and genomic sequence variant data, transcriptional data (mRNA and non-coding RNA), epigenomic data, proteomic data, and metabolomic data. In some embodiments disclosed herein, test performance is improved by augmentation with clinical data, and digital radiographic data can be analyzed.

[0116] In some embodiments disclosed herein, the cloud-based analytics engine 102 contains a repository of clinically important predictive methods and may include multiple containers or vaults within the ML environment. The containers may be organized into a container library, with each container containing a diagnostic method and key-encrypted, locked predictive methods. Containers may be versioned to manage revisions or updates to diagnostic and / or locked predictive methods. The predictive methods accessed for data analysis depend on the test ordered. Each test may be based on complex, potentially high-dimensional data that may require the development of specific predictive methods for test interpretation. In some embodiments disclosed herein, test reports are automatically generated and delivered.

[0117] 4 , in embodiments disclosed herein, a system 400 includes a cloud-based analytical engine 402, one or more user terminals 404, and one or more system terminals or DPPMs 424 connected via a network 408. In some embodiments disclosed herein, each user terminal 404 includes a user dashboard 412, a reporting module 414, and a clinical test recommendation tool. In some embodiments disclosed herein, each DPPM 424 includes a multiplexed analytical platform 426 for analyzing one or more analyte matrices and a system dashboard 410.

[0118] In some embodiments disclosed herein, the clinical test recommendation tool 416 is accessible to users through a secure browser-accessible portal and includes a survey module 418, a clinical test library 420, and a guidance report module 422. In some embodiments disclosed herein, the recommendation tool 416 is linked to the LIS 432, the RIS 430, and / or the EMR 434. The survey module 418 uses a series of questions to help define the patient's diagnostic needs. In some embodiments disclosed herein, the user is posed survey questions, such as, but not limited to, a branching format, where the answers to the questions determine the next question. The clinical test library 420 contains information derived from the survey and is used to match diagnostic needs with the characteristics of tests contained in the library. The clinical test library 420 may be manually curated and updated.

[0119] 5A through 5C illustrate survey questions for some exemplary embodiments disclosed herein. More specifically, FIG. 5A illustrates some initial questions regarding demographics, tissue samples, and disease categories before entering into more detailed survey branches related to specific disease groupings, such as FIG. 5B for cancer and FIG. 5C for respiratory diseases. Referring to FIG. 5A , in block 502, demographic data about the patient, such as age and gender, may be collected. In block 504, information regarding the type of tissue sample being provided, such as biopsy, blood, urine, saliva, or sputum, may be collected. In block 506, information regarding the disease category, such as cancer, blood, skin, rheumatism, gastrointestinal, respiratory, renal, or cardiac, may be collected. Once initial information such as blocks 502 through 506 has been obtained, more detailed survey branches may be entered.

[0120] 5B shows an exemplary embodiment of cancer-related survey branching and result reporting. In block 508, the survey module 418 prompts the user to indicate whether the patient desires or requires cancer screening. If the patient desires or requires cancer screening, specific risk factors are received in block 510, and recommendations and / or screening test results are provided in block 512. In block 514, the survey module 418 prompts the user as to whether the patient is suspected of cancer. If the patient is suspected of cancer, the suspected site of cancer is received in block 516, and recommendations, test results, and / or results regarding whether the tumor is benign or malignant are provided in block 518. In block 520, the survey module 418 prompts the user as to whether the test is related to a recent diagnosis of cancer. If the test is related to a recent diagnosis of cancer, the survey module 418 prompts the user as to whether the type of cancer is known in block 522. If the type of cancer is not known, recommendations and test results identifying the cancer type and subtype are provided in block 524. If the type of cancer is known, information regarding whether the cancer is primary, whether there is cancer histology, and / or whether there is metastasis is received in block 526. The survey module 418 may then prompt the user to indicate whether the patient is disease-free after cancer treatment in block 528. If the patient is not disease-free after treatment, recommendations and predictive tests are provided in block 534. If the patient is disease-free after treatment, recommendations, prognostic test results, and / or test results for early detection of recurrence are provided in block 530. In block 532, the survey module 418 prompts the user regarding whether adjuvant chemotherapy is being considered. If adjuvant chemotherapy is being considered, recommendations and predictive test results are provided in block 532.

[0121] 5C shows an exemplary embodiment of a branch of investigations related to pulmonology. In block 536, the investigation module 418 prompts the user to indicate whether the problem is cancer-related. If the problem is not cancer-related, in block 538, a diagnosis such as pneumonia, pneumonitis, pulmonary fibrosis, and / or pulmonary hypertension is obtained, and in block 540, recommendations and test results are provided to provide insight into subtype, etiology, and / or treatment susceptibility. If block 536 is answered affirmatively, in block 542, the investigation module 418 prompts the user whether the patient is at risk for lung cancer, and if the patient is at risk, recommendations and screening test results are provided in block 544. If the answer to block 542 is no, the survey module 418 prompts the user whether the patient is suspected of having cancer in block 546, and if the patient is suspected of having cancer, recommendations, test results, and / or results regarding whether the tumor is benign or malignant are provided in block 548. If the answer to block 546 is no, the survey module 418 prompts the user whether the test is relevant to a recent diagnosis of cancer in block 550. If yes, the survey module prompts the user whether the cancer histology and / or metastasis is present in block 552. In block 554, the survey module prompts the user whether the patient is disease-free after cancer treatment in block 554. If the patient is disease-free after treatment, recommendations, prognostic test results, and / or tests for early detection of recurrence are provided in block 556, and the user is prompted whether adjuvant chemotherapy is being considered in block 558. If adjuvant chemotherapy is being considered or the answer to block 554 is negative, then in block 560, recommendations and / or predictive test results are provided.

[0122] In some embodiments disclosed herein, the guidance report module 422 generates a report describing diagnostic tests that would potentially be beneficial to the patient. Laboratory tests determined to be relevant to the patient's diagnostic needs are recommended, with priority given to tests for which samples / data are readily available. In some embodiments disclosed herein, the guidance report also indicates which tests are included in the cloud-based analytics engine 402. The user can then access the user dashboard 412 to order the test. Upon submitting the test request, the laboratory or supplier can process the order and obtain the appropriate sample or data set to perform the test.

[0123] In some embodiments disclosed herein, the user dashboard 412 is for ordering tests and includes a portal gateway for ordering diagnostic tests included in the cloud-based analytic engine 402 and tests available in other laboratories. Through the portal, a user (e.g., laboratory personnel, physician, radiologist, etc.) can access a menu of diagnostic tests in the cloud-based analytic engine 402. A user (patient or physician) may be directed to the portal after completing a survey. Alternatively, a user may be directed to the user dashboard 412 without accessing the lab test recommendation tool 416. In either case, the user may be able to specify the diagnostic tests to be performed on the data contained in the data matrix. That is, the diagnostic test selected by the user directs the matrix file to the corresponding prediction method in the key vault. In some embodiments disclosed herein, ordering a test initiates the following processes: collection of personal and demographic information, consent to access related samples and relevant medical information, and organization of the testing process.

[0124] In some embodiments disclosed herein, the DPPM 424 processes raw medical data from analytical devices (e.g., DNA / RNA sequencers, mass spectrometers, and imaging diagnostic devices) to create a data matrix (or other structured file format) that is uploaded to the cloud-based analytical engine 402. In some embodiments disclosed herein, the DPPM 424 connects to user terminals 402 (e.g., clinical laboratories or radiology areas) and the cloud-based analytical engine 402 through a system dashboard 410. The DPPM 424 may be customizable based on specific use cases. For example, the DPPM 424 is customizable based on the analytical platform and device and links to information systems that can be customized. The DPPM 424 may link to the LIS 432, the RIS 430, and / or the EMR 434 to obtain demographic, clinical, and related data, as well as patient identifiers. In some embodiments disclosed herein, the DPPM 424 creates a first data matrix file that includes key clinical information and patient identifiers. The DPPM 424 converts raw medical data from the analytical instrument into a second data matrix file. For example, a transcription FASTQ file from next-generation sequencing is converted into a gene expression matrix, a whole genome sequencing (WGS) or whole exome sequencing (WES) FASTQ file is converted into a VCF annotation file, mass spectrometry peaks are converted into a matrix file (or a condensed version thereof) containing proteins or metabolites, or a DICOM file is converted into a matrix containing variables that inform a prediction method. The second data matrix file is separate from the first data matrix file containing clinical information and patient identifiers. The first and second data matrix files are sent to the cloud-based analysis engine 402 and processed using the selected prediction method.In some embodiments disclosed herein, data quality control (QC) analysis is performed before assembling the data matrix file. The details of the QC checks, including the type and criteria of the checks, are determined based on the type of data and the specific use. Referring to FIG. 6, an exemplary embodiment of DPPM 424 for processing data from a sequencer is shown. Method 600 includes step 602 of reading available FASTQ files from a shared volume or shell as input, step 604 of performing adapter trimming, step 606 of performing quality control on each paired file, step 608 of summarizing the quality control report, step 610 of performing alignment-free mapping, step 612 of aggregating transcript expression vectors into a gene expression matrix, step 614 of importing quality control index data into the gene expression matrix, and step 616 of exporting the gene expression matrix to the shared volume.

[0125] In some embodiments disclosed herein, the system dashboard 410 is for communicating with the user terminal 404 and the cloud-based analytical engine 402. In some embodiments disclosed herein, the system dashboard 410 includes a portal gate to the user terminal 404 and the cloud-based analytical engine 402. In some embodiments disclosed herein, the portal is for uploading data matrix files created by the DPPM 424 to the cloud-based analytical engine 402. Referring to FIG. 7 , an exemplary embodiment of the architecture of the cloud-based analytical engine 720 and its connection with the DPPM 706 is shown. More specifically, the relationship between the DPPM 706, the cloud-based analytical engine 720, the ML function 736, and the key vault library 734 is shown. Referring to FIG. 7 , the system 700 includes a lab 710 and a cloud-based analytical engine 720. The lab 710 includes a system dashboard 712, an order file 714 (the order file is information from an LIS, RIS, or EMR that links the data in the data matrix to a specific individual or patient), and a gene matrix file 716, and is connected to a cloud-based analysis engine 720, as well as a laboratory information management system (LIMS) 702, a DPPM 706, which is interfaced with a next-generation sequencer 704. The cloud-based analysis engine 720 includes cloud file storage 722, a cloud app service 724, a reporting module 726, and a database 728. The cloud-based analysis engine 720 is further connected to an active directory 730, a lab file share 732, and ML functionality 736.

[0126] In some embodiments disclosed herein, a cloud-based analytical engine 402 manages and controls the flow of data to an ML function corresponding to a test ordered by a user. The cloud-based analytical engine 402 may be accessible through a browser-accessible portal and controls and audits data flow to ensure compliance with applicable standards, including ISO 13485, the Health Insurance Portability and Accountability Act (HIPAA), the Personal Information Protection and Electronic Documents Act (PIPEDA), the General Data Protection Regulation (GDPR), and regulatory standards specific to the user's jurisdiction (e.g., the Food and Drug Administration (FDA) in the United States and Health Canada in Canada). The cloud-based engine 402 may be connected to an ML function 442 that contains one or more containers. When a diagnostic test is selected, the container in the ML function 442 identifies which method is relevant and then requests access to the method-specific private key needed to decrypt the method corresponding to the test. Each diagnostic method is "frozen" and may include additional QC checks. In some embodiments disclosed herein, the cloud-based analytical engine 402 includes version control for when diagnostic methods are updated. In some embodiments disclosed herein, the cloud-based analytics engine 402 provides information to the reporting module 414 and instructs the reporting module 414 on what test results should be generated. The ML function 442 is a set of one or more containers (or pods), each capable of performing a prediction or classification task based on a requested test by executing the relevant method of the requested test. In some embodiments disclosed herein, the ML function 442 can access the key vault library 438.

[0127] In some embodiments disclosed herein, key vault library 438 is a library of cloud-based key vaults managed and operated by cloud-based analytic engine 402. Key vault library 438 is invoked by ML functions 442. Access to each key vault is enabled using a private key shared only with ML functions 442, which allows encrypted diagnostic methods to be decrypted for purposes of performing selected diagnostic tests.

[0128] In some embodiments, a network storage device or data warehouse 436 is connected to the network 408 and contains data related to system access and performance, as well as data contained in the data matrix.

[0129] In some embodiments disclosed herein, the reporting module 414 generates a patient-specific test report to provide an interpretation of the results of running the data through the diagnostic method. The reporting module 414 delivers the patient test report to the user and designated recipients.

[0130] In some embodiments disclosed herein, an audit system 440 logs and records the flow of data, including test orders, data processed by the DPPM 424, analysis of data processed by the DPPM, logs of the ML functions 442, logs of the cloud-based analytics engine 402, and generation / delivery of test reports.

[0131] The embodiments disclosed herein are scalable, capable of handling large volumes of tests from a variety of locations, and also capable of accommodating libraries of hundreds of thousands of clinically important predictive methods.

[0132] The embodiments disclosed herein include important security features essential for clinical diagnostics. In some embodiments disclosed herein, the systems and methods comply with ISO 13485, which is required for regulatory approval, and also comply with applicable privacy laws (ensuring data privacy and security to protect medical information). In some embodiments, the locked diagnostic method resides on a secure server and is protected by a unique encryption key.

[0133] 8 is a flow diagram illustrating steps of a method 800 according to one embodiment of the present disclosure. Method 800 begins with obtaining a sample containing medical test data or receiving raw medical test data (step 802). In step 804, the raw medical data is converted into a data matrix. Optionally, in step 806, the data matrix is ​​stored. In step 808, a diagnostic test is performed on the data matrix using ML methods to provide diagnostic test results to assist in assessing the patient's health. Optionally, in step 810, the results of the diagnostic test are stored. In step 812, a report based on the results of the diagnostic test is generated.

[0134] Although several embodiments have been shown and described with reference to the accompanying drawings, it will be understood by those skilled in the art that various changes and modifications can be made to these embodiments without altering or departing from their scope, spirit, or function as defined by the appended claims. The terms and expressions used in the foregoing detailed description are used herein as terms of description and not as terms of limitation, and the use of such terms and expressions is not intended to exclude equivalents of the features shown and described or portions thereof. [Explanation of symbols]

[0135] 100 Computer Network Systems 102 Server Computer or Analysis Engine 103 Other Devices 104 Computing Device or Terminal 106 Computing Device or Terminal 108 Network 120 Hardware Structure 122 Processing Structure 124 Control Structures 126 Non-transitory computer-readable memory or storage device 128 network interfaces 130 Input Interface 132 output interface 134 Other Components 138 System Bus 160 Software Architecture 162 Application Layer 164 Application Programs 166 Operating Systems 168 logical input / output (I / O) interfaces 170 Device Drivers 172 logical memory 400 System 402 Cloud-based analytics engine 404 User terminal 408 Network 410 System Dashboard 412 User Dashboard 414 Reporting Module 416 Clinical Examination Recommendation Tools 418 Survey Module 420 Clinical Test Library 422 Guidance Report Module 424 system terminal or DPPM 426 Multiplexed Analysis Platform 430 RIS 432 LIS 434 EMR 436 Network Storage Device or Data Warehouse 438 Key Vault Library 440 Audit System 442 ML features 600 ways 700 System 702 Laboratory Information Management System (LIMS) 704 Next-generation sequencer 706 DPPM 710 Lab 712 System Dashboard 714 Order File 716 Gene Matrix File 720 Cloud-based Analytics Engine 722 Cloud File Storage 724 Cloud App Services 726 Reporting Module 728 databases 730 Active Directory 732 Lab File Share 734 Key Vault Library 736 ML Features 800 ways

Claims

1. 1. A system for use by users over a network, comprising: a user terminal for obtaining one or more instructions from the user; two or more Data Pre-Processing Modules (DPPMs), each DPPM for receiving raw medical test data and converting the medical test data into a data matrix; 1. An analytical engine comprising a plurality of diagnostic tests, performing a diagnostic test from the plurality of diagnostic tests on each of the data matrices and providing a diagnostic test result; providing one or more recommendations based on said diagnostic test results; an analytics engine for generating a report including the diagnostic test results and the one or more recommendations; The system wherein the user terminal, the DPPM, and the analysis engine each include an interface for communicating over the network, and wherein the user may order one or more medical tests.

2. The system of claim 1 , operating on a cloud-based computer architecture, wherein the analytics engine comprises allocated, scalable computing power.

3. 3. The system of claim 1 or 2, wherein the diagnostic test comprises one or more machine learning (ML) methods.

4. 4. The system of claim 1, wherein the user terminal includes a user interface for user operation of the system.

5. The system of any one of claims 1 to 4, wherein the user terminal includes a clinical test recommendation tool.

6. 6. The system of claim 5, wherein the laboratory recommendation tool is for communicating with at least one of a laboratory information system, a radiology information system, and an electronic medical record database.

7. 7. The system of claim 5 or 6, wherein the lab test recommendation tool includes clinical data and a lab test library of diagnostic tests.

8. 8. The system of claim 5, wherein the clinical test recommendation tool comprises a survey question.

9. 9. The system of claim 5, wherein the one or more recommendations include a recommendation for additional testing.

10. 10. The system of any one of claims 1 to 9, further comprising a network storage device for storing the data matrix and the results of the one or more diagnostic tests.

11. The system of any one of claims 1 to 10, wherein the raw medical test data includes one or more of tissue samples, fluid samples, and imaging data.

12. 12. The system of claim 1, wherein the diagnostic test comprises immunohistochemical staining with one or more selected antibodies.

13. 13. The system of any one of claims 1 to 12, wherein the diagnostic test comprises determining a molecular signature.

14. 14. The system of claim 1, wherein the diagnostic test comprises radiographic image analysis to characterize features of the radiographic image.

15. 15. The system of any one of claims 1 to 14, wherein the diagnostic test comprises mapping the epigenome to assess one or more epigenetic modifications of the genome.

16. 16. The system of any one of claims 1 to 15, wherein the diagnostic test comprises assessing chromatin accessibility.

17. 17. The system of any one of claims 1 to 16, wherein the diagnostic test comprises evaluating the proteome.

18. 18. The system of any one of claims 1 to 17, wherein the diagnostic test comprises assessing the metabolome.

19. receiving raw medical test data; converting the raw medical test data into a data matrix; performing a diagnostic test on the data matrix to provide a diagnostic test result to assist in assessing the patient's health; providing one or more recommendations based on the diagnostic test results; generating a report based on the results of the diagnostic test and the one or more recommendations; A method comprising:

20. 20. The method of claim 19, wherein the diagnostic test comprises a method of ML.

21. 21. The method of claim 19 or 20, wherein providing one or more recommendations includes accessing at least one of a laboratory information system, a radiology information system, and an electronic medical record database.

22. 22. The method of any one of claims 19 to 21, wherein providing one or more recommendations comprises accessing a library of clinical data and clinical tests.

23. 23. The method of any one of claims 19 to 22, wherein providing one or more recommendations comprises asking a survey question.

24. 24. The method of any one of claims 19 to 23, wherein providing one or more recommendations includes recommending additional testing.

25. 25. The method of any one of claims 19 to 24, further comprising the step of storing the data matrix.

26. 26. The method of any one of claims 19 to 25, further comprising storing the results of one or more diagnostic tests.

27. 27. The method of any one of claims 19 to 26, wherein the raw medical test data comprises one or more of data from a tissue sample, data from a fluid sample, and imaging data.

28. 28. The method of any one of claims 19 to 27, wherein the diagnostic test comprises immunohistochemical staining with one or more selected antibodies.

29. 30. The method of any one of claims 19 to 29, wherein the diagnostic test comprises determining a molecular signature.

30. 31. The method of any one of claims 19 to 30, wherein the diagnostic examination comprises radiological image analysis to characterize features in X-rays, computed tomography scans, magnetic resonance imaging scans, and other radiographic examinations.

31. 32. The method of any one of claims 19 to 31, wherein the diagnostic test comprises mapping the epigenome to assess one or more epigenetic modifications of the genome.

32. 33. The method of any one of claims 19 to 32, wherein the diagnostic test comprises assessing chromatin accessibility.

33. 34. The method of any one of claims 19 to 33, wherein the diagnostic test comprises evaluating the proteome.

34. 35. The method of any one of claims 19 to 34, wherein the diagnostic test comprises assessing the metabolome.

Images