Neurofilament light chain biomarker composition and method of use thereof

Abstract

A method, a non-transient computer-readable medium, and a computer system for predicting cardiac disease with improved efficiency and accuracy are disclosed. In some examples, the technique involves obtaining a biomarker profile containing a first level of one or more biomarkers in a first sample taken from a subject. The one or more biomarkers include nerve filament light chains (NfLs). A first machine learning model is applied to the input to generate a biomarker score for the subject. The input includes at least the biomarker profile, and the biomarker score indicates the probability that the subject is at risk of or has cardiac disease. The biomarker score, or an index of cardiac disease determined from the biomarker score, is then output to a computer device via one or more communication networks in response to the biomarker profile.

Description

[Technical Field]

[0001] This application claims priority to U.S. Provisional Patent Application No. 63 / 490,493, filed on 15 March 2023, and incorporates its entire contents by reference.

[0002] The disclosed technologies generally relate to molecular biology and cardiovascular health, and more specifically to neuronal filament light chain biomarker compositions and methods of using them. [Background technology]

[0003] Heart disease is a major health issue in the United States, being the leading cause of death for men, women, and most racial and ethnic groups. For example, one person dies from cardiovascular disease every 34 seconds in the United States. In 2020, approximately 697,000 people died from heart disease in the United States, accounting for one in five deaths. Heart disease cost the United States approximately $229 billion annually from 2017 to 2018, including the costs of healthcare services, medicines, and productivity losses due to death. [Overview of the Initiative] [Problems that the invention aims to solve]

[0004] Current methods for detecting cardiac disease are inconvenient, unpleasant, and / or inaccurate. Therefore, there is a need in the art for compositions, kits, and methods for detecting cardiac disease with an acceptable level of specificity and in a manner that aids patient compliance. [Means for solving the problem]

[0005] This disclosure covers heart disease (e.g., acute coronary syndrome, aortic aneurysm, aortic dissection, aortic stenosis, arrhythmia, atherosclerosis, atrial fibrillation, coronary artery disease (e.g., angina pectoris (e.g., stable or unstable angina) or heart attack), cardiac arrhythmia, cardiomyopathy (e.g., dilated, hypertrophic, proarrhythmic or restrictive cardiomyopathy), cardiitis (e.g., endocarditis, infective endocarditis, myocarditis, pericarditis, pancardiitis, or regurgitant cardiomyitis), congenital heart defects or The present invention provides compositions, kits, methods, and computer systems for detecting and / or identifying heart disease, eosinophilic myocarditis, heart failure, heart murmurs, heart valve disease, heart valve stenosis, hypertension, hypertensive heart disease, inflammatory cardiomegaly, Kawasaki disease, myocardial infarction, Marfan syndrome, metabolic syndrome, peripheral artery disease (PAD), rheumatic heart disease, thromboembolism, transthyretin amyloidosis (ATTR-CM), venous thrombosis, or heart valve disease.

[0006] Furthermore, this disclosure provides a method for classifying patients with cardiac disease compared to normal individuals. Furthermore, this disclosure recognizes that certain biomarkers (e.g., nerve filament light chains (NfLs)) are useful for the detection and / or diagnosis of cardiac disease, and / or the classification of patients with cardiac disease. This disclosure further recognizes that these biomarkers (e.g., NfLs) contained in compositions, kits, and methods are useful for the classification, detection, and / or diagnosis of cardiac disease without the need for invasive or expensive testing, which represents a significant advance in patient care. In particular, cardiac disease can be detected and identified by methods of this technology that are more comfortable for patients, cause less harm to patients, and / or reduce the time required for patient recovery after detection and / or diagnosis.

[0007] Furthermore, this disclosure provides that certain biomarkers (e.g., NfL) are useful for detecting cardiac disease with improved sensitivity. This disclosure provides that certain biomarkers (e.g., NfL) are useful for detecting cardiac disease with improved selectivity. This disclosure further provides that a combination of one or more demographic factors, particularly age and / or sex, and NfL, is unexpectedly effective for detecting cardiac disease with improved sensitivity and / or specificity. Furthermore, this disclosure further provides that a combination of one or more image-based biomarkers, particularly left ventricular septal thickness and / or ejection fraction, and NfL, is remarkably effective for detecting cardiac disease with improved sensitivity and / or specificity.

[0008] The increased sensitivity achieved with the biomarkers described herein (e.g., NfL) reduces the number of false negatives obtained in the detection and / or identification of cardiac disease. This reduction in false negatives helps ensure that more patients with cardiac disease receive early treatment that is crucial for mitigating the signs, symptoms, and conditions associated with cardiac disease, thereby promoting long-term survival for patients with cardiac disease.

[0009] Accordingly, several examples of methods for predicting cardiac disease implemented by computer systems are described herein. In some examples, the technique involves obtaining a biomarker profile containing a first level of one or more biomarkers in a first sample taken from a subject. The one or more biomarkers include nerve filament light chains (NfLs). A first machine learning model is applied to the input to generate a biomarker score for the subject. The input includes at least the biomarker profile, and the biomarker score indicates the probability that the subject is at risk of or has cardiac disease. The biomarker score, or an index of cardiac disease determined from the biomarker score, is then output to a computer device via one or more communication networks in response to the biomarker profile.

[0010] In one example, the method further includes obtaining demographic data corresponding to one or more demographic factors of a subject from a computer device via a communication network. The input in these examples further includes demographic data, the demographic factors including at least one or more of the subject's sex or age. In another example, one or more image-based biomarkers are obtained for a subject from a computer device via a communication network. The input in these examples further includes image-based biomarkers, the image-based biomarkers including at least left ventricular septal thickness or ejection fraction.

[0011] In further examples, one or more acquired images related to a subject are analyzed to determine image-based biomarkers, and one or more measurements are obtained. In these examples, the analysis includes pattern recognition performed on one or more acquired images, or segmentation and classification using a first machine learning model or a second machine learning model with one or more configurations extracted from the acquired images.

[0012] In some examples, machine learning models are derived from bagging, boosting, or additive decision trees, neural boosting, bootstrap forests, boosted trees, or support vector machines. In other examples, the first machine learning model is a cardiac disease classifier, which further includes classifying whether a subject has cardiac disease or not based on biomarker scores. The cardiac disease indicators include binary indicators determined based on the classification in these examples. In some cases, one or more of the classifier's sensitivity or specificity exceeds 80%.

[0013] In further other examples, the subject is a human subject, and the first sample includes one or more of the following: blood, serum, plasma, or cardiac tissue. Furthermore, in some examples, at least one of the following is determined based on a biomarker score: (i) the subject's cardiac disease status, (ii) whether the subject has cardiac disease, or (iii) the probability that the subject has cardiac disease.

[0014] In another example, the first machine learning model can be trained using acquired training data that, in some cases, includes biomarker profiles of multiple subjects and the health status of the subjects corresponding to those biomarker profiles. In these examples, each biomarker profile includes at least two biomarker levels, including NfL.

[0015] In some examples, the first sample is obtained from the subject at a first time, and the method further includes determining whether the subject is at risk of or has heart disease if the level of a second NfL detected in a second sample obtained from the subject at a second time after the first time exceeds the level of the first NfL by a threshold amount. In other examples, the first sample is obtained from the subject at a first time, and the method further includes determining whether the subject is at risk of or has heart disease based on a change in the first NfL level determined based on a comparison of the level of the first NfL with the level of the second NfL detected in a second sample obtained from the subject at a second time after the first time.

[0016] In yet other examples, the first NfL level change is compared to a change in a second NfL determined for the subject over a first period corresponding to the difference between the second time and the first time. The subject in these examples is within 3 years of age and does not have a heart disease. In some examples, the subject is determined to be at risk of having or having a heart disease based on a change in a second NfL level determined over a second period during which a series of samples were obtained from the subject, based on the third NfL levels detected in each of the series of samples including the first sample.

[0017] The present disclosure also provides a non - transient computer - readable medium that, when executed by a processor, causes the processor to execute the methods disclosed herein, including executable instructions.

[0018] The present disclosure also provides a computer system including a memory and a processor, where the memory includes a trained machine learning model and executable instructions that, when executed by the processor, cause the computer system to execute the methods disclosed herein are stored thereon.

[0019] Any two or more of the configurations described in this specification, including this summary section, can be combined to form an implementation of the present disclosure, regardless of whether they are specifically described as separate combinations in this specification.

Brief Description of the Drawings

[0020] [Figure 1] It is a diagram showing a block diagram of an exemplary computer system. [Figure 2] It is a diagram showing a flowchart of an exemplary method of identifying a subject to be further examined and / or medicated as being at risk of having or having a heart disease. [Figure 3]A diagram showing a flowchart of an exemplary method for generating a biomarker score indicative of the risk of heart disease. [Figure 4] A diagram showing a block diagram of an example of a network environment. [Figure 5] A diagram showing a plot demonstrating the performance of a machine learning model for classifying heart disease using various neurofilament light chains (NfL). [Figure 6] A diagram showing a plot demonstrating the correlation between NfL and age in subjects with various disease states.

Mode for Carrying Out the Invention

[0021] Definition Antibody Agents: As used herein, the term “antibody agent” refers to a drug that specifically binds to a particular antigen. In some embodiments, the term encompasses any polypeptide or polypeptide complex containing sufficient immunoglobulin structural elements to confer specific binding. Such polypeptides may be produced naturally (e.g., by organisms that react to the antigen) or by recombinant engineering, chemical synthesis, or other artificial systems or methods. Exemplary antibody agents include, but are not limited to, human antibodies, primate antibodies, chimeric antibodies, bispecific antibodies, humanized antibodies, conjugated antibodies (e.g., antibodies conjugated or fused with other proteins, radiolabeled, or cytotoxic substances), Small Modular ImmunoPharmaceuticals ("SMIP" trademark), single-chain antibodies, cameloid antibodies, and antibody fragments. As used herein, the term “antibody agent” also includes intact monoclonal antibodies, polyclonal antibodies, single-domain antibodies (e.g., single-domain antibodies of sharks (e.g., IgNAR or a fragment thereof)), multispecific antibodies formed from at least two intact antibodies (e.g., bispecific antibodies), and antibody fragments exhibiting desired biological activity. Antibody agents may have antibody constant region sequences characteristic of mouse, rabbit, primate, or human antibodies. In some embodiments, the term “antibody agent” includes stapled peptides. In some embodiments, the term “antibody agent” includes one or more antibody-like conjugated peptide mimetic substances, or one or more antibody-like conjugated scaffold proteins, monobodies, or adnectins. In some embodiments, an antibody agent is or includes a polypeptide whose amino acid sequence includes one or more structural elements containing complementarity-determining regions (CDRs). In some embodiments, the antibody agent is a polypeptide having an amino acid sequence substantially identical to that of a reference antibody, comprising at least one CDR (e.g., at least one heavy chain CDR and / or at least one light chain CDR).In some embodiments, the antibody agent is a polypeptide whose amino acid sequence includes a structural element containing an immunoglobulin variable domain, or encapsulates such a component. In some embodiments, the antibody agent is a polypeptide protein having a binding domain homologous or largely homologous to the immunoglobulin binding domain. In some embodiments, the antibody agent may include covalent modifications (e.g., the attachment of a glycan, a payload (e.g., a detectable portion, a therapeutic portion, a catalytic portion, etc.), or other pendant groups (e.g., polyethylene glycol, etc.)).

[0022] Biomarkers: The terms “biomarker” or “biological marker” are used herein to refer to entities whose presence, level, or form correlates with a particular biological event or condition of interest and are considered “markers” of that event or condition, consistent with their use in the art. To give some examples, in some embodiments, a biomarker may be, or include, a marker of a particular disease condition, or a marker of the likelihood of a particular disease, disorder, or condition developing, occurring, or recurring. In some embodiments, a biomarker may be, or include, a marker of a particular disease or treatment outcome, or the likelihood thereof. Thus, in some embodiments, a biomarker is a prediction, prognosis, or diagnosis of the relevant biological event or condition of interest. In some embodiments, a biomarker is a possible marker of the relevant biological event or condition of interest. A biomarker may be an entity of any chemical class. For example, in some embodiments, a biomarker may be, or include, a nucleic acid, polypeptide, small molecule, or a combination thereof. In some embodiments, a biomarker is a cell surface marker. In some embodiments, a biomarker is intracellular. In some embodiments, the biomarker is found in a specific tissue (e.g., cardiac tissue). In some embodiments, the biomarker is found outside the cell (e.g., secreted, or produced or present outside the cell (e.g., in bodily fluids such as blood, urine, tears, saliva, cerebrospinal fluid)).

[0023] Characteristic Fragment: As used herein, the term “characteristic fragment” refers to a fragment of a biomarker (e.g., NfL) sufficient to identify the biomarker from which the fragment originates. For example, in some embodiments, a “characteristic fragment” of a biomarker includes an amino acid sequence or set of amino acid sequences that collectively enable the distinction of the biomarker from which the fragment originates from other possible biomarkers, proteins, or polypeptides. In some embodiments, the characteristic fragment includes at least 10, at least 20, at least 30, at least 40, or at least 50 amino acids, but other numbers of amino acids may also be used.

[0024] Gene product or expression product: As used herein, the term “gene product” generally refers to RNA transcribed from a gene (pre-processed and / or post-processed) or a polypeptide encoded by RNA transcribed from a gene (pre-processed and / or post-processed).

[0025] Hybridization: As used herein, the term “hybridization” refers to the physical property of a single-stranded nucleic acid molecule (e.g., DNA or RNA) annealing to a complementary nucleic acid molecule. Hybridization can typically be evaluated in a variety of contexts, including when the interacting nucleic acid molecules are studied individually or in more complex systems (e.g., covalently or otherwise bound to a carrier material, and / or within a biological system or cell). In some embodiments, hybridization can be detected by hybridization techniques such as insight hybridization (ISH), microarrays, Northern blotting, or Southern blotting. In some embodiments, hybridization refers to 100% annealing between a single-stranded nucleic acid molecule and a complementary nucleic acid molecule. In some embodiments, annealing is less than 100% (e.g., at least 95%, at least 90%, at least 85%, at least 80%, at least 75%, or at least 70% of the single-stranded nucleic acid molecule anneals to the complementary nucleic acid molecule). Hybridization techniques and methods for evaluating hybridization are well known in the art. For example, see Sambrook et al., 1989, Molecular Cloning: A Laboratory Manual, 2nd edition, Cold Spring Harbor Press, Plainview, New York, and Ausubel, FM et al., 1994, Current Protocols in Molecular Biology, John Wiley & Sons, Secaucus, New Jersey, whose contents are incorporated herein by reference in their entirety.

[0026] Detection agent: In this specification, the term "detection agent" refers to any detectable element, molecule, functional group, compound, fragment, or moiety. In some embodiments, the detection agent is provided or utilized alone. In some embodiments, the detection agent is provided and / or utilized in partnership (e.g., conjugated) with another agent. Examples of detection agents include, but are not limited to: various ligands, radionuclides (e.g., 3 H, 14 C, 18 F, 19 F, 32 P, 35 S, 135 I, 125 I, 123 I, 64 Cu, 187 Re, 111 In, 90 Y, 99m Tc, 177 Lu, 89 Zr, etc.), fluorescent dyes, chemiluminescent agents (e.g., acridinium ester, stabilized dioxetane, etc.), bioluminescent agents, spectrally resolvable inorganic fluorescent semiconductor nanocrystals (i.e., quantum dots), metal nanoparticles (e.g., gold, silver, copper, platinum, etc.) nanoclusters, paramagnetic metal ions, enzymes, colorimetric labels (e.g., dyes, colloidal gold, etc.), biotin, digoxigenin, haptens, and proteins for which antisera or monoclonal antibodies are available.

[0027] Diagnostic Tests: As used herein, the term “diagnostic tests” refers to any process or set of processes performed or carried out to obtain useful information in determining whether a patient has a disease, disorder, or condition, and / or classifying a disease, disorder, or condition into any category that is significant in terms of phenotypic category or prognosis, or that has a high probability of responding to treatment (either general or specific treatment). Similarly, as used herein, “diagnosis” means providing any kind of diagnostic information, including, but not limited to, information relating to whether a subject has or is likely to develop a disease, disorder, or condition, the circumstances, stage, or characteristics of the disease, disorder, or condition appearing in the subject, the nature or classification of a tumor, prognosis, and / or information useful in selecting an appropriate treatment or additional diagnostic tests. Treatment choices include the choice of a specific therapeutic agent or other therapeutic modalities such as surgery or radiation, the choice of whether to withhold or perform treatment, and the choice of administration regimen (e.g., the frequency or level of one or more doses of a specific therapeutic agent or combination of therapeutic agents). The selection of additional diagnostic tests may include more detailed examinations for a given disease, disorder, or condition.

[0028] Sample: As used herein, the term “sample” refers to a biological sample obtained or derived from a human subject, as described herein. In some embodiments, a biological sample includes biological tissue or bodily fluids. In some embodiments, a biological sample may include blood, blood cells, tissue or fine-needle biopsy specimens, bodily fluids containing cells, free suspended nucleic acids, cerebrospinal fluid, lymph, tissue biopsy specimens, surgical specimens, other bodily fluids, secretions, and / or excretions, and / or cells derived therefrom. In some embodiments, a biological sample includes cells obtained from an individual (e.g., a human or animal subject). In some embodiments, the obtained cells are or include cells from the individual from which the sample was obtained. In some embodiments, a sample is a “primary sample” obtained directly from the subject source of interest by any suitable means. For example, in some embodiments, a primary biological sample is obtained by biopsy (e.g., fine-needle aspiration or tissue biopsy), surgery, or collection of bodily fluids (e.g., blood). In some embodiments, a sample is cardiac tissue obtained from a subject. In some embodiments, the term “sample” refers to a preparation obtained by processing a primary sample (e.g., by removing one or more components and / or by adding one or more agents). For example, filtration using a semipermeable membrane. Another example of sample processing is that the sample may be a plasma sample treated with an anticoagulant (e.g., EDTA, heparin, or citrate). Another example of sample processing is that the sample may be processed to isolate one or more proteins (e.g., by capturing proteins with one or more antibodies). A “processed sample” may include, for example, nucleic acids or polypeptides extracted from the sample or obtained by subjecting the primary sample to techniques such as mRNA amplification or reverse transcription, isolation and / or purification of specific components.

[0029] Subject: As used herein, the term “subject” refers to an organism, such as a mammal (e.g., human). In some embodiments, a human subject is an adult, adolescent, or child subject. In some embodiments, a subject is at least 50 years old, at least 55 years old, at least 60 years old, at least 65 years old, at least 70 years old, at least 75 years old, or at least 80 years old, although other threshold ages may be used. In some embodiments, a subject suffers from a disease, disorder, or condition (e.g., a disease, disorder, or condition that can be treated in the manner provided herein). In some embodiments, a subject is susceptible to a disease, disorder, or condition. In some embodiments, a susceptible subject has a predisposition to developing a disease, disorder, or condition and / or a higher risk of developing it (compared to the average risk observed in a reference subject or reference population). In some embodiments, a subject exhibits one or more symptoms of a disease, disorder, or condition. In some embodiments, a subject does not exhibit a specific symptom (e.g., clinical manifestation of the disease) or feature of a disease, disorder, or condition. In some embodiments, the subject does not exhibit any symptoms or characteristics of a disease, disorder, or condition. In some embodiments, the subject is a patient. In some embodiments, the subject is an individual who has been diagnosed and / or treated.

[0030] Therapeutic dose: As used herein, the term “therapeutic dose” refers to the amount that, when administered to a population suffering from or susceptible to a disease, disorder, and / or condition, is sufficient to treat that disease, disorder, and / or condition when administered according to a therapeutic administration regimen. In some embodiments, the therapeutic dose is the amount that reduces the incidence and / or severity of one or more symptoms of the disease, disorder, and / or condition, and / or delays their onset. As used herein, the term “therapeutic dose” does not require that the treatment be successful in a particular individual. Rather, the therapeutic dose may be the amount that, when administered to patients in need of such treatment, produces a particular desired pharmacological response in a significant number of subjects. In some embodiments, the reference to the therapeutic dose may be a reference to an amount measured in one or more specific tissues (e.g., tissues affected by the disease, disorder, or condition) or body fluids (e.g., blood, saliva, serum, sweat, tears, urine, etc.). In some embodiments, a therapeutically effective amount of a particular drug or treatment can be formulated and / or administered as a single dose. In some embodiments, a therapeutically effective drug can be formulated and / or administered in multiple doses (e.g., as part of a dosing regimen).

[0031] Threshold: As used herein, the term “threshold” refers to a value (or set of values) used as a reference for obtaining information about measurement results (e.g., measurement results obtained in an assay) and / or classifying the measurement results. Thresholds can be determined based on one or more control samples. Thresholds can be determined before, simultaneously with, or after the measurement of interest is performed. In some embodiments, thresholds may be a range of values. In some embodiments, thresholds may be values ​​(or ranges of values) reported in the relevant field (e.g., values ​​found in standard tables).

[0032] Detailed description of a specific embodiment heart disease The methods and apparatus disclosed herein can be used, for example, for the identification, diagnosis, monitoring, or treatment of cardiac disease. Cardiac diseases may include, or may include, acute coronary syndrome, aortic aneurysm, aortic dissection, aortic stenosis, arrhythmias, atherosclerosis, atrial fibrillation, coronary artery disease (e.g., angina pectoris (e.g., stable or unstable angina) or heart attack), cardiac arrhythmias, cardiomyopathy (e.g., dilated, hypertrophic, proarrhythmic or restrictive cardiomyopathy), cardiitis (e.g., endocarditis, infective endocarditis, myocarditis, pericarditis, pancardiitis, or regurgitant cardiomyitis), congenital heart defects or heart diseases, eosinophilic myocarditis, heart failure, heart murmurs, heart valve disease, heart valve stenosis, hypertension, hypertensive heart disease, inflammatory hypertrophy, Kawasaki disease, myocardial infarction, Marfan syndrome, metabolic syndrome, peripheral artery disease (PAD), rheumatic heart disease, thromboembolism, transthyretin amyloidosis (ATTR-CM), venous thrombosis, or heart valve disease. Cardiac disease may be a condition or disorder of the heart. Cardiac disease may cause and / or exacerbate one or more neurological symptoms (e.g., neuropathy, e.g., peripheral neuropathy) in a subject. Cardiac disease may directly or indirectly result in (e.g., cause) an increase in the level of NfL in a subject. Such an increase may occur regardless of whether the subject has a neurological disease, disorder, or condition. In some embodiments, the neurological condition of the subject (e.g., whether the subject has one or more neurological diseases, disorders, or conditions) is considered when comparing, monitoring, or determining changes in the level of NfL over time in the subject, for example, when determining whether the subject has (e.g., suffers from) or is at risk of having a cardiac disease.In some embodiments, cardiac disease is treated by administering a therapeutically effective amount of a cardiac drug or treatment, such as angiotensin-converting enzyme (ACE) inhibitors, angiotensin II receptor blockers, angiotensin receptor-neprilysin inhibitors, anticoagulants or blood diluents, antiplatelet agents, aspirin, beta-blockers, calcium channel blockers, cholesterol ester transfer protein (CETP) inhibitors, cholesterol absorption inhibitors, digitalis preparations (e.g., digitalis glycosides), diuretics, dual antiplatelet therapy, fibrates, niacin, statins, vasodilators, or combinations thereof. For example, treatments include apixaban, dabigatran, edoxaban, heparin, rivaroxaban, warfarin, aspirin, clopidogrel, dipyridamole, prasugrel, ticagrelor, benazepril, captopril, enalapril, fosinopril, moexipril, perindopril, quinapril, ramipril, trandolapril, azilsartan, candesartan, eprosartan, irbesartan, losartan, olmesartan, telmisartan, valsartan, sacubitril, sacubitril / valsartan, acebutolol, atenolol, betaxolol, bisoprolol / hydrochlorothiazide, bisoprolol, metoprolol, nadolol, propranol This may include olol, sotalol, carvedilol, labetalol hydrochloride, amlodipine, diltiazem, felodipine, nifedipine, nimodipine, nisoldipine, verapamil, atorvastatin, fluvastatin, lovastatin, pitavastatin, pravastatin, rosuvastatin, simvastatin, niacin, ezetimibe, ezetimibe / simvastatin, digoxin, acetazolamide, amiloride, bumetanide, chlorothiazide, chlorthalidone, furosemide, hydrochlorothiazide, indapamide, metalozone, spironolactone, torsemide, isosorbide dinitrate, isosorbide mononitrate, hydralazine, nitroglycerin, minoxidil, or a combination thereof.

[0033] Cardiac disease biomarkers The embodiments described herein offer many advantages over the prior art discussed herein. For example, the technologies of this disclosure are non-invasive, minimize patient discomfort, are quick to implement, and are relatively cost-effective. Accordingly, the technologies described herein offer advantages over these prior arts, including, but not limited to, providing alternatives to single-marker IVD testing, including non-invasive in vitro diagnostic (IVD) testing for cardiac disease, one or more specific in vitro biomarkers suitable for use in IVD testing, and more than one marker for effectively determining or excluding more expensive and invasive procedure candidates in the diagnosis of disease.

[0034] This disclosure relates, in particular, to NfL as a biomarker for cardiac disease. In some embodiments, the NfL described herein is an NfL protein, a nucleic acid sequence encoding NfL, a characteristic fragment thereof, and / or a variant thereof, or comprises them.

[0035] In this specification, NfL includes the gene product associated with NfL. For example, NfL may include a protein or nucleotides (e.g., RNA or mRNA). NfL also includes the full-length protein of NfL and fragments of NfL (e.g., characteristic fragments). In some embodiments, NfL includes fragments having the same amino acid sequence as a continuous range of at least 10 amino acids, at least 20 amino acids, at least 30 amino acids, at least 40 amino acids, at least 50 amino acids, at least 60 amino acids, at least 70 amino acids, at least 80 amino acids, at least 90 amino acids, or at least 100 amino acids as the amino acid sequence provided in Table 1, but other numbers of amino acids may also be used. In some embodiments, NfL includes fragments having at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, at least 95%, at least 96%, at least 97%, at least 98%, or at least 99% identity to the amino acid sequence provided in Table 1, but other percentages may also be used. In some embodiments, an NfL includes a nucleic acid fragment having the same nucleic acid sequence as a contiguous range of at least 10, at least 20, at least 30, at least 40, at least 50, at least 60, at least 70, at least 80, at least 90, or at least 100 nucleic acids of the nucleic acid sequence provided in Table 2, but other numbers of nucleic acids may also be used. In some embodiments, an NfL includes a fragment having at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, at least 95%, at least 96%, at least 97%, at least 98%, or at least 99% identity to the nucleic acid sequence provided in Table 2, but other percentages may also be used. Variants or alternative forms of an NfL include, for example, polypeptides encoded by any splice variant of the transcript encoding the NfL.

[0036] The biomarkers intended herein also include cleaved forms or polypeptide fragments of NfLs described herein. Cleaved forms or polypeptide fragments of NfLs may include forms with N-terminus deletion or cleavage and forms with C-terminus deletion or cleavage. Cleaved forms or fragments of NfLs may include, but are not limited to, fragments resulting from any mechanism, such as alternative translation, degradation by exoproteases and / or endoproteases, and / or degradation by physical, chemical and / or enzymatic proteases. Biomarkers may include cleavage portions or fragments of NfL and may include at least 10%, at least 15%, at least 20%, at least 25%, at least 30%, at least 35%, at least 40%, at least 45%, at least 50%, at least 55%, at least 60%, at least 65%, at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, at least 95%, at least 96%, at least 97%, at least 8%, or at least 99% of the amino acid sequence of the NfL protein, but other percentages may also be used.

[0037] In some cases, the fragments are cleaved by only 1-20 amino acids at the N-terminus and / or C-terminus compared to the corresponding mature full-length NfL protein, e.g., 1-15 amino acids, 1-10 amino acids, or 1-5 amino acids.

[0038] The NfL proteins of this disclosure (e.g., NfL proteins or fragments thereof) may also include modified forms of NfL having post-expression modifications, including but not limited to modifications such as phosphorylation, glycosylation, lipidation, methylation, selenocystination, cysteinylation, sulfonation, glutathioneation, acetylation, and / or oxidation of methionine to methionine sulfoxide or methionine sulfone.

[0039] In some embodiments, NfL may be a nucleotide (also referred to herein as nucleic acid or polynucleotide). In some embodiments, the nucleotide may be RNA or DNA (e.g., cDNA). In some cases, the corresponding RNA or DNA may exhibit better diagnostic discriminative ability than the full-length protein.

[0040] Nerve filament light chain (NfL) Nerve filaments are cytoskeletal components of neurons that are particularly abundant in axons. Functions of nerve filaments include providing structural support, as well as maintaining the size, shape, and diameter of the axon. Nerve filaments consist of three subunits: the light chain (NfL), the medium chain, and the heavy chain. NfL levels increase in cerebrospinal fluid (CSF) and blood in proportion to the degree of axonal damage in various neurological diseases, including inflammatory diseases, neurodegenerative diseases, traumatic diseases, and cerebrovascular diseases.

[0041] In some embodiments, NfL may be useful for the detection and diagnosis of cardiac disease and is a biomarker for cardiac disease. In some embodiments, the detection of NfL, characteristic fragments of NfL, and / or variants of NfL is used in a method to assess the risk of developing cardiac disease in a subject, to diagnose cardiac disease in a subject, or to recommend additional cardiomyopathy testing for a subject. In some embodiments, NfL is detected in a sample using an anti-NfL agent (e.g., an anti-NfL antibody, probe, etc.). In some embodiments, the detection of nucleotides encoding NfL, nucleotides encoding characteristic fragments of NfL, and / or variants of NfL is used in a method to assess the risk of developing cardiac disease in a subject, to diagnose cardiac disease in a subject, or to recommend additional cardiomyopathy testing for a subject. In some embodiments, nucleotides encoding NfL are detected in a sample using an anti-NfL nucleotide sequence agent (e.g., an anti-NfL nucleotide sequence antibody, probe, complementary nucleic acid, etc.).

[0042] Table 1 below contains an example amino acid sequence of NfL.

[0043] [Table 1]

[0044] Table 2 below contains example nucleic acid sequences of NfL.

[0045] [Table 2-1] [Table 2-2] [Table 2-3]

[0046] In some embodiments, additional biomarkers can be analyzed or evaluated. In some embodiments, the biomarkers may include image-based biomarkers of the subject from which the sample was acquired (e.g., left ventricular septal thickness and / or ejection fraction). In some embodiments, the biomarkers may include, but are not limited to, demographic factors (e.g., one or more of age, weight, biological sex, ethnicity, BMI, medical history, risk factors, family history, or geographical location).

[0047] Exemplary combinations of cardiac disease biomarkers As provided herein, NfL may be a full-length protein or a fragment thereof. In some embodiments, the fragment of NfL is a characteristic protein fragment. In some embodiments, NfL (e.g., in a sample) may include a subset of full-length NfL protein and a subset of characteristic protein fragments of NfL.

[0048] In some embodiments, NfL has a wild-type amino acid sequence. In some embodiments, NfL has a mutant amino acid sequence (e.g., an amino acid sequence containing one or more mutations). In some embodiments, a subset of NfL proteins has a wild-type amino acid sequence, and a subset of NfL proteins has a mutant amino acid sequence.

[0049] As provided herein, an NfL may be a full-length nucleotide (e.g., DNA, cDNA, or RNA) encoding an NfL or a fragment thereof. In some embodiments, the NfL fragment is a characteristic nucleotide fragment. In some embodiments, the NfL (e.g., in a sample) may include a subset of full-length NfL nucleotides (e.g., DNA, cDNA, or RNA) and a subset of characteristic nucleotide fragments of the NfL.

[0050] In some embodiments, the NfL has a wild-type nucleic acid sequence. In some embodiments, the NfL has a mutant nucleic acid sequence, for example, a nucleic acid sequence containing one or more mutations. In some embodiments, a subset of the NfL includes a wild-type nucleic acid sequence encoding the NfL, and a subset includes a mutant nucleic acid sequence encoding the NfL.

[0051] Exemplary Method Exemplary methods of the disclosed technology enable earlier identification of more patients at risk of cardiac disease while minimizing the number of false negatives. In some embodiments, the methods disclosed herein offer the advantage of early screening for the presence of cardiac disease. In some embodiments, the methods disclosed herein can assist in the detection or diagnosis of cardiac disease (e.g., after cardiac disease has been ruled out by genetic testing). In some embodiments, the methods disclosed herein reduce or eliminate the need to initiate cardiac disease screening (e.g., expensive and complex echocardiography, CMR, and / or scintigraphy). In some embodiments, if the methods disclosed herein detect a potential for cardiac disease in a subject, the subject may undergo subsequent confirmatory tests (e.g., echocardiography, magnetic resonance imaging (CMR), scintigraphy, and / or cardiac biopsy).

[0052] In some embodiments, the methods disclosed herein include determining the risk of developing a heart disease in a subject. In some embodiments, the methods disclosed herein include diagnosing a subject with a heart disease, and a sample is obtained from that subject. In some embodiments, the methods disclosed herein include treating a heart disease in a subject that is at risk of developing a heart disease or is suffering from a heart disease. In some embodiments, the methods disclosed herein include determining that a patient does not have a heart disease or is not at risk of developing a heart disease.

[0053] In some embodiments, the methods disclosed herein include selecting a subject to receive one or more doses of a cardiac drug, from which a sample is obtained. In some embodiments, the methods disclosed herein include administering one or more doses of a cardiac drug to a subject. In some embodiments, the cardiac drug includes anticoagulants, antiplatelet agents, ACE inhibitors, angiotensin II receptor blockers, angiotensin receptor-neprilysin inhibitors, beta-blockers, calcium channel blockers, cholesterol-lowering agents, digitalis preparations, diuretics, vasodilators, or combinations thereof.

[0054] In some embodiments, the methods disclosed herein include selecting a subject to undergo one or more cardiomyopathy tests, from which a sample is obtained. In some embodiments, the one or more cardiomyopathy tests include echocardiography, advanced imaging techniques, or both. In some embodiments, advanced imaging techniques include magnetic resonance imaging (CMR), scintigraphy, or both. In some embodiments, scintigraphy includes using a radioisotope conjugate such as 99mTc-pyrophosphate. In some embodiments, scintigraphy is performed using single-photon emission computed tomography (SPECT).

[0055] This disclosure provides a diagnostic test for cardiac disease, characterized by the detection of NfL by the method described and illustrated herein.

[0056] In some embodiments, the methods for detecting, diagnosing, or identifying the risk of cardiac disease disclosed herein have one or more of the following advantages: improved sensitivity for identifying cardiac disease, improved specificity for identifying cardiac disease, improved accuracy for identifying cardiac disease, reduced time to diagnosis of cardiac disease, and / or reduced cost of screening patients with cardiac disease.

[0057] In some embodiments, NfL can be used in in vitro diagnostic (IVD) or screening tests for cardiac disease conditions. In some embodiments, the diagnostic tests taught herein detect whether NfL is present in a sample taken from a subject. In some embodiments, the diagnostic tests disclosed herein can assist in the detection or diagnosis of cardiac disease in a subject.

[0058] In some embodiments, the diagnostic tests disclosed herein are adapted to an immunoassay platform. In some embodiments, such an immunoassay platform includes a semi-automated or automated immunoassay platform. In some embodiments, the diagnostic tests disclosed herein are adapted to a semi-automated test of one or more biomarkers.

[0059] In some embodiments, the diagnostic tests disclosed herein provide an improved diagnosis of cardiac disease compared to standard techniques, including one or more of the following advantages: improved sensitivity in identifying cardiac disease, improved specificity in identifying cardiac disease, improved accuracy in identifying cardiac disease, reduced time to diagnosis of cardiac disease, and / or reduced cost of screening patients with cardiac disease.

[0060] In some embodiments, the diagnostic tests disclosed herein may be plasma-based screening assays. In some embodiments, the diagnostic tests are adapted to, for example, the Siemens Atellica® system or the Siemens Advia Centaur® system.

[0061] In some embodiments, the methods provided herein include detecting the level of NfL present in a sample to obtain a biomarker profile and calculating a biomarker score using the biomarker profile. In some embodiments, the methods provided herein include detecting the level of NfL in a sample to obtain a biomarker profile and calculating a biomarker score using the biomarker profile and demographic factors. In some embodiments, the methods provided herein include detecting the level of NfL in a sample to obtain a biomarker profile and calculating a biomarker score using the biomarker profile and image-based biomarkers. In some embodiments, the methods provided herein include detecting the level of NfL in a sample to obtain a biomarker profile and calculating a biomarker score using the biomarker profile, demographic factors, and image-based biomarkers.

[0062] In some embodiments, the methods provided herein include receiving NfL levels, demographic factors, and / or image-based biomarkers in a sample. In some embodiments, receiving includes electronic receiving. In some embodiments, demographic factors include one or more of age, weight, biological sex, ethnicity, BMI, medical history, risk factors, family history, or geographical location. In some embodiments, image-based biomarkers include left ventricular septal thickness and / or ejection fraction.

[0063] In some embodiments, the methods described herein include selecting subjects for further cardiomyopathy testing using a biomarker score. In some embodiments, the methods described herein include selecting subjects for one or more administrations of cardiac medication using a biomarker score. In some embodiments, the methods described herein include identifying subjects who have or are at risk of cardiac disease using a biomarker score.

[0064] In some embodiments, the methods described herein include comparing a biomarker score with a reference biomarker score. In some embodiments, the methods described herein include administering a cardiac drug to a subject once or more times. In some embodiments, the methods described herein include performing one or more cardiomyopathy tests on a subject.

[0065] NfL detection and biomarker profiling In particular, the methods provided herein include the evaluation of the level of NfL detected in a sample. Exemplary methods for detecting the level of NfL are described herein. However, the level of NfL can also be provided, for example, in electronic form, from the laboratory that detected the level of NfL in the sample.

[0066] In some embodiments, the Disclosure provides techniques for detecting, analyzing, and / or evaluating NfLs in a sample. In some embodiments, NfLs are present in a sample obtained from a subject, and diagnostic or therapeutic decisions are made based on their detection, analysis, and / or evaluation.

[0067] In some embodiments, the NfL levels described herein include the presence or absence of NfL, the amount of NfL, the absolute amount of NfL, the relative amount of NfL, or the concentration of NfL.

[0068] Methods for detecting NfL include detecting biomarkers as proteins. Methods for detecting protein-based biomarkers include, for example, mass spectrometry (MS), immunoassays (e.g., immunoprecipitation), Western blotting, ELISA, immunohistochemistry, immunocytochemistry, flow cytometry, and / or immunoPCR.

[0069] In some embodiments, mass spectrometry methods include MS, MS / MS, MALDI-TOF, electro-spray ionization mass spectrometry (ESIMS), ESI-MS / MS, and ESI-MS / (MS). n Matrix-assisted laser desorption / ionization time-of-flight mass spectrometry (MALDI-TOF-MS), surface-enhanced laser desorption / ionization time-of-flight mass spectrometry (SELDI-TOF-MS), tandem liquid chromatography-mass spectrometry (LC-MS / MS), desorption / ionization on silicon (DIOS), secondary ion mass spectrometry (SIMS), quadrupole time-of-flight (Q-TOF), atmospheric pressure chemical ionization mass spectrometry (APCI-MS), APCI-MS / MS, APCI-(MS), atmospheric pressure photoionization mass spectrometry (APPI-MS), APPI-MS / MS, and APPI-(MS) n These include quadrupole mass spectrometry, Fourier transform mass spectrometry (FTMS), and ion trap mass spectrometry. Often, MS approaches quantify fragments of biomarkers rather than full-length proteins. However, MS approaches may be sufficient to determine the protein levels of biomarkers with sufficient precision for the disclosed methods and / or evaluations.

[0070] In some embodiments, the immunoassay can be a chemiluminescent immunoassay, a high-throughput, and / or automated immunoassay platform. For example, a high-throughput and / or automated immunoassay platform can be used to analyze at least 240 tests per hour, or at least 440 tests per hour, but in other examples, the number of tests per hour can be any other number.

[0071] In some embodiments, a method for detecting NfL in a sample includes contacting the sample with one or more antibody agents against NfL. In some embodiments, such a method also includes contacting the sample with a first set of one or more detection agents. In some embodiments, the antibody agent is labeled with the first set of one or more detection agents. In some embodiments, the first set of one or more detection agents includes one or more acridinium ester (AE) molecules.

[0072] AE molecules can be used to label proteins and nucleic acids. Acridinium-labeled proteins can be used for detection in immunoassays. When AEs are exposed to alkaline H2O2 (hydrogen peroxide), chemiluminescence occurs. The light emits most vigorously at wavelengths in the 430–480 nm range, depending on the specific AE variant. Such light can be detected, for example, with a high-efficiency photomultiplier tube. The luminescence is rapid, completing within 1–5 seconds. The diversity of AE morphologies contributes to better assay performance, including improved sensitivity and robustness. AE molecules can be used to label small molecules, large analytes, and antibodies.

[0073] Additional methods for detecting biomarkers (e.g., NfLs) include methods for detecting biomarkers as nucleic acids. Nucleic acid-based methods for detecting NfLs include performing nucleic acid amplification methods such as polymerase chain reaction (PCR), reverse transcription polymerase chain reaction (RT-PCR), transcriptional amplification (TMA), ligase chain reaction (LCR), strand displacement amplification (SDA), and nucleic acid sequence-based amplification (NASBA). In some embodiments, nucleic acid-based methods for detecting biomarkers include detecting hybridization between one or more nucleic acid probes and one or more nucleotides encoding NfLs. In some embodiments, the nucleic acid probes are complementary to at least a portion of one of the one or more nucleotides encoding NfLs. In some embodiments, the nucleotides encoding NfLs include DNA (e.g., cDNA). In some embodiments, the nucleotides encoding NfLs include RNA (e.g., mRNA).

[0074] Exemplary sample In some embodiments, the sample disclosed herein is a biological sample. In some embodiments, the biological sample is a blood sample (for example, taken from an artery or vein of the subject). The blood sample may be a whole blood sample, a plasma sample, or a serum sample. In some embodiments, the biological sample includes cardiac tissue.

[0075] In some embodiments, the sample is obtained from the subject (e.g., by biopsy). In some embodiments, the subject from whom the sample is obtained is being evaluated for heart disease. In some embodiments, the subject from whom the sample is obtained has heart disease or is at risk of developing heart disease.

[0076] Exemplary examples In some embodiments, the subjects disclosed herein are mammals (e.g., humans). In some embodiments, the subjects disclosed herein are biological males. In some embodiments, the subjects disclosed herein are biological females. In some embodiments, the subjects disclosed herein are overweight. In some embodiments, the subjects have a body mass index (BMI) of 25 or greater. In some embodiments, the subjects have a body mass index (BMI) of 30 or greater, but other threshold BMIs may also be used. In some embodiments, the subjects are at least 50 years old, at least 55 years old, at least 60 years old, or at least 65 years old, but other age thresholds may also be used. Other types of subjects, subject characteristics, BMI, and / or age may also be used in conjunction with the disclosed technology.

[0077] Exemplary methods using thresholds Some methods disclosed herein allow the level of NfL to be compared to a threshold. In some embodiments, the methods disclosed herein include comparing the level of NfL to a respective threshold. In some embodiments, the methods disclosed herein include comparing the level of NfL to a reference threshold.

[0078] The reference threshold may be a threshold from subjects known to have good cardiac health or independently verified, or, in the case of subjects with cardiac disease, a threshold from subjects known to have poor cardiac health or independently verified. Alternatively, or in combination, the biomarker profile of a subject can be compared to a reference threshold determined from multiple subjects in known states (e.g., healthy, not diagnosed with cardiac disease, or diagnosed with cardiac disease). In some embodiments, the reference threshold is the mean of known NfL levels obtained from multiple subjects, or a range defined by the NfL level range observed in the reference subject.

[0079] A more complex evaluation approach may involve comparing the biomarker levels of a subject to reference biomarker levels constructed from a large number of subjects in known states (e.g., healthy, not diagnosed with heart disease, or diagnosed with heart disease), such as at least 10, at least 50, at least 100, at least 500, or at least 1000 subjects, although a different threshold number of subjects may also be used. In some cases, the reference subjects can be evenly distributed between (1) healthy / not diagnosed with heart disease and (2) diagnosed with heart disease. The evaluation may also involve repeatedly or simultaneously comparing the biomarker levels of a subject (e.g., NfL) with multiple profiles of known states.

[0080] Multiple known reference biomarker profiles (e.g., NfL detection levels in reference samples, demographic factors, and / or image-based biomarkers) can also be used to train computational assessment algorithms (e.g., machine learning models), allowing a single comparison between the target biomarker profile and the reference biomarker profile to integrate or aggregate information from a large number of subjects with known health statuses (e.g., healthy, not diagnosed with heart disease, or diagnosed with heart disease), e.g., at least 10, at least 50, at least 100, at least 500, at least 1000, or more individuals, although a different number of individuals may also be used. Generating such reference biomarker profiles can facilitate the assessment of the target's risk of heart disease and / or rapid assessment with reduced computational power usage.

[0081] A reference biomarker profile can be generated from multiple reference biomarker profiles using any of the various computational methods. Machine learning models based on this technology can be constructed using any number of statistical programming languages, such as R, scripting languages ​​and related machine learning packages like Python, or data mining software like Weka, Java, Mathematica, Matlab®, or SAS.

[0082] The subject's biomarker profile can be compared to a reference biomarker profile (e.g., one generated as described above) to generate an output assessment. Many output assessments are consistent with the disclosures herein. Output assessments may include, depending on the case, a single assessment narrowed down by sensitivity, specificity, or sensitivity and specificity parameters, indicating a health status assessment (e.g., the probability that the subject has heart disease, the subject is not at risk for heart disease, the subject is at risk for heart disease, or the subject has heart disease). Alternatively, or in combination, additional parameters are provided, such as demographic factors of the subject (e.g., one or more of age, weight, biological sex, ethnicity, BMI, medical history, risk factors, family history, and geographical location) and / or image-based biomarkers of the subject (e.g., left ventricular septal thickness and / or ejection fraction).

[0083] In some embodiments, the methods disclosed herein further include diagnosing a subject with cardiac disease if the detected NfL level exceeds a threshold. In some embodiments, the methods disclosed herein include diagnosing a subject with cardiac disease if the detected NfL level is at least 1.3 times, at least 1.4 times, at least 1.5 times, at least 1.6 times, at least 1.7 times, at least 1.8 times, or at least 1.9 times greater than the threshold, although other coefficients of the threshold may also be used.

[0084] In some embodiments, the methods disclosed herein include recommending one or more cardiomyopathy tests for a subject if the NfL level exceeds a threshold. In some such methods, one or more cardiomyopathy tests are recommended for a subject if the detected NfL level is at least 1.3 times, at least 1.4 times, at least 1.5 times, at least 1.6 times, at least 1.7 times, at least 1.8 times, or at least 1.9 times greater than the threshold, although other coefficients of the threshold may also be used.

[0085] As described above, in some embodiments, a method for detecting NfL in a sample includes contacting the sample with one or more antibody agents against NfL. In some embodiments, such a method also includes contacting the sample with a first set of one or more detection agents. In some embodiments, the antibody agent is labeled with the first set of one or more detection agents. In some embodiments, the first set of one or more detection agents includes one or more acridinium ester (AE) molecules.

[0086] AE molecules can be used to label proteins and nucleic acids. Acridinium-labeled proteins can be used for detection in immunoassays. When AE is exposed to alkaline H2O2 (hydrogen peroxide), chemiluminescence occurs. The light emits maximum light in the wavelength range of 430-480 nm, depending on the specific AE variant. Such light can be detected, for example, by a high-efficiency photomultiplier tube.

[0087] In some embodiments, detecting binding between NfL and one or more antibody agents to NfL involves determining the absorbance or emission value of a first set of one or more detection agents. For example, the absorbance value indicates the level of binding (e.g., higher absorbance indicates more binding). In some embodiments, the absorbance or emission value of the first set of one or more detection agents exceeds a threshold. In some embodiments, the absorbance or emission value of the first set of one or more detection agents is at least 1.3 times, at least 1.4 times, at least 1.5 times, at least 1.6 times, at least 1.7 times, at least 1.8 times, or at least 1.9 times greater than the threshold, although other coefficients of the threshold may also be used. In some embodiments, the threshold is the average of the absorbance or emission values ​​determined for a second set of one or more detection agents labeling two or more control samples. In some such embodiments, the second set of one or more detection agents is similar to or identical to the first set of one or more detection agents.

[0088] In some embodiments, the methods disclosed herein further include diagnosing a subject with cardiac disease if the detected NfL level exceeds a threshold. In some embodiments, the methods include diagnosing a subject with cardiac disease if the detected NfL level is at least 1.3 times, at least 1.4 times, at least 1.5 times, at least 1.6 times, at least 1.7 times, at least 1.8 times, or at least 1.9 times greater than the threshold, although other coefficients of the threshold may also be used.

[0089] In some embodiments, the methods disclosed herein include diagnosing a subject with heart disease if the absorbance or emission value of a first set of one or more detection agents exceeds a threshold. In some embodiments, the methods include diagnosing a subject with heart disease if the absorbance or emission value of a first set of one or more detection agents is at least 1.3 times, at least 1.4 times, at least 1.5 times, at least 1.6 times, at least 1.7 times, at least 1.8 times, or at least 1.9 times greater than the threshold, although other coefficients of the threshold may also be used. In some embodiments, the threshold is the average of the absorbance or emission values ​​determined for a second set of one or more detection agents labeling two or more control samples. In some such embodiments, the second set of one or more detection agents is similar to or identical to the first set of one or more detection agents.

[0090] In some embodiments, the methods disclosed herein include recommending one or more cardiomyopathy tests for a subject if the NfL level exceeds a threshold. In some such methods, one or more cardiomyopathy tests are recommended for a subject if the detected NfL level is at least 1.3 times, at least 1.4 times, at least 1.5 times, at least 1.6 times, at least 1.7 times, at least 1.8 times, or at least 1.9 times greater than the threshold, although other coefficients of the threshold may also be used.

[0091] In some embodiments, the methods disclosed herein include recommending one or more cardiomyopathy tests for a subject if the absorbance or emission value of a first set of one or more detection agents exceeds a threshold. In some such methods, one or more cardiomyopathy tests are recommended for a subject if the absorbance or emission value of a first set of one or more detection agents is at least 1.3, at least 1.4, at least 1.5, at least 1.6, at least 1.7, at least 1.8, or at least 1.9 times greater than the threshold, although other coefficients of the threshold may also be used.

[0092] Cardiomyopathy tests that can be used according to the methods of this disclosure include, for example, echocardiography or advanced imaging methods. In some embodiments, advanced imaging methods include magnetic resonance imaging (CMR) or scintigraphy (for example, 99m This includes methods using radioisotope conjugates such as Tc-pyrophosphate, and / or single-photon emission computed tomography (SPECT).

[0093] In any of the embodiments described and illustrated herein, the threshold may be the average value of the values ​​detected for two or more control samples. In some such embodiments, the values ​​detected for two or more control samples represent the control level of NfL. In some embodiments, the control samples include recombinant NfL. In some embodiments, the two or more control samples are each samples obtained from subjects without heart disease. In some embodiments, the threshold is the value reported in the standard table.

[0094] An exemplary method of utilizing biomarker profiles The algorithm-based assays and related information provided by any practice of the methods described herein can facilitate improved treatment and decision-making for subjects. For example, the methods described herein can enable physicians or caregivers to identify, for instance, patients who are unlikely to have heart disease and therefore do not require treatment, additional cardiac examinations, or enhanced cardiac monitoring, or patients who are likely to have heart disease and require treatment, additional cardiac examinations, or enhanced cardiac monitoring.

[0095] Biomarker scores can be determined in some cases by applying specific algorithms. In some embodiments, biomarker scores are quantitative. Algorithms used to calculate biomarker scores in some methods disclosed herein can group the expression level values ​​of NfLs or groups of biomarkers containing NfLs. Furthermore, the formation of specific biomarker groups can facilitate the mathematical weighting of the contributions that various expression levels of a biomarker or a subset of biomarkers (e.g., a classifier) ​​have to the quantitative score.

[0096] Several methods described herein, as well as kits and systems provided herein, can utilize algorithm-based diagnostic assays to predict whether a subject from whom a sample is taken is at risk of or has cardiac disease, to select subjects from whom a sample is taken for one or more cardiomyopathy tests, and / or to select subjects from whom a sample is taken for one or more cardiac medication administrations.

[0097] The level of NfL, and optionally one or more demographic factors (e.g., one or more age, weight, biological sex, ethnicity, BMI, medical history, risk factors, family history, and geographical location) and / or image-based biomarkers (e.g., left ventricular septal thickness and / or ejection fraction), can be used alone or organized into functional subsets to calculate biomarker scores used to predict whether a sampled subject is at risk of or has cardiac disease, to select subjects from whom samples are taken for one or more cardiomyopathy tests, and / or to select subjects from whom samples are taken for one or more cardiac medication administrations.

[0098] Some methods disclosed herein involve calculating a biomarker score using a biomarker profile. In some embodiments, calculating a biomarker score using a biomarker profile involves applying an algorithm to the biomarker profile to calculate the biomarker score. In some embodiments, the algorithm is derived from or is derived from decision trees, neural boosting, bootstrap forests, boost trees, K-nearest neighbors, generalized regression forward selection, generalized regression pruning forward selection, fit stepwise, generalized regression lasso, generalized regression elastic nets, generalized regression ridges, nominal logistics, support vector machines, discriminant analysis, naive Bayes, or a combination thereof. In some embodiments, the algorithm is a decision tree, neural boost, bootstrap forest, boost tree, support vector machine, or a combination thereof, or derived from them.

[0099] A machine learning model may use a biomarker profile that includes data corresponding to one or more (e.g., two or more) demographic factors and data corresponding to NfL levels in samples from subjects. In some embodiments, one or more demographic factors include one or more of age, weight, biological sex, ethnicity, BMI, medical history, risk factors, family history, and geographical location. In some embodiments, one or more demographic factors include age, biological sex, or both age and biological sex. In some embodiments, one or more demographic factors are two demographic factors. In some embodiments, the two demographic factors are age and biological sex.

[0100] In some embodiments, the method includes providing data corresponding to one or more demographic factors of the subject and a corresponding biomarker profile as input to a machine learning model. In some embodiments, one or more demographic factors include biological sex, age, or both biological sex and age. The corresponding biomarker profile includes data about the level of NfL in the subject sample characterized by the demographic factors. The performance of the machine learning model (e.g., predictive power, sensitivity and / or selectivity, and / or classification accuracy) may be improved by using a machine learning model that considers NfL and a combination of one or more (e.g., two or more) demographic factors compared to a machine learning model that does not consider demographic factors.

[0101] The methods provided herein may utilize additional algorithms, which are merely examples of the types of algorithms that can be used to generate biomarker scores. Examples of exemplary algorithms include, for example, Duda, 2001, *Pattern Classification*, John Wiley & Sons, New York, pp. 396-408 and 411-412, and Hastie et al., 2001, *The Elements of Statistical Learning*, Springer-Verlag, New York, Chapter 9, each incorporated herein by reference in its entirety. Furthermore, as stated above, combinations of algorithms can be used in the methods provided herein. For example, the boosted tree method may be a combination of the decision tree method and the boosting method. Further combinations are possible and intended for use in the methods provided herein. More example algorithms usable in the methods provided herein are listed below.

[0102] Decision tree One method that can be used to calculate biomarker scores from biomarker profiles is decision trees. Decision trees can be constructed using a training population and specific data analysis algorithms. Decision trees are generally described in Duda, 2001, Pattern Classification, John Wiley & Sons, New York, pp. 395-396. In tree-based methods, the constituent space is divided into a set of rectangles, and then a model (such as constants) is fitted to each.

[0103] Training population data may include biomarker profiles across the training set population (e.g., NfL levels in samples). One specific algorithm that can be used to construct decision trees is Classification and Regression Tree (CART). Other specific decision tree algorithms include, but are not limited to, ID3, C4.5, MART, and Random Forest. CART, ID3, and C4.5 are described in Duda, 2001, Pattern Classification, John Wiley & Sons, New York, pp. 396-408 and 411-412. CART, MART, and C4.5 are described in Hastie et al., 2001, The Elements of Statistical Learning, Springer-Verlag, New York, Chapter 9. Random forests are described in Breiman, 1999, "Random Forests—Random Features," Technical Report 567, Statistics Department, UC Berkeley, September 1999, and their entire contents are incorporated herein by reference.

[0104] The purpose of a decision tree is to derive a classifier (i.e., a tree) from real-world example data. This tree can be used to classify previously unseen examples that were not used in the derivation of the decision tree. Thus, a decision tree can be derived from training data. Exemplary training data includes data on multiple subjects (e.g., a training population). For each subject, a biomarker profile can be provided and / or used. In some embodiments, the training data includes a biomarker profile of the training population.

[0105] Example decision tree The following algorithm describes the derivation of an exemplary decision tree: Create a root node If all Users have the same Class value, assign this label to the route. If Feature is empty, label the route according to the most common value. Otherwise, start the following: Calculate the information gain for each feature. Select Feature A, which has the highest information gain, and make it the root feature. For each possible value v of this feature Below the root, add a new branch corresponding to A=v. Let Example(v) be an example where A=v. If Example(v) is empty, the new branch will be a leaf node labeled with the most common value in Example. Otherwise, the new branch will be the tree created by the following formula. Tree(Example(v),Class,Feature-{A}) End

[0106] In a univariate decision tree, each split is based on a constituent value (e.g., level) of a corresponding biomarker. Furthermore, multivariate decision trees can be implemented by the methods described herein. In some embodiments, splits are based on constituent values ​​corresponding only to demographic factors, or in combination with constituent values ​​of corresponding biomarkers. For example, splits may be based on a combination of constituent values ​​corresponding to sex and one or more biomarker levels, age and one or more biomarker levels, or both age and sex and one or more biomarker levels. Multivariate decision trees are described in Duda, 2001, Pattern Classification, John Wiley & Sons, New York, pp. 408-409. In such multivariate decision trees, some or all decisions involve a linear combination of constituent values ​​(e.g., levels) for multiple biomarkers in a profile. Such linear combinations can be trained using known methods, such as gradient descent or the sum of squared errors criterion in classification.

[0107] As an example, use the following formula: 0.05(X1) + 0.2(X2) < 500. In this example, X1 and X2 refer to two different configurations (e.g., levels) for two different biomarkers (including NfL). To apply this method, obtain the values ​​for configurations X1 and X2 (e.g., as part of a biomarker profile) from measurements taken from unclassified subjects. Then, substitute these values ​​into the formula. If the calculated value is less than 500, proceed to the first branch of the decision tree. Otherwise, proceed to the second branch of the decision tree.

[0108] Biomarker profiles can be used to train machine learning models. Biomarker profiles may come from different subjects, the same subjects at different times, or a combination thereof. Biomarker profiles can be associated with corresponding data indicating health status (e.g., they can be labeled with it). For example, a biomarker profile of a subject used as training data may be labeled with data indicating the health status of that subject. Such corresponding data may be probabilities or scores (e.g., biomarker scores). Health status may be the probability that a subject has heart disease, whether a subject is at risk of heart disease, or whether a subject has heart disease (e.g., healthy, not diagnosed with heart disease, or diagnosed with heart disease). Health status may be determined manually, for example, by a physician. The task of labeling biomarker profiles with corresponding data indicating health status may be performed manually (e.g., by the physician who determined the health status) or automatically based on stored correlations.

[0109] Demographic factor data can be used in combination with biomarker profiles for training machine learning models. For example, a machine learning model can be trained using biomarker profiles and one or more demographic factor data for each of a set of subjects. This set may be, for example, at least 10 subjects, at least 20 subjects, at least 50 subjects, at least 100 subjects, at least 200 subjects, at least 300 subjects, at least 500 subjects, at least 1,000 subjects, at least 1,500 subjects, at least 2,000 subjects, at least 5,000 subjects, or at least 10,000 subjects, but other examples may use a different number of subjects. In some embodiments, training a machine learning model uses a combination of one or more demographic factors for each set of subjects—age, sex, or both—and a biomarker profile. The biomarker profile may include data corresponding to NfL levels.

[0110] In some embodiments, training a machine learning model may involve determining one or more configuration values. Each configuration value may correspond to a biomarker (e.g., NfL). One or more configuration values ​​may be determined based on a biomarker profile and / or a linear combination thereof. One or more configuration values ​​may correspond to an NfL level. Training a machine learning model may involve determining one or more decision rules based on a biomarker profile used as training data. One or more decision rules may be based on one or more demographic factors corresponding to a subject associated with the biomarker profile. One or more decision rules may be based on one or more configuration values ​​(e.g., those determined during training). One or more decision rules may be used in one or more decision trees.

[0111] Bagging, boosting, and additive trees can be combined with decision techniques to improve weak decision rules. These techniques are designed for decision trees, such as those described herein. In some embodiments, a machine learning model may include at least 25 decision trees, at least 50 decision trees, or at least 100 decision trees, but any number of decision trees may be used. Furthermore, such techniques may also be useful for decision rules developed using other types of data analysis algorithms, such as linear discriminant analysis.

[0112] In bagging, a training set is sampled, random, independent bootstrap replicas are generated, a decision rule is constructed for each of these, and the final decision rule is aggregated by simple majority voting. See, for example, Breiman, 1996, Machine Learning 24, pp. 123–140, and Efron & Tibshirani, An Introduction to Boostrap, Chapman & Hall, New York, 1993, the full content of which is incorporated herein by reference.

[0113] In boosting, decision rules are constructed based on weighted versions of the training set, depending on previous classification results. Initially, all constructs under consideration have equal weights, and the first decision rule is constructed based on this dataset. Next, the weights are modified based on the performance of the decision rule. The weights of misclassified constructs are increased, and the next decision rule is boosted with the modified training set. In this way, a series of training sets and decision rules are obtained, and then combined in the final decision rule by majority vote or weighted majority vote. For example, see Freund & Schapire, "Experiments with a new boosting algorithm," Proceedings 13th International Conference on Machine Learning, 1996, pp. 148-156, the full content of which is incorporated herein by reference.

[0114] The measurement data used in the techniques disclosed herein may be normalized. Normalization refers to the process of correcting for differences in the amount of assayed gene or protein levels and variability in the quality of the template used, for example, to eliminate undesirable sources of systematic variation in measurements associated with the processing and detection of gene or protein expression. Other sources of systematic variation may be due to laboratory processing conditions.

[0115] In some cases, normalization methods are used to normalize laboratory processing conditions. Non-exclusive examples of normalizing laboratory processing that can be used with this technique include, but are not limited to, taking into account systematic differences between instruments, reagents, and / or devices used during the data generation process, as well as the date and / or time or elapsed time of data collection.

[0116] The assay can be normalized by incorporating the expression of a specific normalization reference gene or protein that is known to have stable and consistent expression levels in its particular sample type and does not show significant differences in expression levels under relevant conditions. Suitable normalization genes and proteins that can be used in this disclosure include housekeeping genes. For example, see E. Eisenberg et al., Trends in Genetics 19(7):362-365 (2003), the full content of which is incorporated herein by reference. In some cases, normalization biomarkers (genes and / or proteins), also called reference genes, are known not to show significantly different expression levels when comparing subjects with heart disease with controls without heart disease. In some cases, it may be useful to add a stable isotope-labeled standard that can be used as an entity with known properties for use in data normalization and that can represent it. In other cases, a standard fixed sample can be measured in each analytical batch to account for instrument and daily measurement variability.

[0117] Machine learning models for sub-selecting identifying biomarkers and, optionally, characteristics of the subject, and constructing classification models are used in some of the methods and systems herein to determine scores for clinical outcomes. Examples of such algorithms are described above. These algorithms can assist in selecting the composition of key biomarkers and can translate underlying measurements into scores or probabilities related, for example, to classifications of clinical outcomes, disease risk, likelihood of disease, presence or absence of disease, treatment response, and / or disease status.

[0118] A machine learning model can output a biomarker score. The machine learning model can determine whether a subject is at risk of or has heart disease. In some embodiments, the output of the machine learning model is a determination of whether the subject is at risk of or has heart disease. The machine learning model may also be a classifier for heart disease. In some embodiments, the machine learning model can be used to classify whether a subject has heart disease, for example, based on the subject's biomarker profile. In some embodiments, the classifier or classification has a sensitivity and specificity of over 80% (e.g., at least one or both are over 90%).

[0119] Biomarker scores can be determined by comparing a target-specific biomarker profile with a reference biomarker profile. The reference biomarker profile may reflect a known diagnosis. For example, a biomarker profile may represent a positive diagnosis of cardiac disease. Alternatively, a reference biomarker profile may represent a negative diagnosis of cardiac disease. In some cases, an increase in score indicates an increased probability of one or more of the following events occurring: poor clinical outcome, good clinical outcome, high disease risk, low disease risk, complete response, partial response, stable disease, no response, or recommended treatment for disease management. In some cases, a decrease in the quantitative score indicates an increased probability of one or more of the following events occurring: poor clinical outcome, good clinical outcome, high disease risk, low disease risk, complete response, partial response, stable disease, no response, or recommended treatment for disease management.

[0120] When the biomarker profile from the subject and the reference biomarker profile are similar, it often indicates an increased probability of one or more of the following events occurring: poor clinical outcome, good clinical outcome, high disease risk, low disease risk, complete response, partial response, stable disease, no response, or recommended treatment for disease management. In some applications, when the biomarker profiles between the subject and the reference differ, it indicates an increased probability of one or more of the following events occurring: poor clinical outcome, good clinical outcome, high disease risk, low disease risk, complete response, partial response, stable disease, no response, or recommended treatment for disease management.

[0121] The results will be provided to the subjects, healthcare professionals, or other experts. The results may include health recommendations, such as, for example, recommendations to confirm or independently assess the risk of heart disease using one or more cardiomyopathy tests.

[0122] Recommendations may include information related to the treatment regimen. The effectiveness of a regimen can, in some cases, be assessed by comparing the subject's biomarker profiles at a first time point, possibly pre-treatment, and a second subsequent time point, possibly post-treatment. Biomarker profiles can be evaluated by comparing them to each other, to each reference, or otherwise, to determine whether they have demonstrated sufficient effectiveness in addressing the cardiac disease or associated signs and symptoms to warrant continuing, increasing, switching to an alternative regimen, or discontinuing the treatment regimen. Some assessments are based on comparing the subject's biomarker profiles at multiple time points, such as at least one pre-treatment time point and at least one post-treatment time point. Biomarker profiles can be compared to each other, to at least one reference biomarker panel level, or both to each other and to at least one reference biomarker panel level.

[0123] Treatment of amyloid-transthyretin cardiomyopathy In some embodiments, the disclosure includes a method for selecting patients to receive treatment with cardiac drugs, which includes the step of detecting the level of NfL in a sample obtained from a subject.

[0124] In some embodiments, the Disclosure includes a method for treating a subject at risk of or suffering from cardiac disease, comprising administering a therapeutically effective amount of a cardiac agent to the subject. In some embodiments, the subject expresses an NfL level above a threshold. In some embodiments, the method further includes determining that the subject expresses an NfL level above a threshold. In some embodiments, it is determined that the subject expresses an NfL level above a threshold before administration. In some embodiments, the cardiac agent is an angiotensin-converting enzyme (ACE) inhibitor, angiotensin II receptor blocker, angiotensin receptor-neprilysin inhibitor, anticoagulant or hemodiluent, antiplatelet agent, aspirin, beta-blocker, calcium channel blocker, cholesterol ester transfer protein (CETP) inhibitor, cholesterol absorption inhibitor, digitalis preparations (e.g., digitalis glycosides), diuretics, dual antiplatelet therapy, fibrates, niacin, statins, vasodilators, or combinations thereof.For example, cardiac medications include apixaban, dabigatran, edoxaban, heparin, rivaroxaban, warfarin, aspirin, clopidogrel, dipyridamole, prasugrel, ticagrelor, benazepril, captopril, enalapril, fosinopril, moexipril, perindopril, quinapril, ramipril, trandolapril, azilsartan, candesartan, eprosartan, irbesartan, losartan, olmesartan, telmisartan, valsartan, sacubitril, sacubitril / valsartan, acebutolol, atenolol, betaxolol, bisoprolol / hydrochlorothiazide, bisoprolol, metoprolol, nadolol, and propranol. This may include olol, sotalol, carvedilol, labetalol hydrochloride, amlodipine, diltiazem, felodipine, nifedipine, nimodipine, nisoldipine, verapamil, atorvastatin, fluvastatin, lovastatin, pitavastatin, pravastatin, rosuvastatin, simvastatin, niacin, ezetimibe, ezetimibe / simvastatin, digoxin, acetazolamide, amiloride, bumetanide, chlorothiazide, chlorthalidone, furosemide, hydrochlorothiazide, indapamide, metalozone, spironolactone, torsemide, isosorbide dinitrate, isosorbide mononitrate, hydralazine, nitroglycerin, minoxidil, or a combination thereof.

[0125] In some embodiments, the biomarkers disclosed herein can be used to screen patients for effective treatments for cardiac disease. In some embodiments, treatments for cardiac disease include ACE inhibitors, angiotensin II receptor blockers, angiotensin receptor-neprilysin inhibitors, anticoagulants or hemodilutants, antiplatelet agents, aspirin, beta-blockers, calcium channel blockers, CETP inhibitors, cholesterol absorption inhibitors, digitalis preparations (e.g., digitalis glycosides), diuretics, dual antiplatelet therapy, fibrates, niacin, statins, vasodilators, or combinations thereof. For example, therapeutic drugs include apixaban, dabigatran, edoxaban, heparin, rivaroxaban, warfarin, aspirin, clopidogrel, dipyridamole, prasugrel, ticagrelor, benazepril, captopril, enalapril, fosinopril, moexipril, perindopril, quinapril, ramipril, trandolapril, azilsartan, candesartan, eprosartan, irbesartan, losartan, olmesartan, telmisartan, valsartan, sacubitril, sacubitril / valsartan, acebutolol, atenolol, betaxolol, bisoprolol / hydrochlorothiazide, bisoprolol, metoprolol, nadolol, and propranol. This may include olol, sotalol, carvedilol, labetalol hydrochloride, amlodipine, diltiazem, felodipine, nifedipine, nimodipine, nisoldipine, verapamil, atorvastatin, fluvastatin, lovastatin, pitavastatin, pravastatin, rosuvastatin, simvastatin, niacin, ezetimibe, ezetimibe / simvastatin, digoxin, acetazolamide, amiloride, bumetanide, chlorothiazide, chlorthalidone, furosemide, hydrochlorothiazide, indapamide, metalozone, spironolactone, torsemide, isosorbide dinitrate, isosorbide mononitrate, hydralazine, nitroglycerin, minoxidil, or a combination thereof.

[0126] Example kit Kits comprising one or more anti-NfL agents and instructions for use (e.g., for therapeutic, prophylactic, or diagnostic purposes) are also provided by this disclosure. In some embodiments, the kit is used in an in vitro diagnostic assay for diagnosing cardiac disease. In some embodiments, the kit of this disclosure further comprises cardiac agents.

[0127] In some embodiments, one or more anti-NfL agents include an antibody agent. In some embodiments, one or more antibody agents are labeled with a detectable portion. In some embodiments, the kit includes a detection agent (e.g., one or more acridinium ester molecules). In some embodiments, one or more antibody agents are labeled with one or more acridinium ester molecules. In some embodiments, the kit further includes one or more secondary antibody agents that bind to one or more of the anti-NfL antibody agents.

[0128] In some embodiments, one or more anti-NfL agents include nucleic acid probes. In some embodiments, at least a portion of each nucleic acid probe hybridizes to one or more portions of nucleotides encoding NfL. The nucleotides encoding NfL may be DNA (e.g., cDNA) or RNA (e.g., mRNA). In some embodiments, the nucleic acid probes are labeled with one or more detection agents (e.g., the detection agent indicates the presence of nucleotides encoding NfL).

[0129] In some embodiments, the kit includes one or more control samples. In some embodiments, the control samples include one or more standards. In some embodiments, the standards include recombinant NfL. In some embodiments, the standards include synthetic NfL nucleic acids.

[0130] Kits based on the disclosed technology may also include other components such as solvents or buffers, stabilizers or preservatives, and / or agents for treating a condition or disorder. Alternatively, other components may be included in the kit but provided in a different composition or container from the anti-NfL agent. In such embodiments, the kit may include instructions for mixing the anti-NfL agent with the other components, or instructions for using the anti-NfL agent together with the other components. In some embodiments, a kit used in accordance with this disclosure may include a reference sample or control sample, instructions for handling the sample, instructions for performing the test on the sample, instructions for interpreting the results, and / or buffers and / or other reagents necessary for performing the test.

[0131] A single biomarker may be useful for the detection and / or diagnosis of cardiac disease, as described herein. This disclosure further provides insight into why NfL is particularly useful for the detection and / or diagnosis of cardiac disease. Accordingly, the methods, compositions, and kits described herein can be used in assays to assess the risk of cardiac disease, assays to assess whether a subject should undergo further cardiac examination, and / or assays to diagnose cardiac disease, based on the detection or measurement of NfL in a sample (e.g., a biological sample obtained from a subject).

[0132] The methods and kits provided herein can detect cardiac disease in a sample with sensitivity and specificity that is medically practical and reliable enough. The methods and kits described herein for the detection and / or diagnosis of cardiac disease in a subject can detect cardiac disease with sensitivity greater than 75%, greater than 80%, greater than 85%, greater than 90%, greater than 95%, greater than 96%, greater than 97%, greater than 98%, greater than 99%, or about 100%. In some embodiments, the methods and kits provided herein can detect cardiac disease with sensitivity between about 70% and 100%, between about 80% and 100%, or between about 90% and 100%. In some embodiments, the methods and kits provided herein can detect cardiac disease with specificity greater than 70%, greater than 75%, greater than 80%, greater than 85%, greater than 90%, greater than 95%, greater than 96%, greater than 97%, greater than 98%, greater than 99%, or about 100%. In some embodiments, the methods and kits provided herein can detect cardiac disease with specificity between approximately 50% and 100%, between approximately 60% and 100%, between approximately 70% and 100%, between approximately 80% and 100%, or between approximately 90% and 100%. In some embodiments, the methods and kits provided herein can detect cardiac disease with sensitivity and specificity of 50% or more, 60% or more, 70% or more, 75% or more, 80% or more, 85% or more, or 90% or more. In some embodiments, the methods and kits provided herein can detect cardiac disease with sensitivity and specificity between approximately 50% and 100%, between approximately 60% and 100%, between approximately 70% and 100%, between approximately 80% and 100%, or between approximately 90% and 100%.

[0133] Exemplary composition Furthermore, compositions are also provided herein. In some embodiments, the composition comprises NfL and one or more anti-NfL agents. In some embodiments, one or more anti-NfL agents in the compositions provided herein comprise an antibody agent. In some embodiments, one or more antibody agents are labeled with a detectable portion. In some embodiments, the composition comprises a detection agent (e.g., one or more acridinium ester molecules). In some embodiments, one or more antibody agents are labeled with one or more acridinium ester molecules. In some embodiments, the composition comprises one or more secondary antibody agents conjugated to one or more anti-NfL antibody agents.

[0134] In some embodiments, one or more anti-NfL agents included in the compositions provided herein comprise nucleic acid probes. In some embodiments, at least a portion of each nucleic acid probe hybridizes with one or more portions of nucleotides encoding NfL. The nucleotides encoding NfL may be DNA (e.g., cDNA) or RNA (e.g., mRNA). In some embodiments, the nucleic acid probes are labeled with one or more detection agents (e.g., the detection agent indicates the presence of nucleotides encoding NfL).

[0135] In some embodiments, the composition includes one or more control samples. In some embodiments, the control samples include one or more standards. In some embodiments, the standards include recombinant NfL. In some embodiments, the standards include synthetic NfL nucleic acids. Compositions produced by this technique may include other components, such as solvents or buffers, stabilizers or preservatives, and / or agents for treating conditions or disorders.

[0136] Exemplary computer system Figure 1 is a block diagram of a computer system 1100, which can be used in the operations described and illustrated herein and is disclosed according to one embodiment. The system 1100 includes a processor 1110, memory 1120, storage device 1130, and input / output device 1140. The components 1110, 1120, 1130, and 1140 are interconnected using a system bus 1150. The system may include an analytical instrument 1160 for determining the level of NfL in a sample.

[0137] The methods described herein can be implemented by a computer system 1100 having a processor 1110 that executes specific instructions in a computer program stored in memory 1120. In some embodiments, the computer system 1100 may be configured to output a biomarker score based on the reception of a biomarker profile and / or NfL level, demographic factors, and / or image-based biomarkers. The computer program may include instructions for the computer system 1100 to select appropriate next steps, including additional medication, treatment, and / or additional tests (e.g., cardiomyopathy tests) for the subject.

[0138] In some embodiments, the computer program can be configured so that the computer system 1100 identifies subjects for further examination (e.g., cardiomyopathy examination), identifies subjects as being at risk or having heart disease, and / or identifies subjects to receive medication based on received data (e.g., biomarker profiles), and uses that data to calculate biomarker scores. The computer system 100 using this technique can rank the identified subsequent steps based on the biomarker profiles and demographic factors and / or image-based biomarkers. The computer system 1100 can adjust the ranking, for example, based on the clinical response of subjects or their families who have or are suspected of having heart disease.

[0139] Therefore, the processor 1110 can process instructions for execution within the computer system 1100. In one embodiment, the processor 1110 is a single-threaded processor. In another embodiment, the processor 1110 is a multi-threaded processor. The processor 1110 can process instructions stored in memory 1120 or storage device 1130, including those for receiving or transmitting information via input / output device 1140.

[0140] Memory 1120 stores information within the system 1100. In one embodiment, memory 1120 is a computer-readable medium. In one embodiment, memory 1120 is a volatile memory unit. In another embodiment, memory 1120 is a non-volatile memory unit (e.g., a non-temporary computer-readable medium). Storage device 1130 can provide high-capacity storage to the computer system 1100. In one embodiment, storage device 1130 is a volatile or non-volatile computer-readable medium. Thus, the non-temporary computer-readable medium of memory 1120 may, when executed by the processor 1110, include executable instructions that cause the processor 1110 to perform operations including methods provided herein. For example, non-temporary computer-readable medium including executable instructions that, when executed by the processor 1110, cause the processor 1110 to perform operations including methods 1200 or 1300.

[0141] The input / output device 1140 provides input / output operations to the computer system 1100. In one embodiment, the input / output device 1140 includes a keyboard and / or a pointing device. In one embodiment, the input / output device 1140 includes a display unit for displaying a graphical user interface. In some examples, the input / output device 1140 is a touchscreen.

[0142] The computer system 1100 can be used to build databases in several examples. Figure 2 shows a flowchart of an exemplary method 1200 for generating a database used to identify subjects for further examinations (e.g., cardiomyopathy examinations), to identify subjects who are at risk of or have heart disease, and / or to identify subjects who should receive medication. Any of the methods described and illustrated herein, including method 1200 and / or method 1300, can be performed in whole or in part by the computer system 1100. For example, a computer program product may include instructions to cause the processor 1110 to perform the steps of method 1200 or method 1300.

[0143] In step 1205 of this example, the computer system 1100 trains or acquires a machine learning model. In some examples, the machine learning model is trained using training data that includes a biomarker profile of a subject and the subject's health status corresponding to that biomarker profile (e.g., healthy, suffering from heart disease). Each biomarker profile in this example includes at least two biomarker levels, including NfL. In some examples, the training data further includes demographic data and / or image-based biomarkers correlated with each subject, as well as their corresponding biomarker profiles and health statuses.

[0144] In some examples, machine learning models are derived from bagging, boosting, or additive decision trees, neural boosting, bootstrap forests, boosted trees, or support vector machines, but in other examples, other techniques and / or machine learning algorithms or models may also be used. In some examples, the machine learning model is a cardiac disease classifier, trained to provide an output indicating whether a subject has cardiac disease (e.g., based on an output biomarker score), as will be described in more detail below. Thus, the machine learning model is trained to analyze the probability that a subject is at risk of or has cardiac disease, by leveraging, for example, pattern and vector search to correlate input data related to the subject with the subject's health status in the training data, as will also be described in more detail below.

[0145] In step 1210, the computer system 1100 obtains a biomarker profile of the subject (e.g., levels of one or more biomarkers in a sample taken from the subject, including at least NfL levels) from a computer device (e.g., computer device 2404) via one or more communication networks (e.g., network 2408). In some examples, a machine learning model is trained with training data related to a human subject, and / or the biomarker profile obtained in step 1210 is associated with a human subject. In these examples, the sample may include one or more of blood, serum, plasma, or cardiac tissue, but in other examples, other types of subjects and / or samples may also be used. In some cases, the sample may be obtained via, for example, an analytical instrument 1160.

[0146] Depending on the circumstances, the computer system 1100 may acquire additional data, including demographic data and / or image-based biomarkers, for subjects corresponding to the acquired biomarker profiles. The demographic data may correspond to, for example, one or more demographic factors of a subject (e.g., age, sex), but any other demographic factors, including those identified above, may also be used. Furthermore, image-based biomarkers may include at least left ventricular septal thickness or ejection fraction, but other image-based biomarkers may also be used.

[0147] In an example where an image-based biomarker is acquired in step 1210, the computer system 1100 may be configured to determine the image-based biomarker by acquiring one or more measurements through analysis of one or more acquired images associated with the subject. For example, the analysis may include, for instance, pattern recognition performed on the acquired images, and / or segmentation and classification of one or more components extracted from the acquired images, but other types of analysis may also be used in other examples. Depending on the circumstances, segmentation and classification may be performed by a first machine learning model, or a second machine learning model trained to extract components from images.

[0148] In some examples, a computer program of the computer system 1100 may include instructions for displaying a suitable graphical user interface (GUI) on the input / output device 1140, and the GUI may prompt the user to input NfL levels using the input / output device 1140, such as a keyboard. In other examples, the computer system 1100 may provide an application programming interface (API) or other integration to a third-party service or other device for obtaining a biomarker profile of interest, demographic data, and / or image-based biomarkers.

[0149] In step 1220, the computer system 1100 applies the machine learning model trained in step 1205 to the input to generate a biomarker score for the subject. The input includes at least the biomarker profile received in step 1210, but demographic data and / or image-based biomarkers received in step 1210 can also form part of the input to the machine learning model. In this example, the biomarker score indicates the probability that the subject is at risk of or has heart disease. Thus, in some examples, the biomarker score is effectively calculated or determined from the subject's biomarker profile, demographic data or factors, and / or image-based biomarkers. The biomarker score may be binary, indicating that the subject belongs to one of two classes. Alternatively, the biomarker score may include a range, or be a value within a range, and other types of biomarker scores can also be generated in step 1220.

[0150] In step 1230, the computer system 1100 provides the computer device via a communication network with biomarker scores and / or indicators of cardiac disease derived therefrom, according to the biomarker profile, and optionally stores the calculated biomarker scores. The computer system 1100 may, for example, store the biomarker scores in a storage device 1130.

[0151] In some examples, the indicator of heart disease includes a binary index determined based on the classification of a subject into one of two classes corresponding to whether or not it has heart disease, by a machine learning model. In some cases, one or more of the sensitivity or specificity of the classifier in these examples may exceed 80%, although a different percentage may be used or requested before deploying the machine learning model to a production environment. In other examples, a biomarker score is used by a computer system 1100 to determine at least one of the following: (i) the status of heart disease in a subject, (ii) whether or not the subject has heart disease (e.g., a heart disorder or condition), or (iii) the probability that the subject has heart disease.

[0152] Furthermore, or alternatively, the computer system 1100 may provide readout information (e.g., via the input / output device 1140) including biomarker scores and / or indicators of cardiac disease (e.g., the likelihood or probability that the subject has or is at risk of having cardiac disease, as determined from the biomarker scores). The readout information may include suggested next steps for the subject related to the received biomarker profile and / or confidence levels related to the calculated biomarker scores. For example, as described in more detail below with reference to Method 1300, biomarker scores within a first range or above a first threshold may correlate with next steps for re-examination after a recommended period to provide a comparative analysis to determine trends in NfL levels that may indicate an increased risk of cardiac disease.

[0153] In another example, a biomarker score that is within a second range or exceeds a second threshold (e.g., a value higher than the first threshold) may correlate with the next step in confirmatory testing (e.g., echocardiography, magnetic resonance imaging (CMR), scintigraphy, and / or cardiac biopsy). In this example, the biomarker score may indicate a subject who is more likely to have heart disease. In other examples, other types of signs and recommended next steps may also be used.

[0154] In the exemplary method 1300 illustrated in Figure 3, in step 1310, the computer system 1100 detects the level of a biomarker in the sample (e.g., from the subject), including at least the level of NfL. The level of NfL can be detected, for example, using the analytical instrument 1160 of the computer system 1100, but in other examples, other methods may be used to detect the level of the biomarker.

[0155] In step 1320, the computer system 1100 uses NfL levels to obtain a target biomarker profile corresponding to the sample.

[0156] In step 1330, the computer system 1100 calculates a biomarker score from the biomarker profile. The biomarker score can be calculated from (i) the biomarker profile, and (ii) demographic factors and / or image-based biomarkers. In some examples, the biomarker score of interest can be generated using a trained machine learning model, as described in more detail above with reference to method 1200.

[0157] In step 1340, the computer system 1100 stores the calculated biomarker score in, for example, a storage device 1130. Alternatively, the computer system 1100 may provide readout information, including the biomarker score, via, for example, an input / output device 1140. The readout information may also include the next steps proposed for the subject related to the received biomarker profile, as described in more detail above, and / or the confidence level associated with the calculated biomarker score.

[0158] In some examples, the sample in which a biomarker level is detected in step 1310 is a second sample, or a subsequent sample in a series of samples, obtained from the subject and analyzed according to method 1200 or 1300. In this example, the biomarker score and / or index of cardiac disease can be generated based on a comparison of the level of NfL detected in the second sample, obtained from the subject at a second time after the first time the first sample was obtained, with the level of NfL detected in the first sample. In some examples, the biomarker score and / or index of cardiac disease is generated based on the level of NfL detected in the second sample exceeding the level of NfL detected in the first sample (which may be an arbitrary amount, percentage, or coefficient, e.g., by a threshold amount).

[0159] In other examples, the computer system 1100 is configured to compare the NfL level change between two samples with a second NfL change determined for a control over a first period corresponding to the time difference between when the two samples were acquired. In these examples, the control may be less than 3 years in age from the subject and the control does not have heart disease, but other age differences can also be used.

[0160] In yet another example, the computer system 1100 may be configured to analyze trends in NfL levels detected from a series of samples taken from a subject to generate a biomarker score in step 1330 and / or an index of cardiac disease in step 1340. For example, a gradual increase in NfL over time may indicate the type of damage accumulating due to decreased cerebral perfusion caused by impaired cardiovascular function. Other methods for generating the biomarker score in step 1330 and / or the index of cardiac disease in step 1340 may also be used in other examples.

[0161] Exemplary embodiments of the disclosed technology can be executed by computer system 1100 locally or over a network. Figure 4 shows an exemplary network environment 2400 used in the examples of the technology described herein. The network environment 2400 may include one or more resource providers 2402a, 2402b, 2402c (collectively 2402). Each resource provider 2402 may include a computing resource. In some implementations, the computing resource may include any hardware (for example, the computing resource may be computer system 1100) and / or software used to process data in accordance with any of the methods described herein, including methods 1200 and 1300. For example, the computing resource may include hardware and / or software capable of running algorithms, computer programs, and / or computer applications. In some implementations, the exemplary computing resource may include an application server and / or database with storage and retrieval capabilities. Each resource provider 2402 may connect to any other resource provider 2402 within the network environment 2400. In some implementations, resource providers 2402 may be connected via computer network 2408. Each resource provider 2402 can connect to one or more computer devices 2404a, 2404b, 2404c (collectively 2404) via the computer network 2408.

[0162] The network environment 2400 may include a resource manager 2406. The resource manager 2406 may connect to resource providers 2402 and computer devices 2404 via a computer network 2408. In some implementations, the resource manager 2406 can facilitate the provision of computing resources to one or more computer devices 2404 by one or more resource providers 2402. The resource manager 2406 can receive requests for computing resources from a specific computer device 2404. The resource manager 2406 can identify one or more resource providers 2402 that can provide the computing resources requested by the computer device 2404. The resource manager 2406 can select a resource provider 2402 to provide the computing resources. The resource manager 2406 can facilitate connections between resource providers 2402 and a specific computer device 2404. In some implementations, the resource manager 2406 can establish connections between a specific resource provider 2402 and a specific computer device 2404. In some implementations, the resource manager 2406 can redirect a specific computer device 2404 to a specific resource provider 2402 that has the requested computing resources. [Examples]

[0163] Example Performance results of machine learning models A machine learning model was trained and evaluated for its ability to classify subjects based on the presence or absence of cardiac disease. Biomarker profiles, including measured levels of NfL, were obtained for a population of subjects with diverse racial / ethnic characteristics, sex, age, and geographical location. This population included subjects diagnosed with cardiac disease and those without. A bootstrap forest machine learning model was trained and evaluated using k-fold cross-validation (k=5). The results are shown in Figure 6 as a box plot with the area under the receiver operating characteristic (ROC) curve (AUC) on the y-axis (showing 95% confidence interval diamonds). Figure 5 shows that, on average over the training fold, the NfL-based machine learning model was able to distinguish between patients with and without cardiac disease. [Examples]

[0164] The relationship between NfL and age Biomarker profiles, including measured NfL levels, were obtained from a population of subjects with various cardiac disease states. Subjects had either ATTR-CM, non-ATTR-CM HF-PEF, or normal cardiac function. Figure 6 shows that, at least in older subjects (i.e., ~60 years and older), blood NfL levels increased with age in ATTR-CM and non-ATTR-CM HF-PEF subjects, but not in subjects with normal cardiac function. Therefore, changes in NfL levels over time can be used to predict whether an individual has cardiac disease and / or to monitor the progression of cardiac disease. For example, for comparison, subjects' NfL levels may be monitored regularly, such as annually. Monitoring can begin at a specific age at which NfL levels are expected to increase over time, e.g., at least 40, at least 50, at least 60, or at least 65 years, although other threshold ages may be used.

[0165] While various exemplary embodiments incorporating the principles of this teaching have been disclosed, this teaching is not limited to the disclosed embodiments. Rather, this application is intended to cover all variations, uses, or adaptations of this teaching and to utilize its general principles. Furthermore, this application is intended to cover any deviations from this disclosure within known or customary practices in the art to which these teachings belong.

[0166] The detailed description above refers to the accompanying drawings, which form part of this specification. In the drawings, unless otherwise indicated in the context, similar symbols generally identify similar components. The exemplary embodiments described herein are not intended to be limiting. Other embodiments may be used and other modifications made without departing from the spirit or scope of the subject matter presented herein. The various configurations of this disclosure described herein and illustrated in the drawings may be arranged, substituted, combined, separated, and designed in a wide variety of different configurations, all of which are expressly intended herein.

[0167] Aspects of this technical solution are described herein with reference to flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the technical solution. It will be understood that each block in the flowcharts and / or block diagrams, as well as combinations of blocks in the flowcharts and / or block diagrams, can be implemented by computer-readable program instructions.

[0168] These computer-readable program instructions are provided to the processor of a special-purpose computer or other programmable data processing device for the production of a machine, thereby creating means for instructions executed via the processor of the computer or other programmable data processing device to implement functions / operations specified by blocks or multiple blocks of a flowchart and / or block diagram. These computer-readable program instructions may also be stored in a computer-readable storage medium that can instruct a computer, programmable data processing device, and / or other device to function in a particular manner, thereby including a product containing instructions that implements modes of functions / operations specified by blocks or multiple blocks of a flowchart and / or block diagram.

[0169] Computer-readable program instructions can also be loaded into a computer, other programmable data processing device, or other device to produce a computer implementation process by having the computer, other programmable device, or other device execute a series of operations, thereby enabling the instructions executed by the computer, other programmable device, or other device to implement the functions / operations specified by blocks or multiple blocks in a flowchart and / or block diagram.

[0170] The terms “worker,” “algorithm,” “system,” “module,” “engine,” or “architecture” as used herein are not intended to limit any particular implementation to achieving and / or performing any operation, process, etc. resulting from and / or performed by them. An algorithm, system, module, engine, and / or architecture may be, but not limited to, software, hardware, and / or firmware, or any combination thereof, that performs a specified function, including, without limitation, the use of any general-purpose and / or specialized processor in combination with appropriate software loaded into machine-readable memory and executed by the processor. Furthermore, any names associated with a particular algorithm, system, module, and / or engine are for convenience of reference only and are not intended to limit any particular implementation unless otherwise specified. In addition, any functionality resulting from an algorithm, system, module, engine, and / or architecture can be equally performed by multiple algorithms, systems, modules, engines, and / or architectures, which may be incorporated into and / or combined with the functionality of other algorithms, systems, modules, engines, and / or architectures of the same or different types, or distributed across one or more algorithms, systems, modules, engines, and / or architectures of various configurations.

[0171] The flowcharts and block diagrams in the drawings illustrate the architecture, functionality, and operation of possible implementations of the system, method, and computer program product according to various embodiments of the technical solution. In this regard, each block in a flowchart or block diagram may represent a module, segment, or part of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions shown in a block may be executed in an order different from the order shown in the diagram. For example, two consecutively shown blocks may actually be executed substantially simultaneously, depending on the functions involved, or the blocks may sometimes be executed in reverse order. It should also be noted that each block in a block diagram and / or flowchart, as well as combinations of blocks in a block diagram and / or flowchart, may be implemented by a special-purpose hardware-based system that performs a specified function or operation, or executes or implements a special-purpose hardware and computer instruction combination.

[0172] A second action can be said to be "responding" to the first action, regardless of whether the second action arises directly or indirectly from the first action. A second action may still be a response to the first action even if it occurs substantially later than the first action. Similarly, even if an intervening action occurs between the first and second actions, and one or more of the intervening actions directly trigger the second action, the second action can still be said to be a response to the first action. For example, if the first action sets a flag, and the setting of that flag subsequently triggers the second action, the second action may be a response to the first action.

[0173] This disclosure should not be limited to the specific embodiments described in this application, which are intended as examples of various configurations. As will be apparent to those skilled in the art, many modifications and variations can be made without departing from the spirit and scope. In addition to those enumerated herein, functionally equivalent methods and apparatus within the scope of this disclosure will be apparent to those skilled in the art from the foregoing description. It should be understood that this disclosure is not limited to any particular method, reagent, compound, composition or biological system, which may vary. It should also be understood that the terms used herein are intended solely to describe specific embodiments and are not intended to limit them.

[0174] With regard to the use of substantially any plural and / or singular terms herein, those skilled in the art can convert from plural to singular and / or singular to plural as appropriate to the context and / or use. For clarity, various singular / plural permutations are explicitly listed herein.

[0175] Those skilled in the art will understand that the terms used herein are generally intended to be “open” terms (for example, “including” should be interpreted as “including but not limited,” “having” should be interpreted as “having at least,” and “includes” should be interpreted as “including but not limited,” etc.). Various compositions, methods, and apparatus are described using the term “comprising” (meaning “including but not limited”) which includes various components or processes, but these compositions, methods, and apparatus may also “consist essentially of” or “consist of” various components and processes, and such terms should be interpreted as defining essentially closed member groups.

[0176] In this specification, the singular forms "a," "an," and "the" include plural references unless the context clearly indicates otherwise. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art. Nothing in this disclosure shall be construed as acknowledging that the embodiments described herein do not take precedence over prior art disclosures.

[0177] In addition, even if a specific number is explicitly stated, a person skilled in the art will recognize that such a statement should be interpreted as meaning at least the stated number (for example, the statement “twice” without other modifiers means at least two times, or two or more times). Furthermore, when a similar convention is used for “at least one of A, B, and C, etc.,” such a construction is generally intended to be understood by a person skilled in the art (for example, “a system having at least one of A, B, and C” includes, but is not limited to, A only, B only, C only, A and B together, A and C together, B and C together, and / or having A, B, and C together, etc.). Where a similar convention is used for “at least one of A, B, or C,” such a configuration is generally intended to be understood by those skilled in the art (for example, “a system having at least one of A, B, or C” includes, but is not limited to, A only, B only, C only, A and B together, A and C together, B and C together, and / or having A, B, and C together). It will be further understood by those skilled in the art that substantially any separate word and / or phrase representing two or more alternative terms, regardless of the description, the embodiment of the sample, or the drawings, should be understood to be intended to include the possibility of including one of the terms, either of the terms, or both of the terms. For example, the phrase “A or B” will be understood to include the possibilities of “A” or “B” or “A and B.”

[0178] In addition, if the structure of this disclosure is described using the Markush Group, a person skilled in the art will recognize that this disclosure also describes any individual member or subgroup of a member of the Markush Group.

[0179] As will be understood by those skilled in the art, all scopes disclosed herein also encompass any and all possible sub-scopes and combinations thereof for any and all purposes, including in terms of providing written explanations. Any enumerated scope can be readily recognized as sufficiently describing and feasible to divide the same scope into at least equal halves, thirds, quarters, fifths, tenths, etc. As a non-restrictive example, each scope discussed herein can be readily divided into a lower third, a middle third, an upper third, etc. Also, as will be understood by those skilled in the art, all words such as “up to,” “at least,” etc., include the stated numerical values ​​and refer to scopes that can subsequently be divided into sub-scopes, as discussed above. Finally, as will be understood by those skilled in the art, a scope includes each individual member. Thus, for example, a group having 1 to 3 components refers to a group having 1, 2, or 3 components. Similarly, a group having 1 to 5 components refers to a group having 1, 2, 3, 4, or 5 components, and so on.

[0180] Various features and functions disclosed above, as well as other features and functions, or their substitutes, can be combined with many other different systems or applications. Various alternatives, modifications, variations, or improvements not currently anticipated or expected may be made later by those skilled in the art, each of which is also intended to be incorporated by the disclosed embodiments.

Claims

1. A method for predicting cardiac disease, implemented by a computer system, is: Obtaining a biomarker profile that includes a first level of one or more biomarkers in a first sample obtained from a subject, wherein one or more biomarkers include nerve filament light chains (NfLs); Applying a first machine learning model to the input to generate a biomarker score for a subject, wherein the input includes at least a biomarker profile, and the biomarker score indicates the probability that the subject is at risk of or has heart disease; and Outputting biomarker scores, or indicators of cardiac disease determined from biomarker scores, to a computer device via one or more communication networks. Methods that include...

2. The method according to claim 1, further comprising obtaining demographic data corresponding to one or more demographic factors of a subject from a computer device via a communication network, wherein the input further comprises demographic data, and the demographic factors include at least one or more of the sex or age of the subject.

3. The method according to claim 1 or 2, further comprising obtaining one or more image-based biomarkers of a subject from a computer device via a communication network, wherein the input further comprises image-based biomarkers, the image-based biomarkers comprising at least left ventricular septal thickness or ejection fraction.

4. To determine image-based biomarkers, the analysis involves analyzing one or more acquired images related to the subject to obtain one or more measurements, and the analysis may include one or more of the following: Pattern recognition performed on acquired images; or The method according to claim 3, comprising segmentation and classification using a first machine learning model or a second machine learning model having one or more configurations extracted from acquired images.

5. The method according to any one of claims 1 to 4, wherein the first machine learning model is derived from bagging, boosting, or additive decision tree, neural boosting, bootstrap forest, boosted tree, or support vector machine.

6. The first machine learning model is a classifier for cardiac disease, and the method further comprises classifying whether a subject has cardiac disease or not based on a biomarker score, wherein the index for cardiac disease includes a binary index determined based on the classification, and one or more of the sensitivity or specificity of the classifier is greater than 80%, according to any one of claims 1 to 5.

7. The method according to any one of claims 1 to 6, wherein the subject is a human subject, and the first sample comprises one or more of blood, serum, plasma, or cardiac tissue.

8. The method according to any one of claims 1 to 7, further comprising determining, based on a biomarker score, at least one of the following: (i) the cardiac disease status of the subject, (ii) whether or not the subject has cardiac disease, or (iii) the probability that the subject has cardiac disease.

9. The method according to any one of claims 1 to 8, further comprising training a first machine learning model using acquired training data, which includes biomarker profiles of multiple subjects and the health status of subjects corresponding to the biomarker profiles, wherein each biomarker profile includes levels of two or more biomarkers, including at least NfL.

10. The method according to any one of claims 1 to 9, wherein the first sample is obtained from the subject at a first time, and the method further comprises determining that the subject is at risk of or has heart disease if the second NfL level detected in the second sample obtained from the subject at a second time after the first time is above the first NfL level by a threshold amount.

11. The method according to any one of claims 1 to 10, wherein the first sample is obtained from the subject at a first time, and the method further comprises determining whether the subject is at risk of or has heart disease based on a change in the first NfL level determined based on a comparison of the first NfL level with a second NfL level detected in a second sample obtained from the subject at a second time after the first time.

12. The method according to claim 11, further comprising comparing a first NfL level change with a second NfL change determined for a control subject over a first period of time corresponding to the difference between a second time and a first time, wherein the control subject is no more than three years of age and does not have heart disease.

13. The method according to any one of claims 1 to 12, further comprising determining whether a subject is at risk of or has heart disease based on a third NfL level detected in each of a series of samples including a first sample, and on a second NfL level change determined over a second period of time in which a series of samples were obtained from the subject.

14. A non-temporary computer-readable medium comprising, when executed by a processor, an executable instruction causing the processor to perform the method according to any one of claims 1 to 13.

15. A computer system including memory and a processor, wherein the memory includes a first machine learning model, and which stores executable instructions thereon that, when executed by the processor, cause the computer system to perform the method according to any one of claims 1 to 13.

Images