Machine learning algorithms for patients with suspected acute coronary syndromes in the emergency department

Abstract

The present invention relates to a method for assessing a subject suspected of having an acute coronary syndrome. In particular, the present invention relates to a method for predicting the risk of an adverse event in said subject. Furthermore, the present invention relates to a method for predicting the need for myocardial revascularization in patients suspected of having an acute coronary syndrome. The method of the present invention may be implemented as a computer-implemented method.

Description

[Technical Field]

[0001] The present invention relates to a method for assessing a subject suspected of having an acute coronary syndrome. In particular, the present invention relates to a method for predicting the risk of an adverse event in said subject. Furthermore, the present invention relates to a method for predicting the need for myocardial revascularization in patients suspected of having an acute coronary syndrome. The method of the present invention may be performed as a computer-implemented method. [Background technology]

[0002] According to the 2014 National Hospital Ambulatory Care Survey, a total of 6.9 million patients visit the emergency department (ED) in the United States due to chest pain (CP), and another 3.4 million due to shortness of breath (Stoyanov KM, Biener M, Hund H, Mueller-Hennessen M, Vafaie M, Katus HA, Giannitsis E. Effects of crowding in the emergency department on the diagnosis and management of suspected acute coronary syndrome using rapid algorithms: an observational study. BMJ Open. 2020 Oct 8;10(10):e041757. doi:10.1136 / bmjopen-2020-041757. PMID:33033102; PMCID:PMC7545662). Only a fraction of these patients receive a final diagnosis of ACS, requiring hospitalization and invasive treatment strategies. The diagnostic workup requires admission to the ED, registration of a 12-lead electrocardiogram (ECG), blood tests to diagnose or rule out myocardial injury, evaluation of clinical symptoms and medical history, physical examination, and other diagnostic tests to diagnose ACS or differential diagnosis.Current 2020 European Society of Cardiology (ESC) Guidelines (Collet JP, Thiele H, Barbato E, Barthelemy O, Bauersachs J, Bhatt DL, Dendale P, Dorobantu M, Edvardsen T, Folliguet T, Gale CP, Gilard M, Jobs A, Juni P, Lambrinou E, Lewis BS, Mehilli J, Meliga E,Merkely B,Mueller C,Roffi M,Rutten FH,Sibbing D,Siontis GCM;ESC Scientific Document Group.2020 ESC Guidelines for the management of acute coronary syndromes in patients presenting without persistent ST-segment elevation ahead.Eur Heart J.2020 Aug 29:ehaa575.doi:10.1093 / eurheartj / ehaa575.Epub of print.PMID:32860058) recommends monitoring patients for ECG and vital signs unless myocardial injury has been ruled out.

[0003] Currently, the number of emergency department (ED) visits for nonspecific chest pain is increasing worldwide, causing overcrowding in busy EDs. ED physicians must apply strict admission criteria to address the relative shortage of hospital beds. Therefore, accurate diagnosis following risk stratification is essential to guide admission or discharge decisions and to select the appropriate selective invasive strategy, i.e., to identify patients who should undergo an invasive strategy rather than anatomical imaging or functional testing for myocardial ischemia. While the decision to accept and assign an invasive strategy is clear for patients classified as rule-in, less information exists for patients classified as rule-out and most patients in the observation zone. This situation has worsened since the 2020 ESC guidelines moved patients previously assigned to the intermediate risk category with a planned invasive strategy within 72 hours to the low-risk category.

[0004] In the research underlying the present invention, different methods were developed to improve the evaluation of patients with suspected ACS (see also the Examples section). These methods are referred to herein as "prognostic methods" and "predictive methods." The developed methods make it possible to a) predict the risk of adverse events and b) predict the need for myocardial revascularization in patients with suspected acute coronary syndrome.

[0005] As described in the Examples section, the developed methods are comprehensive and provide valuable support for two key challenges. Specifically, they allow for more accurate estimation of individual risk of 365-day and 180-day mortality than the GRACE score, the preferred clinical risk stratification tool (Level IIA of evidence) promoted by the 2020 ESC guidelines. Therefore, ED physicians can use a better risk prediction tool to justify their own decisions regarding discharge, with or without the recommendation of further outpatient diagnostic workup. Furthermore, even patients with unstable angina who are classified as rule-out and considered at low risk for major coronary events may have underlying obstructive coronary artery disease requiring coronary intervention. Currently, few tools provide information on the likelihood of CAD (coronary artery disease). The ESC guidelines for chronic coronary syndromes recommend the use of the ESC Consortium algorithm, a modified Diamond-Forrester algorithm, which has moderate ability to identify patients at risk for significant coronary artery stenosis (Genders TS et al. BMJ: British Medical Journal 2012 June 12, 344:e3485). Thus, the developed method provides a better estimate of the pre-test probability of CAD requiring revascularization, which may facilitate the decision on invasive strategy compared to coronary CT or functional stress testing.

[0006] Prognostic diagnostic methods The study cohort consisted of all consecutive patients who presented with suspected ACS at the Chest Pain Unit of Heidelberg University Hospital between June 30, 2016, and July 1, 2018 and were triaged to the "rule-out" or "observation zone" after retrospective adjudication by three cardiologists not involved in the patient's management. Patients triaged as "rule-in" and patients with STEMI were excluded. 3,928 patients were considered eligible. Outcome variables included only all-cause death occurring within 180 or 365 days to allow for a larger number of outcome events in the cohort of patients with low all-cause mortality. The cohort was stratified by the outcome of interest and then randomly divided into a training set comprising 75% of the entire study population (n = 2,946) and a test set comprising 25% (n = 982; stratified random sampling). Logistic regression with elastic net regularization was used to fit the model to the training data. All estimated parameters were collected strictly on the training dataset only. We used mean imputation for numerical variables and mode imputation for nominal variables (n = 343 data points, with only 0.37% of data entries missing in total). Therefore, mean and mode imputation were trained on the training dataset only. Imputation was not performed on outcome variables. Skewed variables were log-transformed, and all numerical variables were centered and scaled based on the distribution of the training data. Additionally, an interaction term between the patient's baseline troponin and their age was considered relevant and added to the model. Five-fold cross-validation was used to estimate the hyperparameter penalty (representing the amount of penalty applied to the logistic regression model) and mixture (representing the relative amounts of L1 and L2 penalties) required for elastic net regularization. The hyperparameter grid consisted of 1,000 parameter combinations, each of which was fitted in the model selection process during five-fold cross-validation. The best-fit hyperparameters were evaluated based on the ROC-AUC performance against five-fold cross-validation. After finding the best hyperparameters, the model was finally fitted to the entire training dataset to obtain estimates of covariates. As a result of the regularization technique, some estimates may become zero and therefore no longer be relevant for calculating the outcome probability.Due to the unbalanced results in our dataset, we chose ROC-AUC as our primary performance measure. We then calculated the Youden index on the ROC curve to obtain thresholds for reporting the sensitivity and specificity of our model.

[0007] The minimal model consisted of five variables: initial troponin, delta troponin, age, gender, and creatinine. The full model consisted of 10 variables: initial troponin (numeric), delta troponin (numeric), age (numeric), gender (binary), creatinine (numeric), abnormal ECG (binary), CRP level (numeric), sodium (numeric), hemoglobin (numeric), and platelet level (numeric). In addition to the minimal and full models, another 31 models with 6, 7, 8, 9, or 10 parameters were constructed by adding parameters, such as 6, 7, 8, 9, or 10 parameters, to the minimal model or by replacing relevant variables with similar prognostic information, such as, but not limited to, urea instead of creatinine as an indicator of renal function or estimated glomerular filtration rate. All models, including the minimal and full models, encompass the minimum set of variables mentioned above. A total of 33 models were constructed and evaluated against a blinded study set comprising 25% of the total population.

[0008] Methods exist for predicting short- and mid-term outcomes, such as death, in patients suspected of ACS. The overall performance of established clinical scores is suboptimal, necessitating improved risk stratification. The use of clinical scores has been found to be particularly ineffective when using hs-cTn assays in combination with accelerated protocols. When the ESC0 / 1-hour protocol or the High-STEACS pathway was used, some clinical scores did not improve the accuracy or safety of the FAST protocol and significantly reduced the number of eligible patients. Given that patients triaged as "rule-ins" are generally considered high-risk patients with a high pre-test probability of acute myocardial injury or acute myocardial infarction, only patients classified as "rule-outs" or "observation zones" were eligible for predictive tools. In the study underlying the present invention, a sufficient number of events occurred within 365 days (n = 100) and 180 days (n = 65), allowing the establishment of a method for estimating the risk of all-cause mortality at 365 days (and 180 days). The method is based on the complementation of paired high-sensitivity cardiac troponin concentrations and the change in cardiac troponin T concentration in a second blood draw. Other variables include, but are not limited to, age, sex, past medical history, current symptoms, vital signs, ECG parameters, and other laboratory values.

[0009] Accordingly, the present invention provides a computer-implemented method for predicting the risk of an adverse event in a patient presenting with a suspected acute coronary syndrome, comprising: a) receiving in a processing unit data relating to a set of parameters obtained from a patient, said set of parameters comprising the following parameters: i. the amount of cardiac troponin in a first sample obtained from the patient at the time of presentation; ii. the amount of the cardiac troponin in a second sample obtained from the patient within 30 minutes to 6 hours after the first sample, for example, within 1 hour to 3 hours after the first sample, for example, within 1 hour to 2 hours after the first sample; iii. the amount of CRP (C-reactive protein) in a sample from the patient; iv. at least one parameter of the patient's renal function selected from the group consisting of the amount of urea in a sample from the patient, the patient's GFR, and the patient's serum creatinine level, in particular the patient's serum creatinine level; vv The amount of sodium in a sample from a patient, vi. the amount of hemoglobin in the sample from the patient; vii. The patient's platelet levels; viii. Age of the patient; ix. The patient's gender, and x. Presence or absence of a normal ECG in the above patient receiving a parameter set including at least five, e.g., five, six, seven, eight, nine, or ten, of: b) performing, in a processing unit, an analysis of the set of parameters, the analysis comprising calculating a score for predicting a risk of an adverse event for the patient based on the set of parameters received in step a); c) providing information about the score calculated in step b), thereby predicting the risk of an adverse event; and The present invention relates to a computer-implemented method, including:

[0010] In one embodiment, data is received regarding at least five, for example, five, six, seven, eight, nine, or ten of the following parameters: the amount of cardiac troponin in a first sample obtained from the patient at the time of presentation, the amount of cardiac troponin in a second sample (as described above), the amount of CRP (C-reactive protein) in a sample from the patient, the patient's serum creatinine amount, the amount of sodium in a sample from the patient, the amount of hemoglobin in a sample from the patient, the patient's platelet level, information regarding the patient's age, information regarding the patient's gender, and information regarding the presence or absence of a normal ECG in the patient (i.e., information regarding whether the patient has a normal or abnormal ECG).

[0011] In one embodiment, the parameters are from any one of Models 1-33 shown in Tables 3 or 4 in the Examples section.

[0012] In a preferred embodiment, data is obtained on at least five of the following parameters: the amount of cardiac troponin in a first sample obtained from the patient at the time of presentation, the amount of said cardiac troponin in a second sample (as described above), information on the patient's age, information on the patient's gender, and the patient's serum creatinine level.

[0013] Alternatively, the present invention provides a method for predicting the risk of an adverse event in a patient presenting with a suspected acute coronary syndrome, comprising: a) The following steps a1) to a10), i.e., a1) determining the amount of cardiac troponin in a first sample obtained from said patient at the time of presentation; a2) determining the amount of the cardiac troponin in a second sample obtained from the patient within 30 minutes to 6 hours after the first sample, for example within 1 hour to 3 hours after the first sample, for example within 1 hour to 2 hours after the first sample; a3) determining the amount of CRP (C-reactive protein) in a sample from the patient; a4) determining at least one parameter of the patient's renal function selected from the group consisting of the amount of urea nitrogen in a sample from the patient, the patient's GFR and the patient's serum creatinine level, preferably the patient's serum creatinine level; a5) determining the amount of sodium in a sample from the patient; a6) determining the amount of hemoglobin in a sample from the patient; a7) determining the patient's platelet level in a sample from the patient; a8) providing information about the patient's age; a9) providing information about the patient's gender; and a10) providing information regarding the presence or absence of a normal ECG in said patient; performing at least five, e.g., five, six, seven, eight, nine, or ten, of b) calculating a score for predicting the risk of an adverse event for the patient based on the information obtained in step a); c) predicting the patient's risk of an adverse event based on the score calculated in step b); and The present invention relates to a method, comprising:

[0014] Preferred combinations for the steps included in a) are described above in connection with the computer-implemented method.

[0015] In a preferred embodiment of the aforementioned method, the sample is a blood, serum or plasma sample.

[0016] In a preferred embodiment of the aforementioned method, the adverse event is death, such as all-cause death.

[0017] In a preferred embodiment of the method, the risk of an adverse event within about 180 to about 365 days is predicted. For example, the risk of death within about 180 days is predicted. Alternatively, the risk of death within about 180 days is predicted.

[0018] Forecasting Methods The study cohort consisted of all consecutive patients who presented with suspected ACS to the Chest Pain Unit at Heidelberg University Hospital between June 30, 2016, and July 1, 2018 and underwent coronary angiography with or without myocardial revascularization within 30 days of their index presentation. Patients with STEMI were excluded. 1,344 patients were considered eligible. Outcome variables included patients with significant coronary artery stenosis ≥50% assigned to percutaneous coronary intervention (PCI), coronary artery bypass surgery (CABG), or treated conservatively after failed or unsuccessful attempted PCI, or patients with complex coronary anatomy deemed unsuitable for myocardial revascularization. A machine-learning-enabled model for predicting the presence of significant coronary artery disease requiring revascularization was trained using logistic regression with elastic net regularization.

[0019] The cohort was stratified by the outcome of interest and then randomly divided into a training set containing 75% of the entire study population (n = 1,007) and 25% of the test set (n = 337; stratified random sampling). All estimated parameters were collected strictly on the training dataset. We used mean imputation for numerical variables and mode imputation for nominal variables (n = 670 data points, with only 2.62% of data entries missing in total). Therefore, the mean and mode imputed values ​​were trained on the training dataset only. The outcome variable "obstructive CAD requiring revascularization" was present in n = 889 patients and absent in n = 444 patients. There were no missing values ​​for the outcome variable. Skewed variables were log-transformed, and all numerical variables were centered and scaled based on the distribution of the training data. Additionally, an interaction term between patients' baseline troponin and their age was considered relevant and added to the model. Five-fold cross-validation was used to estimate the hyperparameter penalty (representing the amount of penalty applied to the logistic regression model) and mixture (representing the relative amount of L1 and L2 penalties) required for elastic net regularization. The hyperparameter grid consisted of 1,000 parameter combinations, each of which was fitted to the model selection process during five-fold cross-validation. The best-fit hyperparameters were evaluated based on the ROC-AUC performance for five-fold cross-validation. After finding the best hyperparameters, the model was finally fitted to the entire training dataset to obtain estimates of the covariates. ROC-AUC was selected as our primary performance measure. The Youden index on the ROC curve was then calculated to obtain a threshold for reporting the sensitivity and specificity of our model.

[0020] The minimal model consisted of eight variables: initial troponin, delta troponin, age, gender, creatinine, smoking status, history of revascularization, and experienced chest pain. The full model included the following variables: Early troponin (value), Delta troponin (value), Age (numerical), Gender (binary), Creatinine (value), alternatively or alternatively CKD-EPI (value) and / or urea (value), Abnormal ECG (binary), Cardiac risk factors (diabetes mellitus (binary), Smoking status (binary) · History of coronary artery disease (binary), and · Predominant symptom of dyspnea (binary).

[0021] Background on the need for ML models to predict obstructive CAD requiring revascularization within 30 days A challenge for physicians is the decision to perform coronary angiography. The risks of unnecessary coronary angiography, associated with excessive risk of procedure-related and non-procedure-related major bleeding, radiation exposure, and potential renal injury, must be balanced against the risk of missing significant coronary artery disease that would benefit from reperfusion therapy. To this end, the ESC guidelines 2 recommend a selective invasive strategy for low-risk patients who remain asymptomatic for recurrence. Furthermore, the recommendation to perform stress testing, preferably using imaging stress testing, to determine whether low-risk patients should undergo routine coronary angiography (selective invasive strategy) is not practically feasible given the large number of patients who require specialized imaging stress testing.

[0022] Currently, no established predictors or models exist that allow reliable prediction of obstructive coronary artery disease requiring revascularization. As a result, coronary angiography may be overused in low-risk patients and underused in patients with uncertain risk, such as unstable angina, where the decision to select an invasive strategy is based on recurrent symptoms despite optimal medical treatment or on pathological stress testing, preferably stress imaging. Other scenarios include patients with indeterminate symptoms and comorbidities, including heart failure, obstructive airway disease, and arterial hypertension.

[0023] In the research underlying this invention, a method was established to estimate the probability of having obstructive coronary artery disease requiring reperfusion therapy within 30 days of the index hospitalization. This method is useful for predicting the likelihood of obstructive coronary artery disease requiring revascularization. The method is based on patient characteristics, including, but not limited to, age, sex, past medical history, current symptoms, vital signs, ECG parameters, hs-cTnT, hs-cTnT kinetics, and / or other laboratory values. The readout is the percentage probability of the predicted event.

[0024] Accordingly, the present invention provides a computer-implemented method for predicting the need for myocardial revascularization in a patient presenting with a suspected acute coronary syndrome, comprising: a) receiving in a processing unit data relating to a set of parameters obtained from a patient, said set of parameters comprising the following parameters: i. the amount of cardiac troponin in a first sample obtained from the patient at the time of presentation; ii. the amount of the cardiac troponin in a second sample obtained from the patient within 30 minutes to 6 hours after the first sample, for example, within 1 hour to 3 hours after the first sample, for example, within 1 hour to 2 hours after the first sample; iii. the patient's age; iv. The patient's gender, and v. a parameter of the patient's renal function selected from the group consisting of the patient's GFR (such as the GFR according to the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) method), the patient's serum creatinine level, and the patient's blood urea nitrogen (BUN) level, in particular the patient's serum creatinine level; vi.vi The presence or absence of chest symptoms such as chest pain in the above patients; vii. The presence or absence of dyspnea (preferably as a presenting symptom) in the patient; viii. The presence or absence of a normal ECG in the patient; and ix. Information regarding the patient's past medical history, including at least one, and preferably all, of the following: information regarding the patient's diabetes history; information regarding the patient's nicotine smoking history; and information regarding the patient's coronary artery disease history. receiving at least eight, e.g., eight, nine, ten, eleven, twelve, thirteen, fourteen, or fifteen, of b) performing, in a processing unit, an analysis of the set of parameters, the analysis comprising calculating a score for predicting the need for myocardial revascularization of the patient based on the set of parameters received in step a); c) providing information regarding the score calculated in step b), thereby predicting the patient's need for myocardial revascularization; and The present invention relates to a computer-implemented method, including:

[0025] In one embodiment, data is received regarding at least five, e.g., five, six, seven, eight, nine, or ten, of the following parameters: the amount of cardiac troponin in a first sample obtained from the patient at the time of presentation; the amount of cardiac troponin in a second sample (as described above); the patient's serum creatinine level; whether the patient has chest symptoms; whether the patient has dyspnea; information regarding the patient's gender; information regarding the patient's age; information regarding the patient's past medical history, the information including at least one, and preferably all, of the patient's history of diabetes;

[0026] In one embodiment, the parameters are from any one of Models 1-33 in Table 9 of the Examples section.

[0027] In a preferred embodiment, data is obtained regarding at least eight of the following parameters: the amount of cardiac troponin in a first sample obtained from the patient at the time of presentation, the amount of cardiac troponin in a second sample (as described above), information regarding the patient's gender, information regarding the patient's age, information regarding the patient's history of coronary artery disease, and information regarding the presence or absence of chest symptoms (such as chest pain) in the patient.

[0028] Furthermore, the present invention provides a method for predicting the need for myocardial revascularization in a patient suspected of having an acute coronary syndrome, comprising: a) The following steps a1) to a9), namely: a1) determining the amount of cardiac troponin in a first sample obtained from said patient at the time of presentation; a2) determining the amount of the cardiac troponin in a second sample obtained from the patient within 30 minutes to 6 hours after the first sample, for example within 1 hour to 3 hours after the first sample, for example within 1 hour to 2 hours after the first sample; a3) providing information about the patient's age; a4) providing information about the patient's gender; a5) providing information about a parameter of the patient's renal function selected from the group consisting of the patient's GFR, the patient's serum creatinine level, and the patient's blood urea nitrogen (BUN) level, in particular the patient's serum creatinine level; a6) providing information regarding the presence or absence of chest symptoms in said patient; a7) providing information regarding the presence or absence of dyspnea in said patient; a8) providing information regarding the presence or absence of a normal ECG in said patient; and a9) providing information regarding the patient's past medical history, including at least one, and preferably all, of information regarding the patient's diabetes history, information regarding the patient's smoking history, and information regarding the patient's history of coronary artery disease; performing at least eight, e.g., eight, nine, ten, eleven, twelve, thirteen, fourteen, or fifteen, of the b) calculating a score for predicting the need for myocardial revascularization of the patient based on the information obtained in step a); c) predicting the need for myocardial revascularization of a patient event based on the score calculated in step b); The present invention relates to a method, comprising:

[0029] Preferred combinations for the steps included in a) are described above in connection with the computer-implemented method.

[0030] In a preferred embodiment of the aforementioned method, the sample is a blood, serum or plasma sample.

[0031] In a preferred embodiment of the aforementioned method, the myocardial revascularization results from obstructive coronary artery disease.

[0032] In a preferred embodiment of the aforementioned method, the need for myocardial revascularization within about 30 days is predicted.

[0033] In a preferred embodiment of the aforementioned method, the score indicates the likelihood that the patient will require myocardial revascularization.

[0034] In a preferred embodiment of the above method, the calculated score is shown on a display.

[0035] In a preferred embodiment, the method comprises recommending or subjecting the patient to myocardial revascularization. definition

[0036] As can be seen, the present invention encompasses two methods: a prognostic method and a predictive method. The definitions and explanations provided above shall apply to all methods unless otherwise specified.

[0037] As used in this specification and claims, it should be understood that "a" or "an" can mean one or more, depending on the context in which it is used. Thus, for example, reference to a "cell" can mean that at least one cell can be utilized.

[0038] Furthermore, as used herein, the term "at least one" will be understood to mean that one or more of the items mentioned following this term can be used in accordance with the present invention. For example, when this term indicates that at least one feed solution is to be used, this can be understood as one feed solution or more than one feed solution, i.e., two, three, four, five, or any other number of feed solutions. Depending on the item to which this term refers, a person skilled in the art will understand what upper limit, if any, this term can refer to.

[0039] As used herein, the term "about" means that for any number listed after the term, there is an interval of precision within which the technical effect can be achieved. Thus, the term "about" as referred to herein preferably refers to the exact numerical value or a range of ±20%, preferably ±15%, more preferably ±10%, and even more preferably ±5% around the exact numerical value.

[0040] As used herein, the term "comprising" is not intended to be understood in a limiting sense. Rather, the term indicates that there may be more items than the actual items referenced; for example, when referring to a method including certain steps, the presence of additional steps is not intended to be excluded. However, the term "comprising" also encompasses embodiments in which only the referenced elements are present, i.e., it has a limiting meaning in the sense of "consisting of."

[0041] It will be understood that the methods according to the invention are preferably ex vivo methods, i.e., they do not have to be carried out on a human or animal body. Rather, they are based on previously collected existing patient data. For example, it is envisaged that the methods are in vitro methods. Furthermore, they may comprise steps in addition to those explicitly mentioned above. For example, further steps may involve taking a pre-treatment sample or evaluating the results obtained by the method. The methods may be carried out manually or may be assisted by automation.

[0042] In some embodiments, the methods of the present invention are computer-implemented methods, in which typically all steps of the computer-implemented methods of the present invention are performed by one or more processing units of a computer or computer network.

[0043] The phrase "predicting the risk of an adverse event" as used herein means that a subject analyzed by the method of the present invention is assigned to either a group of subjects at risk of suffering from an adverse event or a group of subjects not at risk of suffering from an adverse event. Thus, whether or not a subject is at risk of an adverse event is predicted. As used herein, a "patient at risk of an adverse event" preferably has an elevated risk of suffering from the adverse event, preferably within a prediction window. Preferably, the risk is elevated compared to the average risk in the cohort of subjects. As used herein, a "subject not at risk of an adverse event" preferably has a reduced risk of developing the adverse event, preferably within a prediction window. Preferably, the risk is reduced compared to the average risk in the cohort of subjects. Preferably, the elevated or reduced risk referred to herein is a statistically significant elevated or reduced risk. More preferably, the prediction window according to the present invention for predicting the risk of an adverse event is within about 180 days to about 365 days, such as within about 180 days or within about 365 days, i.e., within one year. The prediction window is typically calculated starting from the date the parameters were obtained from the patient.

[0044] The term "adverse event" as used herein refers to any deterioration that occurs in a patient within a predictive window and severely and negatively affects one or more physiological functions in the patient. More specifically, it affects the physiological function of the cardiovascular system. More preferably, the adverse event is death from any cause. Thus, in this embodiment, the adverse event is death. The term "death" as used herein preferably relates to death from any cause.

[0045] The phrase "predicting the need for myocardial revascularization" as used herein means that a subject analyzed by the method of the present invention is assigned to either a group of subjects requiring myocardial revascularization or a group of subjects not requiring myocardial revascularization. Therefore, whether or not a subject requires myocardial revascularization is predicted. As used herein, a "patient requiring myocardial revascularization" preferably has an increased likelihood of requiring myocardial revascularization, preferably within a prediction window. Preferably, the likelihood is increased compared to the average likelihood of requiring myocardial revascularization in a cohort of subjects. As used herein, a "subject not requiring myocardial revascularization" preferably has a decreased likelihood of requiring myocardial revascularization, preferably within a prediction window. Preferably, the likelihood is decreased compared to the average likelihood of requiring myocardial revascularization in a cohort of subjects. Preferably, the increased or decreased likelihood referred to herein is a statistically significant increased or decreased likelihood. More preferably, the prediction window according to the present invention, within which the need for myocardial revascularization is predicted, is within about 30 days. The prediction window is typically calculated starting from the date the parameters are obtained from the patient. Alternatively, obstructive coronary artery disease may be diagnosed by the methods of the present invention.

[0046] The term "myocardial revascularization" as used herein refers to any therapeutic procedure that allows for revascularization of tissues affected by cardiovascular disease or disorders, preferably by occlusive vascular events such as those caused by occlusive coronary artery disease. Coronary artery disease as referred to herein is preferably defined as any luminal obstruction of a major epicardial coronary artery of 50% or more. Preferably, the therapeutic procedure that allows for myocardial revascularization according to the present invention is either percutaneous coronary intervention with or without stent placement or coronary artery bypass graft (CABG).

[0047] Furthermore, myocardial revascularization is one of the following: Invasive coronary angiography with percutaneous coronary intervention with or without stent placement, b. Invasive coronary angiography with lesion(s) suitable for revascularization and attempted percutaneous coronary intervention with or without stent placement; c. Invasive coronary angiography with lesion(s) suitable for revascularization and attempted but unsuccessful or unsuccessful percutaneous coronary intervention; d. Invasive coronary angiography with associated lesion(s) not suitable for revascularization and subsequent conservative medical treatment without percutaneous coronary intervention; e. Immediate transfer for invasive coronary angiography and coronary artery bypass surgery with lesion(s) suitable for revascularization; f. Recommendation for invasive coronary angiography and coronary artery bypass surgery with lesion(s) suitable for revascularization, or g. Invasive coronary angiography with lesion(s) planned for revascularization, such as percutaneous coronary intervention or coronary artery bypass surgery.

[0048] In a preferred embodiment, the myocardial revascularization is invasive coronary angiography involving percutaneous coronary intervention with or without stent placement.

[0049] In another preferred embodiment, the myocardial revascularization is a coronary artery bypass graft (CABG).

[0050] In another preferred embodiment, when percutaneous coronary intervention has failed or is unsuccessful, or when the lesion(s) are relevant but not suitable for percutaneous coronary intervention or coronary artery bypass surgery, the treatment is conservative pharmacological treatment, such as administration of an effective amount of acetylsalicylic acid, at least one beta-blocker, at least one angiotensin II receptor blocker (ARB), and / or at least one statin.

[0051] As will be understood by those skilled in the art, the above-mentioned assessments made by the methods of the present invention, i.e., prediction or prognosis, are usually not intended to be accurate for 100% of the individuals surveyed. This term typically requires that the assessment be accurate for a statistically significant portion of individuals (e.g., a cohort in a cohort study). Whether a value indicating a difference in risk or induction, a portion of a cohort, or any other difference in values ​​is statistically significant can be easily determined by those skilled in the art using various well-known statistical evaluation tools, such as determining a confidence interval, a p-value, a Student's t-test, or a Mann-Whitney test. Details can be found in Dowdy and Wearden, *Statistics for Research*, John Wiley & Sons, New York, 1983. Preferred confidence intervals are at least 90%, at least 95%, at least 97%, at least 98%, or at least 99%. The p-value is preferably 0.1, 0.05, 0.01, 0.005, or 0.0001.

[0052] The term "sample" refers to a sample of bodily fluid, a sample of separated cells, or a sample from a tissue or organ known or suspected to contain the analyte to be determined as a parameter. It will be understood that the sample may depend on the analyte to be determined. For example, if cardiac troponin is to be determined in the first and / or second samples referred to herein, the sample may typically be a sample containing or suspected to contain the cardiac troponin. A typical sample may be a whole blood sample or a derivative thereof, such as a plasma or serum sample. For other analytes, the sample may be a urine sample or a sample of other bodily fluids, cells, or tissues. Those skilled in the art are well aware of which samples can be used for a given analyte to determine the parameters referred to in accordance with the present invention. Furthermore, those skilled in the art are also familiar with how such samples can be collected from patients, for example, by conventional blood collection devices such as a lancet, biopsy, etc.

[0053] In a preferred embodiment, the sample is a blood, serum or plasma sample.

[0054] In another preferred embodiment, the sample is interstitial fluid.

[0055] The term "cardiac troponin" typically refers to human cardiac troponin T or cardiac troponin I. However, this term also refers to variants of the aforementioned specific troponins, i.e., preferably cardiac troponin I, and more preferably cardiac troponin T. Such variants have at least the same essential biological and immunological properties as the specific cardiac troponin. In particular, they share the same essential biological and immunological properties if they are detectable by the same specific assays referred to herein, for example, by ELISA assays using polyclonal or monoclonal antibodies that specifically recognize the cardiac troponin. Furthermore, the variants referred to in accordance with the present invention are intended to have a different amino acid sequence due to at least one amino acid substitution, deletion, and / or addition, and the amino acid sequence of the variant should still be understood to be preferably at least about 50%, at least about 60%, at least about 70%, at least about 80%, at least about 85%, at least about 90%, at least about 92%, at least about 95%, at least about 97%, at least about 98%, or at least about 99% identical to the amino acid sequence of the specific cardiac troponin. The variants may be allelic variants or any other species-specific homologs, paralogs, or orthologs. Furthermore, the variants referred to herein include fragments of the specific cardiac troponins, or variants of the above-mentioned types, so long as these fragments possess the essential immunological and biological properties referred to above. Preferably, cardiac troponin variants have immunological properties (i.e., epitope composition) comparable to those of human troponin T or troponin I. Therefore, the variants should be recognizable by the aforementioned means or ligands used to determine the concentration of cardiac troponin. Thus, the variants should be recognizable by the aforementioned means or ligands used to determine cardiac troponin concentration. Such fragments may be, for example, degradation products of troponin. Also included are variants that differ due to post-translational modifications such as phosphorylation or myristylation.Preferably, the biological property of troponin I and its variants is that they can inhibit actomyosin ATPase or inhibit angiogenesis in vivo and in vitro, which can be detected, for example, based on the assay described by Mose et al. 1999 PNASUSA 96(6):2645-2650. Preferably, the biological property of troponin T and its variants is that they can form a complex with troponin C and I, bind to calcium ions, or bind to tropomyosin, preferably as a complex of troponin C, I, and T, or as a complex formed by troponin C, troponin I, and a variant of troponin T, if present. Troponin T or troponin I can be determined by immunoassays well known and commercially available in the art, such as ELISA. Particularly preferred according to the present invention is the determination of troponin T with high sensitivity, for example, using the commercially available hs-cTn assay.

[0056] CRP (C-reactive protein) is an acute-phase protein discovered over 75 years ago as a blood protein that binds to the C-polysaccharide of Streptococcus pneumoniae. CRP is known as a reactive inflammatory marker and is produced by distal organs (i.e., the liver) in response to chemokines or interleukins derived from the primary lesion site. CRP is known to consist of five single subunits noncovalently linked and assembled as a cyclic pentamer with a molecular weight of approximately 110-140 kDa. Preferably, the CRP used herein refers to human CRP. The sequence of human CRP is well known and is disclosed, for example, by Woo et al. (J. Biol. Chem. 1985. 260(24), 13384-13388). CRP levels are typically low in normal individuals but can increase by 100-200-fold or more due to inflammation, infection, or injury (Yeh (2004) Circulation. 2004;109:11-11-11-14). It is known that CRP is an independent factor for predicting cardiovascular risk.CRP can be measured by immunoassays, such as ELISA, which are well known in the art and commercially available.Typically, CRP is hsCRP (high sensitivity CRP).

[0057] Urea is the major end product of protein nitrogen metabolism. It has the chemical formula CO(NH2)2 and is synthesized in the liver through the urea cycle from ammonia, which is produced by amino acid deamination. Urea is primarily excreted by the kidneys, but small amounts are also excreted in sweat and are degraded in the intestine by bacterial action. Blood urea nitrogen determination is the most widely used screening test for renal function. Urea can be measured by an in vitro test for quantitative determination of urea / urea nitrogen in human serum, plasma, and urine on the Roche / Hitachi cobas c system. The test can be performed automatically using different analyzers, including the cobas c 311 and cobas c 501 / 502. This assay is a kinetic assay using urease and glutamate dehydrogenase. Urea is hydrolyzed by urease to form ammonium and carbonate. In the second reaction, 2-oxoglutarate reacts with ammonium in the presence of glutamate dehydrogenase (GLDH) and the coenzyme NADH to produce L-glutamate. In this reaction, two moles of NADH are oxidized to NAD+ for each mole of urea hydrolyzed. The rate of decrease in NADH concentration is directly proportional to the urea concentration in the sample and is measured photometrically.

[0058] In some embodiments, the amount of urea may be determined, or alternatively, the amount of blood urea nitrogen (abbreviated BUN) may be determined.

[0059] The marker "creatinine" is well known in the art. In muscle metabolism, creatinine is endogenously synthesized from creatine and creatine phosphate. Under normal renal function, creatinine is excreted by glomerular filtration. Creatinine determination is performed for the diagnosis and monitoring of acute and chronic renal diseases, as well as for monitoring renal dialysis. The creatinine concentration in urine can be used as a reference value for the excretion of certain analytes (albumin, α-amylase). Creatinine can be determined as described by Popper et al. (Popper H et al. Biochem Z 1937;291:354), Seelig and Wust (Seelig HP, Wust H. Arztl Labor 1969;15:34) or Bartels (Bartels H et al. Clin Chim Acta 1972;37:193). Preferably, the amount of creatinine is determined in a serum sample. Thus, the patient's serum creatinine level is determined.

[0060] The term "hemoglobin" as used herein preferably refers to total hemoglobin. Hemoglobin levels can be measured by well-known methods, for example, by oxidizing hemoglobin to methemoglobin with potassium ferricyanide. Hemoglobin levels are proportional to color intensity and can be measured, for example, at a wavelength of 567 nm and 37°C. Hemoglobin levels can also be measured by contacting a sample with an antibody that specifically binds to hemoglobin.

[0061] The parameter "GFR (glomerular filtration rate)" is a well-known parameter that can be determined by clinical chemistry assays and detection methods well known in the art. GFR can be accurately calculated by comparative measurements of substances in blood and urine, or can be estimated by a formula using only blood test results (eGFR). These estimators are typically used in clinical practice, especially in elderly and sick patients where reliable urine collection is difficult. eGFR is related to GFR. Clinical assessment scales of eGFR and GFR can be used interchangeably. In the research underlying the present invention, eGFR was determined. In one embodiment, GFR is GFR according to the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formula.

[0062] The same is true for the amount of "sodium." The amount of sodium can be determined without further difficulty by using routine clinical chemistry and well-known detection techniques.

[0063] Additionally, platelet levels can be determined by well-established clinical laboratory assays. For example, individual cells can be counted manually in a counting chamber. Alternatively, automated instruments, including FACS analyzers, can be used.

[0064] The term "amount" as used herein refers to the absolute amount of the biomarker referred to herein, the relative amount or concentration of the compound, and any value or parameter that can be correlated therewith or derived therefrom. Such values ​​or parameters include intensity signal values ​​derived from any specific physical or chemical property obtained from the compound by direct measurement, such as intensity values ​​in mass spectrum or NMR spectrum. Furthermore, it also encompasses any value or parameter obtained by indirect measurement as specified elsewhere in this specification, such as the response amount measured by a biological readout system in response to a compound, or the intensity signal obtained from a specifically bound ligand. It should be understood that values ​​that correlate with the above-mentioned amounts or parameters can also be obtained by any standard mathematical operation.

[0065] The terms "determining" or "measuring" the level of a marker referred to herein refer to quantifying the biomarker, e.g., determining the level of the biomarker in a sample, using appropriate detection methods described elsewhere herein. In one embodiment, the level of at least one biomarker is measured by contacting the sample with a detection agent that specifically binds to the respective marker, thereby forming a complex between the agent and the marker, and detecting the level of the complex formed, thereby measuring the level of the marker.

[0066] Electrocardiography (abbreviated as ECG) is the process of recording the electrical activity of the heart by a suitable ECG. The ECG device records the electrical signals generated by the heart and spreads from the entire body to the skin. The recording is of the electrical signals achieved by contacting the skin of the subject with electrodes provided by the ECG device. The process of obtaining the recording is non-invasive and risk-free. In some embodiments of the method of the present invention, whether the patient has a normal ECG is evaluated. Therefore, the presence or absence of a normal ECG in the patient is evaluated.

[0067] An ECG is considered normal if all of the following criteria are met: 1. Sinus rhythm without second- or third-degree AV block, and 2. No bundle branch block (LBBB, RBBB) or nonspecific block (QRS ≥ 120 ms), and 3. No ventricular pacing (no single RV / LV paced QRS on the obtained EKG), and No ST depression >4.1 mV (in ≥2 contiguous leads), and 5. No T-wave inversion (in two or more consecutive leads)

[0068] If at least one of the above criteria is not met, the patient's ECG is not normal.

[0069] In some embodiments, the score takes into account information about the presence or absence of chest symptoms in the patient, thus assessing whether the subject has chest symptoms, particularly whether the subject is experiencing chest pain.

[0070] In some embodiments, information regarding the presence or absence of a patient's dyspnea ("shortness of breath") is taken into account in the score. The term "dyspnea" refers to a respiratory disorder that results in an increased respiratory rate and / or an increased respiratory volume. Therefore, shortness of breath may preferably result in hyperventilation. Shortness of breath usually occurs at an oxygen saturation level that is below a normal oxygen saturation level of at least 95%. As used herein, the term "dyspnea" refers to acute shortness of breath, i.e., shortness of breath that occurs non-permanently.

[0071] In some embodiments, the method of the present invention includes obtaining information about the patient's past medical history. The information can be obtained, for example, from the patient's medical records. In some embodiments, the information includes information about the patient's diabetes history. Thus, it is evaluated whether the patient has or has had diabetes. As used herein, the term "diabetes" preferably refers to type I diabetes or type II diabetes. Symptoms and clinical parameters related to type I and type II diabetes are well known in the art. In a preferred embodiment, it is evaluated whether the patient has or has had type II diabetes.

[0072] In some embodiments, the methods of the present invention include obtaining information including information about the patient's history of nicotine smoking. As used herein, the term "smoking" preferably refers to previous or current smoking. In preferred embodiments, it is assessed whether the patient is currently smoking nicotine or has smoked nicotine in the past.

[0073] In some embodiments, the information includes information about the patient's medical history of coronary artery disease, e.g., CAD, with or without a history of myocardial revascularization. Typically, the patient is evaluated as having or not having coronary artery disease (CAD). Patients with a history of coronary artery disease preferably meet at least one of the following criteria: previous myocardial infarction, known CAD, previous percutaneous coronary intervention (PCI), and / or previous coronary artery bypass surgery (CABG).

[0074] In some embodiments, the information regarding coronary artery disease includes information regarding the patient's history of myocardial revascularization. Thus, it is assessed whether the patient has previously undergone myocardial revascularization, such as PCI and / or CABG. The term "myocardial revascularization" is defined elsewhere herein.

[0075] In some embodiments, the information includes information regarding the patient's history of myocardial infarction. Thus, it is assessed whether the patient has previously suffered a myocardial infarction. The term "myocardial infarction" is defined elsewhere herein.

[0076] The "patient" or "subject" referred to herein is preferably a mammal. Mammals include, but are not limited to, livestock animals (e.g., cows, sheep, cats, dogs, and horses), primates (e.g., humans and non-human primates, such as monkeys), rabbits, and rodents (e.g., mice and rats). Preferably, the patient or subject according to the present invention is a human. The patient referred to according to the present invention is preferably a patient who presents in an emergency department with a suspicion of acute coronary syndrome (ACS). Typically, such a patient is suffering from ACS or presents at least one or more symptoms associated with ACS, such as chest pain. In one embodiment, the subject is suffering from unstable angina.

[0077] As used herein, the term "acute coronary syndrome (ACS)" refers to an occlusive event affecting a coronary artery that involves multiple interrelated mechanisms. Preferably, in ACS, plaque ruptures or erodes in response to inflammation, potentially resulting in localized occlusive or non-occlusive thrombosis. Depending on the extent and reversibility of this dynamic obstruction, the clinical manifestations of ACS include a continuum of risk, progressing from unstable angina (UA) to non-ST-segment elevation myocardial infarction (NSTEMI) to ST-segment elevation myocardial infarction (STEMI). NSTEMI is distinguished from UA by ischemia of sufficient intensity and duration to cause myocyte necrosis, as recognized by the detection of cardiac troponin, the most sensitive and specific biomarker of myocardial injury. ACS is typically accompanied by prolonged episodes of chest pain, preferably lasting 20 minutes or longer.

[0078] In a preferred embodiment, the patient being tested is suspected of having a non-ST elevation acute coronary syndrome, including myocardial infarction (NSTEMI) and unstable angina. Thus, the patient does not have ST elevation myocardial infarction (STEMI). STEMI is defined as the presence of at least two consecutive leads or new bundle branch block (right or left bundle branch block) or persistent ST segment elevation in a permanently paced rhythm. A subject suspected of having NSTEMI preferably has normal or non-diagnostic, or ST segment depression or T wave inversion on the ECG, and therefore does not have such ST segment elevation.

[0079] As used herein, the term "data" refers to digital information, such as a numerical value, indicative of a parameter of a set of parameters for which data is to be received in accordance with the present invention. Preferably, the digital numerical value shall represent the amount of a compound to be considered, or a count of blood cell or platelet levels. Other digital information taken into account in the method according to the present invention may be an identifier, for example, an identifier of gender, an identifier of a normal or impaired ECG, an identifier of a particular event in the patient's medical history, such as those mentioned elsewhere in accordance with the method of the present invention, or a numerical identifier of age.

[0080] In one embodiment of the present invention, the method of the present invention is a computer-implemented method. Typically, all steps of the computer-implemented method of the present invention are performed by one or more processing units of a computer or computer network. However, the computer-implemented method may include additional steps, such as determining the amount of a marker in the samples, such as the amount of cardiac troponin in the first and second samples.

[0081] According to the method of the present invention, a set of parameters is evaluated, particularly in "prediction" and "prognosis". The term "set of parameters" as used herein refers to a collection of different parameters selected from the group of different parameters mentioned above, which are taken into consideration for carrying out the method of the present invention. The set of parameters includes at least 6, at least 7, preferably 8, 9 or 10 parameters in the case of a method for predicting the risk of adverse events in patients with suspected acute coronary syndrome, or at least 8, preferably 8, 9, 10, 11, 12, 13, 14 or 15 parameters in the case of a method for predicting the need for myocardial revascularization in patients with suspected acute coronary syndrome.

[0082] In some embodiments, at least five parameters (ie, of the parameters listed above) are evaluated.

[0083] In some embodiments, at least six parameters are evaluated.

[0084] In some embodiments, at least seven parameters are evaluated.

[0085] In some embodiments, at least eight parameters are evaluated.

[0086] In some embodiments, nine parameters are evaluated.

[0087] In some embodiments, 10 parameters are evaluated, thus all parameters are evaluated.

[0088] "Prognostic" method According to the "prognostic" method, at least five, at least six, for example at least seven, eight, nine, or ten of the following parameters shall be evaluated: The amount of cardiac troponin in a first sample obtained from said patient at the time of presentation. the amount of the cardiac troponin in a second sample obtained from the patient within 30 minutes to 6 hours after the first sample, for example, within 1 hour to 3 hours after the first sample, for example, within 1 hour to 2 hours after the first sample; the amount of CRP (C-reactive protein) in a sample from the patient, at least one parameter of the patient's renal function, such as the amount of urea in a sample from the patient, the patient's GFR, and in particular the patient's serum creatinine level; the amount of sodium in the patient's sample, the amount of hemoglobin in the patient sample, The patient's platelet levels, the patient's age, Patient gender, and · Presence or absence of a normal ECG in the above patient.

[0089] In some embodiments of the methods described herein, the presence or absence of a normal ECG may be assessed based on an ECG reading obtained from the subject.

[0090] As described above, at least five, six, or seven of the above parameters are evaluated for prognosis. The full model includes 10 parameters. However, it is envisioned that the prognosis method includes evaluation of at least the following five parameters (out of the 10): The amount of cardiac troponin in a first sample obtained from said patient at the time of presentation. the amount of said cardiac troponin in a second sample obtained from said patient within 30 minutes to 6 hours after the first sample, for example, within 1 hour to 3 hours after the first sample, for example, within 1 hour to 2 hours after the first sample. At least one parameter of the patient's renal function. Preferably, the parameter is the patient's serum creatinine level. Patient's age. Patient's gender. · Presence or absence of a normal ECG in the above patient.

[0091] In one embodiment, all ten parameters i) through x) are evaluated (or a1) through a10)).

[0092] Forecasting Methods At least eight, for example, eight, nine, ten, eleven, twelve, thirteen, fourteen, or fifteen of the following parameters shall be evaluated according to the "prediction" method: the amount of cardiac troponin in a first sample obtained from said patient at the time of presentation; the amount of cardiac troponin in a second sample obtained from the patient within 30 minutes to 6 hours after the first sample, for example, within 1 hour to 3 hours after the first sample, for example, within 1 hour to 2 hours after the first sample; Patient gender Patient's age (years) Patient's serum creatinine level - Presence or absence of a normal ECG in the above patients - Chest symptoms, especially chest pain, in the above patients Information about the patient's smoking history a parameter of the patient's renal function selected from the group consisting of the patient's GFR, the patient's serum creatinine level, and the patient's blood urea nitrogen (BUN) level (or urea), in particular the patient's serum creatinine level The presence or absence of respiratory distress in these patients, and Information regarding the patient's past medical history, including at least one, and preferably all, of the following: information regarding the patient's diabetes history; and information regarding the patient's history of coronary heart disease (such as information regarding the patient's history of myocardial revascularization).

[0093] As described above, at least 8, 9, 10, 11, 12, 13, 14, or 15 of the above parameters are evaluated for the predictive method. However, it is envisioned that the predictive method includes evaluation of the following eight parameters: the amount of cardiac troponin in a first sample obtained from the patient at the time of presentation; the amount of the cardiac troponin in a second sample obtained from the patient within 30 minutes to 6 hours after the first sample, for example, within 1 hour to 3 hours after the first sample, for example, within 1 hour to 2 hours after the first sample; Patient gender Patient's age (years) Patient's serum creatinine level Presence or absence of chest pain as the main symptom Information about the patient's history of coronary artery disease, including a history of myocardial revascularization (and therefore whether the patient has previously undergone myocardial revascularization) · Information about the patient's smoking history (past or current).

[0094] In one embodiment, all eight parameters are evaluated.

[0095] The amount of creatinine in serum is necessary to evaluate the patient's renal function. Alternatively or additionally, at least one parameter of the patient's renal function is the amount of urea. Alternatively, the amount of BUN can be evaluated. Furthermore, at least one parameter of the patient's renal function is the patient's eGFR or GFR.

[0096] According to the predictive and prognostic methods, the amount of cardiac troponin in a first sample and a second sample is evaluated (or information regarding the amount is considered). The first sample is obtained at the time of medical examination. Furthermore, the second sample is preferably obtained from the patient about 30 minutes to about 6 hours after the first sample, more preferably about 1 hour to about 3 hours after the first sample, and most preferably about 1 hour to about 2 hours after the first sample. In some embodiments, the second sample is obtained about 1 hour after the first sample. In some embodiments, the second sample is obtained at least 2 hours after the first sample.

[0097] The amount of cardiac troponin in the second sample is typically used to calculate the difference (delta) between the amount of cardiac troponin in the first sample and the amount in the second sample. Thus, the method of the present invention can include calculation of this difference by a processing unit or the like. The difference can be given as a value. Preferably, the value is used as a parameter in the prediction and prognosis methods referred to herein.

[0098] Data for the set of parameters specified above is received by the processing unit. Typically, the data may be received from a database containing stored data for the parameters referenced in accordance with the present invention. Alternatively or additionally, the data may be received from a measurement device that performs real-time measurements on patient samples. It will be understood that there are parameters that cannot be measured but need to be obtained from the subject by other means and stored in a database. These parameters include, for example, age, gender, and medical history. Data may be received from the database(s) or real-time measurement device via a physical connection or wireless data transfer. Thus, such data transmission may be achieved by a permanent or temporary physical connection, such as coaxial, fiber, optical fiber, or twisted pair cable. However, it may also typically be achieved by a temporary or permanent wireless connection using radio waves, such as Wi-Fi, LTE, LTE Advanced, or Bluetooth®.

[0099] The processing unit referred to in accordance with the method of the present invention typically includes a central processing unit (CPU) and / or one or more graphic processing units (GPUs) and / or one or more application specific integrated circuits (ASICs) and / or one or more tensor processing units (TPUs) and / or one or more field programmable gate arrays (FPGAs), etc. Preferably, the data processing unit is a computer or computer-like device, such as a tablet, a smart device, or a mobile device.

[0100] The data processing unit is to perform an analysis of the set of parameters, said analysis comprising calculating a score for predicting the patient's risk of an adverse event based on the set of parameters received in step a), and therefore the processing unit requires software instructions tangibly embedded thereon that, when executed thereon, perform the analysis of the parameters comprising calculating a score for predicting the patient's risk of an adverse event or a score for predicting the patient's need for myocardial revascularization.

[0101] The processing unit also provides information about the calculated score so that adverse events or the need for myocardial revascularization can be predicted. Typically, the score calculated according to the method of the present invention is compared by the processing unit with at least one identifier stored in a database containing information for predicting adverse events or the need for myocardial revascularization. Thus, if the calculated score can be linked to the identifier by the processing unit, the information linked to the identifier can be linked to the score, and a prediction of adverse events or the need for myocardial revascularization can be provided.

[0102] The term "score" used in accordance with the methods of the present invention, particularly the prediction and prognosis methods, refers to a parameter that integrates the information provided by the set of parameters described above. By using the scoring system described herein, advantageously, values ​​of different dimensions or units of parameters may be used, since the values ​​are mathematically converted into scores. Preferably, a low score is associated with a low risk, and a high score is associated with a high risk. Typically, the score is a single value calculated by applying a mathematical operation based on other values ​​and weighting such other values ​​according to predetermined rules. Therefore, some values, such as the amount of cardiac troponin, may affect the score more than other values, such as gender or age. Preferably, the score according to the present invention can be calculated as described in detail below.

[0103] Preferably, the score allows for evaluation of whether a patient is at risk of an adverse event (in a prognostic method) or whether a subject needs myocardial revascularization (in a predictive method). Preferably, the score is calculated based on a suitable scoring algorithm. The scoring algorithm preferably allows for such evaluation based on a set of parameters. The readout may be the percentage probability of the predicted event.

[0104] For example, the formula for prognosis of all-cause mortality within 365 days using the full model estimates plugged into the penalized logistic regression equation: [Table 1]

[0105] Shown are model equations consisting of intercepts, estimates, penalties, and odds ratios for each of the included variables: sodium concentration, ECG, hemoglobin level, sex, creatinine, platelet count, delta troponin, age, C-reactive protein, and early troponin (from top to bottom). Note that the parameters of the minimal model, i.e., delta troponin, creatinine, sex, age, and early troponin, are conserved in all 33 models and therefore are conserved in the full model.

[0106] For example, the score of a method for assessing the need for revascularization due to obstructive CAD using eight parameters (minimal method, see Examples) can be calculated by plugging the following estimates into a penalized logistic regression model: [Table 2]

[0107] Shown are model equations consisting of intercepts, estimates, penalties, and odds ratios for each of the included variables (from top to bottom): sodium concentration, ECG, hemoglobin level, sex, creatinine, age, delta troponin, smoking history, initial troponin, chest pain as presenting symptom, and history of coronary artery disease including history of myocardial revascularization. Note that the parameters of the minimal model are retained in the full model including 15 variables and in all 33 models.

[0108] The present invention further relates to a computer program comprising computer-executable instructions for performing the steps of the computer-implemented method according to the present invention when the program is run on a computer or a computer network. Typically, the computer program may comprise computer-executable instructions in particular for performing the steps of the methods disclosed herein. In particular, the computer program may be stored on a computer-readable data medium.

[0109] The present invention further relates to a computer program product having program code means stored on a machine-readable carrier for implementing the method according to the invention, such as one or more steps discussed above in connection with the computer program, when the program is run on a computer or a computer network. As used herein, a computer program product refers to a program as a tradeable product. The product may generally exist in any form, such as in paper form, or may reside on a computer-readable data carrier. In particular, the computer program product may be distributed over a data network.

[0110] The invention further relates to a computer or a computer network comprising at least one processing unit, the processing unit being adapted to carry out all the steps of the method according to the invention.

[0111] The present invention also contemplates in principle a computer program, a computer program product or a computer-readable storage medium tangibly embodied with said computer program, which comprises instructions for carrying out the inventive method specified above when said computer program is run on a data processing device or computer. In particular, the present disclosure further encompasses: a computer or computer network comprising at least one processor, the processor being configured to execute a method according to one of the embodiments described herein, a computer-loadable data structure adapted to perform a method according to one of the embodiments described herein while the data structure is being executed on a computer, a computer script, the computer script being adapted to carry out a method according to one of the embodiments described herein while the program is running on a computer, a computer program comprising program means for carrying out a method according to one of the embodiments described herein while the computer program is running on a computer or on a computer network, a computer program comprising program means according to any preceding embodiment, the program means being stored on a computer-readable storage medium; a storage medium on which a data structure is stored and adapted to perform a method according to one of the embodiments described herein after the data structure has been loaded into a main memory and / or a working memory of a computer or a computer network, a computer program product having program code means in which the program code means can be stored or can be stored on a storage medium for performing a method according to one of the embodiments described in this specification when the program code means is executed on a computer or a computer network, a typically encrypted data stream signal containing data of the parameters defined elsewhere in this specification, and a typically encrypted data stream signal containing the scores calculated by the method of the present invention and preferably providing information for the prediction.

[0112] The present invention further relates to an apparatus for predicting the risk of an adverse event or the need for myocardial revascularization, said apparatus comprising a processing unit and a computer program comprising computer-executable instructions (such as the computer program described above) which, when executed by the processing unit, cause the processing unit to perform the computer-implemented method according to the present invention, i.e., to perform the steps of the method. The apparatus may further comprise a user interface and a display, the processing unit being coupled to the user interface and the display. Typically, the apparatus provides a prediction as an output. In one embodiment, the classification is provided on the display.

[0113] All references cited throughout this specification are hereby incorporated by reference in their entirety and for the disclosure content specifically mentioned above. [Brief explanation of the drawings]

[0114] [Figure 1] Discriminatory ability to predict 365-day all-cause mortality. A) The minimal model (left panel) shows moderate discriminatory ability with an AUC of 0.81 for predicting 365-day all-cause mortality. B) The full model shows good discriminatory ability with an AUC of 0.86 for predicting 365-day all-cause mortality. [Figure 2]Calibration plot between estimated and observed all-cause mortality within 365 days using the minimal and full models. The calibration plot shows the agreement between estimated (black line) and observed (gray line) all-cause mortality using the five-variable minimal prognostic model (left panel) or the full prognostic model encompassing 10 parameters. This agreement is linear with the full model systematically overestimating observed events. The agreement for the minimal model is linear over a narrow range from 0.5 to 3.5%, then rises exponentially. [Figure 3] Discriminatory ability to predict all-cause mortality at 180 days The minimal model (left panel) shows moderate discriminatory ability with an AUC of 0.82 for predicting all-cause mortality at 180 days. The full model shows excellent discriminatory ability with an AUC of 0.86 for predicting all-cause mortality at 160 days. [Figure 4] Calibration plot between estimated and observed all-cause mortality within 180 days using the minimal and full models. The calibration plot shows the agreement between estimated (black line) and observed (gray line) all-cause mortality using the five-variable minimal prognostic model (left panel) or the full prognostic model encompassing 10 parameters. [Figure 5] Performance of the GRACE score and comparison of the full model with the minimal model at 180 and 365 days. Comparison of the GRACE score (black line), the minimal model (light gray line), and the full model (dark gray line) for predicting death at 180 days (left panel) and at 365 days (right panel). At day 365 (right panel), the full model outperforms the GRACE score with a statistically significant (p=0.02) gain in AUC (=delta AUC) of 0.07. Thus, the full model demonstrated superior prognostic performance to the state-of-the-art comparator. At day 180 (left panel), there is a trend toward a significant gain in prognostic performance of the full model compared to the GRACE score (p=0.085). [Figure 6]Performance of the Full and Minimal Models for Predicting Obstructive CAD Requiring Revascularization. Receiver operating curves (ROC) show an AUC of 0.71 for the minimal model (left panel) and 0.69 for the full model (right panel). [Figure 7] Calibration plots of the minimal and full models for predicting revascularization within 30 days. The calibration plots of the minimal model (left panel) and full model (right panel) show a nearly linear correlation between the predicted (dark line) and observed (gray line) output values. [Figure 8] Performance of the ESC Consortium Algorithm and Comparison with the Full and Minimal Models This figure shows the AUC of the ESC Consortium algorithm, which includes nine variables. Overall performance is moderate, with an AUC of 0.63 (95% CI: 0.65-0.76). The AUC of the minimal model is 0.71 (95% CI: 0.65-0.76), while the AUC of the full model is 0.69 (95% CI: 0.63-0.75). The minimal model significantly outperforms the ESC Consortium algorithm (p=0.004), with an AUC gain of 0.08. The full model also significantly outperforms the ESC Consortium algorithm (p=0.04), with an AUC gain of 0.06. DETAILED DESCRIPTION OF THE INVENTION

[0115] Example The examples are illustrative of the invention and should not be construed as limiting the scope.

[0116] Example 1: Baseline characteristics of the derivation cohort (training set) and validation cohort (test set) Patients with symptoms suggestive of myocardial infarction who had serial high-sensitivity cardiac troponin T measurements obtained at presentation and later in the emergency department were included. Patients with persistent ST-segment elevation myocardial infarction (STEMI) or with a probable new left bundle branch block, as well as patients deemed ineligible as previously described, were excluded. Both the derivation and validation cohorts were prospective and included serial hs-cTnT concentrations. The final diagnosis of NSTEMI was determined according to the universal definition of myocardial infarction. The diagnosis of NSTEMI required an elevation of hs-cTnT above a prespecified cutoff and its associated elevation and / or depression, along with clinical signs or symptoms suggestive of an ischemic situation. These signs or symptoms included ischemic symptoms (but not exclusively typical chest pain), new or presumed new significant ST-T wave changes excluding ST-segment elevation, the development of pathological Q waves, imaging evidence of new loss of viable myocardium or new regional wall motion abnormalities, and / or the identification of intracoronary thrombus by angiography or autopsy. The algorithm was derived from patients recruited in the emergency department of Heidelberg University Hospital, Germany, between July 1, 2016, and July 1, 2018. The cohort collected 12 months of follow-up data until July 1, 2019.

[0117] The entire cohort included 3,928 patients, of whom 3,018 were classified as rule-out and 910 were triaged to the observation zone. All-cause mortality occurred in 65 (1.65%) and 100 (1.82%) patients within 180 and 365 days, respectively. Briefly, the study cohort was predominantly male (55.91%, n = 2,195), with a mean age of 60.7 years. The overall discharge rate was 70.1% (2,801 of 3,928 patients), most of whom were in the rule-out zone (83.6%) (2,801 of 2,343 patients).

[0118] The rate of coronary angiography within 30 days of the index event was 1,344 (27.2%) of 4,934 eligible patients, of whom 889 (66.1%) patients required revascularization.

[0119] The model parameters were derived and trained on 75% of the entire study cohort, and the model was then tested on 25% of the entire study cohort. A random selection of patients was stratified for outcome events to eliminate bias.

[0120] The 30-day and 90-day all-cause mortality rates were 29 patients (0.74%) and 52 patients (1.3%), respectively. The 180-day and 365-day all-cause mortality rates were 65 patients (1.7%) and 100 patients (2.5%), respectively. Of the 1,344 patients who underwent coronary angiography within 30 days of the index event, a total of 889 (66.1%) required myocardial revascularization for obstructive coronary artery disease.

[0121] Example 2: Algorithm Development Logistic regression with elasticity for binary outcomes was applied to identify predictors and construct prediction and prognostic models in a training set comprising 75% of the entire study population. Models were trained using 5-fold cross-validation. A total of 33 models were developed by permuting a defined set of variables. Post-training performance was measured in a blinded test set comprising 25% of the study population. All estimated parameters were collected strictly on the training dataset only. We used mean imputation for numerical variables and mode imputation for nominal variables, so the mean and mode imputed values ​​were trained on the training dataset only. Skewed variables were log-transformed, and all numerical variables were centered and scaled based on the distribution of the training data. Additionally, an interaction term between the patient's baseline troponin and their age was considered relevant and added to the model. Five-fold cross-validation was used to estimate the hyperparameters penalty (representing the amount of penalty applied to the logistic regression model) and mixture (representing the relative amount of penalties L1 and L2) required for elastic net regularization. The hyperparameter grid consisted of 1,000 parameter combinations, each of which was fitted to a model selection process during 5-fold cross-validation. The best-fitting hyperparameters were evaluated based on the ROC-AUC performance for 5-fold cross-validation. After finding the best hyperparameters, the model was finally fitted to the entire training dataset to obtain estimates of covariates. ROC-AUC was selected as our primary performance measure. The Youden index on the ROC curve was then calculated to obtain thresholds for reporting the sensitivity and specificity of our model. Additionally, algorithm calibration was evaluated on the validation cohort (test set) by plotting calibration curves of predicted and observed means. The ML algorithm allowed us to estimate each patient's individual risk due to the probability of an adverse event within a prespecified follow-up interval.

[0122] Example 3: Algorithm for mortality prediction 3.1. Completeness and duration of follow-up For outcome variables in the prediction models, all-cause mortality was selected according to the Academic Research Consortium-2 Consensus Document (European Heart Journal 2018, 2192-220). Patients were followed for all-cause mortality for a median of 468 (range, 374-670) days, with 99.0% complete follow-up (41 of 3,969 patients lost to follow-up). Death was analyzed as a binary outcome (all-cause mortality at 180 and 365 days). Complete information on vital status at 90, 180, and 365 days was collected for all but 41 patients (99.0%).

[0123] 3.2 Model Selection Step 1. Correlation and discrimination within the derivation cohort. The ML algorithm was calibrated by comparing the predicted (fitted) probabilities for the blinded study population with the actual mean values ​​for the endpoint all-cause mortality at 180 and 365 days. Calibration plots of predicted mortality at 365 days using the full model, as well as the minimal and full models predicting mortality at 180 days, found a nearly linear relationship between predicted and actual mortality in the independent study population, particularly within the mortality range of interest of 0-3% (Figure 1). Overall, the predicted mortality risk slightly overestimated the actual mortality rate. This reduces the risk of inappropriate discharge and is preferable to underestimating the risk of death, which can lead to unsafe discharge and inadequate treatment.

[0124] The following points were taken into consideration when selecting the model:

[0125] 1) Patient population Applicable rules: From a clinical perspective, patients classified as rule-in by the validated hs-cTn protocol are at very high risk for myocardial infarction and require immediate diagnostic and therapeutic measures. Therefore, rule-in patients were excluded from the model. Only patients initially classified in the observation zone or as rule-out for myocardial infarction were considered. 3,018 patients were classified as rule-out and 910 patients were classified as the observation zone (excluding n = 965 rule-outs). This resulted in a sample size of n = 3,969 patients. The outcome of interest was missing in 41 patients. Therefore, the final sample size consisted of n = 3,928 patients.

[0126] 2) Outcome variables: timing and outcome measures Outcome Mortality follow-up was available. Event rates of all-cause mortality within 180 days and 365 days were examined in the rule-out and observation zone study populations.

[0127] Timing of events: Given that the follow-up period for the population was at least 12 months, timing of the binary outcome between 180 and 365 days was considered feasible. In this population of suspected ACS, the 1-month (0.5%) and 90-day (1.1%) event rates were too low to allow for the creation of well-characterized predictive models. The full model for predicting death at 365 days performed better than the GRACE score, with a significant gain in discriminatory power (p = 0.02). The full model and other models based on events at 6 months performed similarly, though not statistically significantly better, than the GRACE score. This is likely due to the small number of events, possibly due to a clinical course that was not directly related to the current presentation and not strongly correlated with the predictor variables at baseline.

[0128] The 180-365 day mortality outcome is clinically reasonable and substantially relevant for further clinical investigation and decision making. Models using 180-365 day mortality also yielded good diagnostic properties and were well calibrated in the region of interest (0-3%).

[0129] 3) Predictor variables: demographic, clinical, vital signs, ECG, and laboratory Different models were tested. The best-performing model was well calibrated and had good diagnostic performance in the blinded study cohort with an AUC of 0.86 (95% confidence interval 0.80-0.92). This consisted of only rule-out and observation zone patients randomly assigned to the 25% study set (n = 982). Predictor variables included age, sex, ECG parameters, and laboratory values ​​(including hs-cTnT and hs-cTnT kinetics, creatinine, sodium, C-reactive protein, hemoglobin, and platelet count).

[0130] Tables 3 and 4 show each parameter with its individual model estimate and corresponding relative weighting (for 180 and 365 days). The minimal model contains five different parameters, and the full model contains 10 parameters. Another 31 models are listed in the tables from top to bottom. [Table 3-1] [Table 3-2] [Table 3-3] [Table 4-1] [Table 4-2]

[0131] Step 2. Model performance on the test set Table legend: This table lists typical performance measures for predicting 365-day mortality using all 33 models. The list includes AUC, accuracy, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), precision, recall, and f1 score. Models are sorted in the same way as described in the previous table legend.

[0132] For example, the performance of the full model for 365-day mortality was associated with an AUC of 0.86 (0.80–0.92), a sensitivity of 83%, a specificity of 81%, a PPV of 14%, and an NPV of 99%. [Table 5] [Table 6]

[0133] Display of AUC and calibration plots for the minimal and full models. The discriminatory ability of the model to predict mortality is displayed using the area under the curve (AUC).

[0134] Performance across test sets for individual prediction of all-cause mortality All models were constructed and trained on 3,928 eligible individuals and then tested on all 982 individuals in the test set (25% randomly selected) to obtain the probability of all-cause mortality at either 365 or 180 days. The predicted probabilities after application of all models, including information on whether the endpoint was present or not, are listed in Tables 7 and 8. [Table 7] [Table 8]

[0135] Case studies Case Study: Case #505 was classified into the "rule out" category according to the ESC0 / 1 time algorithm, and a low risk of subsequent death, defined as a GRACE score of 84, was calculated using the GRACE score version 1.0. The patient was transferred to a local hospital. On day 52, the patient died of noncardiac causes. The ML-based minimal and full models predicted a 1.9% and 3.5% risk of death at 180 days, respectively, demonstrating superior prediction of mortality risk with the full model compared to the established predictive GRACE score.

[0136] Example 3-1: GRACE score vs. new model Several validated clinical scores reflecting individual risk have been proposed. Among these, the 2020 ESC Guidelines on Acute Coronary Syndromes without ST-segment Elevation propose the GRACE score as the preferred clinical score and assign it a Class IIa recommendation (should be considered). The GRACE score integrates patient age, occurrence of prehospital resuscitation, presence of pulmonary congestion, and renal dysfunction into a total score. Total scores of <109, 109–139, and ≥140 are interpreted as low, intermediate, or high risk of death between 180 days and 1 year. This model performance of the GRACE score was tested in a test set for its ability to predict death at 180 days and 365 days. In receiver operating characteristic (ROC) analysis, the area under the curve (AUC) was statistically compared using the method proposed by DeLong et al.

[0137] The AUC of the GRACE score for predicting mortality at 180 days was 0.774 (95% CI: 0.65-0.90) (see Figure 5). The AUC of the minimal model was 0.821, yielding a delta AUC of 0.05. This difference tended to be higher for the minimal model compared to the GRACE score (p=0.085). The AUC of the full model was 0.86 (95% CI: 0.74-0.98), yielding a non-significantly large (p=0.11) delta AUC of 0.15.

[0138] Similarly, the performance of the GRACE score was compared with the minimal and full models for predicting death at day 365. Here, the AUC of the full model was 0.86 (95% CI: 0.80-0.92), and thus significantly higher than the AUC of 0.79 (95% CI: 0.72-0.87) of the GRACE score (p=0.02).

[0139] Example 4: Algorithm Obstructive CAD: Model Selection Step 1. Correlation and discrimination within the derivation cohort. The ML algorithm was calibrated by comparing the predicted (fitted) probability for an independent study population with the actual mean value for the endpoint obstructive CAD requiring revascularization therapy. Calibration found a well-calibrated, approximately linear relationship between the predicted probability of the presence of obstructive CAD and the actual detection of obstructive CAD requiring revascularization across the entire probability space.

[0140] The following points were taken into consideration when selecting the model:

[0141] 1) Patient population Applicable rules: From a clinical perspective, obstructive CAD was present in all patient categories: rule-in, observation zone, and rule-out. All patients who underwent coronary angiography within 30 days of presentation were included (N = 1,344). Patients who did not undergo coronary angiography within 30 days of admission were excluded.

[0142] 2) Outcome variables: timing and outcome measures Outcome The definition of obstructive coronary artery disease requiring revascularization is defined in the previous section. Clinically, the probability of obstructive CAD is a relevant outcome because it may influence further diagnostic testing, treatment, and their timing.

[0143] 3) Predictor variables: demographic, clinical, vital signs, ECG, and laboratory Different models were tested. The best-performing model was well calibrated and showed moderate diagnostic performance in the test set with an AUC of 0.71 (95% CI 0.708-0.781). Predictor variables included age, sex, creatinine, ECG, and laboratory values ​​(including renal function, hs-cTnT, and hs-cTnT kinetics). [Table 9-1] [Table 9-2] [Table 9-3]

[0144] Step 2. Performance of the new model on the test set Risk estimates for the presence of obstructive CAD requiring revascularization within 30 days by ML were generally moderate to moderate, with AUC values ​​ranging from 0.69 to 0.71. A summary of the performance of all ML algorithms, including information on AUC, sensitivity, specificity, negative predictive value, positive predictive value, precision, recall, and f1 score, is listed in Table 10. [Table 10]

[0145] Display of AUC and calibration plots for selected models. The minimal and full models were well calibrated, with AUCs of 0.705 and 0.689, respectively (Figure 6), showing near-linear calibration between predicted and observed values.

[0146] The full model includes the following parameters: patient sex, age, first troponin (c0), delta troponin, creatinine, estimated glomerular filtration rate, urea (Hst), smoking history, history of previous revascularization, history of coronary heart disease, history of diabetes, EKG, and presence of chest pain or dyspnea as presenting symptoms.The minimal model includes the following parameters: patient sex, age, first troponin (c0_Tn), delta troponin, creatinine, smoking history, history of previous revascularization, and presence of chest pain.

[0147] Performance across the test set for individual prediction of the presence of significant coronary artery disease requiring myocardial revascularization within 30 days.

[0148] All models trained on 1007 eligible individuals were tested on 337 individuals (25% randomly selected) from the test set to obtain the probability of the presence of significant CAD requiring myocardial revascularization within 30 days. The predicted probabilities after application of all models, including information on whether the endpoint was present or not, are listed in Table 11. [Table 11]

[0149] Case studies Case Study: Case #304 was a 69-year-old man with atypical chest pain but a history of CAD. He was classified as "rule-out" according to the ESC 0 / 1 time algorithm and was considered to have a low pre-test probability of associated coronary artery disease (25%) based on the modified Diamond-Forrest prediction tool. The latter, if available, integrates age, sex, symptom typicality, presence of cardiovascular risk factors, and coronary artery calcium score. The patient was admitted at the discretion of his attending physician and underwent coronary angiography, which demonstrated multivessel coronary artery disease and intimate coronary lesions requiring percutaneous coronary intervention. The ML-based minimal and full models predicted a 55% and 65% risk of CAD requiring revascularization, respectively, highlighting their superior ability to predict CAD requiring revascularization compared with the established modified Diamond-Forrest prediction tool.

[0150] Example 5: ESC Consortium Prediction Tool vs. New Prediction Models Current guidelines in the United States and Canada recommend using the Diamond and Forrester model (Diamond GA, et al. New Engl J Med 1979 June 14, 300(24):1350-8) or the Duke Clinical Score (Pryor DB, et al. Ann Intern Med 1993 January 15, 118(2):81-90) to estimate the pre-test probability of CAD in patients with stable chest pain. Both models tend to overestimate the pre-test probability of CAD compared with the ESC Consortium calculator (Genders TS, et al. BMJ: British Medical Journal 2012 June 12, 344:e3485). This calculator was developed and validated based on over 5,500 patients from 18 different hospitals in Europe and the United States. The prediction model integrates age, gender, chest pain, diabetes, hypertension, hyperlipidemia, smoking, available coronary artery calcium score, and coronary artery calcium score (if unavailable, enter 0, thus calculating probabilities by age and symptom separately for men and women). This model was tested in a test set for its ability to predict coronary artery occlusion requiring revascularization, and the AUC was compared with all other models. The AUC of the ESC Consortium algorithm demonstrated moderate discriminant performance, with an AUC of 0.63 (95% confidence interval: 0.57-0.69). The AUC of the minimal model was 0.71, resulting in a delta AUC of 0.08 (p=0.004), and the AUC of the full model was 0.69, resulting in a delta AUC of 0.06 (p=0.04), significantly higher than the AUC of the ESC Consortium score (see Figure 8 below). All other models exhibited AUCs between the AUC of the minimal model and the full model.

[0151] Supplementary table [Table 12] [Table 13] Table 14 Table 15-1 Table 15-2 Table 15-3 Table 15-4 Table 15-5 Table 15-6 Table 15-7 Table 15-8 Table 15-9 Table 15-10 Table 15-11 Table 15-12 Table 16-1 Table 16-2 Table 16-3 Table 16-4 Table 16-5 Table 16-6 Table 16-7 Table 16-8 Table 16-9 Table 16-10 Table 16-11 Table 16-12 Table 17-1 Table 17-2 Table 17-3 Table 17-4

Claims

1. 1. A computer-implemented method for predicting risk of an adverse event in a patient presenting with a suspected acute coronary syndrome, comprising: a) receiving in a processing unit data relating to a set of parameters obtained from said patient, said set of parameters comprising the following parameters: i. the amount of cardiac troponin in a first sample obtained from said patient at the time of presentation; ii. the amount of the cardiac troponin in a second sample obtained from the patient within 30 minutes to 6 hours after the first sample, for example, within 1 hour to 3 hours after the first sample, for example, within 1 hour to 2 hours after the first sample; iii. The amount of CRP (C-reactive protein) in a sample from the patient; iv. at least one parameter of renal function of said patient selected from the group consisting of the amount of urea in a sample from said patient, the GFR of said patient, and in particular the amount of serum creatinine of said patient; v. the amount of sodium in a sample from the patient; vi. the amount of hemoglobin in a sample from the patient; vii. the patient's platelet levels; viii. the age of the patient; ix. the patient's gender, and x. Presence or absence of a normal ECG in the patient the receiving step including at least five, e.g., five, six, seven, eight, nine, or ten, of the parameters b) performing, in the processing unit, an analysis of the set of parameters, the analysis comprising calculating a score for predicting the patient's risk of the adverse event based on the set of parameters received in step a); c) providing information regarding the score calculated in step b), thereby predicting the risk of the adverse event; 11. A computer-implemented method comprising:

2. 2. The method of claim 1, wherein the adverse event is death and / or the risk of the adverse event within about 180 days to about 365 days, such as within about 180 days or about 365 days, is predicted.

3. 3. The method of claim 1 or 2, wherein the score indicates the likelihood that the patient will suffer from an adverse event.

4. 4. The method according to any one of claims 1 to 3, wherein in step a) data is received on at least five, such as five, six, seven, eight, nine or ten of the following parameters: the amount of cardiac troponin in a first sample obtained from the patient at the time of presentation, the amount of said cardiac troponin in a second sample (as described above), the amount of CRP (C-reactive protein) in a sample from the patient, the amount of creatinine in a serum sample from the patient, the amount of sodium in a sample from the patient, the amount of hemoglobin in a sample from the patient, the patient's platelet level, information on the patient's age, information on the patient's gender, and information on the presence or absence of a normal ECG in the patient.

5. 4. The method according to claim 1, wherein in step a) data is obtained on at least the following parameters: the amount of cardiac troponin in a first sample obtained from the patient at the time of presentation, the amount of said cardiac troponin in a second sample (as described above), the amount of CRP (C-reactive protein) in a sample from the patient, the creatinine value of the patient, information on the age of the patient, and information on the gender of the patient.

6. 1. A method for predicting the risk of an adverse event in a patient with suspected acute coronary syndrome, comprising: a) The following steps a1) to a10), namely: a1) determining the amount of cardiac troponin in a first sample obtained from said patient at the time of presentation; a2) determining the amount of the cardiac troponin in a second sample obtained from the patient within 30 minutes to 6 hours after the first sample, for example within 1 hour to 3 hours after the first sample, for example within 1 hour to 2 hours after the first sample; a3) iii. The amount of CRP (C-reactive protein) in a sample from the patient; a4) determining at least one parameter of the patient's renal function selected from the group consisting of the amount of urea nitrogen in a (blood) sample from the patient, the patient's GFR, and the patient's serum creatinine level; a5) determining the amount of sodium in a sample from said patient; a6) determining the amount of hemoglobin in the patient sample; a7) determining the patient's platelet level; a8) providing information regarding the patient's age; a9) providing information regarding the patient's gender; and a10) providing information regarding the presence or absence of a normal ECG in said patient; performing at least five, e.g., five, six, seven, eight, nine, or ten, of b) calculating a score for predicting the patient's risk of an adverse event based on the information obtained in step a); c) predicting the patient's risk of the adverse event based on the score calculated in step b); and A method comprising:

7. 1. A computer-implemented method for predicting the need for myocardial revascularization in a patient presenting with a suspected acute coronary syndrome, comprising: a) receiving in a processing unit data relating to a set of parameters obtained from said patient, said set of parameters comprising the following parameters: i. the amount of cardiac troponin in a first sample obtained from said patient at the time of presentation; ii. the amount of the cardiac troponin in a second sample obtained from the patient within 30 minutes to 6 hours after the first sample, for example, within 1 hour to 3 hours after the first sample, for example, within 1 hour to 2 hours after the first sample; iii. the age of the patient; iv. the patient's gender; v. a parameter of renal function of said patient, such as said patient's serum creatinine level, selected from the group consisting of said patient's GFR, said patient's serum creatinine level, and said patient's urea or blood urea nitrogen (BUN) level; vi. the presence or absence of chest symptoms in the patient; vii. the presence or absence of dyspnea in the patient; viii. The presence or absence of a normal ECG in the patient; and ix. Information regarding the patient's past medical history, including at least one, and preferably all, of the following: information regarding the patient's diabetes history; information regarding the patient's smoking history; and information regarding the patient's history of coronary heart disease (such as information regarding the patient's history of myocardial revascularization). receiving at least eight, e.g., eight, nine, of b) performing, in the processing unit, an analysis of the set of parameters, the analysis including calculating a score for predicting the need for myocardial revascularization of the patient based on the set of parameters received in step a); c) providing information about the score calculated in step b), thereby predicting the need for myocardial revascularization in the patient; and 11. A computer-implemented method comprising:

8. 8. The method of claim 7, wherein the myocardial revascularization is due to obstructive coronary artery disease.

9. 9. The method of claim 7 or 8, wherein the need for said myocardial revascularization procedure within about 30 days is predicted.

10. In step a), the following parameters are determined: the amount of cardiac troponin in a first sample obtained from said patient at the time of presentation, the amount of said cardiac troponin in a second sample (as described above), the amount of serum creatinine in said patient, 10. The method of any one of claims 7 to 9, wherein data is received relating to at least eight, such as eight, nine, ten, eleven, twelve, thirteen, fourteen or fifteen pieces of information relating to the patient's past medical history, including at least one, and preferably all, of the presence or absence of chest symptoms in the patient, the presence or absence of dyspnea in the patient, information about the patient's gender, information about the patient's past medical history, the information being information about the patient's history of diabetes, information about the patient's smoking history, information about the patient's history of coronary heart disease, and the presence or absence of a normal ECG in the patient (i.e., whether the patient has a normal or abnormal ECG).

11. 10. The method according to any one of claims 7 to 9, wherein in step a) data is obtained on at least eight of the following parameters: the amount of cardiac troponin in a first sample obtained from the patient at the time of presentation, the amount of said cardiac troponin in a second sample (as described above), information on the patient's gender, information on the patient's age, creatinine, and information on the patient's history of coronary artery disease, information on the patient's smoking history (past or current), information on the patient's presence of chest symptoms, in particular chest pain.

12. The myocardial revascularization a. Invasive coronary angiography with percutaneous coronary intervention with or without stent placement; b. Invasive coronary angiography with lesions suitable for revascularization and attempted percutaneous coronary intervention with or without stent placement; c. Invasive coronary angiography with lesions suitable for revascularization and attempted but unsuccessful or unsuccessful percutaneous coronary intervention; d. Invasive coronary angiography with associated lesions not suitable for revascularization and subsequent conservative medical treatment without percutaneous coronary intervention; e. Immediate transfer for invasive coronary angiography and coronary artery bypass surgery with lesions suitable for revascularization; f. Recommendation for invasive coronary angiography and coronary artery bypass surgery with lesions suitable for revascularization, or g. Invasive coronary angiography with lesions planned for revascularization, such as percutaneous coronary intervention or coronary artery bypass surgery The method according to any one of claims 7 to 11, wherein the method is one of

13. 1. A method for predicting the need for myocardial revascularization in a patient presenting with a suspected acute coronary syndrome, comprising: a) The following steps a1) to a9), namely: a1) determining the amount of cardiac troponin in a first sample obtained from said patient at the time of presentation; a2) determining the amount of the cardiac troponin in a second sample obtained from the patient within 30 minutes to 6 hours after the first sample, for example within 1 hour to 3 hours after the first sample, for example within 1 hour to 2 hours after the first sample; a3) providing information regarding the patient's gender; a4) providing information regarding the patient's age; a5) providing information regarding a parameter of renal function of said patient, such as said patient's GFR, selected from the group consisting of said patient's GFR, said patient's serum creatinine level, and said patient's urea or blood urea nitrogen (BUN) level; a6) providing information regarding the presence or absence of chest symptoms in said patient; a7) providing information regarding the presence or absence of dyspnea in said patient; a8) providing information regarding the presence or absence of a normal ECG in said patient; and a9) providing information regarding the patient's past medical history, including at least one, and preferably all, of the following: information regarding the patient's diabetes history; information regarding the patient's smoking history; and information regarding the patient's history of coronary artery disease. performing at least eight, such as nine, such as ten or eleven, or twelve or thirteen or fourteen or fifteen of the following: b) calculating a score for predicting the need for myocardial revascularization of the patient based on the information obtained in step a); c) predicting the need for myocardial revascularization of the patient event based on the score calculated in step b); and A method comprising:

14. The method of any one of claims 7 to 13, wherein the score indicates the likelihood that the patient will require myocardial revascularization.

15. The method of any one of claims 1 to 14, wherein the sample is a blood, serum or plasma sample and / or the calculated score is shown on a display.

16. 16. An apparatus for predicting a risk of an adverse event or for predicting the need for myocardial revascularization, comprising: a processing unit; and a computer program comprising computer-executable instructions that, when executed by the processing unit, cause the processing unit to perform the computer-implemented method of any one of claims 1 to 5, 7 to 12, 14 and 15.

Images