Provision of decision support for acute myocardial infarction

Details Provision of decision support for acute myocardial infarction Provision of decision support for acute myocardial infarction

2000-02-10

2002-09-03

Winakur, Eric F. (Department: 3736)

Surgery

Diagnostic testing

C128S925000, C706S015000

Reexamination Certificate

active

06443889

ABSTRACT:

TECHNICAL FIELD
The present invention relates to the assessment, management and treatment of patients with suspected acute myocardial infarction (AMI), and more particularly to methods and apparatuses for decision support intended for early assessment of such patients with regard to classification into groups related to different preferred management and treatment options.
BACKGROUND
An early assessment within the first hours after onset of symptoms is essential for the optimal management and treatment of patients with suspected acute myocardial infarction (AMI) as has been documented by e.g. the National Heart Attack Alert Program Coordinating Committee, 60 minutes to Treatment Working Group (in Ann Emerg Med 1994; 23:311-329). In patients with ST-elevation in their initial ECG recording on admission the diagnosis is straight forward. Moreover, the total sum of ST-elevations in all ECG-leads gives a good estimate of the myocardium at risk. However, in at least 40% of patients with AMI the 12-lead ECG is non-diagnostic on admission. In these patients the diagnosis has to be based on clinical data and measurements of biochemical markers, until recently a time consuming procedure that causes a delay in starting the treatment. An early and reliable prediction of the infarct size is also difficult for this group of patients. In the heterogeneous group of patients admitted to the hospital because of chest pain, considerable economic gains might be achieved by early identification of those patients (approximately 60-70%) who are at sufficiently low risk of AMI and its complications to be transferred to a general ward outside the coronary care unit (CCU). Among patients with unstable angina 30-50% have minor elevations of sensitive biochemical markers such as creatine kinase MB and troponin-T. The term minor myocardial damage (MMD) has been proposed for these minor elevations. Since MMD indicates an increased risk of future cardiac events it seems important not only to detect AMI, but also to detect MMD as early as possible.
Thus, in the early management of patients with acute chest pain and with non-diagnostic ECG at admission, there are several important issues to be addressed in order to take suitable actions. The existence of an AMI has to be determined. If an AMI is present, the ultimate infarct size is useful to know for the continued treatment. The time when the infarction actually occurred is also important. If the patient does not have an AMI, it is still useful to decide if the patient have a high or low risk for subsequent cardiac events. The clinician has to consider the patient's history and physical status, the results of ECG and blood tests etc. This evaluation process may be time consuming and is heavily dependent on the clinician's knowledge and experience. The biochemical measurement results form an important basis on which the future management of the patient relies.
There are important differences in the kinetic properties of the biochemical markers of AMI of today, e.g. myoglobin, CKMB and cardiac troponin-T. All these markers are indicative of AMI, however, on different time scales. Myoglobin has a rather quick response, while at least troponin-T has a much slower appearance rate. A combination of measurements of more than one marker is necessary.
Computer methods using artificial neural networks have been applied for AMI detection based on clinical data, measurements of biochemical markers and ECG/VCG parameters. Jörgensen et al, Clin. Chem. 1996, vol. 42(4), pp. 604-612 and 613-617, investigated the diagnostic performance of neural networks trained on various combinations of initial ECG data, and serum concentrations of CKB, LD1 and potassium, on admission, 12 and 24 hours after admission. In a recent paper (Comp. Biomed. Research 1998, vol. 36, 59-69) Sunemark et al investigated an approach to serial VCG/ECG analysis based on artificial neural networks combining two ECG and VCG measurements. In short, the neural network applications by Sunemark et al and Jörgensen et al have not addressed the problem of early AMI size predicition in chest pain patients with non-diagnostic ECG on admission.
In U.S. Pat. No. 5,690,103, methods and apparatuses for detection or exclusion of acute myocardial infarction using artificial neural network (ANN) analysis of measurements of biochemical markers are disclosed. The overall invention categorizes patients with suspected AMI with regard to AMI
on-AMI; infarct size; time since onset of infarction; and non-AMI with/without minor myocardial damage (MMD). The categorization is generally based on frequent blood sampling during the first hours after admission and measurement of selected biochemical markers of AMI with different rates of appearance in circulating blood. The computations are performed by using specially designed artificial neural networks. Early detection/exclusion of acute myocardial infarction is provided, generally within 3 hours from admission of the patient. Furthermore, early prediction of the “major infarct size” and early estimation of the time from onset are also provided.
It is important to understand that the time aspect is very crucial in the present technical field. After 12-24 hours after admission, the patterns of the marker concentrations are generally so pronounced that the clinicians themselves can easily assess AMI therefrom without using any type of neural network support. However, 12 hours after admission, the potential benefit of e.g. thrombolytic treatment has decreased to a very low level. Measurement evaluation and decisions have to be made in the very first hours after admission of a patient.
The methods and apparatuses of the above U.S. patent provide a large development in providing early and reliable detection, prediction and estimation in connection with AMI. Although the new technology of the above patent has worked in a satisfactory manner, there still remain issues that should be addressed regarding the incorporation of these types of artificial neural networks into a medical decision support system, and the transferability of such a system to function in a new environment with its specified medical requirements.
SUMMARY
A general objective of the present invention is to further improve the quality of the decision support for early assessment of patients with suspected acute myocardial infarction (AMI), derived by using artificial neural networks. A further objective of the present invention is to relate measured variables associated with AMI to regions of AMI classification groups related to management options and treatment procedures, and in particular to visualize such relations in an informative way. Another objective of the present invention is to provide artificial neural networks, which are tuned to fulfil clinical requirements on the certainty of predicted infarct size and corresponding AMI classification groups in various prevalence situations. Yet another objective of the present invention is to provide artificial neural networks that make use of variables associated with AMI, which are derived from intermittent and/or continuous ECG/VCG measurements.
The above objectives are accomplished by methods and apparatuses according to the appended claims.
In general words, one aspect of the present invention provides a method and an apparatus, which make use of at least one trained and tuned artificial neural network to generate decision regions in the n-dimensional space of n input variables associated with AMI. The set of measured variables at a certain time instance, i.e. a point in the n-dimensional space, is related to the decision regions, in order to provide decision support. Preferably, these decision regions are projected and graphically visualized as areas in a parameterized two-dimensional diagram, with two major variables as independent variables. The values of the remaining n−2 input variables are the corresponding parameters. Preferably, the two major variables are biochemical AMI markers. The position of the point representing a present set of values

Affiliated with

Also associated with

Rating

Provision of decision support for acute myocardial infarction does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Provision of decision support for acute myocardial infarction, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Provision of decision support for acute myocardial infarction will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2912190