Surgery – Diagnostic testing – Cardiovascular
Reexamination Certificate
2001-06-12
2004-05-04
Schaetzle, Kennedy (Department: 3762)
Surgery
Diagnostic testing
Cardiovascular
Reexamination Certificate
active
06731973
ABSTRACT:
BACKGROUND OF INVENTION
The invention relates to a method and apparatus for processing physiological data, particularly non-electrocardiogram, cardiac-related data, by identifying the physiological data associated with atypical cardiac events.
Much of the physiological data which may be recorded from the human body is related to the cardiac cycle. The most obvious example is the electrocardiogram (ECG). Each time the heart beats, the electrical activity of the heart creates electrical potentials which are detectable with electrodes attached to the body surface. The ECG waveform is so dependent upon the cardiac cycle that it is relied upon to provide significant diagnostic information about the health and functioning of the heart. Moreover, differences in the beat to beat function of the heart closely correlate with beat to beat changes in the morphology of the ECG waveform.
In recent years, the manual analysis of the ECG waveform has been enhanced by the use of ECG analysis software algorithms, such as ECG classification engines. ECG classification engines are used to automatically identify and categorize individual heart beats within an ECG waveform stream based on the morphology of the individual heart beats. ECG classification engines are also used to form associations among the individual heart beats having similar morphologies.
To identify and categorize individual heart beats, the ECG classification engine first samples an acquired ECG waveform stream to create ECG data representing the waveform stream. Various filtering techniques may be applied, either before or after sampling the waveform stream, to eliminate noise, such as power line interference.
Once the ECG data is sampled and filtered, the data is divided into segments. Each segment generally represents a single cardiac cycle or heart beat. In order to identify each heart beat, the ECG classification engine locates the QRS complex within each heart beat. Once the ECG data is segmented, each data segment is classified according to the type of heart beat that may have generated the data segment. For example, the data segments may be classified as typical, ventricular, or paced.
It is not uncommon, however, for the ECG classification engine to group the data segments before classifying the data segments. The data segments may be grouped based on the morphology of the ECG data within each data segment. The data segments with similar morphologies are grouped together. In order to group the data segments, each new data segment is compared, in some manner, to a representative data segment from each of the data segment groups. In order to compare each new data segment to the representative data segments, the two data segments must be aligned according to a fiducial point. Typically, the R wave within the QRS complex of each data segment is used as the fiducial point, although any other easily-recognizable feature in the two data segments may be used as the fiducial point.
Once the data segments are grouped, the ECG classification engine can analyze and classify an entire group of data segments by analyzing either one data segment within the group or by analyzing and classifying a representative data segment for the group. The representative data segment for a group of data segments may be, for example, an average of the data segments or the median data segment. The ability to analyze one data segment that represents an entire group of data segments greatly reduces the computational burden on the ECG classification engine and reduces the data storage requirements of the ECG system.
Once the data segments are analyzed and classified based solely upon their morphology, an additional analysis may be performed. Each data segment may be analyzed based on the morphology of adjacent data segments. This additional analysis may lead to a different classification for the data segment or an enhanced classification. For example, a data segment classified as a ventricular contraction (VC) data segment may be classified as a premature ventricular contraction (PVC) data segment if it is sufficiently close to the preceding heart beat. In addition, a sequence of data segments may exhibit a rhythm or an arrhythmia. For example, a series of three or more ventricular contractions at a high rate represents a ventricular tachycardia arrhythmia.
In summary, the ECG classification engine first groups data segments based on the morphology of each individual data segment. Second, the ECG classification engine groups the morphologies into a classification. Third, the ECG classification engine narrows the classification of each individual data segment based on the data segment's relationship to immediately-adjacent data segments. Finally, the ECG classification engine qualifies the classification of each individual data segment based on the sequence of data segments in which each individual data segment is located.
In addition to ECG data, many other types of physiological data are closely related to the cardiac cycle. Non-ECG physiological data closely related to the cardiac cycle includes, for example, continuous blood pressure data, blood oxygen saturation data, cardiac impedance data, cardiac sound data, and cardiac output data. Non-ECG physiological data closely related to the cardiac cycle also includes cardiac imaging data, which is created whenever the heart is imaged with imaging methods such as X-ray, nuclear magnetic resonance (NMR), or scintiphotography.
Similar to the processing of ECG data, non-ECG physiological data from one cardiac cycle is often averaged or otherwise combined with the data obtained during other cardiac cycles in order to enhance the features of the non-ECG physiological data. In order to combine the non-ECG physiological data from several cardiac cycles, the data must be divided into segments by identifying which portion of the data is from one cardiac cycle and which is from another cardiac cycle. Also, in order to combine the data in a coherent manner, the data segments must be aligned so that the same features are being compared in each of the data segments. However, non-ECG physiological data does not generally include a consistent and recognizable fiducial point like the QRS complex of ECG data.
While dividing and aligning ECG data segments is generally straightforward, there can be difficulties when attempting to divide and align non-ECG physiological data segments, because it may be difficult to identify a consistent fiducial point within the data. This problem is often solved by simultaneously acquiring ECG data which may be temporally associated with the non-ECG physiological data. The non-ECG physiological data is then divided into segments and aligned using the QRS complex of the simultaneously-acquired ECG data as a fiducial point. This method of aligning non-ECG physiological data with simultaneously-acquired ECG data is known as R-Wave Gating. A representative non-ECG physiological data segment is then generated by averaging or combining the segments.
However, the representative non-ECG physiological data segment that is generated by averaging all of the non-ECG physiological data segments may not accurately represent the non-ECG physiological data. The combination of segments relies upon the presumption that each segment represents a cardiac cycle substantially similar to the other data segments that it is being combined with. However, atypical and abnormal cardiac cycles may occur periodically while the non-ECG physiological data is being acquired.
An atypical cardiac cycle may be considered a normal cardiac cycle, i.e. a cardiac cycle may be different from other cardiac cycles, but may not be related to a cardiac disease, malfunction, or abnormality. Several types of normal cardiac cycles may exist within a stream of ECG waveform data. However, each of these normal cardiac cycles differs from the typical cardiac cycles and is thus considered atypical. When averaged with typical data segments, atypical data segments, even though they represent a normal cardiac cycle, can have a negative impact
Droesch Kristen
GE Medical Systems Information Technologies Inc.
Michael & Best & Friedrich LLP
LandOfFree
Method and apparatus for processing physiological data does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for processing physiological data, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for processing physiological data will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3265234