Arrhythmia Detection Using Feature Delineation and ML Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing implantable medical devices (IMDs) face challenges in accurately detecting and classifying cardiac arrhythmias, particularly malignant tachyarrhythmias like ventricular fibrillation, which can lead to sudden cardiac death, due to limitations in feature delineation and power consumption constraints.
Innovation Solution
A medical device system combining feature delineation and machine learning to enhance arrhythmia detection and classification, with IMDs performing initial detection and an external device performing comprehensive analysis to reduce power consumption and improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If comprehensive arrhythmia analysis is performed using both feature delineation and machine learning, then detection accuracy is improved, but power consumption increases
Solution Approach 1:
The patent segments the arrhythmia analysis process into two distinct parts: (1) feature delineation performed by the implantable medical device (IMD) using low-power processing, and (2) machine learning classification performed by an external computing device. This segmentation allows the IMD to perform only the essential initial detection with minimal power consumption, while the computationally intensive machine learning operations are offloaded to the external device, thereby resolving the contradiction between detection accuracy and power consumption.
Solution Approach 2:
The patent introduces an intermediary communication channel between the IMD and the external computing device. The IMD extracts and transmits only the essential cardiac features (such as QRS complexes, T-waves, and other morphological parameters) to the external device, which then performs the machine learning analysis. This intermediary approach allows the system to achieve high detection accuracy through comprehensive analysis while the IMD itself consumes minimal power, as it only performs feature extraction and data transmission.
2Measurement precision
If machine learning model is applied to cardiac electrogram data, then arrhythmia classification accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent segments the computational tasks by assigning feature extraction and preprocessing to the IMD and reserving the computationally intensive machine learning model application for the external computing device. The IMD performs relatively simple feature delineation on cardiac electrogram data, extracting key morphological parameters, and transmits these features to the external device which then applies the complex machine learning model for accurate arrhythmia classification. This segmentation resolves the contradiction by placing the computational burden on the external device while maintaining high classification accuracy.
Solution Approach 2:
The patent applies preliminary action by having the IMD perform feature delineation and preprocessing of cardiac electrogram data before transmission to the external device. The IMD extracts and prepares the essential cardiac features (such as QRS complexes, T-wave morphology, and other relevant parameters) in advance, so that when the data reaches the external computing device, the machine learning model can be applied more efficiently. This preliminary processing reduces the computational complexity at the external device while maintaining high classification accuracy.
3Measurement precision
If feature delineation is used for arrhythmia detection, then detection capability is improved, but false detections may occur
Solution Approach 1:
The patent implements a feedback mechanism where the machine learning model's classification results are used to verify and validate the initial detections made by feature delineation. The system compares the machine learning classification outcomes with the feature-based detection results, and uses this feedback to confirm true arrhythmia events while filtering out false detections. This feedback loop significantly improves reliability by reducing false positive rates while maintaining high detection capability.
Solution Approach 2:
The patent introduces machine learning classification as an intermediary verification step between feature delineation and final arrhythmia detection. The feature delineation process initially identifies potential arrhythmia events based on cardiac electrogram features, and then the machine learning model acts as an intermediary to verify these detections by comparing them against learned patterns from training data. This intermediary verification layer filters out false detections while preserving true arrhythmia events, thereby improving overall detection reliability.
Data Source
AI summary
Techniques are disclosed for using both feature delineation and machine learning to detect cardiac arrhythmia. A computing device receives cardiac electrogram data of a patient sensed by a medical device. The computing device obtains, via feature-based delineation of the cardiac electrogram data, a first classification of arrhythmia in the patient. The computing device applies a machine learning model to the received cardiac electrogram data to obtain a second classification of arrhythmia in the patient. As one example, the computing device uses the first and second classifications to determine whether an episode of arrhythmia has occurred in the patient. As another example, the computing device uses the second classification to verify the first classification of arrhythmia in the patient. The computing device outputs a report indicating that the episode of arrhythmia has occurred and one or more cardiac features that coincide with the episode of arrhythmia.