AI ECG Screening for Early Low Ejection Fraction Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for diagnosing low ejection fraction (EF) rely on expensive and specialized procedures like transthoracic echocardiograms, which are often not performed until symptoms appear, creating a care gap and preventing early intervention for cardiovascular diseases.
Innovation Solution
Utilizing artificial intelligence models, particularly ensemble models, to analyze electrocardiogram (ECG) data to detect low EF, incorporating multiple models to aggregate predictions and adjust for patient characteristics, enabling accurate and equitable detection of undiagnosed cardiovascular conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional diagnostic methods (transthoracic echocardiograms) are used to detect low ejection fraction, then measurement precision is improved, but device complexity and cost increase, and accessibility decreases
Solution Approach 1:
The patent uses AI models to create a virtual copy of the complex echocardiogram diagnostic process. Instead of requiring actual echocardiogram equipment and specialized operators, the system uses trained AI models that can predict ejection fraction from simpler ECG data, effectively copying the diagnostic capability without the complexity
Solution Approach 2:
The patent replaces the mechanical/equipment-based diagnostic system (echocardiogram machines requiring specialized operators) with an information-processing system (AI models analyzing ECG data). This substitution eliminates the need for complex hardware while maintaining diagnostic precision
2Measurement precision
If conventional diagnostic methods are used, then measurement precision is improved, but loss of time increases as procedures are not performed until symptoms appear
Solution Approach 1:
The system performs preliminary detection of low ejection fraction using routinely collected ECG data before symptoms manifest. By continuously monitoring ECG data and applying AI models, the system can identify at-risk patients early, enabling preventive intervention before clinical symptoms appear
Solution Approach 2:
The patent leverages existing ECG data that is already being collected for routine cardiac monitoring, making the diagnostic system self-service. The ECG data serves dual purposes: routine monitoring and early detection of low EF, eliminating the need for separate diagnostic procedures
3Ease of operation
If AI models are used to detect low EF from ECG data, then accessibility and ease of operation are improved, but measurement precision may worsen compared to conventional methods
Solution Approach 1:
The patent merges multiple AI models into an ensemble system where each model processes ECG data independently and their predictions are combined. This merging approach aggregates the strengths of individual models, improving overall accuracy while maintaining ease of operation through automated processing
Solution Approach 2:
The system incorporates feedback mechanisms where the AI models are continuously trained and refined using validated ECG data and clinical outcomes. This feedback loop ensures that the models maintain high precision by learning from actual diagnostic results and adjusting their predictions
4Device complexity
If a single AI model is used to determine low EF score, then device complexity is reduced, but reliability worsens due to model variability and bias
Solution Approach 1:
The patent combines multiple AI models with different architectures and training approaches into an ensemble system. By aggregating predictions from multiple models, the system reduces the impact of any single model's biases or errors, thereby improving reliability and consistency of low EF detection
Solution Approach 2:
The system dynamically adjusts operating parameters and thresholds based on patient characteristics and model performance metrics. This parameter optimization ensures consistent and reliable detection across different patient populations, compensating for potential model biases
Data Source
AI summary
Techniques for detecting low ejection fraction (EF) from electrocardiogram (ECG) data using artificial intelligence (AI) are disclosed. ECG data is obtained for a patient. The ECG data is provided as input to a plurality of trained AI models, wherein each trained AI model is trained to determine a score of low EF based on the ECG data. A score of low EF is determined using each trained AI model based on the ECG data. An overall score of low EF is determined based on a combination of the scores of low EF determined by each of the trained AI models. An operating point threshold is obtained for the patient based on a plurality of patient characteristics. The overall score of low EF is compared to the operating point threshold. A low EF prediction is made for the patient based on the comparison.