Machine Learning Intracardiac Pressure Estimation from Non-Invasive Tests
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for non-invasive monitoring of heart failure do not account for individual patient differences in biological parameters, leading to inaccurate assessments of cardiac function.
Innovation Solution
A method for training a machine learning model using training data derived from invasive and non-invasive tests to estimate intracardiac pressure, involving the acquisition of target and parameter values at different times, and deriving change ratios to optimize the model for remote patient monitoring.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If non-invasive monitoring devices are used for home monitoring, then patient convenience and ease of operation are improved, but measurement precision and accuracy of cardiac function assessment deteriorate due to lack of individual patient calibration
Solution Approach 1:
The patent transforms fixed reference values into dynamic, patient-specific reference values by calculating individual baseline parameters from multiple measurements. This allows the monitoring system to adapt to each patient's unique physiological characteristics, thereby maintaining high measurement precision while preserving the ease of home monitoring.
Solution Approach 2:
The system continuously compares current measurements against individually calibrated reference values and provides feedback on disease state changes. This feedback mechanism enables accurate cardiac function assessment by accounting for each patient's specific baseline, resolving the contradiction between ease of operation and measurement precision.
2Measurement precision
If invasive tests are performed to obtain accurate target values, then measurement precision is improved, but device complexity and difficulty of operation worsen
Solution Approach 1:
The patent introduces machine learning models as intermediaries that translate simple non-invasive measurements into clinically meaningful target values. These models are trained on the relationship between easy-to-obtain parameters and invasive measurements, enabling accurate assessment without requiring complex invasive equipment.
Solution Approach 2:
The system replaces complex mechanical invasive measurement procedures with computational models that process simple non-invasive signals. This substitution maintains measurement precision while dramatically reducing device complexity and operational difficulty.
3Reliability
If frequent hospital visits are conducted to monitor disease state, then reliability of disease status assessment is improved, but loss of time and productivity worsen for both patients and healthcare system
Solution Approach 1:
The patent enables patients to self-monitor their cardiac function at home using simple non-invasive devices, eliminating the need for frequent hospital visits. The system automatically compares current measurements against individual baselines and detects disease state changes, providing reliable monitoring that patients can perform independently in their own environment.
Solution Approach 2:
The system performs preliminary calibration of individual reference values during initial hospital visits, after which patients can continue reliable monitoring at home without repeated visits. This preliminary action establishes the foundation for ongoing self-monitoring, reducing time loss while maintaining assessment reliability.
Data Source
AI summary
A method for training a machine learning model executed to assess a condition of a patient with heart failure, includes generating training data by performing for each patient acquiring a first target value based on an invasive test and a first parameter value based on a non-invasive test at a first time, and acquiring a second target value based on the invasive test and a second parameter value based on the non-invasive test at a second time, deriving a target change ratio based on the first and second target values, deriving a parameter change ratio based on the first and second parameter values, and storing the change ratios as the training data, and training a machine learning model with the training data such that a target change ratio is generated in response to an input of an actual parameter change ratio derived for a patient with heart failure.