LVAD Thrombus Detection via Power Trend Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Determining adverse events in patients with implanted ventricular assist devices (VADs) is challenging due to issues with sensor calibration, power consumption, and sensitivity, often leading to false alarms and increased morbidity.
Innovation Solution
A method using low-pass filters to calculate power consumption trends, determine differences, and generate alarms when thresholds are exceeded, including short-term and long-term trends, and speed-dependent thresholds to detect adverse events such as thrombus and other conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If sensors are implanted into or onto the VAD to detect operating parameters, then adverse events can be detected, but sensor calibration is subject to failure from corrosion and increases power consumption
Solution Approach 1:
The patent replaces physical sensors with a computational method that uses existing power consumption data from the VAD's motor controller. Instead of implanting additional sensors to detect thrombus formation, the system analyzes changes in power consumption patterns through signal processing algorithms, eliminating the need for extra sensing hardware and its associated calibration and power requirements.
Solution Approach 2:
The VAD's existing power consumption measurements, already taken for motor control purposes, are repurposed to detect adverse events. The system uses the motor controller's existing data acquisition capability to monitor power consumption and apply filtering algorithms, allowing the device to self-diagnose thrombus formation without additional sensors or power expenditure.
2Measurement precision
If sensors are implanted to detect adverse events, then detection capability is improved, but the system becomes more complex with calibration requirements
Solution Approach 1:
The patent extracts the detection function from physical sensors and implements it through software-based analysis of power consumption data. By removing the sensor hardware and its calibration infrastructure, the system achieves adverse event detection capability while significantly reducing device complexity and eliminating calibration requirements.
Solution Approach 2:
Instead of using physical sensors to directly measure thrombus-related parameters, the system creates a computational model that correlates power consumption patterns with thrombus formation. The algorithm analyzes deviations in power consumption from expected patterns, effectively copying the detection function through mathematical modeling rather than direct physical measurement.
3Device complexity
If simple detection methods are used, then device complexity is reduced, but sensitivity is insufficient and false alarms increase
Solution Approach 1:
The patent segments the power consumption signal into different frequency components using filtering techniques. By applying low-pass filters with specific cutoff frequencies, the system separates relevant thrombus-related power variations from noise and normal operational fluctuations, enhancing detection sensitivity while maintaining computational simplicity.
Solution Approach 2:
The system dynamically adjusts detection thresholds based on analyzed power consumption trends rather than using fixed thresholds. The algorithm adapts to normal variations in power consumption patterns and only triggers alerts when deviations exceed dynamically determined limits, reducing false alarms while maintaining high sensitivity.
Data Source
AI summary
A method of determining an adverse event within a patient having an implantable blood pump including calculating a plurality of power consumption trends of the blood pump during a plurality of time periods using a low-pass filter, determining a plurality of power trend differences between the plurality of power consumption trends, calculating a total amount of the plurality of power trend differences during a time interval, and generating an alarm when the total amount of the plurality of power trend differences exceeds a pre-determined threshold.