AI-Supported ECG Diagnosis for Rapid LV Dysfunction Prognosis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for assessing left ventricular systolic and diastolic dysfunction (LVSD and LVDD) are laborious, expensive, and sometimes inaccurate, lacking a rapid and easily performed test for cardiac disease evaluation.

Innovation Solution

An apparatus and method using artificial intelligence-supported diagnostic assessment tools that utilize a neural network trained on multi-channel sensor readings, including electrocardiogram data, to provide accurate prognostic data comparable to echocardiograms.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If echocardiograms are used to assess left ventricular systolic and diastolic dysfunction, then diagnostic accuracy is improved, but the test becomes laborious, expensive, and less accessible

Engineering Contradiction:
Improvediagnostic accuracyVSAvoidtest complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent replaces the mechanical echocardiography system with an AI-based diagnostic system that processes electrocardiogram data. The neural network model substitutes the need for complex echocardiogram equipment and operator skill, achieving comparable diagnostic accuracy through computational analysis of electrical heart signals.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent creates a computational model that copies the diagnostic functionality of echocardiograms using electrocardiogram data. The AI system learns to replicate echocardiogram-based assessments by training on paired datasets, enabling the simpler ECG test to provide equivalent diagnostic information about left ventricular dysfunction.

Inventive Principle:
Principle #26Copying

2Reliability

If echocardiograms are used for cardiac disease evaluation, then diagnostic capability is improved, but cost and time requirements increase

Engineering Contradiction:
Improvediagnostic capabilityVSAvoidtest duration
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent performs preliminary training of the neural network model using extensive paired echocardiogram and ECG datasets before deployment. This preliminary action enables the system to quickly process new ECG data without requiring time-consuming echocardiogram procedures during actual patient assessment, achieving both accuracy and speed.

Inventive Principle:
Principle #10Preliminary action

3Loss of information

If echocardiograms are used to assess LVSD and LVDD, then prognostic information is obtained, but accessibility and ease of performance deteriorate

Engineering Contradiction:
Improveprognostic informationVSAvoidtest accessibility
Core Design Contradiction:
Loss of informationVSEase of operation

Solution Approach 1:

The patent makes the electrocardiogram machine universal by enabling it to perform both traditional ECG analysis and AI-based assessment of left ventricular systolic and diastolic dysfunction. This multi-functionality allows a single, widely available device to provide comprehensive cardiac evaluation including prognostic information previously requiring specialized echocardiography.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP4611000A1Apparatus and method for training an artificial intelligence-supported diagnostic assessment tool
Publication Date: 2025.09.03 MAYO FOUNDATION FOR MEDICAL EDUCATION & RESEARCH
  • EP4611000A1 patent drawingFigure 1
  • EP4611000A1 patent drawingFigure 2
  • EP4611000A1 patent drawingFigure 3

AI summary

An apparatus and method for training an artificial intelligence-supported diagnostic assessment tool may provide rapid and accurate prognosis determinations. Apparatus may include at least a processor configured to receive a plurality of multi-channel sensor readings of physiological data, generate training data correlating each of the plurality of multi-channel sensor readings with a plurality of diagnostic labels, train a neural network using the plurality of diagnostic labels, receive a time series input describing user physiological data from at least a sensor, input the time series input into the trained neural network, generate diagnostic data as a function of the time series input and the trained neural network, determine prognostic data as a function of the diagnostic data, and output the prognostic data.