ML Classifier for AF Recurrence Prediction from CT Morphology

Solution Overview

Problem

Current methods for predicting the recurrence of atrial fibrillation (AF) after endovascular pulmonary vein isolation (PVI) are inadequate, as they fail to accurately identify patients likely to experience successful outcomes, despite the importance of pulmonary vein morphology and left atrial size as radiographic markers.

Innovation Solution

A machine learning classifier is trained using radiographic features from chest CT images and clinical features to predict the recurrence or non-recurrence of AF, employing morphometric features such as pulmonary vein and left atrial dimensions, and clinical factors like age and BMI, to generate a prognosis and personalized treatment plans.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional prediction methods are used for AF recurrence after PVI, then the prediction process is simple, but the prediction accuracy is insufficient

Engineering Contradiction:

Improveprediction accuracyVSAvoidprediction system complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent combines multiple data sources including radiographic features from chest CT images, echocardiographic measurements, and clinical parameters into a unified prediction model. This integration of diverse data types enables more accurate prediction of AF recurrence by capturing multiple aspects of patient anatomy and physiology that individually would be insufficient.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The patent introduces machine learning algorithms as intermediaries that process and analyze the collected radiographic and clinical data. These algorithms serve as mediators between the raw data and the prediction output, automatically identifying complex patterns and relationships that would be difficult to detect through traditional manual assessment methods.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Measurement precision

If radiographic features are extracted and analyzed, then prediction accuracy improves, but the complexity of data processing increases

Engineering Contradiction:

Improveprediction accuracyVSAvoiddata processing difficulty

Core Design Contradiction:

Measurement precisionVSDifficulty of detecting and measuring

Solution Approach 1:

The patent performs preliminary processing of radiographic images by automatically extracting relevant morphological features such as pulmonary vein dimensions, left atrial size, and anatomical relationships before analysis. This preprocessing step prepares the data in advance, making it more suitable for subsequent machine learning analysis and reducing the computational complexity during the actual prediction phase.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent replaces manual measurement and analysis of radiographic features with automated computer-based image analysis and machine learning algorithms. This substitution eliminates the need for manual tracing and measurement by clinicians, automatically extracting quantitative features from images and reducing both time and expertise requirements for data processing.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Reliability

If multiple radiographic and clinical features are combined, then prediction reliability improves, but the time required for analysis increases

Engineering Contradiction:

Improveprediction reliabilityVSAvoidanalysis time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent performs preliminary processing of radiographic images by automatically extracting relevant morphological features such as pulmonary vein dimensions, left atrial size, and anatomical relationships before analysis. This preprocessing step prepares the data in advance, making it more suitable for subsequent machine learning analysis and reducing the computational complexity during the actual prediction phase.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent replaces manual measurement and analysis of radiographic features with automated computer-based image analysis and machine learning algorithms. This substitution eliminates the need for manual tracing and measurement by clinicians, automatically extracting quantitative features from images and reducing both time and expertise requirements for data processing.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS11540796B2Prediction of risk of post-ablation atrial fibrillation based on radiographic features of pulmonary vein morphology from chest imaging

Publication Date: 2023.01.03 THE CLEVELAND CLINIC FOUND

AI summary

Embodiments discussed herein facilitate generation of a prognosis for recurrence or non-recurrence of atrial fibrillation (AF) after pulmonary vein isolation (PVI). A first set of embodiments discussed herein relates to training of a machine learning classifier to determine a prognosis for AF after PVI based on radiographic images, alone or in combination with clinical features. A second set of embodiments discussed herein relates to determination of a prognosis for a patient for AF after PVI based on radiographic images, alone or in combination with clinical features.