ICM EGM Triage Using Machine Learning for False Positive Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing implantable cardiac monitors (ICMs) face challenges in managing large volumes of electrogram (EGM) data, including false arrhythmia detections that burden clinicians with manual review and introduce inaccuracies in diagnostic information due to false positives, particularly from conditions like sick sinus rhythms and premature ventricular contractions.
Innovation Solution
A system utilizing machine-learning models to analyze classified EGM datasets from ICMs, determining confidence indicators, ranking scores, and selecting a subset of EGM datasets for display, thereby reducing clinical review burden and improving diagnostic accuracy by mitigating false positives.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all EGM segments are reviewed manually, then diagnostic accuracy is maintained, but clinical workload and resource requirements increase significantly
Solution Approach 1:
The patent extracts and separates EGM segments into distinct categories based on arrhythmia detection confidence levels. High-confidence segments are isolated for automated processing while low-confidence segments remain for manual review, thereby reducing clinical workload while maintaining diagnostic accuracy for critical cases.
Solution Approach 2:
The patent introduces an intermediary machine-learning classification system that acts between the EGM data and clinical reviewers. This intermediary automatically prioritizes and filters segments, presenting only the most relevant data to clinicians and reducing the time and resources required for manual review.
2Loss of time
If EGM data volume is reduced, then clinical review time decreases, but false arrhythmia detections increase
Solution Approach 1:
The patent applies local quality by differentiating EGM segments based on their individual confidence levels. Each segment is evaluated and assigned a priority level, allowing clinicians to focus manual review on low-confidence segments that may represent false detections, while high-confidence segments are processed automatically, thus reducing review time without increasing false detection rates.
Solution Approach 2:
The patent performs preliminary action by using machine-learning models to pre-classify and prioritize EGM segments before they reach clinical reviewers. This preliminary processing filters out obvious false detections and organizes data by urgency and relevance, allowing clinicians to review only the most critical segments efficiently.
3Ease of operation
If machine-learning models are applied to filter EGM data, then clinical review burden decreases, but system complexity increases
Solution Approach 1:
The patent implements self-service by enabling the machine-learning system to autonomously classify, prioritize, and filter EGM segments without requiring continuous human intervention. The system self-adjusts and processes data independently, reducing clinical review burden while the complexity is contained within the automated infrastructure rather than requiring manual management.
Data Source
AI summary
A medical data and diagnostics management system for processing classified EGM datasets includes a server system that receives transmissions of classified EGM datasets, each corresponding to an arrhythmic episode detected by an IMD, and applies a machine-learning model to each classified EGM dataset stored in a database, thereby determining for each of the classified EGM datasets to which the model is applied a respective indicator of whether the IMD classification is a false positive or a true positive. The system is further configured to remove from the database one or more of the classified EGM data sets for which the respective IMD classification is identified using the machine-learning model as being a false positive, thereby creating a plurality of machine-adjudicated patient database records stored in the database. The system may also provide for display a selected one of the machine-adjudicated EGM datasets and/or facilitate reprograming the IMD.