-
ECG Feature Importance Rankings: Cardiologists vs. Algorithms
Authors:
Temesgen Mehari,
Ashish Sundar,
Alen Bosnjakovic,
Peter Harris,
Steven E. Williams,
Axel Loewe,
Olaf Doessel,
Claudia Nagel,
Nils Strodthoff,
Philip J. Aston
Abstract:
Feature importance methods promise to provide a ranking of features according to importance for a given classification task. A wide range of methods exist but their rankings often disagree and they are inherently difficult to evaluate due to a lack of ground truth beyond synthetic datasets. In this work, we put feature importance methods to the test on real-world data in the domain of cardiology,…
▽ More
Feature importance methods promise to provide a ranking of features according to importance for a given classification task. A wide range of methods exist but their rankings often disagree and they are inherently difficult to evaluate due to a lack of ground truth beyond synthetic datasets. In this work, we put feature importance methods to the test on real-world data in the domain of cardiology, where we try to distinguish three specific pathologies from healthy subjects based on ECG features comparing to features used in cardiologists' decision rules as ground truth. Some methods generally performed well and others performed poorly, while some methods did well on some but not all of the problems considered.
△ Less
Submitted 5 April, 2023;
originally announced April 2023.
-
MedalCare-XL: 16,900 healthy and pathological 12 lead ECGs obtained through electrophysiological simulations
Authors:
Karli Gillette,
Matthias A. F. Gsell,
Claudia Nagel,
Jule Bender,
Bejamin Winkler,
Steven E. Williams,
Markus Bär,
Tobias Schäffter,
Olaf Dössel,
Gernot Plank,
Axel Loewe
Abstract:
Mechanistic cardiac electrophysiology models allow for personalized simulations of the electrical activity in the heart and the ensuing electrocardiogram (ECG) on the body surface. As such, synthetic signals possess known ground truth labels of the underlying disease and can be employed for validation of machine learning ECG analysis tools in addition to clinical signals. Recently, synthetic ECGs…
▽ More
Mechanistic cardiac electrophysiology models allow for personalized simulations of the electrical activity in the heart and the ensuing electrocardiogram (ECG) on the body surface. As such, synthetic signals possess known ground truth labels of the underlying disease and can be employed for validation of machine learning ECG analysis tools in addition to clinical signals. Recently, synthetic ECGs were used to enrich sparse clinical data or even replace them completely during training leading to improved performance on real-world clinical test data. We thus generated a novel synthetic database comprising a total of 16,900 12 lead ECGs based on electrophysiological simulations equally distributed into healthy control and 7 pathology classes. The pathological case of myocardial infraction had 6 sub-classes. A comparison of extracted features between the virtual cohort and a publicly available clinical ECG database demonstrated that the synthetic signals represent clinical ECGs for healthy and pathological subpopulations with high fidelity. The ECG database is split into training, validation, and test folds for development and objective assessment of novel machine learning algorithms.
△ Less
Submitted 29 November, 2022;
originally announced November 2022.
-
Whole Heart Anatomical Refinement from CCTA using Extrapolation and Parcellation
Authors:
Hao Xu,
Steven A. Niederer,
Steven E. Williams,
David E. Newby,
Michelle C. Williams,
Alistair A. Young
Abstract:
Coronary computed tomography angiography (CCTA) provides detailed an-atomical information on all chambers of the heart. Existing segmentation tools can label the gross anatomy, but addition of application-specific labels can require detailed and often manual refinement. We developed a U-Net based framework to i) extrapolate a new label from existing labels, and ii) parcellate one label into multip…
▽ More
Coronary computed tomography angiography (CCTA) provides detailed an-atomical information on all chambers of the heart. Existing segmentation tools can label the gross anatomy, but addition of application-specific labels can require detailed and often manual refinement. We developed a U-Net based framework to i) extrapolate a new label from existing labels, and ii) parcellate one label into multiple labels, both using label-to-label map**, to create a desired segmentation that could then be learnt directly from the image (image- to-label map**). This approach only required manual correction in a small subset of cases (80 for extrapolation, 50 for parcella-tion, compared with 260 for initial labels). An initial 6-label segmentation (left ventricle, left ventricular myocardium, right ventricle, left atrium, right atrium and aorta) was refined to a 10-label segmentation that added a label for the pulmonary artery and divided the left atrium label into body, left and right veins and appendage components. The final method was tested using 30 cases, 10 each from Philips, Siemens and Toshiba scanners. In addition to the new labels, the median Dice scores were improved for all the initial 6 labels to be above 95% in the 10-label segmentation, e.g. from 91% to 97% for the left atrium body and from 92% to 96% for the right ventricle. This method provides a simple framework for flexible refinement of anatomical labels. The code and executables are available at cemrg.com.
△ Less
Submitted 18 November, 2021;
originally announced November 2021.
-
Direct Transcription for Dynamic Optimization: A Tutorial with a Case Study on Dual-Patient Ventilation During the COVID-19 Pandemic
Authors:
Eric C. Kerrigan,
Yuanbo Nie,
Omar Faqir,
Caroline H. Kennedy,
Steven A. Niederer,
Jose A. Solis-Lemus,
Peter Vincent,
Steven E. Williams
Abstract:
A variety of optimal control, estimation, system identification and design problems can be formulated as functional optimization problems with differential equality and inequality constraints. Since these problems are infinite-dimensional and often do not have a known analytical solution, one has to resort to numerical methods to compute an approximate solution. This paper uses a unifying notation…
▽ More
A variety of optimal control, estimation, system identification and design problems can be formulated as functional optimization problems with differential equality and inequality constraints. Since these problems are infinite-dimensional and often do not have a known analytical solution, one has to resort to numerical methods to compute an approximate solution. This paper uses a unifying notation to outline some of the techniques used in the transcription step of simultaneous direct methods (which discretize-then-optimize) for solving continuous-time dynamic optimization problems. We focus on collocation, integrated residual and Runge-Kutta schemes. These transcription methods are then applied to a simulation case study to answer a question that arose during the COVID-19 pandemic, namely: If there are not enough ventilators, is it possible to ventilate more than one patient on a single ventilator? The results suggest that it is possible, in principle, to estimate individual patient parameters sufficiently accurately, using a relatively small number of flow rate measurements, without needing to disconnect a patient from the system or needing more than one flow rate sensor. We also show that it is possible to ensure that two different patients can indeed receive their desired tidal volume, by modifying the resistance experienced by the air flow to each patient and controlling the ventilator pressure.
△ Less
Submitted 23 November, 2020;
originally announced November 2020.
