Electrocardiography for the Automatic Analysis of Arrhythmia in Children

Overview

The project is a direct response to the identified lack of ECG diagnostic solutions dedicated to children. There are several tools for automatic ECG signal analysis in adults, but these cannot be used in the diagnosis of heart disorders among children. A digital ECG analysis technology developer, Cardiomatics, and the Medical University of Warsaw team have taken the challenge of developing an internationally innovative tool for automatic assessment, analysis, and interpretation of electrocardiographic signals in pediatric patients. The developed tool will allow cardiac arrhythmias in children to be assessed more effectively and minimize the time needed for cardiologists to evaluate data received from the Holter monitor due to the use of algorithms, which are based on artificial intelligence.

The study is an investigator-initiated, single centre, prospective observational trial. The study will be carried out in in the department of pediatric cardiology of Medical University of Warsaw. Aim of the study: To develop an innovative tool for automatic analysis of cardiac arrhythmias and conduction for pediatric patients The study group will consist of 275 children with various heart rytm disturbances including those with congenital heart diseases following open heart surgery. The study group will be divided into the the age categories as follows: 50 infants under the age of 1 year old, 75 children 1-5 years old, 75 children aged 6-12 years and 75 aged 13-18 years. Control group will consist of 400 healthy volunteers (100 in each age group defined as in the study group). All patients will undergo * 12-lead ECG recording * 24-hour ECG Holter monitoring Together with the ECG signals patients medical history will be acquired including: * identification number * exact date and time of ECG obtaining * age * height * weight * diagnosed comorbidities - especially informations regarding the diagnosed congenital heart disease (if applicable). After obtaining the ECG signals they will be analyzed by an experienced pediatric cardiologist. The obtained signals together with the clinical interpretation will be transferred to Cardiomatics and used to build algorithms that will enable high-quality automatic analysis in children. The algorithms will be built using deep neural network architectures, such as ResNet, and will operate on filtered ECG signals. The created unique database of different arrhythmias in pediatric patients will be used to "train" the algorithm. Once the algorithm is developed its reliability will be tested using signals from a created data base that were not used to "train" the algorithm in the earlier stage of the process. The creation of a reliable system for automatic analysis of ECG recordings using the Holter method in children will not only improve the work of clinicians but also increase the availability and universality of this test, which is of great importance in the detection of rhythm and conduction disorders in pediatrics. Technology will also improve the recognition of broad range of diseases, so it make possible to undertake adequate therapy at an earlier stage.