Skip to Main content Skip to Navigation
New interface

Estimation du risque de mort subite par arrêt cardiaque à l'aide de méthodes d'apprentissage artificiel

Charles-Henri Cappelaere 1 
1 SIGMA - Laboratoire Signaux, Modèles et Apprentissage Statistique
ESPCI Paris - Ecole Superieure de Physique et de Chimie Industrielles de la Ville de Paris, CNRS - Centre National de la Recherche Scientifique : UMR7084
Abstract : Over 40,000 sudden cardiac deaths (SCD) occur per year in France. Implantable cardioverter defibrillators (ICD) have been prescribed for prophylaxis since the early 2000's, for patients at high risk of SCD. Unfortunately, most implantations to date appear unnecessary: according to a recent study, 81% of the implanted ICDs were not used during the first five years following the implantation. This result raises an important issue because of the perioperative and postoperative risks. It has been shown in two studies that 13% to 17% of ICD-implanted patients have undergone at least one inappropriate shock (i.e. a shock generated by the defibrillator although the patient's life was not at risk) during the follow-up periods; these shocks are known to be detrimental to the cardiac muscle. Thus, it is important to improve the selection of the candidates to ICD implantation in primary prevention. Risk stratification for sudden cardiac death based on long-term electrocardiographic (Holter) recordings has been extensively performed in the past, without resulting in a significant improvement of the selection of candidates to ICD implantation. The present report describes a nonlinear multivariate analysis of Holter recording indices. We computed all the descriptors available in the Holter recordings present in our database. The latter consisted of labelled Holter recordings of patients equipped with an ICD in primary prevention; a fraction of these patients received at least one appropriate therapy from their ICD during a 6-month follow-up. Based on physiological knowledge on arrhythmogenesis, feature selection was performed, and an innovative procedure of classifier design and evaluation was proposed. The classifier is intended to discriminate patients who are really at risk of sudden death from patients for whom ICD implantation does not seem necessary. In addition, we designed an ad hoc classifier that capitalizes on prior knowledge on arrhythmogenesis. We conclude that improving prophylactic ICD-implantation candidate selection by automatic classification from Holter recording features may be possible. Nevertheless, that statement should be supported by the study of a more extensive and appropriate database. This is mandatory for decreasing the rate of false negatives (i.e. the proportion of patients who are not deemed to be at risk although they actually are), hence increasing the negative predictive value of our method.
Complete list of metadata
Contributor : Charles-Henri Cappelaere Connect in order to contact the contributor
Submitted on : Thursday, January 30, 2014 - 10:32:26 AM
Last modification on : Sunday, June 26, 2022 - 5:47:21 PM
Long-term archiving on: : Thursday, May 1, 2014 - 1:50:11 AM


  • HAL Id : pastel-00939082, version 1



Charles-Henri Cappelaere. Estimation du risque de mort subite par arrêt cardiaque à l'aide de méthodes d'apprentissage artificiel. Statistiques [math.ST]. Université Pierre et Marie Curie - Paris VI, 2014. Français. ⟨NNT : ⟩. ⟨pastel-00939082v1⟩



Record views


Files downloads